The

Roots of Hinduism
The Roots of Hinduism
The Early Aryans and the Indus Civilization

Asko Parpola
Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in
research, scholarship, and education by publishing worldwide.
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto

With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trademark of Oxford University Press in the UK and certain other countries.

Published in the United States of America by

Oxford University Press
198 Madison Avenue, New York, NY 10016
© Asko Parpola 2015
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or
by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by license,
or under terms agreed with the appropriate reproduction rights organization. Inquiries concerning reproduction outside the
scope of the above should be sent to the Rights Department, Oxford University Press, at the address above.

You must not circulate this work in any other form and you must impose this same condition on any acquirer.
Cataloging-in-Publication data is on file at the Library of Congress
ISBN 978–0–19–022690–9 (hbk.); 978–0–19–022692–3 (pbk.)
eISBN 978–0–19–022693–0
Contents

Preface
Abbreviations

Introduction
1. Defining “ Hindu” and “Hinduism”
2. The Early Aryans
3. Indo-European Linguistics
4. The Indus Civilization
5. The Indus Religion and the Indus Script

PART I The Early Aryans

6. Proto-Indo-European Homelands
7. Early Indo-Iranians on the Eurasian Steppes
8. The BMAC of Central Asia and the Mitanni of Syria
9. The Rigvedic Indo-Aryans and the Dāsas
10. The Aśvins and Mitra-Varu ṇa
11. The Aśvins as Funerary Gods
12. The Atharvaveda and the Vrātyas
13. The Megalithic Culture and the Great Epics

PART II The Indus Civilization

14. The Language of the Indus Civilization
15. Fertility Cults in Folk Religion
16. Astronomy, Time-Reckoning, and Cosmology
17. Dilmun, Magan, and Meluhha
18. Royal Symbols from West Asia
19. The Goddess and the Buffalo
20. Early Iranians and “Left-Hand” Tantrism
21. Religion in the Indus Script

Conclusion
22. Prehistory of Indo-Aryan Language and Religion
23. Harappan Religion in Relation to West and South Asia
24. Retrospect and Prospect

Bibliographical Notes
References
Index
Preface

India’s earliest urban culture, which existed from 2600 until 1900 BCE, was discovered in the
valley of the Indus River in the early 1920s. The Indus civilization had a population estimated
at one million people, who lived in more than 1000 settlements, several of them cities of some
50,000 inhabitants, most notably Harappa and Mohenjo-daro, which display extraordinary
town planning and water engineering, but not such splendid palaces and temples as ancient
Egypt or Mesopotamia. With an area of nearly a million square kilometers, approximately a
quarter the size of Europe, the Indus civilization was more extensive than the other key urban
cultures of the time, in Mesopotamia and Egypt; it stretched eastwards from the Indus Valley
as far as about Delhi and southwards almost as far as Mumbai (Fig. 4.2). It had long-distance
contacts with West Asia, as proved by Indus objects and cuneiform records found in
Mesopotamia. It also largely established a way of life that has continued in the villages of the
subcontinent until the present day.
Yet, after almost a century of excavation and research, the Indus civilization remains
comparatively little understood, especially in regard to its thought. What language did the
Indus people speak? Is it related to the Sanskrit language of north India, to the Dravidian
languages of south India, or to another surviving Indian language family? How might we
decipher the exquisitely carved Indus inscriptions? What deities did the Indus civilization
worship? Did these deities survive to become part of the religion described in the earliest
surviving Indian literature, the Vedas, composed in Sanskrit, which are generally attributed to
the centuries after 1500 BCE? Did the Aryan composers of the Vedas migrate to South Asia
from outside the subcontinent? If so, where did they come from, and during what period did
they migrate? Indeed, are the roots of contemporary Hinduism to be found in the religion of
the Indus civilization as well as in the Vedic religion?
This book proposes answers to these fundamental questions about the beginning of Indian
civilization. Some of the answers occurred to me soon after the start of my career in the study
of Vedic literature and religion half a century ago, while others have struck me only in the last
few years, following decades of research, both in Europe and in Asia. I have published most of
them in the scholarly literature (where the interested reader will find more detail and full
documentation). However, this literature is not well known outside a relatively small circle of
experts. The time is ripe to introduce what I think are the most solid of my conclusions to a
wider readership. My hope is that this book will stimulate further research and thus
contribute to a solution that will be acceptable to many scholars.
Unfortunately, as I shall outline in the Introduction, at present there is little agreement on
many of the issues I discuss. Since the 1980s, the above questions, especially the “Aryan
problem,” have been debated with increasing animosity, colored by the history of modern
colonialism in India, and often by people without pertinent qualifications. Hence, it is vital to
sort out the origins of the Vedic Aryans if we are to establish the true linguistic identity and
religious affinity of the Indus civilization. Fortunately, the prospect for settling this vexed
question has recently improved with the advance of archaeological excavations in Central Asia
—where a major new Bronze Age civilization (Bactria and Margiana Archaeological Complex,
BMAC, or the Oxus civilization), was discovered in the 1970s—and in Eurasia, including
Russia.
Archaeological cultures have long been correlated with linguistic and ethnic groups. To
take one relevant example, the proposal that people of the Andronovo cultures of Kazakhstan
spoke an Indo-Iranian language was put forward by A. M. Tallgren in 1928. Various later
scholars have had the same opinion, but while some speak of Proto-Aryan, others speak of
Proto-Indo-Aryan, others of Proto-Iranian; others (e.g., Lamberg-Karlovsky 2002) deny the
very possibility of such a correlation. I shall not discuss the methodology of archaeo-linguistic
correlation here, but restrict myself only to some fundamental conclusions. Readers interested
in the methodology should consult the publications of J. P. Mallory (1989 and later), who has
developed this methodology remarkably.
The relatively uniform material “cultures” studied by archaeology reflect human
communities, whose members must have communicated with each other by means of a
language (in some cases more than one language). Shared material culture and shared
language are both among the strongest sources of ethnic identity. They may change very little
if people stay in the same place and receive no visitors from beyond their habitat, as has
happened in Iceland since the early Viking Age. Contact with other communities has normally
led to changes, the extent of which depends on the intensity of contact. Trade may introduce
new artifacts and ideas, and loanwords denoting these artifacts and ideas. Conquest and
immigration may lead to a community’s abandoning its earlier culture and language in favor of
a new culture and language. Language shifts involve a period of bilingualism, when part of the
community speaks two (or more) languages.
Such continuities and change can be analyzed using a comparative method. Both
archaeology and linguistics have developed techniques to do this. In principle, when
correlating languages with archaeology, it is better not to assume long-distance migrations,
except under exceptional circumstances, such as the combination of open steppe with wheeled
vehicles or horses (cf. Nichols 1997:369). There must always be tangible archaeological
evidence for assuming the existence of such movements.
The most important methodological principle is that isolated correlations are invalid, only
a web of correlations similar to that of an entire language family can be convincing. Every
piece of evidence available from archaeology and linguistics should be used and should fit
together. With this methodology in mind, I have been searching, since 1973, for an optimal
solution to this puzzle: How can the Eurasian archaeological finds be correlated with the
discoveries of Indo-European historical linguistics and the route of the Aryan migrations to
India be established?

Given the above extremely contentious background, it is perhaps more than usually necessary
for an author to present his credentials—linguistic, anthropological, archaeological, and
otherwise—for writing on this subject in its many and varied aspects. I would therefore like to
relate, fairly briefly, how and why I undertook the relevant research.
After seven years of Latin at school, I went to the University of Helsinki in 1959 to study
classics, especially ancient Greek language and literature, but quickly switched to Sanskrit
and comparative Indo-European linguistics. Unlike the study of Homer and Herodotus, which I
loved, this major seemed to offer vast amounts of material virtually untouched by western
scholarship. For my PhD dissertation my teacher, Pentti Aalto, suggested tackling the
Drāhyāyaṇa-Śrautasūtra (DŚS), which had not been translated into any language. A critical
edition of this text with the medieval commentary of Dhanvin had been started by a Finnish
Sanskritist, J. N. Reuter, but only one-fifth had been published, back in 1904; the rest, in
Reuter’s unfinished manuscript, was kept in the university library.
The DŚS deals with the duties of the chanter priests in Vedic sacrificial rituals. I contacted
Jan Gonda, a distinguished Dutch master of Vedic studies, who kindly got me started. Under
his guidance I prepared an annotated English translation of the published portion of the DŚS
and its commentary (1969), and for my dissertation I studied the relationship of this text to
the closely parallel Lāṭyāyana-Śrautasūtra and other literature of the Sāmaveda (1968). This
was interesting and exciting work, during which I chanced to discover a unique manuscript
containing completely unknown portions of a second parallel text, the Jaiminīya-Śrautasūtra,
at the library of the former maharaja of Tanjore (Thanjavur) in the south Indian state of Tamil
Nadu (1967 [1968]; 2011d; 2012c; in press f). On my first visit to India in 1971, I
photographed this manuscript, and started a search for all existing manuscripts of works
belonging to the less-known Jaiminīya branch of the Sāmaveda (1973). Since then I have
visited Kerala, Tamil Nadu, and Karnataka some twenty-five times, mainly pursuing and
photographing Jaiminīya manuscripts. Ever since 1985, I have collaborated in this work and in
the publication of the Jaiminīya texts with Masato Fujii, my former student (Fujii 2012). I am
grateful to Fujii-san—and to Toshiki Osada—for making it possible for me to do research in
Japan, as well as in India.
Vedic rituals and recitation by Brahmins have survived for some three millennia better in
south India than in the north, which was long dominated by Muslim rulers. Frits Staal had
documented them in Kerala in his book Nambudiri Veda recitation (1961), from which I
learned that these particular Brahmins belong to the Jaiminīya school. Staal wanted to record
this living oral tradition and invited me to join his project, which he inaugurated by tape-
recording the songs of my future guru in studying the Sāmaveda, Śrī Iṭṭi Ravi Nambūdiri. In
1975, I had the opportunity to witness the entire performance of a Vedic soma sacrifice by
sixteen officiating priests, who built the sacred enclosure with its huts and a fire altar of 1000
bricks. The resulting two-volume book edited by Staal, Agni: The Vedic ritual of the Fire Altar
(1983), is a monumental study of Vedic religion. My three contributions include an attempt to
penetrate the pre-Vedic Indian background of Vedic rituals. In 1983–1985, accompanied by my
wife, Marjatta (who eventually wrote a book, Kerala Brahmins in transition [2000]) and two
postgraduate students (Klaus Karttunen and Masato Fujii), I documented domestic rituals of
Jaiminīya Sāmavedins by taking photographs and videos of actual and simulated
performances, and reading with my guru unpublished manuals written in Malayalam, the
Dravidian language of Kerala (2011e). Besides the Vedic heritage of the Brahmins, I also
studied some other exciting Hindu traditions, such as the cult of Kuṭṭiccāttan, a little-known
god of sorcery in Kerala (1999c).
My wish to understand better the formation of Hinduism was what encouraged my initial
interest in the Indus civilization. But the impetus to study the Indus script came from Seppo
Koskenniemi, a childhood friend who was a scientific advisor to IBM in Finland in the mid-
1960s. He asked me, when I had just started working on my PhD, if I would like to use a
computer—quite a new technique in those days!—for any research problem in my field. We
could use IBM’s facilities; Seppo would take care of programming.
Having recently read the Cambridge University classicist John Chadwick’s fascinating
book The decipherment of Linear B (1958), I suggested that we take up the decipherment of
the Indus script as a hobby. Chadwick had explained in some detail the methods applied in this
detective work by Michael Ventris and himself. Compilation of statistics and indexes to Linear
B sign sequences played an important role. I thought a computer could ease this kind of
analysis. My younger brother, Simo, who was studying Assyriology, joined our team and
provided crucial expertise. In our spare time we collected the Indus texts from archaeological
reports, drew up a provisional list of the different Indus signs, allotted a number to each sign,
and punched the texts in numerical form onto cards. After the computer had processed all this
information into lists, we transcribed the numbers back into pictograms, searched the lists for
meaningful patterns, and tested automated methods of decipherment (Parpola, Parpola, &
Koskenniemi 1966; Koskenniemi, Parpola, & Parpola 1970). Later the computer was
programmed to draw in Indus signs the first concordance to all sign combinations, which was
published in 1973. A three-volume revised version was produced in collaboration with Seppo’s
younger brother Kimmo Koskenniemi, a computer linguist, in 1979–1981 (see Fig. 5.1). Kimmo
Koskenniemi (1981) developed automated syntactic methods for the study of the Indus script;
these were implemented and tested in her M.Sc. thesis by my daughter Päivikki Parpola
(1987; 1988).
Pentti Aalto, my Sanskrit teacher, was a code-breaker for military intelligence during the
Second World War. He urged me to test whether some Indus signs could be read on the basis
of our assumption that the underlying language was Dravidian (rather than Sanskrit). Since
our trial substitutions seemed to yield sense, I quickly drafted a booklet, published in 1969,
soon followed by two progress reports (Parpola et al. 1969a,b; 1970). These appeared under
the joint names of our team, even though I could not easily consult my team-mates in Finland,
because I was by now working in Copenhagen. My hurry was partly due to a rumor that a
team of Soviet scholars was trying to decipher the Indus script. I made the mistake of
including unripe ideas and presenting them with overconfidence and without proper checking.
I was also quite unfamiliar with Dravidian languages in the 1960s, which led to some
blunders. Understandably, these “first announcements” were received with severe and quite
deserved criticism. The reviews, though painful reading, were a very useful and sobering
lesson. But they did not shake my belief that our method and some of its fundamental
hypotheses were correct. The decipherment work had to be continued.
Encouragement came from Gerard Clauson and John Chadwick (1969). Pentti Aalto and I
were invited to give lectures in England in 1969, and in 1973 I was one of the speakers at the
Royal Asiatic Society’s sesquicentenary symposium on undeciphered scripts (1975a).
Chadwick, whom I met several times in Cambridge, eventually recommended my book,
Deciphering the Indus script (1994), for publication by Cambridge University Press. There I
developed my early ideas, by proposing specific, carefully checked and tested, Proto-Dravidian
readings for two dozen Indus signs.
Some critics continue to label this whole attempt at a partial decipherment abortive,
because it allegedly has not made any progress. As if nothing has happened since the late
1960s, they still criticize the approach for those first announcements and for hypotheses
abandoned long ago. Yet, a lot of advance was made between 1969 and 1994, and further
steps forward have occurred since then.
To improve my grasp of Dravidian languages, in 1969 I started to learn Tamil in earnest
with two Tamil scholars, R. Panneerselvam and P. R. Subramanian, who were attached for two
years to the Scandinavian Institute of Asian Studies, my institute. They, together with my late
friend Eric Grinstead, a specialist in East Asia, compiled for my personal use a reverse index
to the seven-volume Tamil lexicon—a tool that turned out to be immensely useful in my study
of the Indus script. Later, as Professor of Indology at the University of Helsinki, I arranged for
Tamil and Malayalam to be taught there by native speakers on a regular basis.
During my first trip to South Asia in 1971, I was able to check the original inscriptions
kept in various museums. I discovered hundreds of unpublished inscriptions in South Asian
museums and initiated a major project to publish a comprehensive Corpus of Indus Seals and
Inscriptions (CISI) as an international collaboration under the auspices of UNESCO. After
many bureaucratic difficulties, three of the projected four volumes have appeared (1987,
1991, and 2010). This fundamental research tool is approaching completion.
Apart from Chadwick, my mentors in Cambridge included Bridget and Raymond Allchin,
who specialized in South Asian archaeology. They published a fundamental survey of research,
The birth of Indian civilization: India and Pakistan before 500 B.C. (1968), with later updated
versions. This book, which I studied very closely, gave me basic knowledge of the current
archaeological situation. In 1970, the Allchins established what is nowadays called the
European Association of South Asian Archaeology and Art History, inviting me to be one of the
founding members; and I had the privilege of serving on the board until 2002. Every second
year, the association has arranged an international conference for scholars to present the
results of their recent excavations and research. I attended most of these exciting meetings
(and organized one of them, in 1993 [Parpola & Koskikallio 1994]), which provided a unique
opportunity to follow the development of South Asian archaeology, including, of course, Indus
archaeology.
Study of the iconographic motifs of inscriptions and painted pottery from the Indus Valley,
as well as small statuettes and terracotta figurines, has suggested parallels in Iran and West
Asia that have led to major insights into the history of Indian religions. Archaeological
knowledge is required not only for comprehending the Indus civilization but also the evidence
for Aryan migrations into India. Since the early 1970s (Parpola 1974), I have tried to gain a
better understanding of how the emergence and dispersal of the Indo-European and Uralic
(Finno-Ugric) language families are reflected in the archaeological record—for the prehistory
of the Indo-Iranian languages involves both of these language families. I have learnt a lot from
J. P. Mallory’s important book, In search of the Indo-Europeans (1989), and from close
collaboration with Christian Carpelan, a Finnish archaeologist specializing in northern and
eastern Europe (Carpelan & Parpola 2001). I have also had the privilege of visiting some key
archaeological sites, such as Gonur in Turkmenistan, the main location of the BMAC or Oxus
civilization (two visits at the kind invitation of its excavator, Viktor I. Sarianidi); Botaj in
Kazakhstan, important for the history of horse domestication (one visit thanks to David
Anthony, a leading researcher; Parpola 1997c); and sites of the Abashevo culture in Russia.
For this research I had to acquire a reading knowledge of Russian. My friend Sergej V.
Kuz’minykh, an eminent Russian archaeologist, has kept me up-to-date with Russian
publications.
In a work like this, one must constantly adjust to the advances made in international
research and one’s own better grasp of the problems. Just before the book went to press, I
attended the twenty-second international conference of the European Association of South
Asian Archaeology and Art held in Stockholm in mid-2014. One new piece of information
slightly affecting this book may be mentioned. My dates for the Gandhāra Grave culture
reported in chapter 8, established in the long-continued excavations of Giorgio Stacul, now
require revision. Recent excavations with new radiocarbon dates, communicated by Vidale
and colleagues (2014), suggest c. 1500–1100 BCE for period IV, c. 1200–800 BCE for period V,
and c. 800–400/300 BCE for period VI.
I am immensely grateful to my friend Andrew Robinson, author of excellent books on the
history of writing and decipherment (among many other things), for his willingness to make
my text as accessible to the general reader as possible; and to the Building and Use of
Linguistic Technology (BAULT) research community at the University of Helsinki and its
director Kimmo Koskenniemi for funding Andrew’s work and some of the illustrations. Andrew
was a hard taskmaster, and without his constant prodding and encouragement I would not
have been able to put this book together—or at least not so speedily. The work progressed
chapter by chapter, as I sent my first draft to Andrew, who edited it, asking me to clarify
everything that he could not follow. He also recommended including a short explanation of
Indo-European linguistics and its methods for those readers coming to this subject for the first
time (see chapter 3).
My cordial thanks go to Stephanie Jamison and Joel Brereton for generously allowing me
to use their new translations of Rigvedic verses referring to the Dāsas and Dasyus in chapter
9. For very useful feedback I am indebted to the anonymous expert readers whom the
publisher invited to read the book proposal, and especially the one who read the nearly final
complete manuscript. Harry Falk and Xenia Zeiler also quickly read the text when it went to
press and suggested some adjustments.
The illustrations are vital for understanding the ideas in the text. I am much obliged to the
all the various copyright holders for granting permission to reproduce them. Many friends and
colleagues helped in the task of collecting the permissions: Julian Reade, Dominique Collon,
Mark Kenoyer, Ian Hodder, S. V. Kuz’minykh, N. B. Vinogradov, Yuri Rassamakin, Durre
Ahmed, Gyula Wojtilla, Andrew Robinson, Dipankar Home, Dipak Bhattacharya, Sayan
Bhattacharya, Joyoti Roy, and B. M. Pande. Anna Kurvinen graciously drew two nice maps
according to my specifications.
My best thanks are due to Cynthia Read, my editor at Oxford University Press, USA, and
her team, particularly Marcela Maxfield, Michael Durnin, and Sunoj Shankaran, for bringing
this book out so nicely and in good cooperation.
I dedicate the book to my family: Marjatta, Päivikki and Pekka, and Mette.
Asko Parpola
Helsinki, October 2014
Abbreviations

ASI Archaeological Survey of India

AV Arto Vuohelainen
CISI See CISI in References.
DAMGP Department of Archaeology and Museums, Government of Pakistan
EL Erja Lahdenperä
JL Jyrki Lyytikkä
NMI National Museum of India

VEDIC TEXTS WITH THEIR ABBREVIATIONS

For editions and translations see Gonda 1975; 1977; Jamison & Witzel 1992.
AĀ Aitareya-Āraṇyaka
AB Aitareya-Brāhmaṇa
ĀpŚS Āpastamba-Śrautasūtra
ĀśvGS Āśvalāyana-Gṛhyasūtra
AV(Ś) Atharvaveda(-Saṁhitā) (Śaunaka-śākhā)
AVP Atharvaveda-Saṁhitā, Paippalāda-śākhā
BĀU Bṛhad-Āraṇyaka-Upaniṣad (Kāṇva-śākhā)
BŚS Baudhāyana-Śrautasūtra
BŚu Baudhāyana-Śulvasūtra
ChU Chāndogya-Upaniṣad
DŚS Drāhyāyaṇa-Śrautasūtra
GGS Gobhila-Gṛhyasūtra
HGS Hiraṇyakeśi-Gṛhyasūtra
JB Jaiminīya-Brāhmaṇa
JGS Jaiminīya-Gṛhyasūtra
JŚS Jaiminīya-Śrautasūtra
JUB Jaiminīya-Upaniṣad-Brāhmaṇa
KĀ Kaṭha-Āraṇyaka
KauśS Kauśika-Sūtra
KB Kauṣītaki-Brāhmaṇa
KS Kaṭha-Saṁhitā / Kāṭhaka-Saṁhitā
KŚS Kātyāyana-Śrautasūtra
KŚu Kātyāyana-Śulvasūtra
KU Kaṭha-Upaniṣad
LŚS Lāṭyāyana-Śrautasūtra
MS Maitrāyaṇī Saṃhitā
PB Pañcaviṁśa-Brāhmaṇa
PGS Pāraskara-Gṛhyasūtra
RV Rigveda (-Saṁhitā)
ŚĀ Śāṅkhāyana-Āraṇyaka
ŚB(M) Śatapatha-Brāhmaṇa (Mādhyandina-śākhā)
ŚBK Śatapatha-Brāhmaṇa, Kāṇva-śākhā
ṢB Ṣaḍviṁśa-Brāhmaṇa
ŚŚS Śāṅkhāyana-Śrautasūtra
SV Sāmaveda(-Saṁhitā) (Kauthuma-śākhā)
SVB Sāmavidhāna-Brāhmaṇa
TĀ Taittirīya-Āraṇyaka
TB Taittirīya-Brāhmaṇa
TS Taittirīya-Saṁhitā
VādhA Vādhūla-Anvākhyāna
VādhŚS Vādhūla-Śrautasūtra
VS(M) Vājasaneyi-Saṁhitā (Mādhyandina-śākhā)
The Roots of Hinduism
Introduction
1

Defining “Hindu” and “Hinduism”

Before investigating the roots of Hinduism we must first define what is meant by the terms
Hindu and Hinduism. The etymology of “Hindu” goes back to about 515 BCE, when the Persian
king Darius the Great annexed the Indus Valley to his empire. Sindhu, the Sanskrit name of
the Indus River and its southern province—the area now known as Sindh—became Hindu in
the Persian language. The Ionian Greeks serving the Great King did not pronounce word-initial
aspiration (like French-speakers today) and so in the Greek language Persian Hindu became
Indos (whence, Latin Indus) and its surrounding country became India. But when the Persian-
speaking Mughals conquered northern India in the sixteenth century, they called the country
Hindustan and its people Hindu. During the Mughal empire, in the seventeenth century, the
British, too, began to use “Hindu” (or “Hindoo”) to describe the people living in the
subcontinent. During the British period, in the nineteenth century, the term was adopted by
those Indians who opposed colonialism, in order to distinguish themselves from Muslims. In
the twentieth century, Hindu became the common label for all Indians who were not Muslims,
Buddhists, Christians, Jains, Jews, Parsees, or Sikhs. V. D. Savarkar used the neologism
hindutva (coined with the Sanskrit abstract noun suffix -tva) to mean “the quality of being a
Hindu, Hindu identity” in his influential nationalist-political tract Hindutva: Who is a Hindu?
(1938), first published under a pseudonym in 1923, and still in print.
An East India Company merchant and evangelical Christian, Charles Grant, is known to
have used the term “Hindooism” for the earlier “Hindoo religion” or “Hindoo creed” as early
as 1787. During the nineteenth century, Hinduism became the general name for the native
religion(s) of India, excluding Buddhism, Jainism, and Sikhism (Pennington 2005). Some
Indians object to having a foreign term for their religion, preferring the Sanskrit expression
sanātana dharma, “eternal law or truth,” despite the fact that this expression was not applied
to any religious system in ancient texts.
Another objection is that “Hinduism” artificially bundles together different religions, such
as the Vaiṣṇava, Śaiva, and Śākta traditions of worship, each with its own theology and cult.
One may indeed say that:

Hinduism in toto, with various contradicting systems and all the resulting
inconsistencies, certainly does not meet the fundamental requirements for a historical
religion of being a coherent system; but its distinct entities [the so called “sects”] do.
They are indeed religions, while Hinduism is not. What we call “Hinduism” is a
geographically defined group of distinct but related religions, that originated in the
same region, developed under similar socio-economic and political conditions,
incorporated largely the same traditions, influenced each other continuously, and
jointly contributed to the Hindu culture. (von Stietencron 1989:20)

The classical Hindu gods Viṣṇu and Śiva go back, at least in part, to the polytheistic pantheon
of the older Vedic religion. The Vedas are regarded as the ultimate scriptural authority by
most Hindus, especially the “orthodox” Brahmins, but on the whole the Vedic religion plays a
small role in classical Hinduism, many aspects of which differ fundamentally from the religion
described in the Vedas. The more recent part of the Vedic religion, “Brahmanism,” is
sometimes counted as an older phase of Hinduism, since much of later Hindu philosophy is
based on it. Usually, however, “classical” Hinduism is considered to be of post-Vedic date,
beginning around 400–200 BCE, and epitomized in the epics the Mahābhārata and the
Rāmāyaṇa, and the Purāṇas. In his masterly sketch, Religions of ancient India, the great
French Indologist Louis Renou emphasizes the change that occurred between the Vedic
religion and later Hinduism:
The Vedic contribution to Hinduism, especially Hindu cult-practice and speculation, is
not a large one; Vedic influence on mythology is rather stronger, though here also there
has been a profound regeneration. Religious terminology is almost completely
transformed between the Vedas and the Epics or the Purāṇas, a fact which has not been
sufficiently emphasized; the old terms have disappeared or have so changed in meaning
that they are hardly recognizable; a new terminology comes into being. Even in those
cases where continuity has been suggested, as for Rudra-Śiva, the differences are
really far more striking than the similarities. (Renou 1953:47–48)

But if this analysis is correct, as I believe it is, it prompts the question: where did the many,
indeed dominant, elements of Hinduism that are non-Vedic come from? No doubt many local
religious cults were absorbed into and assimilated by Hinduism in the course of its millennial
development and expansion and a lot of ideological evolution took place, yet even so the Indus
civilization seems a likely original source, in spite of its great antiquity.
Regarding the Aryans, Renou cautiously continues: “It would have been quicker to
enumerate those elements [of Hinduism] that are demonstrably Aryan: they would consist of
perhaps a few functional gods (as it is the fashion to describe them), the soma cult and the
rudiments of a social system: little enough, in all conscience” (Renou 1953:47–48). Regarding
the Indus civilization, he suggests: “If the forms of religion revealed in the seals and figurines
of the Indus have any remote connection with Indian forms, it is not so much with those of
Vedism as with those of Hinduism, a Hinduism which, though known to us only by inference,
must have already existed in Vedic times, and probably considerably earlier” (Renou 1953:3).
Part I of this book (The Early Aryans) examines the evidence for the Aryan migrations and
the formation of the Vedic religion. Part II (The Indus Civilization) is devoted to the possible
legacy of the Indus civilization to the Vedas and to later Hinduism. The treatment is basically
in chronological order, but in some places it proved necessary to depart from this.
2

The Early Aryans

Both the Vedic literature and the fundamental texts of classical Hinduism are in Sanskrit,
which is termed an Old Indo-Aryan language. The Prakrit languages, classed as Middle Indo-
Aryan, were spoken in South Asia between about 300 BCE and 1000 CE. The Prakrits then
changed into early forms of the New, or Modern, Indo-Aryan languages, which include Hindi,
Punjabi, Bengali, Marathi, and many other South Asian languages spoken today. The name
Indo-Aryan was given in order to distinguish all of these languages from the Aryan languages
spoken outside South Asia. The latter consist mainly of the Iranian language family (the name
Irano-Aryan would have been more logical), which has been spoken in modern times
principally in Iran but also in the Caucasus area (the Ossetic language) and in Central Asia,
Afghanistan, and Pakistan (the Tajik, Pashto, and other languages). In addition there is a small
group of so-called Nuristani languages in northeastern Afghanistan (Nuristan), previously
known as Kafiri (“pagan”) languages; their speakers were forcibly converted to Islam by the
Afghans in 1896. The Nuristani languages have both Iranian and Indo-Aryan features, and
their classification is not yet settled.
Old Iranian languages include Old Persian, preserved in the inscriptions of Darius the
Great and other kings of the Achaemenid empire (559–330 BCE), and Avestan, which is the
language of the Avesta, the sacred literature of the Zoroastrians (who include the Parsees of
India). The oldest parts of the Avesta are thought to date from around 1000 BCE, being of
nearly the same age as the oldest Vedic text, the Rigveda. The Avesta and the Rigveda share a
number of religious terms, for example, the words for “worship, sacrifice” (yasna in Avestan,
yajña in Vedic) and “prayer, spell” (manthra in Avestan, mantra in Vedic), and the cultic phrase
“with hands raised up in homage” (ustānazastō . . . n m ŋhā in Avestan, uttānahasto namasā
in Vedic).
Speakers of Old Indo-Aryan and speakers of Old Iranian called themselves by the same
name, that is, ārya (Old Indo-Aryan) and arya (Old Iranian); the Old Iranian genitive plural
aryānām, “(country) of the Aryas,” in the course of time became shortened to the country
name Iran. The meaning of this tribal self-appellation is debated, but many philologists think it
originally stood for “hospitable, noble,” or for “master of the household, lord.” Together, Indo-
Aryan and Iranian form the Aryan group of languages, though this collective is nowadays
usually called Indo-Iranian (to avoid the connotation acquired by “Aryan” from its association
with the racial prejudices and misconceptions culminating in the horrors of Nazi Germany).
Indo-Aryan and Iranian are postulated to descend from a single original language, Proto-Indo-
Iranian, with dialects that drifted apart and became separate languages, which in turn divided
to give rise to the modern Indo-Iranian languages.
After Vasco da Gama reached India in 1498, he was followed by European missionaries,
some of whom acquired remarkable mastery of Indian languages. They included an English
Jesuit, Thomas Stephens (Tomáz Estêvão), who retold the story of the New Testament in
Konkani under the title Krista-purāṇa, covering more than 10,000 verses, and also wrote the
first Konkani grammar. Stephens noted the similarity of Sanskrit and European languages as
early as 1583; so did the French Jesuit Gaston-Laurent Coeurdoux, author of Moeurs et
coutumes des Indiens (pirated by Jean Antoine Dubois [{1825} 1906]), in a letter written in
1767 to A. H. Anquetil-Duperron, describing the relation of Sanskrit to Greek and Latin.
The idea that languages change, and that the ancestor of current related languages may
have died out, was expressed with reference to Sanskrit and its cognates by Sir William Jones
in his speech at the Asiatick Society in Calcutta on February 2, 1786. After studying Sanskrit
with the help of a Bengali paṇḍit, Jones announced that:

The Sanscrit language, whatever be its antiquity, is of a wonderful structure; more

perfect than the Greek, more copious than the Latin, and more exquisitely refined than
either, yet bearing to both of them a stronger affinity, both in the roots of verbs and in
the forms of grammar, than could possibly have been produced by accident; so strong
indeed, that no philologer could examine them all three, without believing them to have
sprung from some common source, which, perhaps, no longer exists; there is a similar
reason, though not quite so forcible, for supposing that both the Gothick and the
Celtick, though blended with a very different idiom, had the same origin with the
Sanscrit; and the old Persian might be added to the same family. (Jones 1788:422)

When published in the widely read journal Asiatick Researches, Jones’s suggestion had a
revolutionary impact. Some of the Romantics of the early nineteenth century imagined that
Sanskrit was the mother tongue of the “Indo-European” language family (a term coined not by
Jones but by another polymath, Thomas Young [1813:255]), and that India was the cradle of
civilization.
But soon Sanskrit was relegated to the same status as the other Indo-European tongues,
along with the area where it was spoken, India. Modern historical linguistics began with Franz
Bopp’s German publication, “On the conjugational system of the Sanskrit language compared
with that of Greek, Latin, Persian and Germanic” (Bopp 1816). The unknown homeland of
Proto-Indo-European was now thought to lie in an area of Eurasia yet to be precisely
determined. Thus the Indo-Aryans were immigrants in South Asia. But when did they come?
Friedrich Max Müller, who taught Sanskrit, comparative philology, and comparative
religion at Oxford University during the latter half of the nineteenth century, edited the
earliest of the four principal Vedas, the Rigveda. In 1859, Max Müller estimated that the
Rigveda was composed around 1200–1000 BCE. He arrived at this date by assigning the Sūtra
texts of the Vedas to 600–200 BCE, then conjecturing that the earlier Brāhmaṇa texts were
composed in 800–600 BCE, and the still earlier Mantra texts (the Atharvaveda and the
Yajurveda Samhitās) in 1000–800 BCE. Although these estimated dates have been much
discussed and criticized, they have turned out to be remarkably sagacious, and today many
Vedic scholars (including myself) more or less agree with them. Moreover, philologists have
long favored 1500–1200 BCE as the date for the migration of Indo-Aryan speakers into South
Asia. One powerful piece of evidence comes from cuneiform documents in Anatolia, West Asia,
and Egypt, according to which the rulers of the Mitanni kingdom of Syria in 1500–1300 BCE
had Indo-Aryan names, swore oaths by Indo-Aryan gods, and trained chariot horses using
Indo-Aryan technical terms (see chapter 8).
All the leading twentieth-century British archaeologists who have studied the Indus
civilization—John Marshall (1931), Stuart Piggott (1950), Mortimer Wheeler (1968), and
Bridget and Raymond Allchin (1982)—have placed the Aryan migrations in the second
millennium BCE. It has proved difficult, however, to pin down specific archaeological evidence
that could be connected with this event, for the Aryans were pastoralists and may therefore be
assumed to have been mobile, with temporary settlements, at least for some time after their
arrival. According to Renou, “In the very primitive architecture which we can infer from
descriptions in the Vedic ritual texts there is nothing that can reasonably be compared with
the buildings of Mohenjo-daro” (Renou 1953:4). Marshall (1931:I, 111) pointed out another
important negative indicator: the horse seems to have been unknown in the Indus civilization,
while prominent in the culture described in the Vedas.
The contrary “out of India” view, that the Indo-Aryan speakers originated in India, has
recently been much propagated, as mentioned in the Preface. Thomas Trautmann writes in his
edited collection The Aryan debate (2005):

In the last several years a number of popular books and websites on the Aryan debate
have appeared, many of them—but not all—by authors who are not scholars by their
training in the skills of ancient history. Moreover, partisan politics and governments of
the day have been making pronouncements about ancient history, and ordering
changes in textbooks. The new writings have pressed various versions of the alternative
view very strongly, arguing that the Aryans are indigenous to India and were the
builders of the Indus civilization. (Trautmann 2005:xvii)

The Aryan debate undoubtedly has multiple causes, including the racial “theories” of the
nineteenth and twentieth centuries endorsed by the Nazis. These have been thoroughly
researched by Edwin Bryant in The quest for the origins of Vedic culture: The Indo-Aryan
migration debate (2001). It is worrying, however, that Bryant does not want to take a stand on
the issue after studying the arguments in depth. He concludes: “how the cognate languages
got to be where they were in prehistory is as unresolved today, in my mind, as it was two
hundred years ago when William Jones announced the Sanskrit language connection to a
surprised Europe” (Bryant 2001:12). In my view, it is possible to go quite far toward
establishing the truth about the Indo-Aryan migrations.
3

Indo-European Linguistics

Since some of this book requires a basic understanding of the concepts and methods of Indo-
European (IE) linguistics, “iron rations” will be provided in this chapter. Those looking for a
fuller account are advised to read a good introductory textbook, such as Lyle Campbell’s
Historical linguistics (2004) or, more particularly, Benjamin W. Fortson’s Indo-European
language and culture (2010).
Since languages are complex phenomena, linguistics has developed various subdisciplines,
each with its own special terminology and methods. Here are some of the important
subdisciplines. Phonetics studies how sounds are produced and classifies their different
qualities. Phonology examines sounds as distinctive units (phonemes) and how phonemes
combine into a sound system. Phonotactics investigates how phonemes behave when
conjoined. Morphology is devoted to the grammatical markers and inflections of words.
Lexicography concerns the vocabulary of a language. Syntactics involves the study of how
words are combined to form phrases and sentences. Semantics grapples with how words
convey meaning. Descriptive linguistics records all aspects of a language as it exists at a given
point in its history.
All languages have a history: excepting the dead ones, they change constantly, in all their
aspects. Especially important is phonological change, which often involves changes in
morphology, too. Historical linguistics (unlike descriptive linguistics) documents how
languages change over time, and tries to reconstruct prehistoric changes. It also aims to
understand how and why linguistic change occurs, and to define how different languages are
related to each other. That is, are the languages part of one family, in other words genetically
related—in which case how closely related? If they are not genetically related, have they been
in contact with each other at some point in their histories? Or are they totally distinct? The
chief method of historical linguistics is therefore comparison of languages. IE linguistics is
essentially historical linguistics.
The ancient Indian linguist Pāṇini’s Sanskrit grammar (known as Aṣṭādhyāyī, “having eight
chapters”), composed around 350 BCE, is a marvel of descriptive linguistics, the finest
achievement in this field for more than two thousand years (Bloomfield 1933:11; Cardona
1976). Historical linguistics came into being only in the early nineteenth century CE. Before
this, as early as the sixteenth century, the similarity of IE languages had been noted, but it
was William Jones’s observation (1788:422) of the similarity of Sanskrit, Greek, and Latin,
among other European languages, that inaugurated modern historical linguistics (see chapter
2). Jones spelled out his new concept of a genetic relationship by suggesting that these
languages had developed from a single original language, which no longer existed—now called
Proto-Indo-European (PIE), from Greek prôtos, “foremost (in place, time, order), first.”
The works of Franz Bopp (1816) and Rasmus Rask (1818) started to establish systematic
phonological and grammatical correspondences between vocabulary words of similar
meaning, in order to prove a genetic relationship between the IE languages; János (in Latin,
Joannes) Sajnovics (1770) had initiated a similar study of the Uralic languages by comparing
Hungarian and Saami.
People without linguistic training may draw radically incorrect conclusions from apparent
similarities in sound and meaning between languages that in reality are unrelated and often
quite distant from each other. For instance, in the Mayan language Kaqchikel, spoken in
Central America, the word mes denotes “mess, disorder, garbage,” which happens to be
similar in sound and meaning to the English word mess. But there is no regular sound
correspondence between words beginning with m- in English and their semantic counterparts
in Kaqchikel for example, man: ači, mouse: č’oy, moon: qati t, mother: nan.
The English–Mayan situation contrasts with the IE languages, where there are regular
sound correspondences. For instance, IE words denoting “father” include the following
cognates (in the nominative singular):

Late PIE *p t r (< Early PIE *pH2t r)

Indo-Iranian: Vedic Sanskrit pit´ā, Old Avestan ptā, Younger Avestan pita, Old Persian pitā
Greek pat r
Latin pater
Celtic: Old Irish athair/athir
Germanic: Gothic fadar, Old Icelandic faðir/ faþer, Old English fæder, Anglo-Saxon fader
Tocharian A pācar, B pācer
Armenian hayr

In the above cognates, the word-initial PIE *p- (the asterisk denotes an unattested
reconstruction) has been preserved as p- in most languages, but in Celtic languages it has
been lost, while in Germanic the change *p- > f- has taken place, and in Armenian so has the
change *p- > h-. (The symbol > means “changes into”; the symbol < means “has developed
from.”)
The important concept of regularity may be illustrated by taking a different word with a
word-initial *p- in PIE, the verb meaning “to ask” (in the first person present singular active).
Here we again find exactly the same sound correspondences at the beginning of the cognates:

PIE *pṛḱsḱō
Indo-Iranian: Vedic Sanskrit pṛcchā´mi, Old Avestan p r sā, Old Persian parsāmiy
Latin: *porkskō > poscō
Celtic: Old Irish arco/arcu
Germanic: Old High German forscōm
Baltic: Lithuanian peršù
Tocharian A praksam, B preksau
Armenian harc’anem

Jakob Grimm (1822) noted that the change of PIE *p into Proto-Germanic *f was not an
isolated phenomenon but part of a major change in the entire sound system, involving all
labial, dental, and velar stops.
Stops are consonants articulated by first making a complete closure of the vocal tract,
stopping the flow of air from the lungs, then abruptly opening the tract with an audible
plosion. Stops can be voiced or voiceless, that is, articulated with or without an accompanying
vibration of the vocal cords. Labial stops are articulated with the lips, dental stops with the
teeth, and velar stops with the soft palate and the back of the tongue; palatals are pronounced
with the blade of the tongue touching the hard palate. Aspirated stops are pronounced with an
audible and forceful release of breath. Fricatives are continuant consonants produced by
partial occlusion of the airstream, causing a friction-like sound.
The PIE voiceless stops changed into corresponding voiceless fricatives in Proto-Germanic:
*p > *f, *t > *θ, *k and *ḱ > χ (written h). At the same time PIE voiced stops changed into
corresponding voiceless stops in Proto-Germanic: *b > *p, *d > *t, *g and *ǵ > *k, *gw > *kw.
And PIE voiced aspirated stops changed into corresponding plain voiced stops in Proto-
Germanic: *bh > *b, *dh > *d, *gh and *ǵh > *g, *gwh > *gw/*w. Before this, the PIE palatals
*ḱ, *ǵ, ǵh had merged with the PIE velar stops *k, *g, *gh. PIE *kw, *gw and *gwh were
labiovelars, that is, velar stops pronounced with a simultaneous rounding of the lips.
Taken together, these Proto-Germanic sound changes are known as the “first Germanic
sound shift,” or Grimm’s law. Such sound laws are expected to apply throughout a language,
in all words and grammatical markers, unless the sound law is “conditioned,” that is,
applicable under restricted conditions. Grimm’s law is conditioned, for it applies neither after
fricatives (thus PIE *sp > Proto-Germanic *sp, not sf) nor after stops (thus PIE *kt > Proto-
Germanic *χt, not χθ). Strict adherence to the rule that sound laws should have no exceptions
is an immensely important methodological principle. Emphasized by the so-called
neogrammarians during the last quarter of the nineteenth century, it led to great advances in
linguistic science.
There are always isolated cases where a sound law does not apply in practice, but even
then there is usually a valid explanation. A common reason for such exceptions is analogy.
Analogical change is of many different kinds, seeking agreement or similarity with existing
patterns or models. For example, the antiquated plural brethren may be replaced with
brothers that is analogous to “normal” plurals. PIE *penkwe- yields English five with Proto-
Germanic *f in accordance with Grimm’s law, but the preceding numeral PIE *kwetwor-
becomes four in English, with a possibly analogical f that imitates the beginning of the
following numeral five, instead of the expected whour. Latin quinque, “five” seems to involve a
similar analogical change (but in the reverse direction), in imitation of quattuor, “four.”
Besides analogy, there may be other reasons why there may be seeming exceptions to
sound laws. One important reason can be that an IE language has not inherited a given IE
word from its own earlier phase(s) of development, but borrowed the word from another IE
language which has different sound laws.
Historical linguistics has progressed over the past two centuries, and the mechanics of
language change are now fairly well understood, so that the same linguistic methodology can
be applied to all languages. Certain sound changes tend to take place in many languages
(even genetically unrelated ones). For example, dental, palatal, and velar stops often change
into affricates when they are followed by front vowels, that is, vowels articulated in the front
part of the mouth, i and e. Affricates are composite sounds consisting of a stop and a fricative
articulated at the same point. Thus Latin centum, “hundred” (with c pronounced as k), has
become cento in Italian (with c pronounced as an affricate, like ch in the English word chess).
The PIE palatal stops became affricated in early post-PIE times in Proto-Indo-Iranian,
Proto-Balto-Slavic, Proto-Armenian and Proto-Albanian—apparently then contiguous
languages spoken somewhere north of the Black Sea. These are called “satem-languages,” as
opposed to other IE “centum-languages.” The word satem (a broad transcription for sat m) is
Avestan for “hundred.” The label derives from the fact that sat m and centum are prominent
examples of this particular sound change.
The words for “hundred”—and for “ten,” from which “hundred” is derived—have also been
important in sorting out the development of PIE “syllabic nasals,” m- and n-sounds that
function like vowels when they occur alone between consonants; they have changed in
different ways in the different branches of the IE languages. In Proto-Indo-Iranian the syllabic
nasals changed into *a, as in most dialects of Greek, whereas the Arcadian and Lesbian
dialects of Greek have o instead of a. In the late phase of Proto-Indo-Aryan and Proto-Iranian,
PIE *e and *o changed into *a, but early Aryan loanwords in Uralic languages show that *e
and *o were still preserved in their early phases. The two words also exemplify the important
role that accent and shift of accent have played in the formation and inflection of IE words:
the vowel e in the accented first syllable of *déḱṃ is reduced and lost when the accent shifts
away from the first syllable to the second syllable in *(d)ḱṃtóm.
Here is a selection of words meaning “hundred” and “ten” in various IE languages:

*(d)ḱṃtóm, “hundred” (possibly “tenth [ten]”)

Old Irish cét [i.e., kēt]
Italic: Latin centum
Greek hekatón (he- < *sem, “one”)
Germanic: *χunda > Gothic hund, English hundred
Indo-Iranian: *ćatám [i.e., tśatam]
> Proto-Indo-Aryan *ćatám/*śatám
> Old Indo-Aryan śatám
> Proto-Iranian *catam [i.e., tsatam]
> Avestan sat m
> Old Persian θata
Baltic: Lithuanian šimtas, Latvian sìmts
Old Church Slavonic sŭto, Russian sto
Tocharian *ḱṇtóm > * käntæ > A känt, B kante
PIE *déḱṃ(t), ‘ten’
Old Irish: deich
Italic: Latin decem
Greek déka
Germanic: *teχun > Gothic taíhun, English ten
Indo-Iranian: *deća [i.e., detśa]
> (A) Proto-Indo-Aryan *deća/*daśa > Old Indo-Aryan daśa;
(B) Proto-Iranian *deca/*daca [i.e., datsa]
> (a) Proto-Nuristani daca [i.e., datsa] > Kati duc,
(b) Avestan dasa
Baltic: Old Prussian dessempts/dessimpts, Old Lithuanian dešimtis, Lithuanian dẽšimt
Old Church Slavonic desętĭ, Russian désyat’
Albanian dhjetë
Armenian tasn
Tocharian *tšäkä(n) > *śäkä(n) > A śäk, B śak
The hypothesis that all existing IE languages go back to a single PIE language once spoken
in a relatively restricted area is now a generally accepted idea. Helpfully, there is an example
in which this assumed process has been documented from the beginning. The Romance
languages have developed out of the Vulgar Latin spoken in the Roman empire in the areas
where nowadays Italian, French, Spanish, Portuguese, Romanian, and Rheto-Romanian are
spoken. We know the Romance protolanguage from copious literary sources: it is Latin, which
was originally spoken in a very restricted region of Latium in northern Italy but then spread
widely along with the expansion of the Roman empire. Latin was just one member of an
“Italic” branch of the IE language family, known from inscriptions since c. 600 BCE, which at
one time included the dead languages Oscan, Umbrian, and Faliscan (known from scanty
sources) but by the beginning of the Christian era consisted of only Latin.
Many other dead languages known from sources inadequate for proper linguistic analysis
have been identified as members of the IE language family, more or less securely. Best known
are Phrygian (c. 800 BCE – 400 CE) from Anatolia, Thracian and Illyrian from the Balkans, and
Venetic and Messapic from Italy. If we ignore these additional members, ten main IE branches
are generally acknowledged (Fig. 3.1). Listed by date of their first documentation (with the
oldest language coming first), they are:

(1) The Anatolian branch in modern Turkey, which became extinct by about 300 BCE. Its
principal members are Hittite (c. 1600–1200 BCE), Palaic (c. 1600–1500 BCE), and Luvian
(c. 1300–750 BCE), all documented in Mesopotamian cuneiform, while Luvian is also
written in its own hieroglyphic script. Luvian was probably the language spoken by the
Trojans of Homer’s Iliad. Lycian and Lydian are important later Anatolian languages,
known from alphabetic inscriptions of the fifth and fourth centuries BCE. (For Anatolian,
see chapter 6.)
(2) The Indo-Iranian or Aryan branch, with many languages now spoken in India, Iran, and
Central Asia, and in antiquity in the Eurasian steppes. Its scanty oldest testimonia are
cuneiform documents from the ancient West Asia related to the Proto-Indo-Aryans of the
Mitanni kingdom, dated c. 1500–1300 BCE (see chapter 8). The Old Indo-Aryan Rigveda
containing 1028 hymns dates from c. 1300–1000 BCE, and the most ancient part of the Old
Iranian Avesta from c. 1000 BCE (see chapter 9). However, loanwords borrowed from
Proto-Indo-Aryan and Proto-Iranian into Proto-Uralic and its immediate successors
between about 2300 and 1600 BCE represent the oldest known phase of Indo-Iranian (see
chapter 7), although they were recorded only in modern times.
(3) The Greek branch with numerous dialects. Its earliest inscriptions in the Mycenaean
Linear B script date from the thirteenth century BCE, Homeric literature from c. 800 BCE.
The ancient Greeks had colonies in Italy, Africa, and Asia, as far as India (see chapter 6).
(4) The Italic branch, discussed above.
(5) The Celtic branch, divided into two groups. Insular Celtic of the British Isles and
Brittany, attested in inscriptions since c. 400 CE, is the only one to survive today in Irish,
Welsh, Cornish, and Breton. Continental Celtic, attested since c. 600 BCE, was the most
widely spoken language in western and central Europe, ranging from Spain to France all
the way east to the Black Sea; the Galatians even reached Anatolia. The spread of the
Roman empire and Latin caused Continental Celtic to die out c. 300 CE.
(6) The Germanic branch, divided into three groups. North Germanic, currently represented
by Icelandic, Norwegian, Danish, and Swedish, is known from runic inscriptions dating
from c. 150 CE and extensive Old Norse literature from c. 900 CE. East Germanic is
synonymous with Gothic spoken in the area north of the Black Sea by Nordic immigrants.
In the fourth century CE, Bishop Wulfila translated part of the Bible into Gothic, which
died out in the sixteenth century. The most important current West Germanic languages
are English (recorded in inscriptions since the fifth century and in literature since the
eighth century CE), German with its dialects, as well as Flemish, Dutch, and Frisian.
(7) The Armenian branch, known since the fifth century from a translation of the Bible into
classical Armenian. This is now spoken mainly in Armenia in the Caucasus and in eastern
Turkey.
(8) The Tocharian branch, an extinct language known from two dialects (A and B) once
spoken in Xinjiang or Chinese Turkestan. It is recorded in Buddhist texts dating from c.
600–800 CE.
(9) The Balto-Slavic branch, divided into the Slavic (or Slavonic) and Baltic languages. A
translation of the Bible into Old Church Slavonic or Old Bulgarian from the ninth century
CE is the oldest record of Slavic languages, which comprise South Slavic (modern
Bulgarian, Slovenian, Serbo-Croatian, and Macedonian), West Slavic (modern Polish,
Czech, and Slovacian), and East Slavic (modern Ukrainian, Belo-Russian, and Russian).
The Baltic languages include the extinct Old Prussian (with some records from 1300–1600
CE), and modern Lithuanian and Latvian (with a literature since c. 1500 CE).
(10) The Albanian branch, spoken in Albania, Bosnia, and northern Italy. This is known in two
dialects and has been recorded since the fifteenth century CE.

FIGURE 3.1 Generalized distribution of the principal Indo-European languages c. 500 BCE (after Mallory and Adams
[1997:300]). Courtesy J. P. Mallory.

Scholars are unanimous that the Anatolian branch was the first to separate from the IE
protolanguage. It is the only branch to preserve the so-called laryngeal phonemes that have
been reconstructed for Early PIE. The other branches appear to have emerged from a Late
PIE linguistic community, which disintegrated suddenly, dispersing in different directions.
Many scholars think that Tocharian is the most archaic of the Late PIE languages.
4

The Indus Civilization

In 1924, Sir John Marshall, Director General of the Archaeological Survey of India, announced
to the general public the discovery of a previously unknown Bronze Age culture. He named it
the Indus civilization, because the finds came from two sites in the Upper and Lower Indus
Valley, Harappa (near Lahore in the Punjab) and Mohenjo-daro (in Sindh), 600 km apart.
However, an archaeological culture is often named after the site of its first discovery. In
this case, the site of Harappa had been described, by Alexander Cunningham, the first
Director of the Archaeological Survey of India, as early as 1875. So “Harappan culture” has
often been used, instead of “Indus civilization,” to embrace both the earlier, developmental
phase of the civilization and its mature phase, which is dated to c. 2600–1900 BCE. This
followed the realization by the archaeologist M. Rafique Mughal in 1970 that urbanization and
many concomitant technologies began before the Mature Harappan phase during the Early
Harappan phase dated to c. 3200–2600 BCE; the term previously used for this earlier period,
Pre-Harappan, did not convey the sense of close cultural continuity between the Early and
Mature Harappan phases. It now appears that during a short transition period, c. 2600–2500
BCE, various Early Harappan cultures were largely transformed into the relatively unified
Indus civilization.
Excavations at the sites of Mehrgarh, Nausharo, Sibri, and Pirak have provided an
unbroken cultural sequence from about 7000 to 1000 BCE (Jarrige et al. 1995). These places
were strategically located close to the Bolan Pass (connecting the Lower Indus Valley with the
highlands of Baluchistan) on the ancient route to Afghanistan, the Iranian plateau, and Central
Asia. The Mehrgarh cultural sequence has made it possible to understand the mutual
relationship of the various small local Neolithic and Copper Age cultures of the Indo-Iranian
borderlands, to follow their gradual evolution, and to sketch their migration to the Indus
Valley and beyond so as to become Early Harappan cultures.
Variant Early Harappan cultures existed in the valley around Quetta (Damb Sadaat
culture); in southern Afghanistan around Kandahar (Mundigak culture); in southern
Baluchistan and southern Sindh (the Kulli, Nal, and Amri cultures), in the eastern plains of the
Punjab, Haryana, and Uttar Pradesh (Ravi-Hakra and Sothi-Siswal cultures); and in nearly the
whole Indus Valley (the Kot Diji culture) (Possehl 1999) (Fig. 4.1). Some of these Early
Harappan cultures continued to exist after the formation of the Indus civilization: the Kulli
culture in southern Baluchistan, the Late Kot Diji culture in the northernmost Indus Valley and
the Anarta tradition of Gujarat (derived from the Amri-Nal cultures). The post-Harappan
Copper Age cultures of Rajasthan, the Deccan, and the Upper Gangetic Valley have a
predominantly Harappan ancestry (Fig. 4.4).
FIGURE 4.1 The Greater Indus Valley during the Early Harappan period, c. 3200–2600 BCE. After McIntosh 2008:
map 1. © ABC-CLIO Inc.

The Indus Valley, with its wide plains watered and fertilized by annual floods, permitted
cultivation on a larger scale than the mountain valleys of the borderlands. Yet the valley was
not colonized immediately, for its environmental conditions also posed difficulties. Habitations
had to survive the floods, which often required building on raised foundations. Villages started
spreading, tentatively, from the piedmont to the plains as early as 3800 BCE. With gradually
increasing social skills and technological advances came walled towns on a grid pattern in the
Early Harappan period. Where before, in the Neolithic and Copper Age villages, people had
stored grain in communal granaries, now they stored it in large storage jars kept in separate
houses. The novel concept of private ownership is also suggested by pot marks, and the
adoption of stamp seals with geometrical motifs for administration. Cattle pulled plows in the
fields: furrows excavated in the Early Harappan level at Kalibangan show the same pattern as
today, running in two directions at right angles to each other (nowadays sown with two
different seeds, horsegram and mustard). Transport was facilitated by ox-carts, still the most
important vehicle in the South Asian countryside; its numerous representations in clay-model
form were the Harappan counterpart of present-day model cars for children.
The Early Harappan cultures had overland trade contacts with southern Turkmenistan and
with the Iranian plateau, where the Proto-Elamite culture (c. 3200–2600 BCE) had spread
widely. Typically Proto-Elamite bevel-rimmed pottery comes even from Mir-i Qalat in Pakistani
Baluchistan (Besenval 2011). Contact with Proto-Elamite people may have given Early
Harappans the idea of creating a writing system, for the foundations of the Indus script were
laid in late Kot Diji times, although the script itself bears scarcely any resemblance to the
Proto-Elamite script.
The Early Harappans already possessed most of the basic cultural constituents of the
Indus civilization. In the Mature Harappan period these became more refined and were
applied on a grand scale. Early Harappan artifacts are clearly distinguishable from Mature
Harappan ones on the basis of form and style. The most characteristic features of the Indus
civilization include the following:
• the fully developed Indus script;
• finely carved stamp seals with writing and/or an animal or some other iconographic motif
mostly, usually made of soft soapstone hardened by heating;
• standardized measures, including cubic weights made of chert carefully cut and polished,
employing a combination of the binary and decimal systems, in the ratios 1/16, 1/8, 1/6, 1/4,
1/2, 1, 2, 4, 8, 16, … 800;
• the large-scale use of burnt brick, standardized in size, with the ratio 1:2:4 the most
effective for bonding;
• exquisite lapidary art, featuring highly developed microdrilling of very long beads made of
hard carnelian, decorated with chemically stained motifs.

We do not know exactly where the Indus civilization came into being. But it is clear that it
spread quickly over almost the entire Early Harappan area and even beyond (Fig. 4.2). The
transitional period (2600–2500 BCE) may not have been altogether peaceful, for several Early
Harappan towns were burned down and resettled in Mature Harappan fashion. Because
rather few bronze weapons have been found, the Indus civilization has been imagined as being
an exceptionally peaceful place, yet it seems unlikely that it could have dispensed with armed
forces to maintain law and order, and as a defense against outside attack. Bows and arrows
made of cane appear in Indus art but are hard to trace in the archaeological record. They
were the weapons of Indian soldiers in the fifth century BCE, according to Herodotus (7,65).

FIGURE 4.2 The Greater Indus Valley during the Indus civilization or Mature Harappan period, c. 2600–1900 BCE.
After McIntosh 2008: map 2. © ABC-CLIO Inc.

As during the Early Harappan period, the cities and their specialized craftsmen were fed
by farmers and pastoralists. Humped cattle, sheep, and goats were kept and wheat and barley
were cultivated from Neolithic times. In the Mature Harappan period the number of cultivars
increased, and two growing seasons were introduced where feasible: barley, wheat, oats,
lentils, beans, mustard, jujube, and linen were grown in winter; millets, cotton, sesamum,
melons, jute, hemp, grapes, and dates in summer during the monsoon.
Mature Harappan metallurgists used copper (alloyed to make bronze mostly with tin),
lead, silver, gold, and electrum. Lapidaries worked many kinds of stone. Analysis has shown
that the Indus people procured their minerals from diverse sources, often far away (Law
2011). A Mature Harappan settlement was established beyond the Hindu Kush mountains in
northeastern Afghanistan at Shortugai to control the mining of lapis lazuli, which was in great
demand in West Asia (Francfort 1989). A major innovation was water transport. Rivers could
be used for moving large and heavy loads long distances. Ships also transformed foreign
trade.
While virtually no object of clearly West Asiatic origin has been discovered in the Greater
Indus Valley, dozens of seals bearing the Indus script have been found in Mesopotamia and the
Persian Gulf, along with Harappan weights and jewelry. Cuneiform texts also speak of sea
trade with foreign countries, and there is nowadays wide agreement that the most distant of
these countries, Meluhha, was the Greater Indus Valley. There are also references to
Meluhhan people living in Mesopotamia, but unfortunately no record of their language, except
that it required an interpreter (chapter 17).
Mohenjo-daro, built by 2400 BCE, is one of the largest Indus settlements, with an area of at
least 100 hectares, possibly more, and a population of perhaps 50,000 people (Marshall 1931;
Mackay 1938; Franke-Vogt 1991 [1992]). The acropolis, or “citadel,” covering 20 hectares,
stands to the west of the residential “lower city” on a 12 m high mud-brick platform. The
purpose of its “public” buildings is largely guesswork: they include an “Assembly Hall”; a
“Granary” or “Warehouse” (once ventilated); a “Great Bath,” with a watertight pool measuring
7 m by 13 m surrounded by verandas, rooms, and a well; and a “College of Priests.” The
“lower city,” too, was artificially raised above the flood plain and surrounded by a wall. Many
houses had two stories and a central courtyard. The main street, 7.6 m wide, ran north–south.
The real wonder of Mohenjo-daro is its water-engineering (Jansen 1993). It is estimated to
have had 700 deep wells ingeniously built with tapering bricks to prevent them from
collapsing. Many houses had waterproof bathing platforms and privies. Waste water was
channeled away through pipes and chutes into covered drains that ran along level streets. The
cesspits of these drains were regularly emptied.
Another major site, the 100-hectare town of Dholavira in Kutch, had three large water
tanks to conserve drinking water (Fig. 4.3). Dholavira looks different from the brick-built
Mohenjo-daro, because here stone was available as a building material. The town has grand
walls and gates, one of which was found to carry a monumental inscription, the only example
of its kind so far discovered: each of its ten signs is about 37 cm high, made of pieces of
crystalline rock inlaid in a wooden plank three meters long.

FIGURE 4.3 The largest of several water reservoirs in the Indus town of Dholavira, Kutch, Gujarat. It is 73.4 m long,
29.3 m wide, and 10.6 m deep, with flights of thirty steps in three corners. Photo Asko Parpola 2008.

Many Indus towns had an acropolis: Allahdino, Banawali, Chanhu-daro, Dholavira,

Harappa, Kalibangan, Kot Diji, Lothal, Mohenjo-daro, Nausharo, Rakhigarhi, Surkotada, and
Sutkagen-dor. The purpose of these varying monumental structures is thought to have been to
impress people and thus reinforce political power. Some of them clearly had a protective
function, too (Eltsov 2008). Kanmer in Kutch is a small site (115 m by 105 m), but its walls are
11 m high by 12–15 m wide (Kharakwal, Rawat, & Osada 2012).
Around 2000–1900 BCE the Indus civilization came to an end. Multiple causes have been
adduced for the gradual decline and eventual downfall of urban life in the Indus Valley, which
resulted in the disappearance of the Indus script and seals. (The latest, definitely
recognizable, samples of the Indus script are from the southernmost site of Daimabad, not too
far from Mumbai, about 1800 BCE. Fig. 14.3) Stagnant water in the irrigation canals of the
civilization may have dissolved ground salts and rendered the fields uncultivable. Thick layers
of silt at Mohenjo-daro attest to catastrophic floods. Rivers, particularly the Indus and the
Ghaggar-Hakra, also changed their courses. These events may have been connected with
earthquakes. Recent palaeoclimatic research (Dixit, Hodell, & Petrie 2014) suggests an abrupt
weakening of the summer monsoon in the northwestern plains of the Indian subcontinent
about 2100 BCE, causing severe decline in summer overbank flooding that adversely affected
monsoon-supported agriculture. (This aridification seems to have been widespread; it
occurred in West Asia too.) All this coincided with the arrival of the first wave of Indo-Aryan
speakers in South Asia (chapter 8), but there is no evidence of a massacre of Indus Valley
populations by “Indra’s hordes,” as was once thought (Wheeler 1947; 1968); rather, the
skeletons lying unburied in the streets of Mohenjo-daro are now known to have belonged to
people who died from malaria and cholera, not armed assault.
The large Harappan population did not all of a sudden disappear. There is a clear
continuity between the relatively unified Indus civilization and the varied regional cultures of
the second millennium, that is, the Jhukar culture of Sindh and the Cemetery H culture of
Harappa, which is distributed in the Punjab and Cholistan. The heritage of the Early and
Mature Harappan cultures likewise continues in the Ochre Coloured Pottery (OCP) culture
associated with the copper hoards found in the Ganges–Yamuna Doab and in the Upper
Gangetic Valley. In Gujarat and Maharashtra the Indus civilization lingered until about 1800
BCE. Harappan traditions continued in the various “Chalcolithic” cultures of Gujarat,
Rajasthan, Madhya Pradesh, and Maharashtra, supplemented by new elements coming from
the west and the east (Fig. 4.4).
FIGURE 4.4 The Greater Indus Valley during the Late and Post-Harappan period, c. 1900–1300 BCE. After McIntosh
2008: map 4. © ABC-CLIO Inc.
5

The Indus Religion and the Indus Script

The Indus civilization was the fourth culture in the world to become literate. Around 3400 BCE,
the Sumerians of the Late Uruk culture in Mesopotamia created the very first writing system,
which inspired the Proto-Elamites of Iran and the Egyptians to devise their respective scripts
by about 3200 BCE. But whereas abundant textual material sheds light on the Mesopotamian
and Egyptian civilizations, only very short Harappan inscriptions survive, written in signs that
are extremely difficult to decipher in the absence of bilingual texts. The archaic Sumerian
script and the earliest Egyptian hieroglyphic are understood (on the basis of their later forms)
to a considerable extent, Proto-Elamite very poorly.
The seal inscriptions, and the tiny clay, faience, and stone tablets labeled as Indus
“amulets,” are written in pictograms, often in conjunction with some iconographic motif,
usually depicting an animal, but also sacred trees and anthropomorphic figures likely to be
deities. In addition there are small statuettes and figurines of human beings and animals and
what have been described as either gaming pieces or phalli. Furthermore, as has already been
remarked, among the most noteworthy public structures at Mohenjo-daro are a remarkably
advanced water drainage system and the so-called Great Bath, suggesting that the Indus
people were preoccupied with ritual bathing.
These discoveries together persuaded Marshall, the leader of the first major excavations of
the Indus civilization, to posit a connection with Hinduism. In a paper published in 1931,
Marshall even identified as “Proto-Śiva” a human-like figure depicted on a seal as three-faced,
seated on a throne in what appears to be a yogic posture, and surrounded by animals (fig.
16.6). Marshall’s paper (Marshall 1931a), the most quoted study of the Indus religion, has
been accepted by many scholars in its general conclusions, despite criticism of its details,
including his identification of “Proto-Śiva.” However, some scholars remain doubtful about any
connection between the Indus civilization and later Indian religions. “We should not impose
the meaning of the later icons upon the earlier images,” writes a skeptical Wendy Doniger in
one of the best current assessments of the Indus religion. Yet even Doniger admits:

once we have explored the meaning of the Indus representations within the context of
their own limited context, we can go on to speculate how they may have contributed to
the evolution of a later iconography that they sometimes superficially resemble. For the
resemblances between some aspects of the IVC [Indus Valley Civilization] and later
Hinduism are simply too stunning to ignore. (Doniger 2009:82–83)

I of course agree with this, while accepting the need for caution in my speculations. In the
second half of this book, I attempt to elucidate the roots of Hinduism in the Indus civilization
with novel interpretations of Indus iconography—in the seals, tablets, painted pottery, and
other objects, such as the well-known statuette of the “priest-king”—based not only on later
Indian parallels but also on the contemporaneous evidence from Mesopotamia and elsewhere
in West Asia. The examination of the Indus iconography occupies most of part II of the book,
where I suggest that the Western Asian parallels yield unprecedented insights into the
Harappan religion. This evidence is largely independent of the results achieved in the study of
the Indus script, summarized in chapter 21 before the conclusion.
But first we must ask whether the Indus script was actually writing, in the sense of a
system for visually recording spoken language, that wonderful means of communication
capable of expressing the full range of thought. The earliest writing systems started as fairly
primitive “proto-writing,” which normally did not record the grammatical forms of words, and
mostly operated with “logographic” pictograms (symbols that in principle could be read as
words in any language); yet all of the systems used punning to express words or syllables
phonetically. They were thus at least partially language-based, although they required
centuries to evolve a writing system sufficiently sophisticated to express all parts of a spoken
language. There are also nonlinguistic sign systems, such as mathematical or musical
notation, which can be understood without reference to language, but they can express only a
limited range of thought.
This issue of whether the Indus script is writing has been hotly debated after a team
consisting of a historian, Steve Farmer, a computer linguist, Richard Sproat, and an Indologist,
Michael Witzel, claimed that the Indus signs do not constitute a language-based writing
system, but are merely nonlinguistic symbols of political and religious significance (Farmer,
Sproat, & Witzel 2004; see also Lawler 2004). The authors admit that the Indus people were in
contact with Mesopotamia and were thus acquainted with its cuneiform script; but they
suggest that the Indus people made a deliberate decision not to adopt a writing system,
probably in order to keep their esoteric knowledge secret. However, the existence of esoteric
or religious knowledge—which was transmitted orally—was no obstacle to the adoption of
writing by Indian people from Persia for administrative and economic purposes during the
latter half of the first millennium BCE.
As to their other arguments, plain statistical tests, such as the distribution of sign
frequencies, can prove neither that the signs represent writing nor that they do not represent
writing, as Sproat admits. The claim that there is no sign repetition within any one Indus
“inscription” (beyond reduplications, which they exclude) is simply false—there are many
cases of sign repetition. What about the authors’ principal claim, that all real writing systems
have produced longer texts than found in the Indus script? The average length of an Indus
inscription is five signs. The shortest texts have just one sign; the longest has twenty-six signs
divided on three sides of a tablet; the longest single-side text has fourteen signs divided into
three lines (Fig. 21.3a); and the longest single-line texts have fourteen signs. (The trio promise
to pay a large sum of money to anyone who finds a genuine Indus text with more than fifty
signs.) This thesis stumbles on the fact that in the earliest phase of Egyptian hieroglyphic
writing, texts consisting of one or two signs can be read phonetically and meaningfully on the
basis of later Egyptian hieroglyphic writing.
It appears that a phoneticized writing system was in use in Egypt in Pre-Dynastic times as
early as 3200 BC, but it was not until the beginning of the Third Dynasty of the Old Kingdom,
some 500 years later, that the Egyptians started to write complete sentences and longer
documents. “Many inscribed artifacts are preserved from the first two Dynasties, the most
numerous categories being cylinder seals and sealings, cursive annotations on pottery, and
tags originally attached to tomb equipment, especially of the First Dynasty kings. Continuous
language was still not recorded,” writes the Egyptologist John Baines (1999:883). Fuller
administrative documents are assumed to have existed in this early period but have not
survived. It is perfectly possible that the same could have happened in the Indus civilization.
This also takes care of the argument of Farmer and his colleagues that any Indus texts
assumed to be lost never actually existed.
During the early half of the third millennium BCE, systems of “potter’s marks” existed at
several locations in the Indus Valley and the Iranian Plateau (Quivron 1980; Vidale 2007).
Each used between about thirty and seventy different incised or painted symbols, which were
probably nonlinguistic, as they never form longer sequences. Some of the Early Harappan
“potter’s marks” excavated at Harappa resemble some of the approximately 400 standardized
signs of the Indus script. Farmer and colleagues do not take into account the fact that many
groups of three Indus signs recur in the same sequence having the very same order at several
different sites (Fig. 5.1), which is hard to explain unless they represent language rather than
decorative designs. By contrast, the graffiti on pots excavated at the Megalithic site of Sanur
in southern India are nonlinguistic symbols, for three different signs can be seen occurring in
the following sequences: 123 three times, 132 twice, 231 four times, 312 four times, 213 once
(Parpola 2008c). In addition, there are Indus inscriptions where the signs are squeezed
together to fit into the end of a line, implying that the signs were written sequentially in one
direction. “If the sequence matters, it’s a language,” writes Doniger (2009:69).
FIGURE 5.1 A sample page from the concordance to Indus inscriptions by K. Koskenniemi & A. Parpola 1982.

The Indus inscriptions are more than 1500 years older than the earliest presently readable
texts of the subcontinent, the stone inscriptions of the Emperor Asoka, written in the third
century BCE in the Kharoṣṭhī and Brāhmī scripts. The potential of the Indus inscriptions to
shed new light on the dark prehistory of South Asia has attracted more than 100 published
claims of decipherment since the 1920s, none of which has been widely accepted.
The most formidable obstacle to decipherment is the absence of any bilingual inscription
written in Indus signs and translated into a known script. Such translations have usually
provided the key to unknown scripts, most famously in the case of the Rosetta Stone, where
its Greek alphabetic inscription provided the key to its Egyptian hieroglyphic equivalent.
Although no bilingual came to the aid of the decipherers of the many varieties of the
cuneiform script, the names and genealogies of Persian kings, known from Herodotus and the
Bible, provided a clue to Old Persian cuneiform. But this kind of information, too, is lacking for
the Indus civilization, which collapsed before historically documented times.
Attempts to decipher the script started even before the ruins of Harappa and Mohenjo-
daro were investigated and the Indus civilization was recognized in 1924. As early as 1875, Sir
Alexander Cunningham reported the first known Indus seal, which was collected from the site
of Harappa. He assumed that the seal was a foreign import because the bull depicted on the
seal did not have the hump of the Indian zebu. A few years later, he suggested that the unique
seal might bear signs of the Brāhmī script in a hitherto unknown early form.
After Cunningham, many other scholars connected the Indus script with the Brāhmī script,
which is the ancestor of no fewer than about 200 South Asian and Southeast Asian scripts. To
take one example, G. R. Hunter studied the Indus inscriptions at first hand in the 1920s and
analyzed them structurally in a valuable doctoral dissertation (published as Hunter 1934),
where he compared the script with other early writing systems. However, it has long been
quite clear that the Kharoṣṭhī script is based on the Aramaic variant of the Semitic
consonantal alphabet, which was brought to the Indus Valley by the bureaucracy of the
Persian empire around 500 BCE, while the Brāhmī script may have been influenced in addition
by the Greek script, which came to India with Alexander the Great in 326 BCE (Salomon 1996;
Falk 1993). These two scripts, Kharoṣṭhī and Brāhmī, thus have nothing to do with the Indus
script, the last traces of which date from around 1800 BCE.
Nonetheless, to take a second example of a published decipherment, the archaeologist S.
R. Rao tried to assert that the Indus script is the basis of not only of the Brāhmī characters but
also of the Semitic alphabet. In The decipherment of the Indus script (1982) and elsewhere
(Rao 1991), Rao proposes that the Indus script developed thirty-four basic cursive signs,
reduced to twenty-four in the Late Harappan period. He then finds somewhat similar Semitic
parallels to seventeen of these basic cursive signs and so reads these seventeen Indus signs
with their “equivalent” Semitic phonetic values. Other, pictographic, Indus signs, however,
Rao reads phonetically on the basis of the Sanskrit names for the pictograms, but rather
arbitrarily: the “man” pictogram is read as ṛ from Sanskrit nṛ- “man,” whereas some other
signs are read acrophonically. It requires a lot of trust to believe that Rao’s proposed readings
make sense in an Aryan language close to Vedic Sanskrit, as he claims. The consensus view is
that the Semitic alphabet was created in the mid-second millennium BCE on the basis of the
uniconsonantal signs of the neighboring Egyptian hieroglyphic script.
Others have looked to the Mesopotamian scripts for clues to the Indus script. Immediately
after the discovery of the Indus civilization was announced, the British Assyriologists A. H.
Sayce (1924) and C. J. Gadd and Sidney Smith (1924) pointed to its resemblance to the
Elamite and Mesopotamian civilizations. They compared the Indus signs with the pictograms
of the Proto-Elamite and archaic Sumerian scripts. The resemblance proved to be superficial,
yet in 1925, the amateur Tibetologist L. A. Waddell published The Indo-Sumerian seals
deciphered, with a subtitle summarizing a radical revision of the early history of mankind:
Discovering Sumerians of the Indus Valley as Phoenicians, Barats, Goths and famous Vedic
Aryans, 3100–2300b.c. Waddell simply read the Indus signs with the phonetic values of
somewhat similar-looking Sumerian signs and identified the inscriptions as “revised forms” of
the names of Sumerian and Aryan kings.
Waddell’s is an example of the most common approach used in would-be decipherments.
But such sign comparison works only when the scripts involved are closely related, and even
then there are pitfalls. For example, the Roman script is closely related to the Greek script but
the Greek P has the phonetic value of the Roman R. The Indus script has no obvious genetic
affinity with any other known script. It is true that some Indus signs have close formal
parallels in many other ancient scripts, for example, the sign looking like a mountain
resembles signs in the Sumerian, Egyptian, Hittite, and Chinese scripts. But each of these
signs has a different phonetic value, even if their meaning may be the same or similar.
Perhaps the strangest Indus “decipherment” was that of a Hungarian engineer, Vilmos
Hevesy (Hévesy 1933). In 1932, he suggested a connection between the Indus inscriptions and
the rongorongo tablets of Easter Island in the Pacific Ocean. The two scripts are separated by
more than 20,000 kilometers and some 3500 years, yet Hevesy’s comparative lists of similar-
looking Indus and rongorongo signs have been taken seriously by some scholars. It is,
however, a particularly useless speculation, given that the rongorongo tablets are to a large
extent undeciphered.
In my view, the most productive line of enquiry has proved to focus on identifying the
underlying language of the Indus script. As far back as 1924, Marshall suggested that the
Indus people may have spoken a Dravidian language, because the Indus civilization is pre-
Aryan and a Dravidian language, Brahui, is even today represented in the Indus Valley and
Baluchistan by a substantial number of Brahui speakers. Piero Meriggi, later an acknowledged
authority on the Hittite and Proto-Elamite scripts, agreed with Marshall about the likelihood of
this Dravidian linguistic affinity (Meriggi 1934), but refrained from a phonetic decipherment
because he thought that reconstruction of the ancient Indus language was impossible. The
Dravidian hypothesis was the basis of Father Henry Heras’s ambitious attempt to decipher the
Indus script, culminating in a large book in 1953. Heras was, I think, right in his readings of a
couple of signs, but these were fatally undermined by a mass of nonsensical interpretations.
My own approach, launched in 1964 with Seppo Koskenniemi and Simo Parpola, was also
based on the Dravidian hypothesis (see the preface and Parpola et al. 1969ab; 1970). The
“rival” Soviet team agreed that the underlying language was probably Dravidian. The team
was led by Yuri Knorozov, who had initiated the successful decipherment of the Mayan script
in 1952, and it included a Dravidian specialist, Nikita Gurov (Knorozov et al 1968). (I
appreciate Gurov’s work—see chapters 14 and Chapter 24—but doubt the Soviet team’s
claims published in their final report [Knorozov, Albedil, & Volchok 1981], which confidently
reads all of the Indus signs in Dravidian.) In India, Iravatham Mahadevan, who has done
remarkable work in the field of Old Tamil epigraphy (Mahadevan 2003), has also published
several papers offering Dravidian readings for Indus signs. Mahadevan started working on
Indus material in 1971, and brought out his useful Indus computer corpus and concordance in
1977.
John Marshall underlined that success in study of the Indus script would require a many-
sided professional approach. Effective interpretation “can be accomplished, now or in the
future, by specialists conversant with the subject in all its bearings. I cannot refrain from
stressing this point here, because the antiquities from Mohenjo-daro and Harappa … have
been made the subject of much nonsensical writing, which can be nothing but a hindrance in
the way of useful research” (Marshall 1931:I,ix). I believe Marshall would have extended the
same advice to all those involved in the Aryan debate, if this had raged in his time as it does
today.
PART I
The Early Aryans
6

Proto-Indo-European Homelands

In searching for the origins of the Indo-Iranian languages we naturally want to identify the
homeland of the Proto-Indo-European speakers. The vocabulary common to the different Indo-
European languages permits some conclusions about the natural and cultural environment of
the PIE language. More than a century ago, Theodor Benfey, one of the leading Indo-
Europeanists of his time, proposed that the PIE homeland was likely to have been in southern
Russia (Benfey 1869:597–600). Otto Schrader, author of the still valuable encyclopedia of IE
antiquity (2nd ed. in two volumes, revised by A. Nehring, 1917–1929) and who wrote a major
work on comparative linguistics and prehistory (3rd ed. 1907), came to the conclusion that the
Pontic steppes north of the Black Sea provided the most likely solution to the PIE homeland
problem. Schrader noted that this location could explain the differences in the agricultural
and environmental vocabulary of the IE languages spoken in Europe and Asia. Placing the
homeland between the areas where the Uralic and Semitic languages are spoken also fits the
assumed prehistoric contact of these language families with PIE. One of the most esteemed
archaeologists of the first half of the twentieth century, V. Gordon Childe, opted for the same
solution in his book The Aryans: A study of Indo-European origins (1926).
Marija Gimbutas correlated the PIE language with what she called the “Kurgan culture” of
the south Russian steppes, and traced its impact on the archaeological cultures of Europe on a
broad scale (Gimbutas 1997). Gimbutas’s work was insightful, but her extensive and often
pioneering syntheses were bound to be somewhat controversial. Her studies have been
conservatively revised and carried further by her student J. P. Mallory. In his classic book In
search of the Indo-Europeans (1989) and numerous other studies, Mallory has in many ways
established—in my opinion convincingly—the thesis that the Early PIE homeland indeed was
in the Pontic-Caspian steppes in southern Ukraine and southern Russia. In recent decades,
Valentin Dergachev (1998; 2007) and David Anthony (1995; 2007) have contributed important
support to this view, which is most widely accepted by both archaeologists and linguists.
The Pontic-Caspian steppes account for more than 40 percent of the wild-horse remains in
Europe from the period c. 5000 BCE. The people of the Late Neolithic and Copper Age cultures
of the Pontic-Caspian steppes not only hunted and ate the horse (at one site in the Mid-Volga
River region 66 percent of the 3,602 identified bones were from horses), but also sacrificed it
(cultic deposits of horse heads and hooves have been found) and made bone-plaque horse
figurines. Not surprisingly, Early PIE had a word for horse, *éḱwos. Moreover, the horse is the
only animal to figure prominently in the personal names of several peoples speaking early IE
languages, such as Old Indo-Aryan, Old Iranian, ancient Greek, Gaulish Celtic, and Old
English, who in addition worshipped deities associated with the horse and offered horse
sacrifices.
The horse was probably at first kept for meat and milk and only later used for traction.
Many archaeozoologists consider the evidence for horse domestication uncertain before about
2500 BCE. I also doubt that the steppe pastoralists of the Copper Age rode domesticated
horses in order to control their herds and make raids as early as 4100 BCE, as claimed by some
scholars, notably Anthony (2007). If they had done so, then surely we would see pictures of
horse-riders from this time? In fact, the first picture of a horse with a human rider dates from
c. 2040 BCE and comes from Ur III dynasty Mesopotamia (Oates 2003), while those from the
steppes date from later still, c. 1500 BCE (Figs. 9.4, 9.5), after which horse-rider images
become increasingly prominent, especially in the cultures of mounted pastoralists, such as the
Scythians. The handles of the Sejma-Turbino knives of the Asiatic steppes and European
forest-steppe in the early second millennium still depict a single plain horse or a pair of horses
(Fig. 7.8), except in one case. In the Eurasian steppes, a man is for the first time represented
with a horse in the handle of a knife from the Rostovka cemetery. But this man is neither
riding the horse, nor, as Anthony (2007:446) prefers, “roping” it; he is actually leaning
backwards while standing on skis, using the horse for traction on snow (Fig. 7.10). Could
horse-riding really have existed to any significant extent for two thousand years without being
depicted? By contrast, the first wagons were depicted almost immediately after the invention
of wheeled vehicles. They were so heavy—a wagon with four solid wheels could weigh 800 kg
—that they could be pulled only by oxen, which are often depicted. Horses were harnessed to
pull light, two-wheeled, chariots, but this did not happen before c. 2100 BCE (see chapter 7). It
also seems significant that words for “riding” and “rider” cannot be reconstructed for PIE, nor
even for Proto-Indo-Iranian.
When people of the Late Neolithic culture called Khvalynsk (c. 5000–3800 BCE) of the
Volga steppes moved westwards and took control of the Dnieper-Donets culture, they fused
with this earlier local population to form the Copper Age culture called Srednij Stog II (c.
4700–3400 BCE) in the steppes between the Don and Dnieper rivers north of the Black Sea.
The Khvalynsk and Srednij Stog II peoples were pastoralists, who traded with the thriving
Carpatho-Balkan cultures of agriculturalists. Together all these cultures formed the world’s
first great “metallurgical domain,” where production (mining and working) was in the hands
of agriculturalists (Chernykh 1992; 2007). The pastoralists were the receiving partners. Their
closest neighbors, just beyond the steppes, were the farmers of the Cucuteni–Tripolye culture
(known as the Cucuteni culture in Romania and Moldavia, the Tripolye culture in Ukraine),
hereafter simply termed the Tripolye culture from which the eastern pastoralists received
their metal and other prestige goods (Fig. 6.1). The Tripolye culture is divided into three
major phases, A, B, and C, spanning c. 5700–2900 BCE, but the last sub-phase C2 (c. 3400–
2900 BCE) is nowadays considered Post-Tripolye.

FIGURE 6.1 The “Carpatho-Balkan Metallurgical Province” of the fifth millennium BCE, according to E. N. Chernykh
(1992:49, fig. 15). Courtesy E. N. Chernykh and Cambridge University Press. The steppe area (zone V)—here
supposed to be the Early PIE homeland—was just importing metal from the west, mainly from the Cucuteni–Tripolye
culture (the northern half of zone IV), which is here supposed to have been the Late PIE homeland c. 4200–3400
BCE.

An early phase of the Srednij Stog II culture (represented by the sites of Stril’cha Skelya,
Chapli, Mariupol’, Novodanilovka, and Suvorovo) was named the Skelya culture (c. 4300–4000
BCE) by Yuri Rassamakin (1999; 2004). During the last quarter of the fifth millennium, Skelya
pastoralists invaded the Balkans. Widely distributed Skelya-type burials with characteristic
Skelya grave goods are newly introduced cultural items in this area. Some 600 sites of the
Carpatho-Balkan cultures were abandoned, many after burning down, and their voluminous
metal production stopped almost completely.
The surviving local cultures of the eastern Balkans were transformed into the Cernavoda
culture (c. 4000–3200 BCE) with notable new cultural components brought from the steppe in
its economy, religion and social hierarchy. It is likely to have spoken the Early PIE of the
Skelya invaders. About a millennium later, the Cernavoda culture fused with steppe invaders
from the Yamnaya culture to form the Ezero culture (c. 3300–2700 BCE). This latter culture has
ceramics supposed to be ancestral to the pottery of the Troy I culture (c. 3000–2600 BCE) of
Anatolia. The most widely accepted date for the intrusion of IE speakers from Europe into
Anatolia is c. 2700 BCE. Then, in western and southern Anatolia, more than 100 Early Bronze
Age II (c. 2600–2250 BCE) sites were abandoned and all major sites destroyed, while Troy V-
type ceramics (c. 2000–1800 BCE) spread eastwards along with the megaron-type architecture
that characterizes early Greece.
This succession of archaeological cultures provides a credible trail for migration from the
north Pontic steppes to Anatolia. Indo-Europeanists are agreed that the first branch to
separate from PIE was the Anatolian one, comprising Hittite, Luvian, and Palaic, which has
been recorded since the sixteenth century BCE (as noted in chapter 3, all of these IE languages
of Anatolia died out by the end of the first millennium BCE).
Early PIE was in all probability spoken in the Khvalynsk and Srednij Stog II cultures.
However, I do not agree with Mallory and Anthony when they both locate the Late-PIE-
speaking homeland in the same place, that is, the Yamnaya cultures (c. 3300–2500 BCE, Fig.
6.6) that succeeded the Khvalynsk and Srednij Stog II. Instead, I suggest correlating Late-PIE-
speaking with the people of the Late Tripolye culture. By “Late Tripolye,” I mean phases B2
and C1 of the Tripolye culture, dated from c. 4100–3400 BCE. Late PIE disintegrated with the
Tripolye culture c. 3400 BCE.
At the same time as the Skelya pastoralists invaded the Balkans, they also subdued their
western neighbors, the Tripolye people, but without destroying their culture. In this Tripolye
B1 phase (c. 4300–4000 BCE), a large number of Tripolye settlements were fortified, which
suggests that they were under attack. At the Tripolye B1 site of Drutsy 1, more than 100
steppe-type flint arrowheads were found around the walls of the three excavated houses.
Drutsy 1 also had “Cucuteni C-type” ceramics decorated with cord impressions; these are
identical with the Skelya ceramics of the steppe, and differ in every respect from the
traditional beautiful Tripolye painted pottery. This intrusive ware appeared at first
sporadically in a number of settlements, then with increasing frequency until it became the
predominant ceramic in the final phase of the Tripolye culture.
Some of the earliest Tripolye sites to yield this steppe-type pottery also produced knobbed
maces (a few of them in the shape of a horse’s head), which spread from the steppes to the
Balkans and the Tripolye area in 4300–4000 BCE (Fig. 6.2). These weapons were symbols of
power, very probably reflecting the spread of a new type of social organization—strongly
hierarchical chieftainship—instrumental in the spread of Early IE. Effective new leadership
seems to have invigorated the Tripolye culture, which expanded northwards and eastwards,
creating ever-larger towns, some covering hundreds of hectares. There was now internal strife
between the rival eastern and western parts of the Tripolye culture. Perhaps this was the age
of warring heroes described in IE epic poetry with poetic formulae, such as Greek kléos
áphthiton (Iliad 9,413) and Sanskrit śrávo . . . ákṣitam (RV 1,9,7), both of which mean
“imperishable fame” and derive from PIE *ḱlewes *ṇdhgwhitom (Schmitt 1967; Watkins 1995;
West 2007).

FIGURE 6.2 Skelya-type scepters with stone mace-heads sometimes shaped like a horse-head, c. 4300–4000 BCE.
These come from the following sites: (1) Casimcea; (2) Terekli-Mekteb; (3) Salcuta; (4) Suvorovo (this one is 17 cm
long); (5) Fedeleseni; (6) Rzhevo. After Mallory 1989:234, fig. 129. Courtesy J. P. Mallory.

The most important means of relating archaeological cultures with the speaking of Late
PIE is the PIE vocabulary associated with wheeled vehicles. On the one hand, these technical
terms imply that Late PIE-speakers used wheeled vehicles, on the other, they date the
disintegration of Late PIE to a period after the invention of wheeled vehicles—a date that can
in principle be deduced from the archaeological record. That Late PIE had at least five terms
associated with wheeled vehicles has been very widely agreed for more than a century,
although in my view as many as twelve terms are plausible. The five more-or-less certain
terms are now discussed to show the extent and solidity of the evidence for Late PIE-speaking
use of wheeled vehicles and the origin of these words.
(1) PIE *kwé-kwl-o-/*kwekwl-á, “wheel” is the origin of Greek kúklos, Tocharian A kukäl / B
kokale, “wagon,” Sanskrit cakrá-, Old Norse hvél, hjól, Old English hwēol. There is even a
PIE poetic phrase that speaks of “the wheel of the sun”: Vedic sūryasya cakrá-, Old Norse
sunnu . . . hvél, Greek hēlíou kúklos. This word is derived by intensifying reduplication
from the PIE root *kwel-, “to turn (around), move in a circle.” PIE *kwol-o-/*kwel-es-
derived from the unreduplicated root is represented by Old Irish cul, “wagon”; Latin
colus, “distaff, the rod on which wool is wound before spinning”; Old Norse hvel (with a
short e), “wheel”; Old Church Slavonic kolo, “wheel” (pl. kola, “wagon”); Greek pólos,
“axle.”
(2) The PIE verbal root *weǵh- is the basis of PIE nominal derivatives denoting “vehicle,
wagon” (the number of which attests to the importance of wagon building and wagon
traffic in Late PIE and early post-PIE-speaking times):
(a) PIE *wóǵh-o-, “wagon”: Greek ókhos, Old Church Slavonic voz; Sanskrit vāhá,
“vehicle, draught animal”;
(b) PIE *wóǵh-a-: Mycenaean Greek wo-ka- (i.e., wokha-), “two-wheeled wagon”;
(c) PIE *wéǵh-os (n.), “wagon”: Greek (Homer, Pindar) ókhos (n., pl. ókhea); Sanskrit
váhas- (n.), “shoulder of a draught animal”;
(d) PIE *weǵh-no-, “wagon”: Old Irish fēn, Welsh gwain;
(e) PIE *woǵh-no-, “wagon”: Proto-Germanic *wagnaz > Old Norse vagn, etc.; cf. Sanskrit
vāhana-, “vehicle, draught animal”;
(f) PIE *weǵh-e-tlom, “device for transport”: Latin vehiculum, Sanskrit vahitram, “ship,
vehicle”; cf. Greek (Hesychius) ókhetla, “carriages.”

The verbal root *weǵh- is securely reconstructed to mean “convey in a vehicle, drive”: Luwian
Hieroglyphic wa-za-, “drive”; Sanskrit váhati, “conveys”; Avestan vazaiti, “conveys”; Greek
ékhō, “I bring” (Pamphylian Greek wekhétō), ókheomai, “I convey”; Latin vehō, “I convey,
bring”; Old Norse vega, “to move”; Lithuanian vežù, “I convey”; Russian vezu, “I convey.”
Before the invention of wheeled vehicles, this native PIE root is likely to have meant
mainly “moving a heavy load” by means of a lever, litter, travois, sledge, or draught animal.
For this older meaning (which agrees with the generally accepted hypothesis of how the
wheeled vehicles were invented) we have the following attestations: Proto-Germanic *wegan,
“move”; Old Norse v g, “lever,” pl. vagar, “sledge,” v gur, vāgir, “litter, travois,” vagn,
“wagon, sledge, litter,” vāg, “lever, weighing machine, weight”; Old Prussian wessis,
Lithuanian vãžis, važỹs, Latvian važus, važas, “sledge”; Greek okhleús, “lever,” okhlízō, “to
move by a lever, heave up.”

(3) PIE *aḱs-, “axle” survives with various extensions in different branches:
(a) PIE *aḱs-o- in Vedic Sanskrit ákṣa-;
(b) PIE *aḱs-i- in Latin axis, Lithuanian ašìs, Old Church Slavonic osь;
(c) PIE *aḱs-on- in Greek áksōn (gen. áksonos), Mycenaean Greek a-ko-so-ne;
(d) PIE *aḱs-a in West Germanic *ahsō, Old High German ahsa, German Achse, Old
English eax; English axle from Old Norse xull from *ahsula; also Greek hámaksa,
“cart with a single axle,” from PIE *sm-aḱs-a.

PIE *aḱs-, “axle,” is probably an *-es-derivative from PIE *aǵ-, “to drive, move, lead” (Sanskrit
aj-, ájati, Latin agō, Greek ágō, Armenian acem, Tokharian āk-; intransitive in Old Norse aka,
“to drive,” Welsh af, “to go”). Lithuanian ašìs denotes, besides “axle,” also “fathom” =
“outstretched arms,” while Avestan aša- (which formally corresponds to Sanskrit ákṣa-, “axle”)
means “shoulder,” like the Old Norse xl, German Achsel. The axle enables the movement of
wheels as the shoulders enable movement of the arms and wings. Thus the axle is what makes
the wagon “move” (*aǵ-). The term could also denote the axle as the “moving part,” for in its
most primitive form, the axle was fixed to the wheels, which did not rotate independent of it.

(4) PIE *nebh-/*nobh-/*ṇbh-, “navel, nave”: the older meaning is the more widely distributed
“navel,” but the word is attested in the meaning “nave of the wheel” in Sanskrit nábhya-,
nābhi-; Old Prussian nabis; Latvian naba; Old High German naba, German Nabe, Old
Norse nof, Old English nafu and of course English nave.
(5) PIE *yug-ó-, “yoke”: Hittite iukan, Sanskrit yugám, Greek zugón, Latin iugum, Gothic juk,
Old Norse ok, Old High German juch, joch, Lithuanian jùngas, Russian igo, Welsh iau,
Armenian luc. This word is derived from the PIE verbal root *yeug-, “yoke, join, unite,
harness”: Sanskrit yuj-, yunákti, Greek zeúgnumi, Latin iungō, Lithuanian jùngti. The yoke
was needed not only for wheeled vehicles but also for the plow, which is likely to have
been a slightly earlier invention. In both cases PIE-speakers used a native term.

Anthony, who locates the Late PIE homeland in the Pontic-Caspian steppes in his graphic map
of the distribution of PIE vehicle terms (Fig. 6.3), has difficulties with the origin of this
terminology. He explicitly denies proposing “that wheeled vehicle technology originated in the
PIE homeland,” admitting “only that most of the IE vocabulary for wheeled vehicles originated
in PIE” (Anthony 1995:558).

If they learned about the invention of the wheel from others they did not adopt the
foreign name for it, so the social setting in which the transfer took place probably was
brief, between people who remained socially distant. The alternative, that wheels were
invented within the Proto-Indo-European language community, seems unlikely for
archaeological and historical reasons, though it remains possible. (Anthony 2007:34)

FIGURE 6.3 Distribution of select Proto-Indo-European terms referring to wheeled vehicles, according to David W.
Anthony. Modified from Anthony 1995:557, fig. 1. Courtesy David W. Anthony. In the 1995 map, Anthony placed the
“Proto-Indo-European homeland” in the Pontic-Caspian steppes, as he did in the map of Anthony 2007:84 (where he
dates this homeland to “about 3500–3000 BCE”).

All of the earlier discussed five terms are native PIE words: they have been formed by deriving
them from PIE verbal roots. This is important because, with any new technology, the inventor
of the technology is the one who names it, and thereafter the inventor’s name usually spreads
with the spread of the technology.
For example, in the Uralic (Finno-Ugric) languages, the traditional vehicle terminology is
predominantly borrowed from IE languages. All the five common words for “wagon” or “cart”
in Finnish are IE loanwords:

• vaunut (pl.) from North Germanic: Old Norse vagn, etc.;

• rattaat (pl.)/ratas (sg.), “wheel,” from Baltic: Lithuanian rãtai (pl.)/rãtas (sg.), “wheel”;
• kärryt (pl.) from Old Swedish kärra (Latin carrus);
• kiesit, keesit, kääsit (pl.) from Swedish schäs (French chaise);
• vankkurit (pl.), perhaps from Swedish vagn or German Wagen.

Likewise, the Finnish terms for the parts of the cart or wagon are IE loanwords: akseli, “axle,”
aisa, “thill,” napa, “nave,” and ies (gen. ikeen), “yoke”; an exception is the native word pyörä,
“wheel” (also “bicycle”), “whorl, whirl,” with such derivatives as pyöreä, “round,” and pyöriä,
“roll, turn.”
Following this logic, the answer to the important question—who did the PIE speakers get
their wheeled vehicles from?—can be answered with fair certainty: from themselves. In my
view, it was PIE-speakers who invented the wheeled vehicle.
Until recently, it has been assumed that the wheeled vehicle was invented in the Late Uruk
culture of Mesopotamia c. 3500–3300 BCE. However, wheeled vehicle finds of comparable date
have been made not only in West Asia but also in many places in Europe. Furthermore,
Johannes Renger (2004) and Josef Maran (2004b) observe that the marshlands of Sumer were
not favorable terrain for wheeled vehicles; sledges would have worked in ordinary life much
better than wheeled vehicles in marshy Mesopotamia, and indeed stayed in use there long
after the Late Uruk period. It is true that the Uruk pictograms show sledges with four wheels;
however, these “wheels” may depict rolling logs over which the sledges ran. Logs rolling
beneath sledges were probably the initial stage in the invention of the wheel for carts and
wagons (Littauer & Crouwel 1979).
Maran (2004a,b) suggests the Late Tripolye culture as the most likely place of origin for
wheeled vehicles. Late Tripolye is the only culture to show evidence of wagons predating 3500
BCE (Burmeister 2004), in the form of drinking cups provided with rotating model wheels and
with ox foreparts protruding from the front of the cup. In addition to these wagon-shaped
drinking cups, there are numerous Late Tripolye drinking cups in the shape of an ox-pulled
sledge, which is thought to be the immediate predecessor of the ox-pulled wagon (Fig. 6.4).

FIGURE 6.4 (a) Pottery cups in the shape of a sledge with bent runners, some with the head(s) of one or two oxen,
and figurines of oxen wearing painted straps, from five different Tripolye culture sites of phase C1 (c. 3800–3400
BCE): Nezvis’ko [1]​, Majdanets’ke [2], Tal’yanky [3–6], Chychyrkozivka [7] and Volodymyrivka [8]. After Ryzhov
2003:56–57, figs. 1, 2. Courtesy Sergej M. Ryzhov. Many Tripolye C1 sites have also produced round or rectangular
pottery cups with or without ox protomes, and pairs of front and hind legs often perforated for an axle and two
wheels (see Gusev 1998; Parpola 2008:15–17, fig. 4e–k), as well as separate perforated model wheels of clay (Gusev
1998:18 figs. 3:3–3:10; Parpola 2008:14, fig. 4a). (b) A pottery cup in the shape of a wheeled wagon from the
Baden/Pécel culture site, Budakalász, Hungary (late fourth millennium BCE). After Piggott 1983:46, fig. 15. Courtesy
Ferenczy Múzeum, Szentendre, Hungary.

Between 4000 and 3400 BCE, the Late Tripolye culture was the most thriving and populous
agricultural community in the entire Copper Age world, cultivating extremely fertile black soil,
in villages covering hundreds of hectares and housing up to 15,000 people (Fig. 6.5). These
agriculturalist people needed transport, whether by sledge or wheeled wagon. The local
forest-steppe provided enough trees for the construction of primitive solid wheels but also
sufficient open and level fields for the movement of wheeled traffic, unlike the forested and
hilly landscape that covered most of Europe.
FIGURE 6.5 (a) Layout of the Tripolye C1 town of Majdanets’ke (180 ha), based on geomagnetic measurements.
After Videiko 1995:49 Abb. 4. Courtesy Mykhajlo Yu. Videjko. (b) Reconstruction of buildings and fortifications at
Majdanets’ke: 1–2 part of the town excavated in 1987–1991. After Videiko 1994:23, fig. 13. Courtesy Mykhajlo Yu.
Videjko.
I am connecting Maran’s hypothesis that wheeled vehicles were invented in the Late
Tripolye culture with the hypothesis that the Tripolye culture was taken over by PIE speakers
by c. 4000 BCE. The PIE speakers would have largely assimilated the earlier Tripolye
population linguistically by the time wheeled vehicles were invented, probably c. 3600 BCE.
The location of Late Tripolye culture makes sense as the geographical center for the spread of
the wheeled vehicles; it is also very near the middle of the IE-speaking area and is a good
candidate for being the Late PIE homeland from this point of view.
Vehicle technology was probably transmitted to West Asia from the Tripolye culture via the
Caucasus, where the Pontic-Caspian and West Asian cultural spheres interacted with each
other during the fourth millennium BCE. From both the south and the north there was great
interest in possessing the copper resources of the Caucasus. This led to the formation of the
south Caucasian Kura-Araxes culture and the north Caucasian Majkop culture (c. 3950–3300
BCE). While the Kura-Araxes culture continued the local traditions with heavy influence of the
Uruk expansion from Mesopotamia, the Majkop culture has long been considered a splendid
mixture of the steppe and West Asian traditions. The pastoralists of the east European steppes
had received their copper mainly from the Balkans during the Copper Age, but after the
collapse of the Balkano-Carpathian “metallurgical domain,” around 4000 BCE, the Caucasus
became their main source of metal during the Early and Middle Bronze Age (Chernykh 1992;
2007). Out of the approximately 300 graves belonging to the last phase (c. 3500–3300 BCE) of
the Majkop culture, two elite burials under a barrow contain a wagon, one at
Starokorsunskaya in the Kuban steppe, the other at Koldyri on the Lower River Don. From the
immediately succeeding Novotitarovskaya culture (c. 3300–2800 BCE) of the Kuban steppe,
116 wagon graves are known. The wagons apparently reached the Caucasus from the west,
from the forest-steppe region between the Prut and the Bug rivers. Several clay models of
wheels are known from the associated post-Tripolye phase C2 sites.
Mallory leaves the origin of wheeled vehicles an open question but comments:

Tomas Gamkrelidze and Vyacheslav Ivanov [1995:I,622] … have noted that … Proto-
Indo-European *kwekwlo- bears striking similarity to the words for vehicles in Sumerian
gigir, Semitic *galgal-, and Kartvelian *grgar. With the putative origin of wheeled
vehicles set variously to the Pontic-Caspian, Transcaucasia or to Sumer, we may be
witnessing the original word for a wheeled vehicle in four different language families.
Furthermore, as the Proto-Indo-European form is built on an Indo-European verbal root
*kwel-, “to turn, to twist”, it is unlikely that the Indo-Europeans borrowed their word
from one of the other languages. This need not, of course, indicate that the Indo-
Europeans invented wheeled vehicles, but it might suggest that they were in some form
of contact relation with these Near Eastern languages in the fourth millennium BC.
(Mallory 1989:163)

Sumerian gigir, inscribed in the cuneiform tablets of the third millennium BCE, may indeed
provide the earliest written testimony for an originally PIE word.
While there is fair agreement concerning the Anatolian branch as the first language group
to separate from the PIE unity, the same cannot be said of the construction of a family tree for
the other branches of the IE language family. The reason seems to be that Late PIE
disintegrated suddenly in all directions. When the Late Tripolye culture dissolved, around
3400 BCE, it gave way not only to local Post-Tripolye cultures (phase C2, c. 3400–2900 BCE),
but also created new cultures all around, cultures that shared components of Tripolye and
steppe origin. And these new cultures spread to the very regions where the various IE
languages first made their appearance, or a fair distance along the route leading to their final
destinations (Iran and India in the case of Indo-Iranian). It is assumed that climatic change
influenced the transformation of the Late Tripolye culture of thriving settled farmers (with
significant animal husbandry) into several cultures of more mobile pastoralists. Tens of
thousands of people started moving with their cattle and ox-drawn wagons in all directions
(Kohl 2007), mixing with the various earlier local populations and eventually bringing about a
language shift.
Starting about 3200 BCE, probably from the Post-Tripolye Middle Dnieper and Sub-
Carpathian cultures, the Corded Ware (or Battle Axe) cultures (c. 3100–2300/2000 BCE) spread
within a couple of centuries over a vast area, spanning from the Netherlands to the coasts of
Finland and to the region of the Upper Volga River. In Finland, their beginning is dated 3200
or 3000 BCE. In the Upper Volga region, the local Corded Ware variant, the Fatyanovo culture
(c. 2800–1900 BCE), belongs visibly to the Bronze Age from about 2300 BCE, when it developed
a metal-working Balanovo extension as far east as the area between the mouths of the Oka
and Kama rivers (Fig. 6.6). The language of the Corded Ware people, who were mobile
pastoralists, can be assumed to be “Northwest IE,” still quite close to Late PIE, the common
ancestor of the later Celtic, Germanic, and Balto-Slavic branches of the IE language family,
which all emerge into history within the area occupied by the Corded Ware cultures.
FIGURE 6.6 The Yamnaya (Y and diagonal lines) and Corded Ware cultures (stippled) (c. 3300–2000 BCE). Within the
Corded Ware: circle = Sub-Carpathian culture; MD = Middle Dnieper culture; S = Scandinavian Corded Ware (from
c. 2800 BCE); F = Fatyanovo culture (c. 2800–1900 BCE); B = Balanovo culture (c. 2300–1900 BCE). After Carpelan
& Parpola 2001:65, fig. 6, modified.

On the basis of their rapid spread to the areas where various branches of IE were spoken
in historical times, the Corded Ware cultures have long been connected with the early
expansion of IE. But it has been difficult to reconcile this with the assumption that Late PIE
was spoken in the Yamnaya cultures, as these two cultures cannot easily be derived directly
from one another. There is no such difficulty if Late PIE is correlated with the Late Tripolye
culture.
In its final C phase, the Tripolye culture expanded to the Pontic steppes and eventually (in
C2 phase) differentiated into a number of regional cultures (Horodistea-Foltesti, Kasperovtsy,
Usatovo, Gorodsk, and Sofievka) (c. 3400–2900 BCE). In the steppes these Post-Tripolye
cultures fused with the various Late Srednij Stog II pastoralist cultures (Stogovska, Kvitanska,
Dereivka, and Nizhne-Mikhailovka) that had flourished in the Pontic-Caspian steppes during
the preceding period, c. 4000–3400 BCE. This resulted in the formation of the Yamnaya
cultural complex (c. 3300–2500 BCE), which eventually extended from the Danube River to the
Urals (Fig. 6.6). A large number of wagon graves belonging to the Yamnaya archaeological
horizon have been found, attesting to a pastoral economy based on mobility (Fig. 7.1). The
Yamnaya cultures can be assumed to have spoken variants of “Southeast IE,” which was still
close to Late PIE, but undoubtedly soon split into a number of local dialects. Its eastern
extension and successors connected with the Indo-Iranian branch of IE language family will be
discussed in chapter 7.
A spearhead of the earliest Yamnaya-related eastward movement reached as far as
southern Siberia (the Minussinsk basin) and Mongolia, founding there the Afanasyevo culture,
dated c. 3100–2500 BCE. This culture is very similar to the early Yamnaya cultures in many
respects, including the burial of the body in supine position with legs flexed and sprinkled
with ochre, pointed-base vessels, “censer” bowls, copper knives, and awls. The roof over an
Afanasyevo grave at the recently excavated site of Khurgak-Govi in Mongolia, radiocarbon
dated to c. 2900–2500 BCE, is interpreted as a wagon chassis. The sites of the Afanasyevo
culture are some 1500 km distant from the easternmost Yamnaya sites in the Urals, but a few
related sites (such at Karagash near Karaganda) have been found in the intervening area.
Afanasyevo-like new cultural components coming from the northern steppes have also been
found at Sarazm in the Zeravshan Valley of Tajikistan (in the fourth cultural phase dated to c.
2500–2000 BCE).
The Afanasyevo culture has been correlated with the Tocharian languages spoken in the
first millennium CE in Chinese Turkestan (Sinkiang/Xinjiang) (Mallory & Mair 2000). Tocharian
has retained many archaic features of PIE, and is widely considered to be the next oldest PIE
branch after Anatolian. Particularly noteworthy is that Tocharian has not participated in the
satemization that affected among others the Indo-Iranian languages before they spread widely
to the Asiatic steppes around 2000 BCE (as will be argued in chapter 7), so Tocharian must
have come to its distant speaking area before the satemization happened. There are some
difficulties with this Afanasyevo–Tocharian correlation, nevertheless the archaeological record
bears witness to an early offshoot of the Late Tripolye/Yamnaya cultures of southeastern
Europe far off in the Asiatic steppes not too far from the Tocharian-speaking areas.
Penetrating deep into the Danube Valley and the Balkans, the western extensions of the
Yamnaya cultural community formed an interaction sphere, which can be correlated with the
later emergence in these regions of various branches of the IE language family. The Greek
speakers are supposed to have migrated to Greece in two or three waves. The most widely
suggested slot for the first wave is the archaeological shift from Early Helladic II to Early
Helladic III, c. 2300 BCE, during which EH II sites in Corinthia and the Argolid were destroyed
and abandoned, and new types of architecture (apsidal houses), burials (tumuli), weapons
(perforated stone hammer-axes), and pottery (Minyan Ware) were introduced.
As discussed in chapter 7, the horse-drawn chariot was developed in the Sintashta culture
of the south Ural steppes at the end of the third millennium BCE, from where it spread quickly
east and west. Steppe-style cheek-pieces for driving chariot horses are known from Mycenae,
too (Usachuk 2004; 2013). The horse-chariot, swords, and status burials under a barrow
mound were probably introduced via Albania by a small body of intruding warriors, who gave
rise to the sudden emergence of Mycenaean chiefdoms, which flourished during the Late
Helladic period, c. 1700–1100 BCE. They might have originated from the late phase of the
Catacomb Grave cultures of the Pontic steppes, where it was the custom to bury a chief with a
face-mask made of clay: an artifact considered to be a possible model for the famous golden
face-masks in the shaft graves of Mycenae (see chapter 7).
The last major wave of migration to Greece is associated with the collapse of the
Mycenaean civilization around 1300 BCE. The general unrest of these times pushed speakers
of Dorian dialects into Greece and Proto-Armenian speakers into Anatolia. (In Anatolia many
sites were destroyed around 1200 BCE; a connection has been suggested between the Proto-
Armenians and the people called Muski, who were said to have had an army of 20,000 men on
the Upper Euphrates, according to Assyrian sources in 1165 BCE.) One factor undoubtedly
connected with the unrest was the large-scale adoption of riding, which seems to have started
among the probably Proto-Iranian speaking roller-pottery cultures of the Pontic-Caspian
steppes around 1500 BCE (see chapters 7 and Chapter 9). The Dorians introduced riding into
Greece along with their cult of the divine horsemen, Kastor and Poludeukes (chapter 10).
My goal in this chapter has been to demonstrate that the PIE terminology for wheeled
vehicles is the most secure criterion for locating the Late PIE-speaking homeland. The
terminology is of native PIE origin and suggests that PIE speakers were the inventors of
wheeled vehicles. On this basis, Late PIE was spoken in the Late Tripolye culture of
agriculturalists, where the wheeled vehicles were most likely invented. This hypothesis is in
agreement with the assumption that Early PIE was spoken by the Skelya pastoralists of the
Pontic Caspian steppes, who invaded and took control of the Tripolye culture c. 4300–4000
BCE. At the same time another group of Skelya pastoralists invaded the Balkans, with the
result that their Early PIE eventually came to Anatolia. Both the Corded Ware cultures of
northeastern Europe and the Yamnaya cultures of southeastern Europe came into being when
the tens of thousands of Late Tripolye people became mobile pastoralists and spread in all
directions with wheeled vehicles around 3400 BCE. From these two extensive cultural
communities of the Early Bronze Age (c. 3300–2500 BCE and later) it is possible, through
successive cultures with one ultimate origin, to trace the trail of the major branches of the IE
language family to the areas where they are first historically attested.
7

Early Indo-Iranians on the Eurasian Steppes

The Yamnaya cultural community is the umbrella name for the different regional cultures that
prevailed from about 3300 to 2500 BCE in the Pontic-Caspian steppes north of the Black Sea,
spanning some 3000 kilometers from the Danube to the Urals (Fig. 6.6). The name derives
from Russian, in which yama means “pit, grave”; indeed, the Yamnaya cultural community was
formerly known as the “Pit Grave culture.” Its constituent cultures are identifiable almost
exclusively from their graves, which were fairly simple pits. The dead were normally placed on
their backs, but sometimes on their sides, in each case with their legs flexed and their head
facing east or northeast, lying either directly on the ground or on wooden planks or reeds. The
body was often sprinkled with red ocher, to represent life-giving blood. Early burials are poor
in grave goods, but later graves include pots, flint or copper knives, flint spear-heads, stone
axes, bone harpoons, sickles with flint blades, and ornaments such as hammer-headed bone-
pins and boar’s-tusk pendants. Any animal bones probably come from food provisions for the
dead, partly from cultic sacrifices; the skulls and forelegs of sheep or more rarely horse occur,
as do sheep knucklebones, which may have served as gaming dice. The burial site might be
covered with planks or stone slabs or surrounded by a stone circle, but as a rule, above it was
a barrow or kurgan, an earthen mound, which could be used for further burials. Sometimes,
anthropomorphic stone stelae top the mound.
The Yamnaya cultures came into being when people of the Late Tripolye culture, who very
probably spoke Late Proto-Indo-European, rapidly spread with their wheeled vehicles to the
Pontic-Caspian steppes, which had until this time been occupied predominantly by the
Skelya/Srednij Stog II and (in the eastern parts) Late Khvalynsk people, whose language was
probably based on Early PIE. These Early PIE substrata may have had an archaizing effect,
leaving traces of laryngeal phonemes which were probably lost in the assumed Late PIE of the
Tripolye area.
It seems that the ancestors of the Greeks and Armenians occupied the westernmost part of
the Yamnaya continuum, and that the future Armenian speakers, whose language is close to
Greek but shares the satem innovation (see chapter 3) with Indo-Iranian, stayed longer in the
steppes than the future Greek speakers. The steppes east of the Dnieper River were probably
the ancestral homeland of the Indo-Iranian speakers. Greek and Indo-Iranian have some
common post-PIE linguistic innovations (such as several derivational suffixes and the
augment, i.e., a prefixed past tense marker of the verb).
FIGURE 7.1 A wagon grave of the Yamnaya culture at Tri Brata. After Piggott 1983:55, fig. 23. Courtesy Alison and
Stewart Sanderson.

Numerous Yamnaya graves contain the remains of entire wagons or carts, with two or four
solid wheels, pulled by oxen (Fig. 7.1). They were essential for the exploitation of the open
steppe, not only for transport purposes but also as mobile homes for a family. This mobility
explains why occupational sites of the Yamnaya culture are rare, although a number of
temporary camps typical of mobile pastoralists have been found. A few settlements are known,
however, notably at Mikhajlovka in the Lower Dnieper River region. This settlement overlies a
layer representing the earlier Skelya culture. It eventually covered 1.5 hectares. The first
houses were oval in shape and half dug into the soil; later they were larger and rectangular,
with three rooms, built of wood on stone foundations. Mikhajlovka was surrounded by a moat
and 2.5-meter stone fortification. Warfare is suggested by the discovery of weapons:
arrowheads and knives of flint, mace-heads, and battle-axes of stone.
The animal bones excavated at Mikhajlovka attest to animal husbandry and hunting.
Besides domestic animals—chiefly sheep, goats, and cattle, with a fair number of horses and
some pigs—there are the bones of wild asses, wild bulls, wild boar, red deer, saiga antelopes,
and, in small numbers, those of hares, foxes, wolves, otters, beavers, and other wild animals.
Agriculture on the open steppe is limited by salty and sandy soils and the absence of water,
but its practice in the river valleys is implied by the discovery of sickles with flint blades and
querns for grinding grain. Metallurgy was practiced locally, though the metal itself and the
forms of the tools (knives, awls, chisels, and adzes) came from the Caucasus, apart from in the
Volga and Ural regions, which had their own copper. Flint, stone, bone, and antler remained
important materials for tools and weapons. Handmade pottery continued earlier local
traditions, and was decorated with cord and comb impressions. An important diagnostic item
is the “censer,” a ceramic bowl with ceramic feet probably used for burning hemp, in the
manner of the much later Scythians mentioned by Herodotus (4,73-75).
Between the Dnieper and Volga rivers, the Yamnaya cultural community was succeeded
around 2500 BCE by the Catacomb Grave cultural community, which continued until about
1950 BCE. Its economy was very similar to the Yamnaya culture, though it included specialized
crafts, particularly bronze working, weapons manufacture, and weaving. Most of the sparse
Catacomb Grave settlements are seasonal camps; but a fortress with stone walls surrounded
by a moat is known. The name derives from a catacomb-like side-niche at the bottom of the
grave pit, into which the dead person was placed in the flexed position on his or her right side.
The catacomb is ascribed to cultural influence from the Caucasus, as is the style of the metal
objects. Some graves contain animal offerings, such as the head and hooves of sheep, goats,
and cattle. Such diagnostic cultural items as a hammer-headed bone-pin and the “censer”
continue to be used. Presumably elite graves under barrows contained wagons, scepters,
stone and metal axes, stone mace-heads with knobs, daggers, arrows, and, remarkably, face-
masks made of clay (about a hundred examples are known), supposed to be prototypes for the
golden face-masks of the Mycenaean shaft-grave burials. There is also evidence of the practice
of skull deformation of young people.
The Catacomb Grave cultures are likely to have formed the ancestral speech community of
the “Iranian” branch of Indo-Iranian. Of course “Iranian” (or “Irano-Aryan”) and “Indo-Aryan”
are anachronistic terms for groups of languages spoken in the Eurasian steppes and the forest
steppes. These terms simply reflect the fact that these languages were ancestral to the two
chief branches of the Aryan languages nowadays spoken mainly in Iran and India, respectively.
Millennial cultural and linguistic continuity in the Pontic-Caspian steppes is suggested by the
fact that after the Catacomb Grave cultures the same large area was occupied by very similar
and genetically related pastoralist cultures.
West of the Don River, the Catacomb Grave cultures were transformed into a culture with
multiroller ceramics (in Russian, kul’tura mnogovalikovoj keramiki = KMK), also called the
Babino III culture (c. 2100–1850 BCE). (“Roller” here denotes a rope-like band of clay attached
to the outer surface of a pot before firing.) The Srubnaya culture (c. 1850–1450 BCE) is also
known as Timber Grave culture, from Russian srub, “timber work,” which is sometimes found
with the burials. The Srubnaya culture succeeded not only the KMK culture but also the
Abashevo culture (discussed below) over the latter’s entire area, and expanded to the
southern Urals where it coexisted with the contemporary Andronovo culture (also discussed
below). The Pozdnyakovo culture (c. 1850–1450 BCE) of the Volga-Oka interfluve represents a
Srubnaya expansion into the forest zone of central Russia.
In the Late Bronze Age (c. 1500–800 BCE), the Srubnaya culture was followed by cultures
distinguished by roller pottery (Russian valikovaya keramika) (Fig. 9.7). With their adoption of
riding about 1500 BCE, the peoples of the roller-pottery cultures were powerful warriors who
enlarged the earlier Srubnaya area westwards, northwards (to the Kama River basin), and
eastwards. In the east, roller-pottery cultures came to cover the whole of Kazakhstan and
Turkmenistan (previously occupied by the Andronovo cultures); in chapter 9 it is argued that
their language was Proto-East-Iranian. In the steppe and forest steppe of northern Kazakhstan
between the Tobol and Ishim Rivers, the roller-pottery tradition is represented by the
Alekseevka, alias Sargary, culture (c. 1500–900 BCE) (Fig. 9.7, no. 8). All of these roller-pottery
cultures were the immediate ancestors of the Iranian-speaking mounted horsemen known as
Scythians, Sarmatians, and Sakas, who ruled the Eurasian steppes from the Danube to
Mongolia in the first millennium BCE and the early first millennium CE. Many of the river
names in the European steppe have an Iranian etymology; for example, Don (and its
diminutive, Donets) comes from Old Iranian (Avestan) dānu, “river,” which is also behind the
initial part of the river names Dnieper and Dniester.
It has been debated which way the speakers of West Iranian languages came from their
homeland in the northern Pontic steppes to the Iranian Plateau and West Asia—via the
Caspian Gates in the eastern Caucasus, which according to Herodotus was the route taken by
the Scythians who in 652–625 BCE ruled the Median empire, or via the eastern route around
the Caspian Sea and through Khorasan. The cultural change associated with the shift from Yaz
I to Yaz II in southern Central Asia around 1000 BCE is likely to be connected with the arrival
of the West Iranian speakers. The Medes and Persians are first mentioned in Neo-Assyrian
royal inscriptions dating from 881–835; they contain Median names and announce that
twenty-seven kings of Parsuva situated in the area around Lake Urmia paid tribute to the
Assyrian king Shalmaneser III in 835 BCE. Cuyler Young (1985) connects the early Medes and
Persians with the Late West Iranian Buff Ware (c. 900–700 BCE) and traces this ceramic back
to southern Central Asia where it first appears around 1100 BCE.
In any case, the West Iranian languages have been spoken in most parts of Iran since the
early first millennium BCE. The Median empire flourished in northern Iran (with its capital
Ecbatana at modern Hamadan) c. 678–549 BCE. Conquered by Cyrus the Great in 549 BCE, it
was succeeded by the Old Persian-speaking Achaemenid empire (559–330 BCE), which
stretched from Egypt and Anatolia to the Indus Valley. After the conquest of Alexander the
Great, the empire was divided by his generals, of whom Seleucus received the eastern part
from Syria to the Indus Valley. The Hellenistic Seleucid empire that thus came into being
lasted from 312 to 63 BCE. In 247 BCE Arsaces of the Parna tribe (see chapter 9) conquered
Parthia in northern Iran from the Seleucids, founding the Arsacid or Parthian empire (247
BCE–224 CE); it was greatly expanded by Mithridates I (ruled c. 171–138 BCE), who took Media
and Mesopotamia from the Seleucids. The Parthians were succeeded by the Sasanids, who
ruled Iran c. 224–651 CE, until the Muslim conquest of the country.
The West Iranian language called Baluchi is closely related with Kurdish spoken in eastern
Turkey, northwestern Iran, and northern Iraq. According to the Baluchi tradition—for which
there is some confirmation in historical sources—the Baluchi speakers in the seventh or eighth
century CE migrated from the west to their present habitat that extends from Iranian Makran
to Pakistani Baluchistan.
Having sketched the early history of the Iranian branch, which probably derives from the
Catacomb Grave culture of the Pontic-Caspian steppes, we return to the split of Proto-Indo-
Iranian. A dialectal differentiation that eventually resulted in the formation of the two
branches seems to have started about 2500 BCE. The “Indo-Aryan” branch probably originated
in the Late Yamnaya culture of the Upper Don and the Poltavka culture of the Volga-Urals,
which are contemporary with Early Catacomb graves, c. 2500–2100 BCE.
The Abashevo culture (c. 2300–1850 BCE), named after a site in the Chuvash Republic on
the Mid-Volga, extended along the border of the forest steppe and the forest zone from the
Upper Don to the Upper Tobol River in western Siberia (Fig. 7.3). That it had its origin in the
eastern Late Yamnaya cultures is indicated by its burial customs (the dead placed on their
backs with their legs flexed, beneath a barrow) and its pottery, which contains crushed shell
mixed with clay to strengthen it, as do all the steppe ceramics from Neolithic times onward.
Metal sickles and stone querns testify to the existence of agriculture. The main means of
subsistence was animal husbandry; almost all the bone finds belong to domesticated animals:
cattle, sheep, goats, and small numbers of horses and pigs.
FIGURE 7.2 Copper deposits in the Urals: (a) area of copper sandstone ores, (b) groups of mines, (c) copper deposit.
After Koryakova & Epimakhov 2007:29, fig. 1.1.B (based on Chernykh 1970). Copyright © Ludmila Koryakova and
Andrej Epimakhov. Reprinted with the permission of Cambridge University Press.

The eastward expansion of the Abashevo culture was motivated by the deposits in
sandstone of rich, pure copper on the Mid-Volga and the Lower Kama and Belaya rivers (Fig.
7.2). The Abashevo people fought for copper with the Balanovo people. An Abashevo burial at
Pepkino contained twenty-eight men, some of whom had been hit on the head with an ax. The
Balanovo culture (c. 2300–1900 BCE) forms the easternmost branch of the Fatyanovo culture
(c. 2800–1900 BCE), which constituted the eastern variant of the Corded Ware/Battle Axe
cultures (Fig. 6.6). The Fatyanovo-Balanovo people almost certainly spoke Proto-Balto-Slavic,
an early branch of Indo-European that is known to have been in early contact with Indo-
Iranian. Balto-Slavic and Indo-Iranian share some early linguistic changes, in particular
satemization, and (partially) the so-called ruki rule (change of *s into *š after r, u, k, and i).
While Abashevo burials are similar to Yamnaya burials but differ from the flat burials of the
Fatyanovo people (which are in shafts, with men on their right side, heads facing the
southwest, women on their left side, heads facing northeast), the early Abashevo ceramics and
metal objects, especially the ax forms, are very similar to those of the Fatyanovo-Balanovo
Corded Ware culture. This indicates a fair amount of interaction between the two cultures,
including presumably linguistic interaction.
The Abashevo culture was an important early center of metallurgy. Its metal production
was initially based on the pure copper ores of the Volga-Kama-Belaya area, but afterwards
moved just east of the Urals, where the arsenical copper was more suitable for producing
harder weapons and tools. Around the twenty-second century BCE, part of the Abashevo
people moved southwards to take possession of the rich metal ores and pastures until then
occupied by the Poltavka culture. This resulted in the emergence of the Potapovka culture in
the Mid-Volga (c. 2100–1700 BCE) and the powerful Sintashta culture (c. 2100–1800 BCE) in
the southeast Urals, where huge copper mines have been located (Fig. 7.3).
FIGURE 7.3 Find spots of artifacts distributed by the Sejma-Turbino intercultural trader network, and the areas of
the most important participating cultures: Abashevo, Sintashta, Petrovka. After Parpola 2012 (2013):157, fig. 8
(based on Chernykh 2007:77). © The Finno-Ugrian Society.

The Sintashta culture had fortified settlements surrounded by an earth wall and a moat,
the best known being Sintashta and Arkaim. The walls are usually formed in a circle about 150
meters in diameter, defending houses that taper inwards so as to create the impression of a
spoked wheel (Fig. 7.4). At Sintashta there are two concentric defense lines and between them
rectangular houses half sunken into the ground, many of them with metallurgical furnaces.
There are several cemeteries near the settlements; at Sintashta one cemetery contained forty
graves. The burials under the barrows have wooden rooms, where chiefs were placed with
their weapons, horse-drawn chariots, and often animal offerings, pottery, and other grave
goods, as well as fireplaces (Fig. 7.5). While it seems that the Catacomb Grave culture already
used ox-drawn chariots, the Sintashta culture seems to be the first to have used horse-drawn
chariots, with both solid and spoked wheels.
FIGURE 7.4 The Sintashta culture fortified settlement of Arkaim in the southern Urals c. 2000 BCE. (1) outer
defensive wall; (2) inner defensive wall; (3) dwellings; (4) circular street with drainage; (5) central square; (6) main
entrance; (7) a moat; (8) yards; (9) rooms inside defensive walls; (10) supplementary entrances; (11) basement of
gate tower. The circles within the rooms are hearths and well-pits. After Zdanovich & Zdanovich 2002:256, fig. 16.3.
Courtesy Gennady B. Zdanovich.
FIGURE 7.5 Horse-chariots in the graves of the Sintashta culture of the southern Trans-Urals, Russia, c. 2100–1800
BCE. (a) Reconstruction of the burials 10 and 16 in the area SM south of the Great Kurgan at Sintashta. A warrior
with his weapons lies in a chariot and a charioteer or groom with two horses. After Gening, Zdanovich, & Gening
1992:154, fig. 72. Courtesy Gennady B. Zdanovich.
FIGURE 7.5 (b) Grave 1 of kurgan 9 in the cemetery of Krivoe Ozero had the following finds on the bottom: (1–3)
ceramics; (4) arrow-heads made of stone and bone; (5–6) pairs of cheek-pieces made of horn; (7) bronze ax-head; (8)
bronze dagger; (9) whetstone; (10–11) chariot wheels (in section view; (14) view of the wheel navel); (12) skulls and
leg bones of two horses; (13) bones of sheep or goat. After Vinogradov 2003:85, fig. 34. Courtesy Nikolaj B.
Vinogradov.

The “daughter branch” of the Sintashta culture, the Petrovka culture (c. 2000–1800 BCE),
expanded eastwards into the Tobol-Ishim interfluve of northern Kazakhstan. Evidently its goal
was the Altai Mountains, rich in not only copper but also tin, needed for alloying copper into
bronze. Together the three cultures—Abashevo, Sintashta, and Petrovka—formed the
backbone of the “Sejma-Turbino transcultural network” (c. 2100–1600 BCE). Within this
network armed traders transmitted high-quality weapons from Siberia to the Urals and even
as far west as Finland (Fig. 7.3). This network probably played a major role in the early
dispersal of the Uralic languages.
The Sintashta and Petrovka cultures also gave rise to the extensive Andronovo cultural
community, which also possessed horse-drawn chariots and was likewise engaged in
metallurgy and stockbreeding. Dispersed all over the Asiatic steppes between the Ural and
Altai mountains, the Andronovo community was divided into two main branches: the Alakul’
Andronovo culture (c. 2000–1700 BCE), distributed mainly in the steppe and forest steppe of
the Trans-Urals and northern, western, and central Kazakhstan, and the Fëdorovo Andronovo
culture (c. 1850–1450 BCE), which covered practically the whole of Turkmenistan and
Kazakhstan; in the north it is connected especially with the forest steppe, from the Trans-
Urals in the west to eastern Kazakhstan and the Upper Yenissei River. The symbiosis of the
Fëdorovo Andronovans with the probably Proto-East-Uralic speaking people of the Cherkaskul’
culture will be discussed shortly (Fig. 7.6).

FIGURE 7.6 The distribution of Srubnaya (Timber Grave, early and late), Andronovo (Alakul’ and Fëdorovo variants),
and Cherkaskul’ monuments. Map drawn by Virpi Hämeen-Anttila for Parpola 1994a:146, fig. 8.15, based on
Chlenova 1984: map facing p. 100; reproduced here after the slightly revised version of Parpola 2012 (2013):165,
fig. 10. © The Finno-Ugrian Society, Asko Parpola, and Virpi Hämeen-Anttila.

On the basis of linguistic criteria, the Uralic, alias Finno-Ugric, languages form three
major groups. Western Uralic comprises Saami, Finnic (Finnish, Estonian, Livonian, Karelian,
etc.), and Mordvin; Central Uralic consists of Mari (formerly called Cheremiss) and the Permic
languages (Udmurt, formerly called Votyak, and Komi, formerly called Zyryan or Zyryene); and
East Uralic (with shared innovations established by Jaakko Häkkinen 2009) consists of the
Samoyed languages distributed very widely in Siberia and the Ugric languages (Hanti or
Khanty, previously called Ostyak; Mansi, previously called Vogul; and Hungarian, which
arrived in Hungary from the southern Urals by the tenth century CE).
The region around the Kama River valley has long been suspected to be the original
homeland of the Uralic (Finno-Ugric) language family. The Proto-Uralic vocabulary suggests a
Mesolithic/sub-Neolithic hunter-gatherer economy, with no indication of agriculture or animal
husbandry. The present distribution of the Uralic languages, and what is known of the history
of the peoples speaking Uralic languages, suggests their origin lies in the forested
northeastern part of European Russia. The homeland should be in the neighborhood of the
Ural Mountains, because Proto-Uralic has words for the Cembra pine, *sïksi, and the Siberian
fir, *ńulka, and these trees do not grow west of the Kama and Pechora rivers. In addition,
Proto-Uralic has native terms for metals, the most important being *wäśka, “copper or
bronze,” which suggests primitive native metallurgy, such as that of the Chalcolithic Garino-
Bor culture of the Kama Valley before the arrival of the Abashevo people (Fig. 7.2).
But Proto-Uralic also had terms for “tin” and “lead,” *äsa and *olna / *olni, respectively;
the former occurs in the compound *äsa-wäśka, which can only denote the tin-bronze
produced in, and imported from, the eastern wing of the Sejma-Turbino network. In addition,
Proto-Uralic has a loanword *ora, “awl,” from Proto-Indo-Aryan *ārā, “awl,” a word not
attested in any Iranian language. Another loanword for a metal object is Proto-West-Uralic
*vaśara, “hammer, ax,” from Proto-Indo-Aryan *vaj’ra-, “weapon of the war-god”; it probably
originally denoted the ax or mace of the Sejma-Turbino warriors, but later acquired the
meaning “hammer” from the Nordic war-god Thor. The Proto-Indo-Aryan compound *madhu-
śišta-, preserved in Sanskrit madhu-śiṣṭa-, “beeswax,” literally “what is left over of honey,”
survives in Komi ma-śis, “beeswax,” while the latter part of the compound is found also in
several other Uralic languages (Mordvin, Mari, and Udmurt) with the meaning “beeswax.”
Beeswax is very important for metal-casting, and the word (or even the verb *śiš-, “to leave,”
from which it is derived) has no Iranian cognate.
Some Aryan loanwords in Proto-Uralic attest to an early stage of language development,
between Proto-Indo-European and the Proto-Indo-Iranian reconstructed on the basis of the
attested Indo-Aryan and Iranian languages, that is, predating sound changes typical of Proto-
Aryan; compare Proto-West-Uralic *kekrä, “circular thing, cycle,” with Sanskrit cakra-, “wheel,
cycle” (see chapter 6). The PIE word *medhu-, “honey,” also still retained its original *e (not
yet changed into *a) in that early form of Proto-Aryan from which it was borrowed into Proto-
Uralic as *mete-. In PIE *medhu- denoted not only “honey,” but also an alcoholic drink made
from honey known as “mead” or “honey-beer” in the historic cultures of Celtic, Germanic, and
Baltic speakers; in Greek méthu came to mean “wine.” The loanwords from Proto-Aryan in
Uralic languages include Proto-Permic *sur, “beer,” corresponding to Sanskrit surā, “beer”
and Avestan hurā, “fermented mare’s milk, kumiss”; the Komi compound ma-sur, “honey-
beer,” shows that it was a kind of mead. In Vedic India, the reception of a respected guest
included a “drink mixed with honey” (madhu-parka, madhu-mantha). Important religious
occasions were celebrated with feasts that involved drinking, called peijas [i.e., peiyas] in
Finnish, which derives from Proto-Indo-Aryan *paiya-s > Old Indo-Aryan peya-, “drink,
drinking; libation.” Alcoholic drinks are assumed to have played a vital part in the life of the
early Indo-European and Aryan elites. This uppermost layer of the society is represented in
Proto-Uralic *asera / *asira, “prince, lord,” from Proto-(Indo-)Aryan *asura-, “lord.” Proto-
Uralic contains also Aryan loanwords that indicated dealings with valued objects: PU *arva-,
“price, value,” from Proto-Aryan *argha-, “price, value,” and the numerals for 100 (PU *śata <
PIA *śata-) and 1000 (PU *śosra < Proto-Aryan *źhasra-).
Several Proto-Aryan religious terms have become part of Proto-Uralic, too. From Proto-
Indo-Aryan *stambha-s, “pillar, world-pillar,” comes Proto-Finnic *sampas, “pillar, world
pillar”; it figures as a much coveted treasure in ancient Finnic epic poetry in the form of a
magic mill called Sampo that has a star-speckled cover. (The vault of heavens rotates around
the axis of the world-pillar supporting the heavens; its being conceived as a magic mill that
grinds riches for its owner is due to Germanic mythology.) Even more important is *juma- (in
this spelling j is the semivowel corresponding to y in English yoke), “god, highest god, heaven”
(Jumala is the word for “God” with a capital G in modern Finnish), from Proto-Indo-Aryan
*dyuma(n)t-, “heavenly, shining; epithet of Indra, the god of thunder and war.”
In view of this religious intercourse and interethnic collaboration of peoples who probably
spoke Proto-Indo-Aryan and Proto-East-Uralic in the steppes, even as far as southern
Turkmenistan (see below), I have ventured to propose that the name of the Aryan god Indra,
which has defied earlier etymological explanations, might come from that of the Proto-Uralic
god of weather, thunder and sky, *Ilmar / *Inmar. Ilmarinen, one of the chief war heroes in
ancient Finnic epic poetry, is also mentioned as the smith who made the vault of heavens and
Sampo; in Finnish, ilma is the usual word for “air, atmosphere, weather.” The change of the
Udmurt form Inmar into Indra is not too difficult to imagine: after a metathesis ar > ra,
occasioned by the fact that so many Aryan words end in -ra (including *vaj’ra, which denotes
Indra’s weapon), the insertion of d between the nasal and r is to be expected (compare Greek
anēr, “man,” genitive an-d-ros).
The best explanation of the many early Aryan loanwords in Uralic languages seems to be a
lengthy symbiosis between the local Kama Valley population and the Abashevo people. Some
loanwords come specifically from the Indo-Aryan branch. They presumably originated with the
Abashevo people, ancestors of the many above-mentioned cultures that dominated the Asiatic
steppes until about 1500 BCE, when they were overrun by horsemen (who in all likelihood
spoke Iranian) from the Pontic-Caspian steppes. Even the Abashevo culture was succeeded by
people who probably spoke Proto-Iranian when its area was taken over by the early Srubnaya
culture about 1850 BCE.
I have suggested that those Proto-Indo-Aryan-speakers who remained in the Volga-Kama
area of pure copper sandstone (Fig. 7.2) eventually started speaking Proto-Uralic, yet retained
an ability to communicate in Proto-Indo-Aryan with the Siberian side of the Sejma-Turbino
international network of metal production and marketing. In this way Proto-Uralic would have
become the language of the European side of Sejma-Turbino metal production and its network
of warrior traders (Fig. 7.3). The westward spread of the Uralic protolanguage with the
Sejma-Turbino network is suggested by its temporal closeness to the disintegration of Proto-
Uralic, which must postdate its adoption of many Proto-Indo-Aryan loanwords.
It is currently thought that the PIE language probably to a large extent spread through
language shifts initiated by local leaders who came to the side of small but powerful groups of
PIE-speaking immigrants. The latter would have an effective, strongly hierarchical social
system, good weapons, and a military backing, plus coveted luxury goods as external markers
of their power. Joining such a network of foreign chiefs would guarantee local leaders both
their existing positions and additional advantages. The loyalty of new network members could
be guaranteed by matrimonial alliances (Mallory 2002a). That the Uralic languages very
probably spread by similar means, with language shifts initiated by Sejma-Turbino warrior-
traders, can be substantiated: a detailed demonstration of how this assumed dispersal of the
Uralic languages to their present areas of speech matches the archaeological record is
available (Parpola 2012 [2013]), but cannot be included here.
While on this topic, I would however like to mention that the collaboration between Uralic
and Aryan speakers continued even into later times, especially between the probably Proto-
East Uralic speaking people of the Cherkaskul’ culture (c. 1850–1500 BCE) of Bashkiria and
the Mid- and South Trans-Urals, and the Proto-Indo-Aryan speaking Fëdorovo Andronovans
(Chlenova 1984). Part of the Cherkaskul’ people adopted pastoralism and accompanied the
Andronovans in long migrations in the steppe, even as far east as the Altai Mountains and as
far south as southern Central Asia (Fig. 7.6). In these migrations the Cherkaskul’ people, like
the later Hungarians who adopted mobile pastoralism, could retain their original language:
some of them apparently remained near the Altai Mountains and became the ancestors of the
speakers of Samoyed languages.
In the light of the evidence presented in this chapter, the early Proto-Indo-Aryan speakers
were not only metalsmiths and arms merchants but also pastoralists and inventors of the
horse-drawn chariot. Remains of chariots and chariot horses with their cheek-plates (Fig. 7.7)
have been found in the elite graves of the Sintashta and Petrovka cultures (2100–1800 BCE)
and those of the succeeding Andronovo cultures (2000–1450 BCE), spreading widely over the
steppes of southern Siberia, Kazakhstan, and Turkmenistan. Hundreds of petroglyphs in
Central and Inner Asia depict the horse-drawn chariot. The horse is the favorite animal
depicted in the knives and scepters of the Sejma-Turbino intercultural network of warrior-
traders (Figs. 7.8, 7.9, ). It is remarkable, however, that these objects never represent a horse-
rider, although the image of a ski-jorer pulled by a horse decorates the top of one Sejma-
Turbino knife (Fig. 7.10). The horse-drawn chariot, first attested in the Sintashta culture, also
diffused to the European steppes in the early second millennium BCE (Usachuk 2004; 2013),
and it went on to revolutionize warfare in the ancient West Asia, where the first horse-drawn
chariots appeared in the twentieth century BCE.

FIGURE 7.7 A horse skull and cheek-plates from grave 11 of the monument SM in Sintashta, southern Trans-Urals.
After the excavation of the grave had been documented, the cheek-plates were placed next to the skull in order to
show students how they had been used. The excavation report, however, failed to note that the photograph does not
represent the finds as they were excavated. After Gening, Zdanovich, & Gening 1992:157, fig. 74,1. Courtesy Nikolaj
B. Vinogradov.
FIGURE 7.8 Sejma-Turbino type metal knives with horse-tops. (a) Sejma. (b) Omsk. (c) Ust’-Muta, Altai. (d) Elunino
1. (e) Rostovka. After Kovtun 2013: t. 2.2; 2.4; 1.2; 1.1; 2.3. Courtesy Igor V. Kovtun.

FIGURE 7.9 Sejma-Turbino-related horse-head stone scepters from the steppes. (a) Shipunovo 5, Chelyabinsk. (b)
Omsk. (c) Semipalatinsk. (d) Bukhtarma, Irtysh. (e) Kizil’, Chelyabinsk. After Kovtun 2013: t. 14.1, 14.3, 14.2, 14.4,
14.5. Courtesy Igor V. Kovtun.
FIGURE 7.10 A ski-jorer pulled by a horse. The top of a Sejma-Turbino knife (see Fig. 7.8e) from Rostovka. After
Kovtun 2013: photo 49.3. Courtesy Igor V. Kovtun.
8

The BMAC of Central Asia and the Mitanni of

Syria

Afghanistan and the surrounding areas of southern Central Asia, through which the Rigvedic
Aryans must have passed on their way to South Asia, were until quite recently a blank on the
Bronze Age archaeological map. In the 1970s, a previously unknown civilization was
discovered in this area, mainly by Viktor I. Sarianidi, a Russian archaeologist of Greek
descent, who excavated it unceasingly until his death “on the spot” at Gonur in 2013. As a
token of respect for a good friend, this book uses Sarianidi’s name for the new civilization, the
Bactria and Margiana Archaeological Complex (BMAC). Bactria and Margiana are derived
from the ancient Greek names for, respectively, northern Afghanistan around modern Balkh
and southeastern Turkmenistan, the satrapy of Marguš in the Persian empire of the
Achaemenids, around the modern city of Merv. It should be noted, however, that an alternative
and perhaps more elegant name, the Oxus civilization (from the ancient Greek name of the
Amu Darya River whose upper reaches flow through the BMAC), has been suggested by Henri-
Paul Francfort.
Francfort (2005) distinguishes three principal periods of the BMAC:

• 3000–2500/2400 BCE is the formative phase, containing handmade ceramics painted with
many colors, mainly with cross and rhombus motifs; stone objects made of alabaster and
steatite; and the first compartmented metal seals. The wide formative area comprised the
piedmont sites of the Kopet Dagh range of southwestern Turkmenistan (Namazga Depe,
Altyn Depe, Ulug Depe, Khapuz Depe), the delta of the Tedzen River (Geoksyur), western
Tajikistan (Sarazm in the Zeravshan Valley), central and northern Afghanistan (Mundigak
and Taliqan), and Iranian Seistan (Shahr-i Sokhta). A particularly important influence was
that of the Proto-Elamite civilization (attested by an inscribed tablet at Shahr-i Sokhta and
cylinder seals in Bactria and Sarazm), manifested in monumental architecture (Mundigak
and Sarazm) and seal motifs. There was influence, too, from the Early Harappan cultures of
the Greater Indus Valley and from the northern steppes, manifested in the Afanasyevo-like
burial at Sarazm, where ceramics of the Kelteminar culture have also been found.
• 2500/2400–1700 BCE is the mature “urban” phase, largely contemporaneous with the
mature Indus civilization, with its most dynamic period, c. 2200–1750 BCE. Its area, the core
area of the BMAC, consists of the oases of southern Turkmenistan (Taip, Gonur, Togolok, and
Kelleli in Margiana; and Namazga, Ulug, Khapuz, Altyn in Kopet Dagh), northern
Afghanistan (Daulatabad, Dashly, and Farukabad), and southern Uzbekistan (Sapalli and
Dzharkutan). Some 300 sites are known altogether, representing a considerable population.
The economy was based on irrigation agriculture using wheat and barley, and commanded
considerable mineral resources, including copper, tin, lead, gold, silver, lapis lazuli, and
turquoise. Numerous settlements were strongly fortified with walls up to 10 meters wide,
employing towers and gates, sometimes with two further inner walls. Inside were residential
houses, workshops, and “palaces” or “temples” (Fig. 8.1).
FIGURE 8.1 The BMAC fortress of Togolok-21 in southern Turkmenistan (ancient Margiana). After Sarianidi
1987:52. Courtesy Margiana Archaeological Expedition.

The “palace” at Gonur, the capital of Margiana, measures 150 × 140 meters, along with a
slightly smaller “temple.” A necropolis of 10 hectares situated 200 m to the west of Gonur
contains more than 1500 excavated graves; a smaller graveyard with royal burials was
discovered south of the “temple.” The latter contained, among other things, a wagon with
bronze-rimmed wheels exactly paralleling those found in Susa. There are many remarkable
finds at Gonur, but I mention here only a few datable to the late third millennium. These
include ivory objects (combs and pins) of the Indus civilization, a large square Harappan
stamp seal with an image of an elephant and an inscription in the Indus script. A cylinder seal
that according to its cuneiform inscription belonged to the cup-bearer of a Mesopotamian king
(whose name is unfortunately illegible) is concrete evidence of relations with Mesopotamia,
where the BMAC was probably known as the far-off country Šimaški (Potts 2008 [2009]).
The material culture comprises a wide variety of beautiful artifacts, including fine wheel-
made ceramics, bronze tools and weapons such as swords and hatchets, and impressive
symbols of power in the form of decorated ritual axes, mace heads, and long stone scepters.
The numerous uninscribed metal and stone seals (Fig. 20.2) form an important category of
objects (Baghestani 1997; Sarianidi 1998; Winkelmann 2004; Franke 2010). The eagle
(sometimes with two heads), snake, scorpion, mountain goat, camel, wild boar, and lion are
among the animals represented on the seals, toilet pins, axes, and maces. Wine cups of gold
and silver give a rare insight into the life of the BMAC people. They depict men plowing fields
with teams of oxen, driving ox-pulled wagons and chariots, feasting on meat and drink,
listening to harp music, hunting game, fighting battles, becoming wounded, and being buried
(Fig. 8.2). These men wear the Sumerian flounced kaunakes garment, as do the statuettes of
seated females, who are likely to be goddesses, made of soft stone (steatite and alabaster),
which is the material of many other objects too. Semiprecious stones (lapis lazuli, turquoise,
and carnelian) are used for jewelry. Pierre Amiet (1986) calls the iconography “Trans-
Elamite”; it has a lot in common with the art of Kerman (Jiroft), ultimately going back to the
Proto-Elamite art of western Iran and Uruk-period Mesopotamia.
FIGURE 8.2 Four BMAC silver cups illustrating the life of the elite. (a–d) A group of goats, archers fighting each
other, and the burial of a hero hit with an arrow. Private collection. After Francfort 2005: fig. 28 a–d. Drawing J.
Svire, courtesy H.-P. Francfort.

FIGURE 8.2  (e) A procession in which an archer wearing boots carries an unrecognized object, nude men carry
heavy round objects, and a one-man chariot and a four-wheeled wagon with two seated men are pulled by pairs of
galloping oxen. Musée du Louvre (AO 28518). After Francfort 2005: fig. 6. Drawing H. David, courtesy H.-P.
Francfort and Musée du Louvre/Béatrice André-Salvini.

FIGURE 8.2  (f–g)A banquet with drinking, eating meat, and harp-playing. Private collection. After Francfort 2005:
fig. 26 a–b. Drawing J. Svire, courtesy H.-P. Francfort.

FIGURE 8.2  (h) Cylindrical cup with an agricultural and ceremonial scene in the Miho Museum, Shiga, Japan
(SF03.055), showing banqueting and plowing with oxen. After Jansen, Mulloy, & Urban 1991:169, pl. 148(h).
Courtesy © Miho Museum.

• 1700–1500 BCE is the late or post-“urban” phase of the BMAC, which is represented by the
Takhirbaj period in Margiana, the Kuzali, Mollali, and Bustan periods in northern Bactria,
and the Namazga VI period in the Kopet Dagh region. In this impoverished phase, the small
finds distinctive of the mature phase of the BMAC such as seals and ax-heads disappear, and
architecture continues only to a limited extent. At Gonur, the building complexes appear to
have been abandoned suddenly at the end of the “urban” phase: valuable grinding stones
were left behind, along with unbaked ceramics near pottery kilns. The situation appears
similar at Togolok 1 and Togolok 21. The desiccation that occurred around 1750 BCE is
thought to be a major cause of the collapse of the BMAC at the beginning of its post-“urban”
phase; it is probably associated with another likely contributing cause, namely the arrival of
steppe nomads representing the Fëdorovo variant of the Andronovo cultures.

Neither Andronovo pottery nor barrows typical of the steppe culture are found south of the
line defined by the Kopet Dagh, Hindukush, and Pamir mountains. Yet, Aryan languages
originally spoken in the northern steppes (see chapter 7) had reached Syria and South Asia by
the middle of the second millennium. In Parpola (1988a), I proposed that the BMAC must have
expanded in these two directions as the archaeological counterpart of this later linguistic
spread. This proposal implied that an elite of Aryan-speaking pastoralist warriors from the
steppes took over the leadership in the BMAC. I compared this with a takeover of power by an
immigrant Proto-Indo-Aryan-speaking elite in the Mitanni kingdom a few centuries later. The
takeover of the BMAC could have occurred relatively peacefully, as there is little evidence of
violence; it could have given the BMAC a more effective hierarchical leadership, which in turn
would provide a plausible explanation for the BMAC’s increased dynamism and
aggressiveness in the later part of its “urban” phase. Fred Hiebert agrees that the expansion
of the BMAC people to the Iranian plateau and the Indus Valley borderlands at the beginning
of the second millennium BCE is “the best candidate for an archaeological correlate of the
introduction of Indo-Iranian speakers to Iran and South Asia” (Hiebert 1995:192). J. P. Mallory
embraces this hypothesis, too, and has developed his well-known Kultur-Kugel model to
support it up theoretically (Mallory 1998; 2002a). Francfort (2005:298), however, excludes the
presence of Aryan-speakers from the “urban” period between 2500 and 1700 BCE, when the
BMAC was ruled by a Trans-Elamite elite, according to him; the Indo-Aryan speakers were in
charge during the post-“urban” period 1700–1500 BCE, while the Iranian speakers came to
southern Central Asia around 1500 BCE.
At present, given the BMAC’s links with South Asia and especially with northern Iran and
Mitanni, and the necessity of having the Rigvedic Aryans as the second (later) wave of Indo-
Aryan immigrants in South Asia (chapter 12), I remain convinced that an early wave of Proto-
Indo-Aryan speakers took charge of the BMAC around 2000 BCE. Some may have come to the
BMAC even earlier, when they had not yet yoked the horses to the chariot but were racing
with chariots pulled by bulls (cf. Fig. 8.2e), though certainly the culture remained thoroughly
“Trans-Elamite.” The mature “urban” phase, however, saw the introduction of the horse and
camel. Horse, camel, and donkey bones have been identified. The necropolis of Gonur was
found to contain two horse burials, which in both cases lacked the head, and which can
plausibly be linked with the prehistory of the Vedic horse sacrifice (chapter 11). Along with the
camel, the domesticated horse is represented in BMAC weapon-scepters (Fig. 8.3) and some
stamp and cylinder seals. By its late phase the BMAC people appear to have gained great
mobility and spread widely both westwards and eastwards. They were in contact with the
Indus civilization, with eastern and western Iran, with the Gulf, with Syria and Anatolia, and
with the steppes, where typical BMAC metal seals have been found as far east as the Ordos
Plateau in the Inner Mongolia autonomous region of China (Baghestani 1997).
FIGURE 8.3 Horse-topped BMAC pin and scepter-maces. (a) A copper pin from Bactria. After Kovtun 2013: photo
112.1. Courtesy Igor V. Kovtun. (b) A mace head from Bactria. After Sarianidi 1986:211. Courtesy Margiana
Archaeological Expedition. (c) A copper mace head from Gonur Depe, burial 2380. After Sarianidi 2007:127, fig. 3.
Courtesy Margiana Archaeological Expedition.

FIGURE 8.3 (d) An unprovenanced copper scepter head from Afghanistan in the Metropolitan Museum of Art, New
York (1989.281.39, jpgt of Norbert Schimmel Trust). When Holly Pittman (1984:70, fig. 32) published this object
thirty years ago, it was united with a bronze mace head. Subsequent conservation work has determined that the
mace head and the horse protome, while both ancient objects, were not originally part of the same piece, and they
have been separated. Image courtesy © The Metropolitan Museum of Art.

A rich grave accidentally found at Zardcha Khalifa in the Zeravshan Valley in Tajikistan
(Bobomulloev 1997) contained, besides ceramics and other grave goods typical of the BMAC
(Sarianidi 2001:434), the remains of a chariot and two horses, which once had cheek-plates
(psalia) of precisely the same type as at Sintashta culture, and mouth bits (Fig. 8.4; cf. Fig.
7.7). In addition, there was a bronze pin, whose horse-top is paralleled by the horse-tops of the
Sejma-Turbino knives and scepters (Figs. 7.8, 7.9). This find suggests that Indo-Aryan
speakers had come from the southern Urals and entered the ruling elite of the BMAC probably
as early as the twentieth century BCE. Sintashta-type cheek-plates have recently been
excavated also in grave 2 at Sazagan in the Zeravshan Valley, here too with BMAC ceramics of
the Sapalli phase (c. 2000–1800 BCE) (Avanesova 2010). Teufer (2012) points out that the
metal bits of Zardcha Khalifa do not come from the Sintashta culture but from West Asia; he
suggests that the spoke-wheeled war chariot was elaborated in southern Central Asia, which
formed a contact zone between the northern (steppe) and southern traditions about 2000
BCE.

FIGURE 8.4 Finds from a BMAC elite grave at Zardcha Khalifa in Tajikistan. (1) A horse-topped bronze pin. (2)
Bronze horse-bits. (3) Sintashta-type cheek-pieces for chariot-horses made of bone. After Carpelan & Parpola
2001:138, fig. 37, based on Bobomulloev 1997:126, Abb. 3.14 (1), 3.12–13 (2) and 128, Abb. 4.1–4 (3).
On the other hand, surface surveys in Margiana (Gubaev, Koshelenko, & Tosi 1998)
revealed that in the late phase of the BMAC, up to the fifteenth century BCE, practically all
BMAC settlements in south Turkmenistan were surrounded by pastoralist campsites with late
Andronovo ceramics, attesting to an ever-increasing presence of steppe pastoralists who very
probably spoke Proto-Indo-Aryan.
During 2000–1900 BCE, the BMAC spread to Pakistani Baluchistan. A rich BMAC graveyard
was accidentally discovered at Quetta, while other BMAC graves were found at Mehrgarh VIII
and Sibri near the Bolan Pass. That BMAC people (not just traders) moved to Pakistani
Baluchistan is evident from the fact that the entire cultural complex was imported, including
burials (Jarrige 1991). BMAC-type seals and seal impressions have been found in small
numbers in the late phase of the Indus civilization at Mohenjo-daro and Harappa (Parpola
2005d; Franke 2010), and in post-Harappan times in the southern Indus Valley (the so-called
Jhukar seals in Chanhu-daro) (Mackay 1943), Gujarat (Somnath, alias Prabhas Patan), and
even Rajasthan, where about a hundred seal impressions were collected in a pot at Gilund
(Shinde, Possehl, & Ameri 2005) (Fig. 8.5). The Gangetic copper hoards have antennae-hilted
swords similar to those coming from plundered BMAC sites of Afghanistan, which dates the
hoards to about the twentieth century BCE or later (Fig. 8.6; Fig. 4.4).

FIGURE 8.5 Expansion of the BMAC, c. 1900–1750 BCE. After Hiebert 1994: fig. 10.8. Courtesy Fred Hiebert.

FIGURE 8.6 Antennae-hilted swords: (a) from Bactria, northern Afghanistan, length 52 cm, and (b) from Fatehgarh,
Uttar Pradesh, India, length 63.5 cm. After Parpola 1988a: 285, fig. 10, based on (a) Sarianidi 1986: 198, fig. 75 and
(b) Gordon 1960: pl. XXVII b.

The Gandhara Grave culture of the Ghālegay IV phase (c. 1700–1400 BCE) in the Swat
Valley starts with the arrival of a new culture, one of the first in South Asia to show clear
evidence for domesticated horse, in motifs of painted pottery as well as in bone finds. Two
well-preserved horse skeletons were excavated in the Kātelai graveyard of Swat (surface
stratum). The black-gray, burnished pottery introduced from the beginning of phase IV is
present throughout Swat and has been connected with the pottery of the late “urban” phase of
the BMAC culture widespread in the surrounding areas at this time, for example at Dashly in
Afghanistan and Tepe Hissar and Tureng Tepe in northern Iran. In the Ghālegay IV Period
graveyard at Kherai in Indus Kohistan, the inhumed bodies are placed on their sides in a
flexed position, this burial custom being typical of the BMAC in southern Bactria. Besides the
gray-burnished ware, the Ghālegay IV Period had black-on-red painted pottery related to the
Cemetery H culture of the Punjab plains (Fig. 4.4). The horse is depicted on several shards of
this kind of ceramics at Bīr-kōṭ-ghwaṇḍai in Swāt. Very typical of the Gandhāra Graves are the
metal pins for fixing clothing; such pins are known from the BMAC necropolis at Gonur and as
imported foreign objects from graves in Syria (Chagar Bazar). The terracotta female figurines
have close parallels in the BMAC necropolis of Gonur, at Tepe Hissar III in northern Iran, and
in Syria. A compartmented metal seal of the BMAC type was recently found in Chitral.
The Gandhāra Grave culture is mainly known from graveyards belonging to the following
Ghālegay V Period, dated to c. 1400–1000 BCE, and the following Iron Age Period of Ghālegay
VI, dated to c. 1000–600 BCE. The Gandhāra Graves are rectangular pits, often surrounded by
a ring of stones. About two-thirds of them are interment burials, with the bodies in a flexed
position, men lying on their right sides and women on their left sides. Some skeletons are
anatomically intact, but there is also a large number of graves where the bones were placed in
heaps, yet with the head always in the upmost position. These inhumation graves are covered
with stones. About one-third of these later Gandhāra Graves are cremation burials. After a
transition period with approximately equal number of inhumation and cremation burials,
cremation became the dominant mode of burial in the Ghālegay V Period (the characteristic
“face urns” are discussed in chapter 11), while during the Ghālegay VI Period cremation
gradually disappeared. Both inhumation and cremation burials contain both single persons
and (quite often) two persons, usually couples that mostly seem to have been buried at
different times.
The Gandhāra Graves of the Ghālegay V period have been compared with the cemeteries
of the Vakhsh and Beshkent Valleys of southern Tajikistan. These represent a fusion of the
local north Bactrian variant of the BMAC in its late Molali-Bustān Phase (c. 1700–1400 BCE)
and of steppe nomads of the Fëdorovo Andronovo culture, which is attested at several sites in
southern Tajikistan. In both Vakhsh and Beshkent interment in a flexed position is the
predominant mode of burial, but at Beshkent secondary and cremation burials are also found.
At the necropolis of Bustān 6 in southern Uzbekistan it is quite clear that the cremation
tradition was introduced by the Fëdorovo Andronovo culture (recognized from ceramics and
terracotta human and horse figurines). This late variant of the Andronovo culture practiced
cremation widely in the steppes of Kazakhstan.
When the BMAC people moved to South Asia, they probably spoke both the original BMAC
language and Proto-Indo-Aryan, which may have been influenced by the original BMAC
language. This situation should be reflected in the available linguistic evidence. But how could
one know anything about the BMAC language which had no written documents, and how
could one judge its influence on Proto-Indo-Aryan? Alexander Lubotsky (2001) has collected
“all Sanskrit etyma which have Iranian correspondences, but lack clear cognates outside Indo-
Iranian,” numbering about 120. He concludes that some of these words may have been
borrowed from Uralic languages, but that the major portion is likely to have come from the
language of the BMAC, which he considers non-Indo-European. One criterion is the non-Proto-
Indo-European structure of many of these loanwords. He refers to trisyllabic words in which
the first and third syllable have a short vowel and the medial syllable a long vowel (∪−∪).
According to Lubotsky, “the Indo-Aryans were presumably the first who came in contact with
foreign tribes” in Central Asia “and sometimes ‘passed on’ loanwords to the Iranians.” He also
points out that these Indo-Iranian loanwords, likely to have been borrowed mainly from the
language of the BMAC, are phonologically and morphologically similar to many loanwords
found in Sanskrit alone. Indeed, a large part of the 383 “foreign words” from the Rigveda
listed by F. B. J. Kuiper (1991) probably comes from the BMAC language.
I would like to note some characteristics of this hypothetical BMAC language not specified
by Lubotsky. The vocalic ṛ occurs in the first syllable in several of Lubotsky’s loanwords, such
as ṛbīsa-, “oven,” and śṛgāla-, “jackal” (both having the trisyllabic structure with long middle
syllable), *gṛda-, “penis” (here Lubotsky notes the two unaspirated voiced stops in the same
word, a root structure impossible for an IE word); moreover, the vocalic ṛ in the initial syllable
is anomalously accented in *ṛši -, “seer,” and this is the case too in several proper names
recorded in the Rigveda as names of the Dāsas, Dasyus, and Paṇis (see chapter 9), namely
Dṛbhīka, Sṛbinda, Bṛsaya (with the anomalous dental s after ṛ, also found in later Vedic bṛsī-,
“roll of twisted grass, cushion”); an unaccented ṛ in the first syllable is found in the Dāsa
proper names Bṛbu (but accented in bṛbūka-, “thick”) and Pipru Mṛgaya (from *mṛga-, which is
listed as a likely loanword). Bṛsaya is particularly important, because it can be located in the
regions west of the Indus Valley with the help of Greek sources (see chapter 9).
Another peculiarity in the above BMAC/Dāsa words is the phoneme b (cf. ṛbīsa-, Sṛbinda,
Bṛbu), which in Indo-Aryan only rarely goes back to PIE *b, but occurs in many words of
probably foreign origin. Among these words are further Dāsa names, Balbūtha, and Ilībiśa.
These again have the phoneme l not found in the main dialect of the Rigveda (chapter 12).
Both b and l occur in kilbiṣa-, “injury, transgression, injustice, sin, guilt.”
Some probable BMAC loanwords occur outside the Indo-Iranian material discussed by
Lubotsky. For example, the following Sanskrit words have the trisyllabic structure with a long
middle syllable as well as the phoneme l:

• kulāla-/kaulāla-, “potter,” “potter’s ware”; this word is not attested in the Rigveda but does
appear in the Yajurveda (Vājasaneyi-Saṃhitā), with cognates in modern Dardic languages of
the northwest (Pashai and Kashmiri), as well as in the modern Iranian language Parachi.
Both the BMAC and the Indo-Aryan speakers coming from the northern steppes to the
BMAC produced pottery. But whereas the BMAC potters threw their pots on a wheel, the
people of the northern steppes made their new pots using inverted cloth-covered old pots as
molds upon which the wet clay for the new pot was patted. All clay vessels needed in the
Vedic ritual had to be made by hand only by an Āryan. In the Vedic texts, a potter who uses
the wheel is described as being a Śūdra, of low social standing; wheel-turned pottery is said
to belong to the worshippers of demons (asura).
• kīlāla-, “some kind of milk product” (RV 10,91,14, kīlāla-pa-, “drinking kīlāla”; kīlāla-
appears seven times in the Atharvaveda), Classical Sanskrit (Suśruta) kilāṭa-; of modern
languages, the etymon is found almost exclusively in Nuristani and Dardic languages, where
it means mostly “fresh cheese”; in addition comes only Sindhi kiroṭu, kirūṭu, “cheese made
from skimmed milk.” Although the word is found also in Burushaski as kīlāy and in Dravidian
(Tamil kilāan, “curds,” etc.), a BMAC origin seems likely to me.
• palālī- (Atharvaveda) / palāla-, “stalk, straw.”

The BMAC people expanded not only eastwards to South Asia and south to Afghanistan and
eastern Iran, but also west, to northern Iran, Syria, and Anatolia. We shall now examine this
western extension of the BMAC. Locally manufactured seals from Anatolia and Syria testify
that the horse-chariot reached these areas in the twentieth century BCE. According to the
cuneiform archives found at the Assyrian trading colony Kanesh, present Kültepe in
Cappadocia, the Assyrian merchants settled there imported textiles and enormous amounts of
tin from Assur, bartering these goods for silver and gold in Anatolia. This trade, which
flourished for seventy years in 1920–1850 BC, affected all areas involved (Brisch & Bartl
1995). The source of the tin was in the east, probably in southern Central Asia (Cierny,
Stöllner, & Weisgerber 2005).
The trade contacts are reflected in the seals of the various regions. From about 1850 BC
Syria was in close contact with Egypt, and many Egyptian motifs started appearing in the
Syrian glyptics (Eder 1996). Among them is the “twist,” which appears on locally made but
clearly Syrian-inspired stamp-cylinders and their impressions excavated at BMAC sites in
Margiana, such as at Taip-depe (Collon 1987:142–143). Conversely the two-humped Bactrian
camel of Central Asia, which figures on BMAC seals and ax-scepters, is attested also on Syrian
seals.
The BMAC expanded to the Gorgan region near the southeastern shore of the Caspian Sea
in northern Iran, where it is represented by the Tepe Hissar IIIC culture (c. 1900–1750 BCE).
The large number of its bronze weapons suggests that this rich culture was ruled by a warring
aristocracy (Schmidt 1937). An alabaster cylinder seal excavated from the IIIB phase at Tepe
Hissar depicts a two-wheeled horse-drawn chariot (Fig. 8.7). Roman Ghirshman (1977)
suggested that the trumpets made of gold, silver, and copper, now also found in Bactria and
Margiana—altogether some fifty trumpets are known (Lawergren 2003)—were used for giving
signals in both training chariot horses (as depicted in an Egyptian wall painting) and for
directing chariots in battles (Fig. 8.8). Ghirshman also proposed that the Indo-Aryan-speaking
nobles of the Mitanni kingdom of Syria came from the Tepe Hissar IIIC culture.
FIGURE 8.7 An alabaster cylinder seal depicting a horse-drawn chariot with cross-bar wheels. Excavated at Tepe
Hissar, Gorgan, northeastern Iran, phase III B. After Schmidt 1937:198, fig. 118. Courtesy University of
Pennsylvania Museum of Archaeology and Anthropology.

FIGURE 8.8 Trumpets and the horse-chariot. (a) Golden and silver signal trumpets from Tepe Hissar III C. After
Mallowan 1965:124, fig. 139, based on Schmidt 1937:209, fig. 121. Courtesy University of Pennsylvania Museum of
Archaeology and Anthropology. (b) An order being given to the horses with a trumpet when the pharaoh inspects the
royal stables. Relief in the temple of Ramses III (ruled 1194–1163 BCE) at Medinet Habu, First court, south wall,
west end. After Nelson et al. 1932: fig. 109, upper right corner. Courtesy of the Oriental Institute of the University of
Chicago. On the trumpet in Egypt see Hickmann 1946.

The beginning of Indo-Aryan rule in Mitanni can be narrowed down to the sixteenth
century BCE. Cuyler Young (1985) has plausibly linked the arrival of these Indo-Aryans in Syria
with the sudden appearance of the Early West Iranian Gray Ware in great quantities all along
the Elburz Mountains, in Azerbaijan, and around Lake Urmia around 1500 BCE. Young sees
this ceramic, which is new to western Iran, as an evolved form of the Gorgan Gray Ware of the
Tepe Hissar IIIC phase, which continued in an impoverished form at Tureng Tepe until about
1600 BCE. Young’s reconstruction provides a chronologically and typologically acceptable link
between the Mitanni Aryans, the Tepe Hissar III culture of Gorgan, and the BMAC.
In the sixteenth century, the Syrian kingdom of the Mitanni or Mittani dramatically
overpowered Assyria and became one of the greatest powers in West Asia, rivalling the
Hittites of Anatolia and the Egyptians (Fig. 8.9). While the population of Mitanni consisted of
local Hurrian people (whose language seems to be a member of the Caucasian family), its
ruling dynasty had Indo-Aryan names. Exactly how this Proto-Indo-Aryan-speaking aristocracy
took control is unclear. A cuneiform letter (L.87–887) recently recovered from the archives of
Tall Leilan in the Khabur Valley, as early as about 1760 BCE, speaks of maryanni soldiers that
served Hurrian rulers as an elite troop. This may have been a vanguard of Proto-Indo-Aryan-
speaking mercenaries. More might soon become known from the ongoing excavations at Tall
Fakhariya and Tall Hamidiya, since there are good reasons to suspect that these mounds
contain the ancient Mitanni capitals Vaššukanni and Taidu, which are expected to have royal
archives of cuneiform tablets (Eidem 2014). In any case, the Proto-Indo-Aryan takeover
resulted (as in the PIE takeover of the Tripolye culture; see chapter 6) in the vitalization of the
Hurrian power, though we of course reject earlier racist interpretations. It is clear that the
new Indo-Aryan-speaking rulers totally adjusted to the local culture, for instance, they
communicated with other rulers in the Akkadian language, written in cuneiform by the local
scribes.

FIGURE 8.9 The Mitanni empire and its neighbors. Map © Anna Kurvinen and Asko Parpola.

The Mitanni dynasty introduced a mastery of warfare with horse-drawn chariots, although
both the horse and the chariot had come to West Asia (including Anatolia) centuries earlier
(Fig. 8.10). At Alalaḥ there is even evidence for the use of horse-drawn chariots in warfare in
the seventeenth century BCE, but chariotry did not become a regular component of the West
Asian and Egyptian armies before the fifteenth century BCE. That the impetus came from
Mitanni is suggested by the associated terminology that spread all over West Asia and Egypt.
Thus, the Indo-Aryan elite was denoted after the fifteenth century BCE by the term maryanni,
which consists of the Indo-Aryan word márya- plus the Hurrian derivational suffix -nni (the
Hurrian plural is maryannina). The word also occurs in the Semitized form maryannu, plural
maryannūma (in West Semitic), or mariyannū, plural maryannūtu (in Akkadian)—the actual
spelling in cuneiform varies (for example, ma-ri-ya-an-nu). In the Egyptian language the word
is maryana, spelled in various hieroglyphic texts as ma-ra-ya-na, m-ra-ya-na, ma3-ra-ya-n, etc.
Everywhere this term denotes a “nobleman” of a high status, close to the local ruler, and in
the possession of a war chariot (von Dassow 2008). In the Rigveda, marya- means “young man,
young warrior, suitor, lover, husband,” corresponding to Avestan mairiia-, “rascal” (this
pejorative sense refers to those who opposed the Zoroastrian religion). The early history of
this word is illustrated by the fact that it was borrowed from Proto-Indo-Aryan into the
ancestor of a Uralic language, Mari (previously called Cheremiss), a language now spoken by
people who live just north of the Mid-Volga in Russia; there mari (< *marya-) means “man,
husband” and is now the ethnic name of this people.
FIGURE 8.10 A chariot warrior and a charioteer in battle. Impression of a cylinder seal from Alalaḥ V of early
Mitanni rule (c. 1525–1460 BCE). After Collon 1987:160, no. 731. Courtesy Dominique Collon.

The royal archives at El Amarna in Egypt preserved 382 letters of the fourteenth century
BCE where the rulers of Egypt, Babylonia, Assyria, Mitanni, Arzawa (to the west of Cilicia),
Alašia (Cyprus), and Hatti (the Hittite country) correspond with each other more or less on the
basis of equality, addressing each other as “brothers.” The great importance attached to the
chariot by the kings is evident from the fact that it is, along with horses, regularly included in
the stereotyped salutation formula. For example, in a letter from Tušratta, the king of Mitanni,
to Nimmureya (i.e., Amenophis III), the king of Egypt (ruled c. 1390–1352), the formula is as
follows: “For me all goes well. For you may all go well. For your household, for my daughter
[the wife of Nimmureya], for the rest of your wives, for your sons, for your chariots, for your
horses, for your warriors, for your country, and for whatever else belongs to you, may all go
very, very well” (El Amarna 19 = BM 29791, translated Moran 1992:43).
Tušratta also details many maninnu necklaces (meant to decorate chariot horses) that he
has sent to his son-in-law Nimmureya, the king of Egypt, among many other jpgts on the
occasion of his wedding to Tušratta’s daughter. For instance: “1 maninnu-necklace, cut: 37
genuine lapis lazuli stones, 39 (pieces of) gold leaf; the centerpiece a genuine ḥulalu-stone
mounted on gold” (El Amarna 25, translated Moran 1992:73).
The Akkadian word maninnu (with its Akkadian suffix -nnu) contained in this text is one of
the few Mitanni Indo-Aryan words that have been etymologically identified: it corresponds to
Vedic maṇí-, “necklace” (Rigveda 1,122,14, maṇi-grīvá-, “wearing a maṇi-necklace on one’s
neck”). Some other identified Mitanni Indo-Aryan words are maganni = Vedic maghá-,
“present, jpgt” (with the Hurrian suffix -nni); martianni = Vedic mártiya-, “man, warrior”; and
(with the Akkadian suffix -nnu) mištannu, “bounty” = Proto-Indo-Aryan *miždha- > Vedic
mīḍhá-, “booty.” A marriage document from Alalaḥ IV contains Proto-Indo-Aryan vadūranni,
“bride-price” = Sanskrit *vadhū-rā-.
The royal archive of the Hittite kings was excavated at Boğasköy (ancient Hattuša) in
Turkey in 1906–1907. It included a manual for training chariot horses. (Raulwing 2005
[2006].) The text consists of four cuneiform tablets written in the Hittite language, which are a
thirteenth-century copy of a lost original written in the fifteenth century. It contains a training
program for the chariot horses stretching over at least 184 days, with some differences in the
autumn, winter, and spring: it specifies how many days a given training unit lasts; at what
time of the day the exercise is performed (the time frame extending from early morning to
midnight); how many rounds in the stadium the horses are made to run; how exactly defined
rations of fodder and water are to be given and sometimes withheld; how the horses are kept
in the stables, massaged, greased, covered with a blanket, and if necessary fitted with a
muzzle; and how they are let out to graze on pasture. The manual opens with the words: “Thus
(speaks) Kikkuli (Ki-ik-ku-li), a horse trainer from the country of Mittanni (Mi-it-ta-an-ni).” The
term used for “horse trainer” is aššuššanni = Indo-Aryan aśva-śā-ḥ, “one who tires or exhausts
[śam-] the horse (during training to use up all its strength).” The text contains some other
technical terms in Proto-Indo-Aryan:

va-ša-an-na and Indo-Aryan genitive case va-ša-an-na-ša-ya, “racecourse, stadium” =

Proto-Indo-Aryan *vaźhana-, gen. *vaźhanasya > Sanskrit vahana-, gen. vahanasya, “the
act of driving”
va-ar-ta-an-zi, “to turn” = Indo-Aryan vart-, “to turn”
va-ar-ta-an-na, “turn” or “round” (in the following compounds) = Indo-Aryan vartana-, “the
act of turning; wheel-track, path”; this word is compounded with five numerals:
a-i-ka-va-ar-ta-an-na = Proto-Indo-Aryan *aika-, “one” > Sanskrit eka- (in contrast, “one”
was in Proto-Iranian *aiva-)
ti-e-ra-va-ar-ta-an-na = Indo-Aryan tri-, “three”
pa-an-za-va-ar-ta-an-na = Indo-Aryan pañca-, “five”
ša-at-ta-va-ar-ta-an-na = Old Indo-Aryan sapta- (whence Middle-Indo-Aryan satta-), “seven”
na-a-va-ar-ta-an-na and Hittite dative-locative case na-va-ar-ta-an-ni (with haplology, i.e.,
omission of one of two identical va syllables in *navavartana-) = Indo-Aryan nava-, “nine”

Cuneiform texts from Mitanni-ruled Nuzi in Syria contain Indo-Aryan color terms connected
with horses (with the Akkadian suffix -nnu, and, importantly, with the rhotacism *l > r
characteristic of the original dialect of the Rigveda, see chapter 12):

paprunnu / babrunnu = Vedic Sanskrit babhrú-, “brown”

pinkarannu = Vedic Sanskrit piṅgalá-, “reddish brown”
parittannu = Vedic Sanskrit palitá-, “gray”

“A man of Uniga, son of Kukki,” is mentioned as “a maker of an a-šu-wa-ni-in-ni chariot” in a

tablet from Alalaḥ; the word written as a-šu-wa-ni and aš-wa-ni (plus the Hurrian suffix -nni)
almost certainly contains the Indo-Aryan word aśva-, “horse.”
The Alalaḥ tablet mentions a chariot with six-spoked wheels, where the term for “chariot”
is the Hurrian word aradiyanni, probably based on Indo-Aryan rathya-, “chariot.”
The kings of Mitanni bear Indo-Aryan names, although it has not been possible to provide
all with an etymology.

Kirta (Ki-ir-ta) (father of:)

Šuttarna (Šu-ut-tar-na) I (son of Kirta)
Parsatatar (Par-sa-ta-tar) (father of Sauštatar)
Saušsatat(t)ar, Sauštatar (Sa-uš-sa-ta-(at-)tar in Alalaḥ, Sa-uš-ta-at-tar in Nuzi) (son of
Parsatatar) ruled c. 1450–1425
Artatama (Ar-ta-ta-a-ma) (son or descendant of Sauštatar) = Vedic ṛtá-dhāman-, “having
Law as his abode” (for Arta-, cf. also the Old Persian names in Greek sources, such as
Artabazos, Artaphernes)
Šuttarna (Šu-ut-tar-na) II
Artaššumara (Ar-ta-aš-šu-ma-ra) (son of Šuttarna) = *Ṛta-smara-
Tušratta (Tu-uš-rat-ta, Tu-iš-e-rat-ta, Tu-uš-e-rat-ta) (younger son of Šuttarna) =
Proto-Indo-Aryan *Tvaišaratha > Vedic tveṣá-ratha-, “having a rushing chariot”
Parattarna (Par-ra-at-tar-na) I
Šattivaza (KUR-ti-ú-a(z)-za) = *Sāti-vāja-, “obtaining booty” (cf. Vedic vāja-sāti-,
vājasya sāti-)

Two previously unexplained royal Mitanni names, Sauštattar and Parsatattar, can be
explained, I believe, from the Indo-Aryan terms savyasthātar and prasthātar, which both
denote “chariot-warrior” (Fig. 8.11). Mayrhofer (1974) has mentioned *Su-sthātar-, “provided
with good horse drivers,” as the most likely earlier proposal, noting that such a compound
would be semantically suitable as the name of a nobleman; the problem is that this compound
is not attested in Vedic or Avestan. Mayrhofer has suggested reading Parsatattar as Puraḥ-
sthātar-, “one who stands in front, leader,” attested in Rigveda 8,46,13. As Sauštattar’s
descendants’ names are all connected with Vedic chariotry, I propose *Savya-šthātar-, “one
who stands on the left (in the chariot), chariot-warrior” (as opposed to the charioteer, who
stands on the right in the chariot, cf. TB 1,7,9,1 dvau savyeṣṭha-sārathī; other parallels in
Vedic texts are savya-ṣṭhā- in AV 8,8,23, and savya-ṣṭha- in ŚB 5,2,4,9; 5,3,1,8; 5,4,3,17–18).
Parsatattar could be *Pra-sthātar-, “chariot-fighter who stands in front.” In the Veda, the
“adhvaryu-pair” of priests, that is, the Adhvaryu and the Pratiprasthātar, are equated with the
two Aśvins, the divine charioteers (Taittirīya-Āraṇyaka 3,3 aśvinādhvaryū). These two priests
are actually in charge of the gharma ritual, in which hot milk is offered to the Aśvins (chapter
11). The word adhvaryu- literally means “one connected with the road(s)” and may originally
have meant “charioteer,” who keeps the chariot on the road. Thus the word prati-prasthātar-
may originally have denoted the “chariot-warrior who stands in front opposite (to the
charioteer).” The unexplained name of Kirta, Sauštattar’s father, might stand for *kṛta-,
“praised, famed,” from the Indo-Iranian root kar- carkarti, “to praise,” for Sanskrit
grammarians mention kṛta- besides kīrṇa- as the past participle of the homonymous root kar-,
“to kill.”
FIGURE 8.11 Impression of the seal of “Sauštattar, son of Parsatattar, King of Mitanni” (sa-uš-ta-at-tar DUMU par-
sata-tar LUGAL ma-i-ta-ni), from c. 1420 BCE, preserved on clay tablets found at Nuzi and at Tell Brak. After Stein
1994:297, fig. 2. Courtesy Diana L. Stein.

The royal archive of the Hittite kings at Boğasköy also contained a peace treaty between the
Hittite king Suppiluliuma I and the Mitanni king Šattivaza. This contract, made around 1350
BCE, reduced the Mitanni empire to a vassal kingdom. Four Indo-Aryan deities are asked to
be divine witness of the treaty: Mitra (Mi-it-ra-), Varuṇa (A-ru-na-), Indra (In-da-ra-), and the
Nāsatya twins (Na-aš-ša-at-ti-ya-). After the text has mentioned about a hundred divine
witnesses on the Hittite side follow two dozen divine witnesses of the Mitanni king, the Indo-
Aryan gods being in the middle. These same deities are mentioned, in the same order, in
Rigveda 10,125,1: aham mitrāvaruṇobhā bibharmy aham indrāgnī aham aśvinobhā; the only
difference is that instead of Indra in the treaty, Indra-and-Agni are mentioned here, and the
divine charioteer twins are mentioned by a different name; yet the name Nāsatya, “rescuer,
savior” is very commonly used in the Rigveda for the Aśvins (Thieme 1960).
In addition to the Indo-Aryan gods mentioned in this treaty come two further Indo-Aryan
deities, the fire-god Agni (A-ak-ni-iš), mentioned in Hittite ritual texts, and the sun-god Sūrya
(Šu-ri-ya-áš), equated with the Akkadian sun-god Šamaš in a glossarial cuneiform tablet
from Babylonia during the Kassite rule (c. 1570–1155 BCE). The Kassite rulers were a
minority elite (like the Indo-Aryans of Mitanni) whose homeland since the eighteenth
century BCE seems to have been Luristan in the Zagros Mountains east of Babylonia; their
language is without known relatives.
Besides the names of the Mitanni kings, a number of other Proto-Indo-Aryan personal names
are known, many of them being maryanni. Several names are compounds ending in atti, that
is, *atthi, which corresponds to Vedic atithi-, “guest.” Five names formed with atithi as the
last part, “having X as his guest,” usually with a god’s name as the first part, occur in the
hymns of the Kāṇva poets of the Rigveda: Medhyātithi, Medhātithi, Nīpātithi, Mitrātithi, and
Devātithi.

Aššuratti, Ašuratti = *Asurātthi

Biriyatti, Biryatti = *Priyātthi
Intaratti = Indrātthi
Maryatti = *Maryātthi
Mitaratti = Vedic Mitrātithi, as a proper name in the Rigveda
Šauššatti
Šuwatti, Šuwatiti = *Suvatthi
Šuryatti = Sūryātthi
Paratti = *Prātthi

Other Proto-Indo-Aryan proper names are

Ar-ta-am-na = *Ṛta-mna-, “having the mind of Law” (cf. Vedic ṛtásya . . . mánas and the
Persian name Artamenēs in Greek sources)
Aššuzzana = *Aśva-canas- (cf. Aspa-canah- as a personal name in Old Persian)
Bi-ri-ya-aš-šu-va = Indo-Aryan *Priyāśva-
Biridašva = *Prītāśva-
Bi-ir-ya-ma-aš-da = Proto-Indo-Aryan *Priya-mazdha- > Vedic Priyá-medha-, as a personal
name in the Rigveda
Indaruta (In-tar-ú-da, En-dar-ú-ta) = Vedic Indrotá- (= Indra-ūtá-)
Šattavaza = *Sāta-vāja- (cf. Sāti-vāja as the name of a Mitanni king)
Šu-ba-an-du = Vedic Sanskrit Subandhu-, as a personal name in the Rigveda
Zantarmiyašta = Proto-Indo-Aryan *-miyazdha- > Vedic -miyédha-, “sacrificial meal”
Zitra = Vedic Citrá-, ‘brilliant, radiant’, occurring as a personal name in the Rigveda.

To conclude, the culture and proper names of the Mitanni Indo-Aryans were dominated by the
horse and the chariot. Indeed, the Mitanni Indo-Aryans seem to have been the prime motors in
the introduction of chariotry into the West Asian warfare around 1500 BCE. This raises the
question whether or not Proto-Indo-Aryan speakers were involved in the Hyksos conquest of
Lower Egypt and the introduction of chariotry there. The Mitanni rule came to an end in the
thirteenth century BCE, by which time there is reason to assume the Proto-Indo-Aryan
language had ceased to be spoken in West Asia. Aryan speakers are not mentioned in
cuneiform documents until the middle of the ninth century BCE, when the West Iranian-
speaking Medians and Persians emerge (see chapter 7). The BMAC settlements in southern
Central Asia were for many centuries surrounded by steppe pastoralists, and so Proto-Indo-
Aryan, being constantly replenished, could prevail there; but in Syria it was assimilated to
local languages, undoubtedly because the Mitanni Indo-Aryans, after their takeover of power,
eventually became isolated from their linguistic relatives further east.
Linguistically, Mitanni Indo-Aryan (1500-1300 BCE) represents an older stage of
development than Rigvedic Indo-Aryan. Though the available Mitanni vocabulary is very
limited, four cases can be cited where the Mitanni word equals the Proto-Indo-Aryan
reconstruction, while a subsequent sound change has taken place in the Rigvedic counterpart
(Mayrhofer 1966:18–20):

Table 8.1
9

The Rigvedic Indo-Aryans and the Dāsas

When did the Indo-Aryan speakers associated with the Rigvedic hymns move from Central
Asia to South Asia? This has been, and continues to be, one of the chief problems of Vedic
research and Indian history as a whole. If we ignore the impossible hypothesis that the Vedic
Aryans were indigenous to South Asia, and some implausibly ancient dates for their existence,
we find that current scholarly estimates for the date of their migration vary between about
2000 and 1000 BCE. Often, comparison is made with the fragments of Mitanni Indo-Aryan
discussed in chapter 8 and the immigration is dated at around 1400–1200 BCE. My own
approach relies on studying the ethnic and linguistic identity of the people known as Dāsa,
Dasyu, and Paṇi, against whom the Rigvedic Aryans were fighting.
First, some basic data about the Rigveda are required. The Rigvedic hymns were collected
from the priestly clans, who had composed them, and systematized at the time of the Kuru
kingdom around 1000 BCE (see chapter 13). The Rigveda is divided into ten books (1–10)
called “circles” (maṇḍala). The so-called “family books” (2–7) are each ascribed by later
tradition (confirmed by internal indications) to a particular family of poets (kavi) or sages (ṛṣi).
Each family descends from a sage: Gṛtsamada (Book 2), Viśvāmitra (3), Vāmadeva (4), Atri (5),
Bharadvāja (6), and Vasiṣṭha (7). Within each book, the hymns are arranged according to three
hierarchical criteria: by deity (first Agni, then Indra, then other gods, in the order of the
frequency of hymns addressed to them), by meter and by the number of verses in the hymn (in
decreasing order).
The family books represent the oldest stratum of the Rigveda. Almost equally old is the
next layer of hymns, the second half (hymns 51–191) of Book 1, where the hymns are arranged
according to the same principles: first according to their authors (nine poets or families), then
gods, meters, and the number of verses. The next addition is Book 8, composed by poets of the
Kaṇva (hymns 1–66) and Aṅgiras (hymns 67–103) families. The first half of Book 1 (hymns 1–
50), the subsequent addition, is closely connected with Book 8, most of its hymns being
ascribed to the Kaṇva and Aṅgiras families; it shares with Book 8 a predilection for tristichs,
song units consisting of three verses, and some other characteristics. Book 9 was created by
extracting all the hymns addressed to Soma (the deified sacrificial drink) from Books 1–8, for
liturgical purposes. Book 10 is by far the most recent addition to the Rigveda, the first half
(hymns 1–84) being older than the second (hymns 85–191): the total number of hymns is
similar to that of Book 1. Finally, in addition to the Rigveda proper, there are a few
“supplementary verses” (khila).
The Rigveda mentions names of countries, mountains, and rivers. The river names are
especially important as their referents can often be identified with certainty. Particularly
prominent is the Indus River, along with its main tributaries, which are intended in the phrase
sapta sindhavaḥ, “the seven streams,” much repeated in the Rigveda. Hymn 75 of Book 10, “In
praise of streams,” lists all the major rivers of the northwest Indian plains, particularly of the
Indus system. In verses 1–4, Sindhu (= Indus); in verse 5, Gaṅgā (= Ganges), Yamunā (=
Jumna), Sarasvatī (= Sarsuti and Ghaggar-Hakra), Śutudrī (= Sutlej), Paruṣṇī (= Irāvatī >
Ravi), Marudvṛdhā with Asiknī (= Chenab), Vitastā (= Jhelum), and Ārjīkīyā with Suṣomā (=?
Sohan); and in verse 6, the western tributaries of the Indus are listed as Tṛṣṭāmā, Susartu,
Rasā, Śvetyā, Kubhā (= Kabul), Gomatī (= Gomal), Krumu (= Kurum, Kurram), and Mehatnū.
Suvāstu (= Swat) is missing from this list, but is mentioned in RV 8,19,37; it may be one of the
listed but unidentified western tributaries of the Indus. The geographical names connect the
Rigveda mostly with the northwest and the Punjab, whereas eastern references occur rarely in
the Rigveda (Gaṅgā only here, in 10,75,5; Yamunā also in 5,52,17 and 7,18,19). This is in
contrast with later Vedic literature, where the geographical horizon widens eastwards and
southwards.
It seems that several other names of rivers (or river systems) have, or had, referents in
Central Asia, Afghanistan, and Baluchistan; and that some of these names were transferred to
the Indian subcontinent by early Rigvedic immigrants. They include:

• Rasā (= Avestan Raŋhā =?Oxus / Amu Darya; Rasā becoming *Rahā in Iranian is the most
likely source for Rhā, which Ammianus Marcellinus has recorded as the ancient name of the
Volga, one of the main rivers of the assumed Indo-Iranian homeland) (see chapter 7);
• Sarayū (=?Avestan Haroiva, “name of a country” = Old Persian Haraiva, “province of the
Persian empire,” Aria, “Ariana” = the Herat province of Afghanistan);
• Sarasvatī (= Avestan Haraxvaiti, “name of a country” = Old Persian Harahuvati, “province of
the Persian empire, Arachosia” = the Kandahar province of Afghanistan, named after its
principal river, present-day Arghandab + Helmand; later Sarsuti and Ghaggar-Hakra in
India);
• Yavyavatī (=?Zhob in northern Baluchistan);
• Bhalāna, “name of a tribe” (= ?Bolan River in Baluchistan).

The Rigveda mentions by name some thirty Aryan tribes and clans. A term meaning “five
peoples” is used throughout the Rigveda to refer to the major tribes, of which four are
regularly paired: Yadu with Turvaśa, Anu with Druhyu. These four tribes seem to have been
among the first wave of Indo-Aryan-speaking immigrants to the northwest of the subcontinent
from Afghanistan. A fifth tribe, Pūru, together with its ally or subtribe Bharata, appears to
have arrived later, again from Afghanistan, overpowering the earlier tribes. The principal
agents in the Rigvedic hymns are Bharata and Pūru dynasties of three to five generations
(Bharata: Atithigva > Divodāsa > Sudās; Pūru: Girikṣit > Purukutsa > Trasadasyu > Tṛkṣi), so
the family books seem to reflect their times. The latest Bharata or Pūru king mentioned in
Books 2–7 is Trasadasyu’s son, Tṛkṣi, so probably the family books were collected fairly soon
after his reign (Witzel 1995a,b).
The Pūru king Purukutsa and the Bharata king Divodāsa seem to have led this immigration
over the Hindukush mountains to the Indus Valley. In the celebrated “battle of ten kings”
(Book 7, hymns 18, 33, and 83) on the river Paruṣṇī (modern Ravi), the Bharata king Sudās,
son or grandson of Divodāsa, defeated the Pūru king (probably Trasadasyu, the son of
Purukutsa) and his many allies, becoming the ruler of the Punjab. The first Kuru king
mentioned in the Rigveda is Kuruśravaṇa, a descendant of Trasadasyu, suggesting that the
Pūrus moved to the area later called Kurukṣetra (the Kuru kings are discussed in chapter 13).
Alexander the Great defeated King Pōros, whose Middle Indo-Aryan name Pora comes from
Sanskrit Paurava, “descendant of Pūru”; his realm lay between the Jhelum and Ravi rivers.
Indra and his protégés, particularly kings Divodāsa and Purukutsa, are said to have
vanquished “black-skinned” enemies called Dāsa, Dasyu, and Paṇi, who neither worshipped
Indra nor sacrificed soma but had their own observances (vrata) and were rich in cattle. Indra
and Agni helped the Aryans to destroy the numerous enemy strongholds, in particular those of
Dāsa Śambara situated among mountains. Though human foes are undoubtedly meant, many
of these enemies are presented as demons. The following anthology of verses about the Dāsas
and Dasyus, taken from a new translation of the entire Rigveda by Stephanie Jamison and Joel
Brereton (2014), will give the reader some idea of the evidence for the hostile nature of these
peoples.

1,51,5 With your wiles you blew away the wily ones, who, according to their own customs,
poured (their offering) “on the shoulder.”
You broke through the strongholds of Pipru, O you of manly mind; you helped Ṛjiśvan
through in the smashing of Dasyus.
6 You helped Kutsa in the smashing of Śuṣṇa, and you made Śambara subject to Atithigva.
With your foot you trampled down Arbuda, though he was great. Indeed, from long ago you
were born to smash Dasyus.
8 Distinguish between the Āryas and those who are Dasyus. Chastising those who follow no
commandment, make them subject to the man who provides ritual grass.
Become the potent inciter of the sacrificer. I take pleasure in all these (deeds) of yours at our
joint revelries.

1,103,3 He who by nature provides support, being trusted for his power, roved widely,
splitting
apart the Dāsa strongholds.
As knowing one, O possessor of the mace, cast your missile at the Dasyu; strengthen Ārya
might and brilliance, O Indra.
2,12,4 By whom all these exploits have been done: who has put the Dāsa tribe below and
hidden away,
who has taken the riches of the stranger, as a winning gambling champion does the wager—
he, O peoples, is Indra.
10 Who has struck with his arrow those constantly creating for themselves great guilt, the
unthinking ones;
who does not concede arrogance to the arrogant man; who is the smasher of the Dasyu— he,
O peoples, is Indra.
11 Who in the fortieth autumn discovered Śambara dwelling in the mountains;
who smashed the serpent displaying its strength, the son of Dānu, (thereby) lying (dead)—he,
O peoples, is Indra.

2,20, 6 The god famed as Indra by name, he the most wondrous, rose upright for Manu.
The able, independent one carried away the Dāsa Arśasāna’s very own head.
7 Smasher of Vṛtra, splitter of fortresses, Indra razed the Dāsa (fortresses) with their dark
wombs.
He gave birth to the earth and the waters for Manu. In every way, he makes the sacrificer’s
laud powerful.

4,16,10 Drive here to the home (of Uśanā Kāvya) with your Dasyu-smashing mind. In
companionship with you, Kutsa will become eager.
Do you two, having the same form, sit down each in his own womb. She is trying to distinguish
between you two—she is a woman who distinguishes the truth.
12 For Kutsa you laid low insatiable Śuṣṇa, who brings bad harvest, with his thousands, before
the day’s first meal.
Immediately crush out the Dasyus with (the weapon) that is Kutsa, and then tear off the wheel
of the Sun at the moment of encounter.
13 You subjugated Pipru Mṛgaya, swollen with power, to Ṛjiśvan, the son of Vidathin.
You scattered down the dark fifty thousand. You shredded their fortresses, like worn-out age a
cloak—

5,34,6 Very energetic in the clash, affixing the wheel (to the chariot?), he is antagonistic to the
non-presser, but strengthener of the presser.
Indra is the dominator of all, spreading fear; the Ārya leads the Dāsa as he wishes.

6,20,10 Might we win anew through your help, Indra. The Pūrus start up the praise with this
(hymn) along with sacrifices.
When he split the seven autumnal strongholds, their shelter, he smote the Dāsa (clans), doing
his best for Purukutsa.

6,26,5 You made that hymn (endowed) with might, Indra, so that you could tear out hundreds
and thousands (of goods), o champion.
You struck the Dāsa Śambara down from the mountain and furthered Divodāsa with glittering
help.

6,47,21 Day after day he drove off from their seat the other half, the black kindred all of the
same appearance.
The bull smashed the two Dāsas, mercenaries, Varcin and Śambara, at the moated place.

7,6,3 Down with those of no intelligence, those tying in knots, those of disdainful words: the
Paṇis, not giving hospitality, not giving strength, not giving sacrifices.
Onward and onward Agni has pursued those Dasyus. The first has made the last to be without
sacrifices.

7,19,2 Just you, O Indra, helped Kutsa, while seeking fame for yourself with your own body in
the clash,
when for him you weakened the Dāsa Śuṣṇa bringing bad harvest, doing your best for Arjuna’s
offspring.
4 You—whose mind is inclined towards men in their pursuit of the gods—along with men you
smash many obstacles, you of the fallow bays;
you put to sleep the Dasyu Cumuri and Dhuni, easy to smash, for Dabhīti.
5 Yours are these exploits, you with the mace in hand—that nine and ninety fortifications at
once
along with the hundredth you worked to the end, in bringing them to rest [= collapse]. You
smashed Vṛtra, and moreover Namuci you smashed.

8,70,10 You are the one who seeks the truth for us, Indra. You find no satisfaction in him who
reviles you.
Gird yourself in between your thighs, O you of mighty manliness. Jab down the Dāsa with your
blows.
11 The man who follows other commandments, who is no son of Manu, no sacrificer, no
devotee of the gods—
him should your own comrade, the mountain, send tumbling down; the mountain (should send
down) the Dasyu for easy smiting.

10,22,8 The Dasyu of non-deeds, of non-thought, the non-man whose commandments are
other, is against us.
You smasher of non-allies, humble the weapon of this Dāsa.

10,138,3 The sun unhitched his chariot in the middle of heaven. The Ārya found a match for
the Dāsa.
The firm fortifications of the crafty lord Pipru did Indra throw open, having acted together
with Ṛjiśvan.
4 Defiant, he threw open the undefiable (fortifications); the unbridled one pulverized the
ungodly treasure-houses.
Like the sun with the moon, he took for his own the goods found in the fortress. Being sung,
he shattered his rivals with his flashing (weapon).

When Mortimer Wheeler exposed the mighty walls of Harappa in 1946, he identified the
fortified Indus cities with the forts of the Dāsas broken up by Indra and the Aryan kings,
according to the Rigveda (Wheeler 1947). However, as Wilhelm Rau (1976) pointed out in his
examination of the Rigvedic descriptions, the Dāsa forts are described as having many
concentric and circular walls—which do not match the layout of the Indus cities. In 1988, I
spotted an archaeological counterpart to this description of the Dāsa forts in the so-called
“temple-fort” of the BMAC in Dashly-3, excavated in the 1970s in northern Afghanistan
(Sarianidi 1977): it has three concentric circular walls (Fig. 20.3a). This suggested to me that
the major fights between the Aryans and the Dāsas probably took place not in the Indus Valley
but in the Indo-Iranian borderlands, en route to the Indus Valley. Dashly-3 is dated to about
1900 BCE and therefore belongs to the Proto-Indo-Aryan phase of the BMAC. However, the
tradition of building such fortresses with three concentric walls continued in Afghanistan until
historical times, as is shown by the Achaemenid fort of Kutlug Tepe in Bactria (Fig. 20.3b).
There are in the Rigveda as many as twenty-four references to a Dāsa called Śambara. The
Aryan king Divodāsa Atithigva, Indra’s protégé, is said to have vanquished Śambara in the
fortieth autumn, breaking his ninety-nine fortifications. The descriptions of the fights between
Śambara and Divodāsa are the most realistic, and apparently the oldest, in Book 6; it has been
suggested that the enmity between the Dāsas and Aryans was at its greatest in the period
represented by Book 6, which contains several indications that its poems refer to places near
to or west of the Hindukush. The repeated references to (great) mountains as Śambara’s
habitat are a significant pointer to the geographical location of the Dāsa forts, ruling out the
Indus Valley.
A second important pointer appears in RV 6,61,1–3. Here the mighty River Sarasvatī is
said to have given the powerful Divodāsa as a son to Vadhryaśva, who worshipped Sarasvatī
with offerings. It is most likely that Divodāsa, the son of Vadhryaśva, is the same person as
Divodāsa Atithigva, and that it was his son, or more likely his grandson, Sudās, who fought the
famous battle of ten kings in the Punjab, in which both Aryans and Dāsas participated. The
description of Sarasvatī in RV 6,61,2 does not at all fit the sacred stream Sarasvatī in India
(both the Sarsuti and the Ghaggar, with which it is identified, descend from the low Siwalik
Mountains): “By means of her gushing and powerful waves, this (Sarasvatī) has crushed the
ridge of the mountains, (breaking river banks) like a man who digs for lotus roots; with praises
and prayers, we solicit Sarasvatī for her help, (Sarasvatī) who slays the foreigners.” It is
widely accepted that the Sarasvatī mentioned here is the river that gave the name Haraxvaitī-
(in Avestan) or Harahuvati- (in Old Persian) or Arachosia (in Greek) to the province of the
Persian empire in southeastern Afghanistan that is chiefly watered by this river. It is generally
identified with the Arghandab that descends from a height of nearly four kilometers down to
about 700 meters, when it joins the Helmand River, which eventually forms shallow lakes; and
the name Sarasvatī means “(river) having ponds or lakes.” The Helmand’s present Pashto
name comes from Avestan Haētumant-, “having dams.” In the dry season the Arghandab, too,
may become a series of lakes.
In RV 6,61,3, the Sarasvatī is asked to throw down the deva revilers, the descendants of
“every Bṛsaya.” In Alexander the Great’s time, the satrap of Arachosia and Drangiana
(Seistan) was a man called Barsaéntēs (Arrian, Anabasis 3,8,3; 3,21,1), a name that resembles
Bṛsaya. The latter appears to have been a hereditary royal title, as suggested by the
expression “every Bṛsaya.” By way of comparison, every king of Taxila was called Taxiles,
according to Q. Curtius Rufus (8,12,14), and Pōros, the name of the king whom Alexander the
Great defeated in the Punjab, is clearly a dynastic title, meaning “descendant of Pūru.” Bṛsaya
has features (b- and -ṛs-) suggesting that it may come from the original language of the BMAC
(see chapter 8).
Now I return to the Rigveda’s repeated references to Śambara’s mountainous domicile. If
Divodāsa was born in Arachosia and his son or grandson Sudās fought in the Punjab, Divodāsa
is likely to have crossed the Afghan highlands roughly from Kandahar to Kabul and from there
entered Swat. The highest mountains are in the northern part of this itinerary, in Waziristan,
now well known as one of the principal bases of today’s Taliban insurgents.
This very region where Divodāsa may have fought against the Dāsa is the only area of
modern-day Afghanistan where the habitations consist of large traditional farm compounds
surrounded by massive walls of kneaded mud and provided with defense towers. These
fortified manors are presently inhabited by Pashto-speaking farmers and called qala,
“fortress” (Figs. 9.1, 9.2).

FIGURE 9.1 Distribution of the fortified manors with massive pakhsa walls in eastern Afghanistan. After Szabo &
Barfield 1991:140. Courtesy Thomas J. Barfield.

There is an evident similarity between these present-day qalas and their possible Bronze Age
predecessors, the fortified manors of the BMAC (Fig. 8.1). The BMAC-type “qala” architecture
started in Margiana around 2300 BCE, spread to Bactria about 1900 BCE and was continued in
the “post-urban” Takhirbaj period (c. 1750–1500). It is important to note that the BMAC-type
citadels and manors survived until the Yaz I–related cultures of the Early Iron Age, for
instance at Tillya Tepe in northern Afghanistan (Fig. 9.3), and indeed until medieval and
modern times. “Massive exterior walls made of mud brick are typical of the later (Parthian,
Sasanian, and medieval) architecture of the Margiana oasis … and can be used as an analogy
for the deflated remains of the massive exterior walls in Bronze Age Margiana to suggest that
these walls functioned both for defence and for insulation against summer heat and winter
sandstorms” (Hiebert 1994:115).
FIGURE 9.2 A fortified manor (qala) of eastern Afghanistan. After Szabo & Barfield 1991:188. Courtesy Thomas J.
Barfield.

FIGURE 9.3 The fort of Tillya Tepe in eastern Afghanistan, c. 1500–1000 BCE. After Sarianidi 1989:19, fig. 5:1.
Courtesy Margiana Archaeological Expedition.

According to the Zoroastrian text Vīdēvdād (2,21–43), Ahura Mazda ordered the first
Aryan/Iranian king Yima to build a fortress (vara-), “long as a riding ground on every side of
the square” in the “Aryan expanse” (airyān m vaējō), which contained plenty of grass for
cattle but offered no shelter during the long winter. The text describes the method of
construction, which also matches that of the current Afghan qala with its thick pakhsa walls:
Yima had to crush earth and knead it with his fingers, as the potter kneads clay, to build the
vara, establishing there “dwelling places, consisting of a house with a balcony, a courtyard,
and a gallery,” and then stock it with an inexhaustible supply of seeds.
Michael Witzel (2000) has discussed the old problem of locating this “home of the Aryans”
[i.e., early East Iranians], where Yima built his fort, on the basis of the climatic conditions and
geographical distribution of the various Iranian tribes mentioned in the first chapter of the
Vīdēvdād. The highlands of Afghanistan lie in the center of the places mentioned, even though
the precise identification of many geographical references remains hypothetical. Witzel’s
proposal of locating the homeland in Afghanistan is supported by the transhumance that for
thousands of years has brought together nomadic tribes in the highlands of Afghanistan from
all directions to utilize fresh pastures during the summer months. In addition to grazing their
animals, the tribes use this time for celebrating marriages.
In the Rigveda, dāsá- normally appears to mean “member of an enemy tribe” or “enemy
demon.” These meanings are implicated in the corresponding adjective or derived noun dā´sa-
, which in classical Sanskrit (as dāsa-) means “slave”; this meaning occurs four times in the
Rigveda. We may note many historical parallels in which the name of a people captured in war
comes to mean “slave,” for instance, Finnish orja, “slave,” derived from ārya-, “Aryan,” and
also English slave, which derives via French esclave and late Latin sclavus from Byzantine
Greek esklabēnós, “Slav taken as war-captive, slave.”
In Iranian languages, Proto-Iranian *s became h before a following vowel at a relatively
late period, perhaps around 850–600 BCE. Old Persian Ū a- < *Hūža- “(the country of) Elam”
(which corresponds to Parthian Χūzestān, “Elam”) is supposed to have come from the Elamite
name of the city of Susa, *Sūša(n), and since the Persians could not have been in contact with
the Elamites before the ninth century BCE, the sound law *s > h is supposed to have been
operating at the time when the Persians came to Elam. Another piece of evidence for a
relatively late date is a Neo-Assyrian cuneiform tablet from the library of Assurbanipal (c.
668–630 BCE) that records a god’s name, dAs-sa-radMa-za-áš; this may render Median *Asura-
Mazdās, where the change Asura- > Ahura- had not yet taken place. (It has been argued,
though, that this Neo-Assyrian tablet might be a copy of a Middle Assyrian text from the
second millennium BCE; Hintze 1998; Schmitt 2000.)
Sanskrit Dāsa- as an ethnic name thus has an exact counterpart in Avestan Dåŋha-, which
stands for *Dāha-. The corresponding Old Persian ethnic name is Daha-. The plural form Dahā
is included among the subjects of the Great King in the “empire list” of Xerxes, immediately
before the two kinds of Sakā. Herodotus (1,125) includes the Dā´ai / Dā´oi (intervocalic h is
omitted in Greek) among the nomadic tribes of the Persians. According to Q. Curtius Rufus
(8,3), the Dahae lived on the lower course of the river Margus (modern Murghab) in
Margiana, where they are also located in Ptolemy’s Geography (6,10,2). Eratosthenes, quoted
by Pomponius Mela (3,42), notes that the great bend of the Oxus towards the northwest
begins near the Dahae. Tacitus (Annales 11,10) places the Dahae on the border river (Sindes,
modern Tejend) between Areia and Margiana.
The same Rigvedic verse which mentions Divodāsa’s birth as a jpgt of the Sarasvatī (of
Arachosia), RV 6,61,1, also says that the river took away from the (enemy) Paṇi his
nourishment (pasture or cattle). The Paṇis are often mentioned as enemies of Rigvedic kings,
sometimes along with the Dāsas and Dasyus, and in very similar terms. Paṇí is supposed to be
a Prakritic development of *Pṛṇi, a reduced-grade variant of an ethnic name that has a full
grade in Parṇáya, the name of an enemy of King (Divodāsa) Atithigva in RV 1,53,8 and 10,48,8.
Strabo (11,9,2) describes the foundation of the Parthian empire of the Arsacids around 240
BCE as follows: “Arsakēs, a Scythian man, who had (in his command) some men of the Dāai,
namely nomads called Párnoi, who were living along the River Okhos (modern Tejend),
invaded Parthia and conquered it.”
A Roman author characterized the Parthian language as being between Median and
Scythian (i.e., Saka) and as a mixture of both tongues (Iustinus 41,2,3: “sermo his inter
Scythicum Medicumque medius et utrimque mixtus”). A number of words in Parthian have
been identified as phonetically and lexically East Iranian and ascribed to the language of these
Parna people belonging to the Dahae, who therefore apparently spoke Saka.
According to RV 6,21,11, gods “made Mánu superior/successor to Dása.” This is the only
occurrence of the word dása- in Sanskrit; it may denote the ancestor of the Dāsas, while Mánu
is the mythical ancestor of the Vedic Aryans. Sanskrit mánu- means “man.” Dasa- is the
original Proto-Iranian form which, with the (relatively late, cf. above) Iranian sound change *s
> h, became in Khotanese Saka daha-, “man.” Khotanese Saka was spoken in Khotan, in
western Sinkiang or Chinese Turkestan.
In Khotanese Saka there is no doubt about the meaning. The word daha- is used in the
sense “man, male person” to render Sanskrit puruṣa-, “man,” and nara-, “man,” contrasting
with strīyā-, “woman, female”; in compounds and derivatives it has the sense of “manliness,
virility, bravery.” These Saka meanings have a counterpart in nearby Wakhi, an East Iranian
language spoken nowadays in the Hindukush, where δay < *daha- means “man, male person,
human being.” The only further Indo-Iranian language to have a cognate is Ossetic, for Ossetic
läg, “man,” may go back to Alanian *dahaka-. This Saka language spoken in northern
Caucasus is the closest known relative of the Khotanese-Wakhi group: Ossetic and Khotanese
Saka share some seventy vocabulary entries (Bailey 1959 [1960]).
It is clear that Proto-Saka *dasa, “man,” is the etymology of the ethnic name which in Old
Persian appears as Daha-, for many ethnic self-appellations go back to words with this
meaning; for instance the native ethnic name of the Mari, who speak a Uralic (Finno-Ugric)
language, goes back to Proto-Indo-Aryan *marya-, “man” (literally “one who has to die,
mortal”).
It is true that *dasa- / *dāsa- and *dasyu- have a limited distribution in Indo-Iranian. “There
are several suggestions for an Indo-European etymology, but they are all doubtful,” observed
Alexander Lubotsky (2001:313). I shall not go into these earlier etymologies here, but repeat
only my own observation (Parpola 2002b). These words could go back to the PIE root *dens-,
“to be(come) clever or wise,” if the accent was originally on the derivative suffix (*-á- and *-yú-
), as in Vedic das-rá- (< *dṇs-ró-), “clever, skillful,” which occurs especially as an epithet of the
wonder-working Aśvins. In the Rigveda, the Aśvins are called dasrā´ nine times, dásrā nine
times, and dásrau four times, with the variation in accent attested in a single standing formula
(dasrā´ híraṇyavartanī in RV 1,92,18b, dásrā híraṇyavartanī in RV 5,75,2c, RV 8,5,11b, RV
8,8,1c, and RV 8,87,5c). Therefore, the accent in the unique RV VI,21,11 dása-, instead of the
expected *dasá-, is no obstacle to this etymology. For dásyu- instead of the expected *dasyú-,
one may compare RV 7,18,5c śimyú- against RV 1,100,18a śímyu-, this being likewise a name
of an enemy people.
Semantically it makes good sense to derive a word meaning “man” from a root expressing
cleverness and skillfulness. One can compare the possible derivation of Sanskrit mánu-,
“man,” from the root man- < PIE *men-, “to think.” An even closer parallel is the ethnic name
Śaka-, which is most likely derived from the root śak-, “to be able, powerful, skillful” (the
Rigveda has su-śáka-, “well able,” as an epithet of “men”; one may also compare the use of
śak-rá-, “strong, mighty,” as an attribute and later as a proper name of god Indra; the
synonymous ugrá-, “strong, mighty,” is used as an ethnic name). In Iranian the root sak- means
“to understand, to know,” which makes Śaka < śak- an even closer parallel to Dasa- < PIE
*dens-.
The royal inscriptions of Darius contain five versions of the “empire list” enumerating the
countries known to be subjects of the Great King; in addition there is one version by Xerxes.
The two shortest lists of Darius (from the Behistun and Susa foundation charters) contain
twenty-three names, in which the eleven last ones relate to the eastern part of the empire:
Parthia, Drangiana, Aria, Chorasmia, Bactria, Sogdiana, Gandhara, Saka, Sattagydia,
Arachosia, and Makran or Quadia. The list of twenty-four names from Persepolis adds Indus
(after Arachosia and before Gandhara and Saka). In the lists of twenty-seven and twenty-nine
countries, after Gandhara and Indus follow the Sakas now divided into two varieties, Sakā
haumavargā, “Sakas with hauma cult,” and Sakā tigraxaudā, “Sakas wearing pointed caps.” In
the list of Xerxes, the thirty-one (or thirty) names are in a rather haphazard order in the
following sequence (Cappadocia being out of place): Gandhara, Indus, Cappadocia, Dahae,
Sakā Haumavargā, and Sakā Tigraxaudā. Here the Dahae figure for the first time in Old
Persian.
Herodotus (7,64), describing the army of Xerxes, lumps the two kinds of Saka together:
“The Sacae, or Skyths, were clad in trousers, and had on their heads tall stiff caps rising to a
point. They bore the bow of their country and the dagger: besides which they carried the
battle-axe, or sagaris. They were in truth Amyrgian Scythians, but the Persians called them
Sacae, since that is the name they give to all Scythians” (translated Rawlinson [1860]
1942:524). The Amyrgian Scythians are considered to correspond to Sakā haumavargā.
Hauma is Avestan haoma, the Iranian counterpart of Sanskrit soma, Indra’s sacred drink.
The Khotanese and the Wakhis are considered descendants of the Scythian tribes of
eastern central Asia, probably the Scythians with pointed caps. It seems to me that the
Khotanese Saka and the Wakhis originally called themselves Dasa/Dāsa, and that this ethnic
name, although Dahae were added as a separate people by Xerxes, may in fact be coterminous
with Sakā tigraxaudā. The following evidence suggests that the immigrants to southern
Central Asia associated with the Yaz I and related cultures (c. 1500–1000 BC) were horsemen
and that they wore pointed caps.
There are only two places in the Rigveda (1,162,17 and 5,61,2–3) that clearly refer to
riding; in four other verses the reference is dubious (Falk 1994). Compared to the prominence
of the horse chariot (a creation of the Proto-Indo-Aryans of the Sintashta culture), horse-riding
played a very marginal role among the Rigvedic Aryans. But they knew of it, which suggests
the presence of mounted horsemen in their neighborhood. Among animal remains, there is a
very significant increase in the numbers of horse (7–18 percent) in the Early Iron Age of
southern Central Asia compared to the immediately preceding Late BMAC (Takhirbaj) period
(only 2.6 percent). There is one Late BMAC cylinder seal depicting a horse-rider (Fig. 9.4).
From the first two periods (c. 1750–1300 and 1300–1100 BCE) of Pirak in Baluchistan near the
Bolan Pass come terracotta figurines of horse-riders, with bowed legs to fit them on the back
of the horse, armless torsos and heads with faces ending in a bird-like beak (Fig. 9.5). I detect
a striking resemblance to the rider’s bird-head in the pointed felt caps topped by beaked bird
heads that have survived in frozen tombs of mounted Saka nomads in the Altai Mountains,
dated to c. 500–200 BCE (Fig. 9.6).

FIGURE 9.4 A BMAC cylinder seal from tombs looted in Afghanistan around 1985, showing a horse-rider. Probably
sixteenth century BCE. After Sarianidi 1998b:169, no. 1482. Courtesy Margiana Archaeological Expedition.

FIGURE 9.5 A clay figurine of a bird-headed horse-rider from Pirak, Pakistani Baluchistan, c. 1500 BCE. After Jarrige
1985:245. Courtesy Mission Archéologique de l’Indus, CNRS/Musée Guimet, Paris.

FIGURE 9.6 Pointed felt caps topped by a bird head, belonging to Saka mounted nomads, from frozen tombs in the
Altai Mountains. After Polos’mak 2001:124, fig. 21. Courtesy Nataliya V. Polos’mak.

That the Early Iron Age cultures of southern Central Asia related to Yaz I (Fig. 9.7) are linked
to Iranian-speakers is suggested by the fact that so far no graves have been found in this vast
area from about 1500 BCE to Hellenistic times (Teufer 2013). This absence of graves points to
the practice of exposure burial, which is associated with the Zoroastrian religion. Burials are
not permitted for the Zoroastrians, because they would defile earth, which is one of Ahura
Mazdā’s creations.
FIGURE 9.7 The distribution of the “Roller Pottery” cultures, among them the late Srubnaya cultures (6), the
Sargary culture (8), and the Yaz I culture (12). After Chernykh 1992: 236, fig. 79. Courtesy E. N. Chernykh and
Cambridge University Press.

Important linguistic evidence supporting the identification of the Dāsas as Proto-Sakas is

provided by the Nuristani languages spoken in northeastern Afghanistan. The Nuristani
languages are unique in their preservation, in etymologically correct words, of the sounds
called dental affricates (ts, dz, dz) that have been reconstructed for Proto-Iranian, but which
have developed further and differently in Avestan (s, z, z) and Old Persian (th, d, d); these
Proto-Iranian sounds are the expected developments of Proto-Indo-Iranian palatal affricates
(ć, j, jh) from PIE palatal stops (k´, g´, g´h), the Iranian development being different from
Proto-Indo-Aryan (ś, j, h). If the ancestors of the Nuristani speakers belonged to the very first
Iranians in Central Asia, and became isolated from other Aryan speakers early on, they could
well have preserved this very archaic stage of phonological development. The exact affiliation
of the Nuristani languages is an old problem, but Almuth Degener (2002) has suggested a
plausible solution: the Nuristani languages are likely to be a mixture of the Proto-Indo-Aryan
spoken in the BMAC and the Proto-Iranian that came to southern Central Asia with the Yaz I-
related cultures.
Some of the Dāsa proper names, in particular Balbūtha Tarukṣa (RV 8,46,32), Bṛsaya (RV
1,93,4; 6,61,3), Bṛbu (RV 6,45,31 and 33), and Sṛbinda (RV 8,32,2), seem to come from the
original, pre-Aryan language of the BMAC (chapter 8), suggesting that its speakers had
survived the Proto-Indo-Aryan rule (c. 2000–1500 BCE) and were now part of the Saka
community of the Dāsas, Dasyus, and Paṇis. While there are several Aryan-looking Dāsa
names, it is important to note that there is an elite proper name that is undoubtedly of Saka
affinity. In RV 8,46,21 and 24 a chief called Pṛthuśravas, “widely famed,” is praised as having
presented to the poet 60,000 horses, 2,000 camels, 1000 dark brown mares, 10,000 cows, and
a golden chariot. This liberal sponsor has the patronym Kānītá, which has been compared, no
doubt correctly, to the Scythian personal names Kanī΄tēs and Kánitos recorded in the Pontic
steppes in the second century BCE (Hoffmann 1940:142). Dāsa Balbūtha Tarukṣa mentioned
above is praised in the same hymn for a similar generosity, and it is noteworthy that in both
cases camels are given—an indication that the location lies somewhere close to Central Asia.
The Dāsa proper name Śambara is yet another important argument in favor of my proposed
solution to the Dāsa problem, to be discussed in chapter 20.
The absence of graves in the Yaz I–related cultures suggests that Proto-Zoroastrianism
may have developed in southern Central Asia in 1500–1000 BCE (Teufer 2013). In my opinion
there is one lexical feature in the Avesta which suggests that its Old Iranian language
belonged specifically to the Saka branch. The word used in Khotanese Saka to translate
Sanskrit deva, “god,” is gyasta-, jasta-, and in Tumshuq Saka jezda-; these words come from
the principal Avestan word for “god,” yazata-, which has an etymological counterpart in
Rigvedic yajata-, “worthy of worship or sacrifice” (in the Rigveda, however, it is not used
specifically to mean “god”). That this word for “god” goes back to Proto-Saka is suggested by
the use of the cognate word zäd or izäg for “god” in Ossetic in the pre-Christian period: there
is no reason to consider that this usage is due to any Zoroastrian influence. Parthian yazd,
“god,” besides bg-, “god,” may also go back to the Saka language of the Dahae. By contrast, in
Sogdian Buddhist texts deva- is translated with βγ- from baga-, which denotes “god” in Old
Persian, with cognates in most Iranian languages, but is rare in the Avesta. That baga, “god,”
comes from a dialect of Proto-Iranian which originally had a more western location than Proto-
Saka is suggested by its semantic borrowing into Old Slavic bogǔ, “god.”
The religion of the early Iranians of the Saka branch coming to Central Asia around 1500
BCE must surely have been much influenced by the Proto-Indo-Aryan religion prevailing there
during their arrival. This question, among others, will be considered in the next chapter. What
this chapter has endeavored to show is that my ethnic and linguistic identification of the
Dāsas, Dasyus, and Paṇis of the Rigveda with the Yaz I–related cultures has provided a new
and forceful argument that the migration of the Rigvedic tribes of Pūru and Bharata into
northwestern South Asia in all likelihood took place c. 1400–1200 BCE. This agrees with the
estimated date most often proposed by others.
10
The Aśvins and Mitra-Varuṇa

The Rigvedic hymns are addressed to gods, whom they praise, whom they invite to a sacrificial
meal and to whom they pray for long life, sons, cattle, victory in battle, or fame. The central
myths are repeated over and over, but the Rigveda abounds in references to myths, legends,
and happenings supposedly well known to the audience but obscure to us. Only later Vedic
literature provides more coherent accounts of the gods’ doings.
Indra, the god of war and thunder, is described as a robust and bearded warrior, who
wields the lightning bolt as his mace (vajra), is victorious in battle, and has an insatiable thirst
for his soma drink. Indra is the chief divinity of the Rigvedic people, to whom more than one-
quarter of the collection of 1028 hymns is addressed. Indra’s most important feat is the
release of the waters by killing the dragon Vṛtra, whose name means “enclosure, obstruction.”
Maruts, the storm winds, assist Indra. Indra also helps in finding the cattle captured by the
enemy, especially the demon Vala, who keeps the cows in a cave; in this Indra is helped by
Bṛhaspati with magic spells. Indra helps Aryan kings to destroy the fortifications of the Dāsas.
He is also a demiurge who creates the sun, the heaven, and the dawns.
Second in popularity is Agni, the god of fire (agni-, related to Latin ignis, “fire,” and
Russian ogon’, “fire”). As the ritual fire into which libations of ghee and soma are poured and
other offerings placed, Agni is the mouth of the gods, and the divine priest and messenger. He
assists Indra in burning the enemy forts. Third most popular (as measured by the number of
hymns devoted to him) is Soma, the god praised in Book 9. Soma is the invigorating divine
juice that enhances a warrior’s performance and keeps a poet inspired and vigilant. It is
pressed, using stones, from the stalks of the soma plant, after they have been soaked in hot
water. Soma is both sacrificed to gods, especially Indra, and imbibed by the priests and poets.
Vedic soma corresponds to Avestan haoma, both from older *sauma, derived from the verb
*sau-, “to press out.” The botanical identity of the plant has been much debated. Rhubarb,
hops, fly-agaric mushroom, and harmel or wild rue are some of the alternative candidates
proposed, but nowadays the identification of the soma plant as belonging to the genus
Ephedra is widely accepted. Ephedra is the source of ephedrine, a banned drug among
modern sportsmen, thus fit to enhance Indra’s martial feats. Moreover, Ephedra has been
used for many centuries for haoma by Zoroastrians in both central Iran and India. In Nepal
Ephedrais called soma-latā, “soma creeper,” which suggests its use for soma in North India. In
South India, where Ephedra does not grow, Sarcostemma brevistigma, a leafless climber
similar to Ephedra, has been used as a substitute (Flattery & Schwartz 1989; Falk 1989;
Nyberg 1995).
Important deities, to whom relatively few hymns are addressed in the Rigveda, are Mitra
(“contract”), the divine guardian of pacts, and Varuṇa (perhaps “true speech”), the upholder of
the cosmic and social order, who punishes sinners; these two usually appear as a dual divinity,
Mitra-Varu ṇa; they are the foremost in the group of divinities called Ādityas, to be discussed
below. Also important are the horse-owning twins, Aśvins or Nāsatyas, whose sister and bride
is the most important goddess of the Rigveda, the beautiful goddess of Dawn, Uṣas (whose
name is related to Greek Ēōs and Latin Aurora). Other nature gods are the sun god, Sūrya; the
wind god, Vāyu or Vāta; the rain god, Parjanya; Dyaus, “Heaven, Sky”; and Pṛthivī, “Earth.”
The latter two are the father and mother of the gods and the world; they were inherited from
the Proto-Indo-European pantheon, although Dyaus is less important than Greek Zeus and
Roman Jupiter. Minor divinities are many, including Pūṣan, a pastoral deity who helps in
finding paths, the creator god Tvaṣṭar (“carpenter, fashioner”) and Savitar (“impeller,
commander”). Viṣṇu and Śiva, who dominate the later Hindu pantheon, are to some extent
already present in the Rigveda, the latter under the name Rudra (see chapter 12), but as
rather marginal divinities. Viṣṇu is a friend and helper of Indra. His Rigvedic feat of measuring
the universe in three steps is often referred to in later Vedic literature and forms the basis of
his dwarf (Vāmana) incarnation in classical Hinduism.
Vedic scholars have long noted a marked contrast between the somewhat barbaric main
hero god, Indra, and Varuṇa, the majestic guardian of the universe, who through his spies
watches out that people speak truth. Varuṇa has been considered a remarkably “ethical” and
“civilized” deity; yet, in spite of his high rank, he is the deity of relatively few Rigvedic hymns.
A certain antagonism and rivalry between Indra and Varuṇa is seen in the Rigvedic hymn 4,42,
where these two deities in turn laud themselves as kings, emphasizing their achievements and
prerogatives. Particular importance has been given to RV 10,124,5, a verse in which Indra
addresses Varuṇa as follows: “Those Asuras have just now become devoid of their magic
power; If you desire me, come to my overlordship of dominion, distinguishing Falsity from
Truth, O king!” (translated Proferes 2007:108). According to Karl Geldner and many others,
Indra here declares that the older generation of gods called asuras have now lost their
struggle with the new generation of gods called devas, and Indra offers Varuṇa, the leading
Asura, a chance to join the ranks of the devas if he accepts becoming a vassal of Indra. In my
opinion, the verse could indeed reflect such a compromise between Varuṇa and Indra
worshippers at the time when the first and second waves of Proto-Indo-Aryan immigrants to
the BMAC mingled in Central Asia, and then again in the Indus Valley, when the Rigvedic
Aryans mingled with the earlier Indo-Aryan immigrants.
In the Rigveda, Indra is the king (Varuṇa anoints Indra in hymn 10,124); but in the
Brāhmaṇas, Varuṇa is generally the divine king. In the Rigveda, Indra is addressed as deva,
“god,” but Varuṇa as asura (a title borrowed from Proto-Indo-Aryan into Uralic languages in
the meaning “prince, lord”); in the Brāhmaṇas, asura means “demon,” which often is an
appellation of enemy gods; some contexts indeed show that there were non-Vedic peoples in
eastern India who worshipped asuras. Indra’s drink is soma, Varuṇa’s beer (surā); the barley
from which surā is made belongs to Varuṇa. Vedic texts usually condemn surā as a
manifestation of untruth. In the Vedic vājapeya ritual, which includes a chariot race, soma is
offered to Indra and other deities and consumed by the priests, but surā is prepared and given
to the charioteers, while the winner receives a bowl of honey (madhu). Honey is associated
with the Aśvins; honey-beer was the ritual drink of the PIE speakers, and figures also among
the early Aryan loanwords of the Uralic languages (chapter 7).
This chapter discusses the Aśvins, alias Nāsatyas, and the dual divinity Mitra-and-Varu ṇa,
whom I see as a double of the Aśvins. I explain the above-mentioned cultic conflicts within
Indra worship—which appear to have been partially resolved by various compromises—by
postulating a fundamental religious difference between the two waves of Proto-Indo-Aryan
speakers who successively took over power in the BMAC. Archaeological evidence suggests
that the first wave came from the south Uralic Sintashta culture, where the horse-drawn
chariot first appeared (see chapter 8, and Figs. 7.7 and 8.4). I see the Aśvins as the divinized
chariot-team, who were probably the highest gods representing their mundane counterparts,
dual kings, although in the Vedas they are divinities of a low rank. The second wave of Proto-
Indo-Aryan speakers was that of the Fëdorovo Andronovans, steppe pastoralists whose arrival
more or less coincided with the beginning of the impoverished “post-urban” phase of the
BMAC. Their main god was Indra, whose name may come from *Inmar, the name of the sky
and thunder god of their Uralic-speaking companions (see chapter 7). Soma, Indra’s cultic
drink, does not have an ancient Indo-European background like the madhu of the Aśvins, and
was probably adopted by the Fëdorovo Andronovans in Central Asia.
Thus I see the cult of the Nāsatyas as being not of Proto-Indo-European origin (as is often
so maintained) but as going back to the times when the horse-drawn chariot evolved, that is,
the end of the third millennium BCE. The chariot was a prestigious and effective new
instrument of war and sport, which was quickly adopted by the elites of neighboring peoples.
Together with the chariot, the mythology and cult of the deified chariot team also spread.
Placing its origin in the steppes of the south Urals best explains the distribution of the early
chariot lore among the Aryans, Greeks, and Balts.
In the Rigveda, the Aśvins are several times called “sons of the sky,” divó nápātā or dívo
napātā. It relates them historically to the horse-riding divine twins of early Greece who are
called Dioskouroi, “youths of Zeus” (i.e., sons of the sky god), and to the horse-riding “sons of
the God” (Latvian Dieva dēli, Lithuanian Dievo sūneliai) in the pre-Christian religion of the
Balts. Moreover, all three sets of equestrian twins have a sister or wife or bride either
associated with the dawn or called the daughter of the sun (Uṣas or Sūryā in India, Helénē <
*svelénā, “sunshine,” in Greece, and in the Baltics, Latvian saules meita, “maiden or daughter
of the sun,” and Lithuanian saules dukryte, “daughter of the sun”).
The Aśvins of the Ṛgveda move in a chariot, but the Greek Dioskouroi and the Baltic “sons
of the God” are horse-riders. The difference is understandable. In much of the ancient world,
throughout the second millennium BCE, the chariot drawn by a pair of horses was “the vehicle
of prestige—the only approved conveyance for the chieftain and his noble entourage in
ceremony and ritual, hunting and its counterpart, warfare” (Piggott 1992:48). But the
situation changed in the early first millennium BCE, when the Dioskouroi first make their
appearance (eighth or seventh century BCE):

The beginnings of regal horse-riding were tentative. In the ancient Near Eastern
tradition the king, if he did not appear in a chariot, might on occasion ride on a mule or
a donkey. … In the early second millennium a well-known letter to Zimri-Lin, King of
Mari, gives him advice on this matter—“Let my lord not ride horses. Let him mount
only chariots or mules and honour his kingly head.” … But … cavalry was taking over
chariotry in Assyria by the ninth century BC and the king Shalmaneser III (858–823 BC)
is depicted as riding on horseback. Thenceforward the monarch as a warrior on
horseback became the accepted convention in the ancient Orient. (Piggott 1992:69–70)

Significantly, the Assyrian mounted warriors of the ninth century “always go in pairs,
translated directly from the chariot: the bowman and the driver, who still holds the reins of
both horses, both riders with cramped and unhorsemanlike seats. These Assyrians use the
weapon they used in the chariot and, no matter how awkwardly at first, they persist in fighting
mounted, thus continuing their chariot-fighting tradition” (Littauer in Littauer & Crouwel
2002:63–64).
The existence of a two-man team associated with the chariot in the Sintashta-Arkaim
culture of the southern Urals (c. 2200–1800 BCE) is suggested by a burial at Sintashta (Fig.
7.5a). Here the warrior was buried together with his weapons and his car at the bottom of the
grave, while another man was buried together with a pair of horses and a burning fireplace in
an upper chamber. In the Vedic religion, the charioteer and the chariot-fighter are expressly
equated with the Aśvins. When, in the royal consecration, the king goes to the house of his
charioteer (samgrahītár-), he “prepares a cake on two potsherds for the Aśvins; for the two
Aśvins are of the same womb; and so are the chariot fighter [savyaṣṭhá-] and the driver [sā
´rathi-] of the same womb (standing-place), since they stand on one and the same chariot:
hence it is for the Aśvins” (ŚB 5,3,1,8, translated Eggeling 1894:III,62).
Of the Greek Dioskouroi, too, one was a fighter and the other took care of horses:
according to their standing Homeric attributes (e.g., Iliad 3,237), Poludeukes was good at
fistfighting (pùks agathós), while Kastor was good at taming horses (hippódamos). The
Rigveda, too, differentiates between the two Aśvins: “one of you is respected as the victorious
lord of Sumakha, and the other as the fortunate son of heaven” (RV 1,181,4, translated Insler
1996:183). This passage suggests that divó nápātā is an elliptic dual, that is, based on the
name of just one member of the pair, “the son of heaven/sky,” just like nā´satyā, derived as it
seems to be from the charioteer member of the team. A similar case is the elliptic plural
Castores, “Castor and Pollux” of Latin (which lacks the dual) for the Greek dual tô Kástore.
Nā´satya- is a derivative of *nasatí-, “safe return home,” and belongs to the same PIE root
*nes- as the Greek agent noun Néstōr—known from Homer (Iliad 23,301–350 and 638–642) as
a hippóta, “horseman,” and a masterly charioteer; it refers to the charioteer’s task of bringing
the hero safely back from the battle (Frame 1978; 2010). In the Indian epics, “the rule of
protecting the knight is formal. ‘In battle the knight, if confused, must be guarded by the
charioteer’; or, ‘ever must the man of the war-car be guarded’; and when the charioteer risks
his life in saving his master, he does so because he ‘bears in mind the rule’ ” (Hopkins
1889:196). In the Rigvedic verse just quoted the “victorious lord of Sumakha” appears to be
the chariot-warrior. The meaning of the words sú-makha- and makhá-, -makhas- is debated,
but this context suits the old etymology that connects them with Greek mákhē, “battle,
combat,” and makhésasthai, “to fight.”
All over the ancient West Asia at the times of the Mitanni empire, the chariot was the
prerogative of the king and the noblemen close to the ruler, and the same was the case in
Vedic India. When the horse-drawn chariot had just been invented, it must have symbolized
royal power in the highest degree, and is indeed found only in elite graves. The chariot crew
normally consisted of two men, the warrior and the charioteer, and the two were normally of
equal aristocratic status: in the Mahābhārata, it is usual for kings to serve as charioteers of
kings in the battle, as in the case of Kṛṣṇa driving Arjuna’s car, or Śalya driving Karṇa’s.
Putting all these things together, I argue that among the inherited features that the Aśvins
share with their Greek counterparts we should include dual kingship. Among the Dorian
Greeks, the Dioskouroi were the greatest gods, and they were widely worshipped in Greece as
“the two kings” (ánake). According to Herodotus (5,75), their images accompanied the two
kings of Sparta on war expeditions. In the Buddhist tradition of India, the universal emperor
wielding supreme political power (rājā cakkavattin) is paralleled by the buddha-, who wields
supreme spiritual power (Aṅguttara-Nikāya I,76 and III,150). The idea of such a dual kingship
manifests itself above all in the integral connection of kṣatrá-, “political power” and bráhman-,
“sacred power,” these two being represented by the king and the royal chief priest, the
puróhita-.
This dual kingship is associated with the chariot (and therewith the Aśvins), for, according
to the Jaiminīya-Brāhmaṇa (3,94), “formerly the kings’ chief priests used to be their
charioteers so that they could oversee that the king did not commit any sin.” The Āśvalāyana-
Gṛhyasūtra (3,12) details the royal purohita’s duties in battle. Standing behind the chariot, he
makes the king put on the coat of mail, hands the bow and the quiver to the king, blesses the
weapons and the chariot with its horses. Then the purohita mounts the chariot and makes the
king repeat the Rigvedic hymn 10,174, in which the king asks Bṛhaspati to help him
overpower his rivals. In the battle hymn called ápratiratha-, which the purohita recites next,
Bṛhaspati, the charioteer and purohita of Indra, the king of gods, is asked to “fly around” in
his chariot, warding off enemies and helping the chariots of the purohita’s side.
In RV 2,24, Bṛhaspati is mentioned as an excellent charioteer and winner of races. In the
chariot race of the vājapeya, the sacrificer announces he will win the race with the help of
Bṛhaspati. Indeed, the Śatapatha-Brāhmaṇa (5,1,1,11) states that “this [vājapeya] is a sacrifice
of the Brāhmaṇa, inasmuch as Bṛhaspati sacrificed with it; for Bṛhaspati is brahman, the
priesthood, and the Brāhmaṇa is brahman, the priesthood.” The vājapeya was prescribed for
the royal purohita (LŚS 8,11,1); both before and after the vājapeya, the sacrificer was
supposed to perform the sava rite of Bṛhaspati (LŚS 8,11,12), and it was by performing the
bṛhaspatisava sacrifice that Bṛhaspati became the purohita of the gods (PB 17,11,5–6).
The charioteer and the royal priest were expected to be wise and crafty. Dasrá-, “having
marvelous skill,” is one of the most distinctive epithets of the Aśvins, often referring especially
to their skill in chariot driving. By the time of the epics, Dasra became the proper name of one
of the Aśvins. The Aśvins are also called purudáṁsa(s)-, “having many skills,” which is
etymologically related to the Greek epithet polud nēs, “having many counsels, plans or arts;
very wise.” One is reminded of the crafty designs and plots of Kṛṣṇa, the charioteer of Arjuna,
in the Mahābhārata (see chapter 13).
There is a problem with the dual kingship of the Aśvins proposed here. In his book on
kingship in the Rigveda and the Atharvaveda (1960), Bernfried Schlerath finds no evidence for
the Aśvins being themselves kings, although they are mentioned as bestowers of royal power
—especially because they make the king’s chariot victorious. However, on one occasion (in RV
3,38,5) “the two sons of the sky” are addressed as two kings (rājānā). According to Rigveda
8,35,13, the two Aśvins are “possessed of Mitra and Varuṇa as well as of Dharma” (mitrā
´váruṇavantā utá dhármavantā). The primary meaning of mitrá- is “contractual alliance, pact
of friendship,” and of váruṇa- probably “true speech.” Thus these personified notions of social
nature—important in illiterate tribal societies—were associated with the Aśvins. The Nāsatyas
are, together with Mitra and Varuṇa, and Indra, Proto-Indo-Aryan oath deities sworn by in the
Hittite-Mitanni treaty of about 1350 BCE (chapter 8). In RV 8,35,12 and 1,120,8, the Aśvins are
invoked to guard against the breach of a treaty. An etymology of the name Nāsatya quoted by
Yāska on Nirukta 6,13 shows that they were regarded as protectors of the truth: they are
“true and not false (na-asatya),” says Aurṇavābha; “they are promoters of the truth,” says
Āgrāyaṇa (satyāv eva nāsatyāv ity aurṇavābhaḥ / satyasya praṇetārāv ity āgrāyaṇaḥ). The
Dioskouroi, too, were oath deities.
The winged solar disk in Achae, menid art, representing the highest god of Zoroastrianism,
Ahura Mazdā, has been taken over from Assyrian art, where it is one of the principal symbols
of the Assyrian high god, Aššur. The conception of Ahura Mazdā as a monotheistic god with
the Ameša Sp ṇtas, “Holy Immortals,” representing his qualities or powers, is also strikingly
similar to that of Aššur. The cuneiform texts present also the “great gods” of the Babylonian
pantheon as functions, powers, and attributes of the Babylonian national god Marduk: as his
“kingship, might, wisdom, victory, strength, counsel, judgment,” and so on. It is widely agreed
that this conception was not a creation of Zarathuštra. The Iranian god Mithra, too, was
surrounded by friendship, obedience, justice, courage, and divine grace.
The Zoroastrian Ameša Sp ṇtas have long been compared to the Ādityas of the Veda: these
“personalized powers … represented the principles on which human action depended”
(Brereton 1981:ix). Two of the Ādityas, Mitra (“Alliance”) and Varuṇa (“Commandment” or
“True Speech”), are invoked in the Mitanni treaty. In addition, the Ādityas comprised such
abstract deities as Aryaman (“Civility”) and Bhaga (“Fortune”), Aṁśa (“Share”), and Dakṣa
(“Ability”) (Brereton 1981). I have proposed that the creation of the Ādityas was inspired by
the Assyrian religion, where deified social virtues are aspects of Aššur, the highest god,
comparable to the advisors who surround the great king. The Assyrian religion is likely to
have influenced Proto-Indo-Aryan religion during the late twentieth and early nineteenth
centuries BCE, when Assyrian traders operating from Syria and Cappadocia imported tin from
Central Asia. Ideological influence is evidenced by the Syrian and Egyptian motifs adopted in
seals of the BMAC (see chapter 8).
Mitra-and-Varu ṇa early on seem to have overtaken the Aśvins’ royal function. The
Śatapatha-Brāhmaṇa (4,1,4) actually describes the relationship between Mitra and Varuṇa as
that prevailing between the king and his high priest. Yet the Aśvins, as the deified chariot-
team corresponding to the Dioskouroi, seem to be the original deities of dual kingship.
According to RV 3,58,8, the chariot of the Aśvins goes around the heaven and earth in one
day. In the āśvinaśastra, a text praising the Aśvins, the same is said of the sun’s horses. In
another verse (RV 1,115,5), the sun is said to show the colors of Mitra and Varuṇa in the lap of
heaven: the one appearance is infinitely white, the other one is black. Here the sun is
conceived of as one divinity having two forms, the white day-sun and the black night-sun, and
these two forms are connected with Mitra and Varuṇa. Some verses of the Rigveda suggest
that the sun is the chariot of the Aśvins. According to RV 6,9,1, “the white day and the black
day—(the pair of) light and darkness—manifestly turn around.”
If the white day and the black night are the two Aśvins, their association with the red dawn
as their sister-wife is most natural: the three are mentioned together for example in RV 7,80,1,
where the Vasiṣṭhas praise Uṣas as one who turns around the darkness and the light, the two
contiguous ones. In RV 10,39,12, the Aśvins are asked to come with their chariot
manufactured by the Ṛbhus, which is quicker than the mind and at the yoking of which is born
the daughter of the sky (i.e., the dawn) and Vivasvant’s two beautiful days (i.e., the white day
and the black day = night). Vivasvant is the sun god, who according to RV 10,17,2 is the father
of the Aśvins. In the Atharvaveda (13,3,13), Mitra and Varuṇa are connected with the two
forms that the fire god Agni has during the day and night: “This Agni becomes Varuṇa in the
evening; in the morning, rising, he becomes Mitra.” Thus Mitra and Varuṇa are the sun and
the fire, the deities of the agnihotra sacrifice—which, in my opinion, is an early variant of the
gharma offering to the Aśvins. The agnihotra is performed at sunset and sunrise, to Agni (Fire)
and Sūrya (Sun). In RV 10,88,6, “Agni is the head of the earth in the night, of him is the rising
sun born in the morning.” Both Mitra and Varuṇa and the two Aśvins are equated with day and
night in the Brāhmaṇa texts.
The sun and the fire—the day-sun and the night-sun—thus seem to be the cosmic and
atmospheric phenomena that the two Aśvins were originally conceived of as representing.
Instead of the fire, the moon could conceivably represent the nocturnal counterpart of the day-
sun; in addition, the rising sun and the moon are not infrequently seen together in the sky in
the morning. The Atharvaveda (7,81,1) offers a characterization of the sun and moon (to whom
the hymn is addressed) that fits well with the youthful Aśvins: “these two playing young ones
by their magic power move eastwards and westwards around the ocean.” In comparison to the
sun and moon, the morning and evening star are much less significant phenomena to qualify
as royal symbols; though they are connected with the Dioskouroi in the classical (but not the
earliest) Greek tradition as well as with the “sons of God” in Baltic folk songs, this can be seen
as a natural later development with these deities of the dawn and dusk, a development that
could have taken place independently in Greece and the Baltics. In India, a different astral
identification took place: the Aśvins were associated with one of the calendrical asterisms
consisting of two stars, called aśvayújau. They have also been equated with the zodiacal stars
of Gemini.
Mithra and Ahura are the Avestan counterpart of Vedic Mitra-and-Varu ṇa. The original
meaning of Asura, the principal epithet of Varuṇa, can be seen from *asera-, “lord, prince,”
attested in Uralic languages as an early loanword from Proto-Indo-Aryan (in Proto-Uralic,
labial vowels could occur only in the first syllable, so *u was replaced with *e). In
Zarathuštra’s pantheon, Asura Varuṇa became Ahura Mazdā, “Lord Wisdom,” whose supreme
symbol is fire. The Vedic word medhā´-, “wisdom,” which is the etymological counterpart of
Avestan mazdā- (both go back to Proto-Aryan *mṇz-dhā´-), is connected especially with the fire
god Agni. As we have seen, Varuṇa is the night/fire partner of the pair Mitra-and-Varu ṇa, alias
Day-Sun-and-Night-Sun, alias Sun-and-Fire.
This agrees with the fact that Apąm Napāt, “son of the waters,” is called Ahura and is
coupled with Mithra in the Avesta, while in the Veda, Apāṃ Napāt is another name of the fire
god Agni and is conceived of as a horse-shaped sun-fire in the waters. Varuṇa, too, is
connected with the waters as their lord, Apāṃ pati. According to the Kauṣītaki-Brāhmaṇa
(18,9), the sun becomes Varuṇa after it has entered the waters. The horse (áśva-) is often said
to belong to Varuṇa.
In Iran, Mithra’s name eventually came to mean the sun. In the Avesta, Mithra is a warrior
god wielding the vazra as his weapon, and in front of him runs the god of victory, V r θraγna,
in the shape of a wild boar. In the Veda, the vajra-wielding Indra, slayer of Vṛtra (vṛtrahan-), is
the counterpart of Mithra, the chariot-warrior, the solar god of the day. In the Veda, Indra’s
charioteer is Bṛhaspati, the purohita of the gods. Bṛhaspati or Brahmaṇaspati, “Lord of the
Magic Song,” was originally an epithet of Indra himself, whose kingship thus also comprised
the priestly function. Bṛhaspati’s becoming a separate purohita-figure, and the creation of the
dual divinities Indra-Bṛhaspatī and Indra-Agnī, was undoubtedly in imitation of the old pair
Mitra-and-Varu ṇa (Gonda 1974:52–53, 276, 281, 291–292, 325–330). Bṛhaspáti- is also an
epithet of Agni, the purohita of the gods.
Indra’s charioteer in the Mahābhārata is called Mātali. In the form Mā´talī, this name is
attested for the first time in the funeral hymn RV 10,14. One of the three occurrences in the
Atharvaveda mentions the chariot-brought immortal medicine known to Mā´talī that is in the
waters; it resembles the Rigvedic verse (by Medhātithi Kāṇva) 1,23,1, which mentions “the
nectar in the waters, the medicine in the waters”—this is recited at the vājapeya rite before
the chariot race, when the horses are bathed in water. Bheṣajá-, “medicine,” and the chariot
associate Mā´talī with the Aśvins, the divine charioteers and divine physicians. Mā´talī is a pet
name abbreviation of Mātaríśvan-, who is mentioned twenty-seven times in the Rigveda
(mostly in Books 1 and 10) and twenty-one times in the Atharvaveda. In several passages
Mātaríśvan is the name of Agni, and once (RV 1,190,2) he is a form of Bṛhaspati (the purohita
and charioteer of Indra). Otherwise Mātaríśvan is mostly spoken of as an Indian counterpart of
the Greek Prometheus, who brought the hidden fire (Agni) to men from heaven or who
produced it by friction.
Stanley Insler has convincingly proposed that Mātaríśvan- comes from earlier *Ātaríśvan-,
“master of fire” (see Parpola 2005a:26–27). The latter part of the compound, íśvan-, has
preserved the original unreduplicated root iś-, “to master”; it has a direct equivalent in
Avestan isvan-, “master.” In about one-third of the occurrences of the name Mātariśvan, it is
preceded by a word ending in -m, and the sequence … m *ātaríśvāwas reinterpreted as …
mātaríśvā, folk-etymologically associating the name with mātar-, “mother.” The reason for this
reworking was that the word *ātar-, “fire,” had become obsolete in the Vedic language, where
agní-, “fire,” has completely supplanted *ātar-, just as ātar- has replaced agní- in Avestan.
There is no myth among the Iranians about the theft of fire from the gods, because fire is the
creation and protected offspring of Ahura Mazdā. I suspect that *Ātar-iśvan originally was a
Proto-Iranian deity and came to South Asia with the Dāsas.
Mātariśvan has a close Greek parallel in Prometheus (Promātheús in the Dorian dialects),
whose name is likely to have originally meant “robber,” etymologically related to the root
math-, “to steal, rob,” which is often used in connection with Mātaríśvan (cf. also Sanskrit
pramātha-, “robbery”). I have no doubt that the myth is also connected with the homophonous
root math- / manth-, “to whirl round, to rotate” (so as to produce cream by rotating a
churning-stick in milk, or fire by rapidly rotating a dry wooden stick in a hole of a dry wooden
plank, that is, a “fire-drill”). Mātariśvan also kindled the hidden fire for the Bhṛgus, who are
spoken of as ancient sacrificers along with the Atharvans and Āṅgirasas. The Bhṛgus are not
only a priestly clan who discovered fire for mankind, they are also mentioned as chariot
builders in the Rigveda. In the Brāhmaṇa texts, Bhṛgu is the son of Varuṇa. The word bhṛgu- is
considered etymologically related to bhárgas-, “effulgence,” which characterizes Agni when
the fire is born out of “power” (i.e., the strong friction of the kindling stick); these words seem
to have a cognate in Greek phlóks (gen. phlogós) and Latin flamma. Mātaríśvan’s double
association with the origin of the fire and with the chariot (as Indra’s charioteer) is paralleled
by the chariot-driving Aśvins’ association with the two pieces of wood used for producing fire
by friction in the Rigveda (10,184,3). Generation of fire with a fire-drill is compared to
generation of offspring, and also in this sense connected with the Aśvins (see the end of
chapter 11).
Agni is the divine priest, the purohita of the gods. As Agni conveys the offerings to the
gods, he is “the charioteer of the rites” (TS 2,5,9,2–3 rathī΄r adhvarā´ṇām). “Fire” is therefore
called váhni-, “driver, charioteer,” from the root vah-, “to drive in a chariot, convey by
carriage.” Varuṇa’s connection with the charioteer is apparent from an episode in the royal
consecration: when the king goes to the house of the sūtá-, “herald, bard,” he makes an
offering to Varuṇa and gives a horse as a sacrificial jpgt (ŚB 5,3,1,5). In the Indian epics, the
charioteer gives the hero advice and encourages him in battle by singing of the feats of his
ancestors; hence sūtá- means both “charioteer” and “bard.”
Varuṇa has the title samrā´j-, generally understood to denote “universal ruler,” but in my
opinion literally “co-ruling” (a title which in itself suggests dual kingship) in the sense of
“priest-king.” According to the Śatapatha-Brāhmaṇa (5,1,1,12–13), “the royal consecration
(rājasūya) is only for the king. For he who performs the royal consecration becomes the king;
and unsuited for kingship is the Brāhmaṇa. . . . By performing the vājapeya [which is a
sacrifice of the Brāhmaṇa and of Bṛhaspati, the purohita of the gods, see ŚB 5,1,1,11], one
becomes the samrā´j; and the office of the king is the lower, and that of the samrā´j the
higher. . . .” The performer of the vājapeya, who becomes samrā´j, is not supposed to stand up
in front of anybody—the symbol of the samrā´j is the throne (āsandī, ŚB 12,8,3,4). Here we
can see how the Brahmins are making the priest-king higher than the mundane king. Varuṇa
was originally the samrā´j in the sense of the priest-king, but came to be considered as the
ruling king: in Śatapatha-Brāhmaṇa 4,1,4,1–6, Varuṇa represents kṣatrá-, the ruling power, and
Mitra is the purohita and bráhman-, representing priesthood. This led to Mitra’s being
practically eclipsed from the dual kingship by Varuṇa; and in Zoroastrianism, too, Ahura
Mazdā ousted Mithra.
However, in Śatapatha-Brāhmaṇa 11,4,3,10–11, while Varuṇa is called samrā´j, it is Mitra
who is connected with the ruling power (váruṇaḥ samrā´ṭ samrā´ṭpatiḥ . . . mitráḥ kṣatráṃ
kṣatrápatiḥ). Moreover, in RV 6,68,3, Varuṇa is called a vípra, “[sacred] poet,” in contrast to
the warrior Indra, who slays Vṛtra with his mace. “There can be no doubt whatever that
Mitra’s characteristic role is that of a king and not that of a priest: the evidence of the RV is
overwhelming and confirmed as genuine by the Avesta” (Thieme 1957:8). Thus originally
Mitra represented mundane kingship, Varuṇa priesthood. In the Atharvaveda, Varuṇa is a
master of the magic, which was the domain of the royal purohita.
11
The Aśvins as Funerary Gods

The Aśvins and Mitra-Varu ṇa stood for the two sides of kingship and of the solar god, the sun
and the fire, day and night. In the Vedas, the night, darkness and Varuṇa are all associated
with death. That the Aśvins were linked not only with day and night but also with life and
death as early as Proto-Aryan times is suggested by Greek evidence. According to Homer, the
Dioskouroi “have this honour from Zeus, albeit in the nether world, they pass from death to
life and life to death on alternate days, and enjoy equal honors with the Gods” (Odyssey
11,298–304, translated Farnell 1921:181). Sometimes one of the Dioskouroi is depicted on a
white horse and the other on a black horse—corresponding with the idea that one is immortal,
belonging to the celestials, the other mortal and belonging to the deceased.
The Dioskouroi were “saviors” (sōtêres) assisting in all kinds of trouble, notably at sea, on
the battlefield and during illness; they also figure in funerary inscriptions and frequently
appear on Roman sarcophagi. In The myth of return in early Greek epic (1978), Douglas
Frame argues that nóstos, “homecoming, return,” in early Greek religion, primarily meant
“return from (darkness and) death,” “coming back to (light and) life,” and that the miracles
performed by the Dioskouroi as saviors largely denote revivals from death. “In the Iliad … the
role of the chariot seems to be chiefly as a means of transport to and from the battlefield.
Generally the warrior leaps down to fight—his charioteer standing by to carry him out of
danger if things go badly” (Wace and Stubbings 1962:521). The rescue function thus belonged
to the charioteer, and the term nāsatya-, “effecting safe homecoming, rescuer, savior,”
therefore denotes this member of the chariot team, although it is used in the Rigveda in an
elliptic dual (which omits the other member) of both Aśvin twins. Sanskrit nāsatya- is related
to Greek néstōr, which in Homer is the name of the old king of Pylos, famed as a horseman
(hippóta) and an experienced charioteer (Frame 2010). Both words go back to a Proto-Indo-
European root *nes-, “to come home safely,” from which also derive Greek nóstos, Sanskrit
astam-aya-, “homecoming, sunset,” Gothic ganisan, “to be saved,” and Modern German
genesen, “to be healed.”
In India, the equestrian twins were conceived of as saviors, too. That this is due to an Indo-
Iranian heritage is suggested by the fact that the twins were also invoked by people in peril at
sea, despite the Vedic people no longer having any direct contact with the sea. A funerary
function for the Aśvins is suggested by the stories of Atri, Kakṣīvant, Cyavāna, Vandana, and
several other persons whom the Aśvins rescued from distress or rejuvenated, even though
they were lying buried, as if dead. The rejuvenation accomplished by the Aśvins is several
times compared to the renovation of an old chariot. That the Aśvins were often funerary
divinities who resurrected the dead is clear from their assistance to Vandana. Vandana had
become decrepit with age; his regeneration out of the ground (also: womb) is compared to the
skillful repair of an old chariot that threatens to fall into pieces (RV 1,119,7). Vandana had
been buried and was like one sleeping in the lap of the goddess of destruction (i.e., a dead
person); he rested like the sun in darkness; the Aśvins dug him up like a buried ornament of
gold, beautiful to look at (RV 1,117,5). In another hymn, the dug-up Vandana is compared to a
dug-up hidden treasure (RV 1,116,11). The Aśvins lifted him up so he could see the sun (RV
1,112,5), that is, live. The Aśvins dug Vandana up from a pit (RV 10,39,8), that is, a grave.
In the Sintashta culture of the southern Urals, deceased chieftains were buried with their
horses and chariots. The chariot was intimately involved with burial rites, and was probably
assumed to take the dead to the next world. In the twenty-third song of the Iliad, Homer,
describing the funeral of Patroclus, reports (verses 171–172) that four horses were cast upon
his pyre. The chariot was involved, too, but in a different way. In athletic contests to honor the
dead hero, his belongings were divided up as victory prizes. The most important of these
contests was the chariot race. Describing it, King Nestor refers to another funeral chariot
race, in which he was beaten by the sons of Aktor, “Siamese twins,” one of whom held the
reins while the other brandished the whip. A comparable horse race, performed by riders on
the day of the burial, occurred in the pre-Christian traditions of the Baltic people (Caland
1914). There the prize consisted either of money placed on top of the goal post, or of property
of the deceased, placed at certain intervals along the route. The day ended in a drinking bout.
One would expect a funeral chariot race to have survived in ancient India, but despite
considerable research, no such connection has been found (Sparreboom 1985). I believe that a
reference has survived in a hymn to the Aśvins, RV 1,116,2: “O you two, who had triumphed
with (your) strong-winged (horses), urged to a fast course, or through the incitements of the
gods, (your) ass won a thousand (cows) in Yama’s prize-contest, O Nāsatyas.” Many scholars
have commented upon this verse, but none has interpreted the phrase “Yama’s prize-contest”
(ājā´ yamásya pradháne) as referring to a funeral chariot race, although Yama is the god of
death and the Yama hymns of Book 10 of the Rigveda were used in funeral rites. Éric Pirart
(1995), however, significantly observes that the verb śad-, “to triumph,” used here of the
Aśvins with regard to the chariot race, is derived from the same Indo-European root as the
Greek name Kástōr, born by one of the Dioskouroi twins, the “tamer of horses.”
The Rigveda specifies that the Aśvins won a thousand cows or verses with their ass in
Yama’s prize-contest. Such a race won by the Aśvins with asses is described in Aitareya-
Brāhmaṇa 4,7–9. The context is the 1000-versed āśvina-śastra- (“praise to the Aśvins”). The
āśvina-śastra- is recited at dawn, when ending an “overnight” (atirātrá-) soma sacrifice, which
has lasted a whole day (representing a full life) and continued throughout the following night.
Night represents death, and dawn rebirth. The text stresses the generative power of the ass,
and this seems to be an important reason why the Aśvins drive a chariot pulled by asses in the
funeral race of Yama; another reason is undoubtedly that the wild ass, which is stronger than
the horse, is associated with death (Parpola & Janhunen 2011).
The symbolism of the āśvina-śastra- suits the funeral context well, for “the āśvina is … the
chariot of the gods. With this chariot of the gods he attains in safety the world of heaven. (The
śastra) should include the suparṇa [the ‘good-feathered’ bird]; the suparṇa is a bird; becoming,
like it, a winged one, a bird, he attains in safety the world of heaven.” (KB 18,4) The suparṇa,
which flies to heaven, is Varuṇa’s messenger at the seat of Yama (RV 10,123,6), and thus
associated with death. In funerary pots of the Late Harappan Cemetery H culture of the
Punjab (c. 1900–1300 BCE), the peacock is depicted as carrying the dead (who lie horizontal
inside the peacock) to the stars in heaven (Fig. 15.9). Birds are connected with the Aśvins, too,
for the horses pulling their airborne chariot are compared to birds (RV 6,63,7).
The āśvina-śastra- contains a thousand verses. The Aitareya-Brāhmaṇa (2,17) explains their
meaning: “A thousand [verses] should be recited for one desiring heaven; the world of heaven
is at a distance of a thousand journeys of a horse hence; (they serve) for the attainment of the
world of heaven, the securing, the going to (the world of heaven).” The number is also
connected with the sun, which is said to have a thousand rays—and the sun’s rays are often
understood as cattle; one thousand cows or verses was the prize in Yama’s chariot race won
by the Aśvins.
The funeral context of the āśvina-śastra has gone unnoticed, because the Rigveda
Brāhmaṇas introduce its exposition by associating it with the marriage of Soma (the Moon)
and the Solar Maiden: “Now when Savitṛ gave Sūryā to Soma, the king, he made over to his
daughter whether she was Prajāpati’s (or his own) on marriage this thousand (of verses) that
was in the possession of these deities; they said, ‘Let us run a race for this thousand’; they ran
the race, then the Aśvins were victorious by means of the ass” (Kauṣītaki-Brāhmaṇa 18,1,
translated Keith 1920:444–445).
Actually, the marriage context is also very relevant, for the conclusion of funeral rituals
aiming at rejuvenation and the attainment of heaven coincides with the beginning of new life
in the impregnation that occurs at a wedding. The mythical explanation of the cup of soma
offered to the Aśvins (āśvinagraha-) at the morning pressing of the soma plant also links
marriage and rejuvenation. Cyavana, a sage decrepit with old age, lay ghost-like on the
ground, and was offended by the people of Śaryāta Mānava—they threw clods of earth at him
(loṣṭa, the word used here for “clod,” is also used of the earthen “bricks” of the funeral
monument, loṣṭa-citi-). The angered sage caused problems for Śaryāta, who gave his beautiful
daughter Sukanyā to the sage in atonement. Through Sukanyā, Cyavana made the Aśvins
rejuvenate him: Cyavana was thrown into water (or: pool of youth in the River Sarasvatī), and
emerged young again. Finally, he sacrificed with a thousand cows (the prize won by the Aśvins
in the race).
Regeneration implies reentering the womb: “A son is a light in the highest heaven. The
husband enters the wife; having become a germ (he enters) the mother; having becoming
renewed in her, he is born in the tenth month” (Ślokas of the Śunaḥśepa legend recited at the
royal consecration, according to Aitareya-Brāhmaṇa 7,13). The Aśvins are deities of both death
and (re)birth, saving people by helping them make the dangerous, liminal passage. They
appear in the morning and evening, at the junctures between night and day, or death and life:
Janus-like, their white-and-black appearance unites these opposites. In this, they resemble the
Dioskouroi.
Among the singer families of the Rigveda, the Kāṇvas and Atris are prominent worshippers
of the Nāsatyas. Of the fifty-four complete hymns addressed to the Aśvins, as many as thirty-
three are in “Kāṇva” books: sixteen in Book 1, twelve in Book 8, and five in Book 10. Six hymns
are in the Atri book, Book 5. There are eight Aśvin hymns in Book 7, which belongs to the
Varuṇa-worshipping Vasiṣṭhas. The contrast with the remaining “family books” is great: Book 2
has just one hymn, Book 3 just one hymn, Book 4 has three hymns, and Book 6 only two. The
references to the gharma rite, which is special to the Aśvins, are of similar distribution: they
occur most frequently in Books 1, 5, 7, 8, and 10.
Kāṇva poets of the Rigveda resided in Gandhāra, the area of the Gandhāra Grave culture
(c. 1700–900 BCE), the first culture in northern Indus Valley to possess the domesticated horse
(Fig. 11.1; Fig. 4.4). Suvāstu, the ancient name of the Swāt River, is mentioned in the Rigveda
only in the Kāṇva hymn 8,19 (verse 37). Women from Gandhāra are mentioned once in book I.
Besides toponyms, cultural criteria back up the location of the Kāṇva poets. Thus, in contrast
to the family books, the Kāṇva hymns repeatedly refer to shepherding and to plowing, which
were both practised in Gandhāra. Out of the five Rigvedic occurrences of the word uṣṭra,
“camel,” four are in Book 8 and one in Book 1. The Kāṇva hymns are the only ones in the
Rigveda to have proper names of the type “having god NN as his guest,” such as Medhātithi.
This name type connects them with the Proto-Indo-Aryans of Mitanni, who took northern Syria
in their control by 1500 BCE and fairly certainly came from the north Iranian branch of the
BMAC (chapter 8). This migration took place around the same time as the Ghālegay IV Period
immigration of the Gandhāra Grave culture from the BMAC into the Swāt Valley and
surrounding regions, probably bringing the Kāṇvas to Gandhāra.

FIGURE 11.1 The location of the Gandhāra Grave culture in relation to the Andronovo cultures, the BMAC (“Central
Asian towns”), and the Late Indus civilization. After Mallory 1989:230, ill. 127. Courtesy J.P. Mallory.

The great majority of the Rigvedic Aryans connected with the family books are likely to
have crossed the Gandhāra and pushed on to the plains. The Atri clan associated with Book 5,
however, remained in Gandhāra. Atri hymn 5,53 mentions the Kābul and Kurum rivers, along
with other rivers of Afghanistan and northwest Pakistan. The Atris were in close and friendly
contact with the Kāṇvas, which suggests that they resided in each other’s neighborhood. The
gharma rite appears to have originally been an independent rite connected with the worship
of the Aśvins that was developed by the Atri clan. Even the Atris predominantly worship Indra.
We may assume that they introduced the gharma rite into the Indra cult, as a result of coming
under the influence of the Aśvin-worshipping Kāṇvas in Gandhāra. The post-Rigvedic legends
telling how the Aśvins obtained a share in the soma sacrifice are likely to refer to this stage of
development. The incorporation of the Aśvins’ gharma offering as a minor component in the
soma sacrifice, one of its many heterogeneous elements, and the Aśvins’ obtaining a share of
soma signal the Aśvins’ submission to Indra. The Aśvins were now second-rank deities of the
Vedic pantheon, mainly associated with healing. They are expressly called physicians, and
were looked down by other gods on account of their too close relationship with human beings.
According to the Maitrāyaṇī Saṁhitā (4,6,2), the Aśvins were not originally soma-drinkers,
and asked for the āśvina cup of soma (offered in the morning outside the actual place of
sacrifice) as a reward for healing the originally “headless” soma sacrifice. The Aśvins healed
the sacrifice by providing it with a “head” in the form of the introductory pravargya rite,
which originally belonged to them. Consequently, a Vedic sacrificer was not supposed to
perform the pravargya rite in his first soma sacrifice (which was at first “headless”), only in
his second sacrifice and thereafter.
The account of the origin of the pravargya rite in the Śatapatha-Brāhmaṇa (14,1,1) may be
summarized as follows. (1–5) The gods, excluding the Aśvins, performed a sacrificial session in
order to attain glory. They agreed that whoever among them would first reach the end of the
sacrifice would be the most excellent among them, and the glory would then be common to
them all. Makha-Vi ṣṇu did so, and became the most excellent. (6) Viṣṇu, who personified the
sacrifice (makha) and the sun, was unable to contain his love of glory. (7–10) Taking his bow
and three arrows, he stepped forth and rested his head on the bow. The other gods did not
dare attack, but instead asked the ants to gnaw Viṣṇu’s bowstring. When the ends of the bow
sprang asunder, they cut off Viṣṇu’s head, which fell and became the sun. His headless trunk
lay stretched out with its top part pointing toward the east. (11) The gods said: Our great hero
(mahān vīraḥ) has fallen. That is why the pravargya vessel is called Mahāvīra. With their hands
the gods wiped up (sam-mṛj-) the vital sap that flowed out of him, and that is why the
Mahāvīra is Emperor (samrāj). (12) The gods rushed forward to him, eager to gain his glory.
Indra reached him first, applied himself to him, limb by limb, and became possessed of his
glory. (13) Indra became makhavat-, “possessed of Makha,” this being (mystically) the same as
maghavat-, “possessed of booty” (Indra’s usual epithet). (14–17) The gods rewarded the ants,
divided among themselves Viṣṇu’s body into three parts (i.e., the three services of the soma
sacrifice), and continued their worship with a headless sacrifice. (It failed, however, because
the sacrifice was headless.)
(18) Dadhyañc Ātharvaṇa knew this śukra, “essence, seed,” the secret of how the head of
the sacrifice is put on again, how the sacrifice becomes complete again. (19) Indra (who had
obtained only the glory of the headless body) threatened to cut off Dadhyañc’s head if he told
the secret to anybody. (20–22) The Aśvins heard that Dadhyañc possessed this secret
knowledge and asked to become his disciples, but Dadhyañc refused to teach them,
mentioning Indra’s threat. (23) The Aśvins suggested the following scheme: after they became
his pupils, they would cut off Dadhyañc’s head and replace it with a horse’s head. Dadhyañc
would then teach them with this horse’s head, and when Indra cut the head off, the Aśvins
would put Dadhyañc’s own head back again. (24–25) This scheme was then realized, as told in
Rigveda 1,116,12 (where it is said that horse-headed Dadhyañc Ātharvaṇa taught madhu,
“honey,” to the Aśvins; in ŚB 4,1,5,18, which deals with the āśvina cup of soma to the honey-
loving Aśvins, Dadhyañc is said to have told them “the secret knowledge called Honey,” madhu
nāma brāhmaṇam).
Also the pravargya chapter of the Taittirīya-Āraṇyaka (5,1) calls the decapitated hero
Makha Vaiṣṇava, but in the Kaṭha-Āraṇyaka (chapter 3) it is Rudra whose head is cut off. This
suggests that Makha is the original name, and indeed the hero is called just Makha in the
oldest version in the Maitrāyaṇī Saṁhitā (4,5,9) as well as in the Pañcaviṁśa-Brāhmaṇa (7,5,6).
Makha reminds one of the name Sumakha associated with one of the Aśvins and its probable
etymological connection with Greek mákhē, “fight.” Sumakha would have personified the
chariot-warrior; this Aśvin would have become the god Mitra, whose name later in Iran
denoted the sun. The body of the hero fell with its top to the east, and its cut-off head became
the sun. In the Bṛhad-Āraṇyaka-Upaniṣad (1,1), the sacrificial horse is equated with the
cosmos, and its head with the dawn. The fact that the secret of the pravargya is told by a
decapitated sage wearing the head of a decapitated horse seems most significant.
From the oldest source available on the Vedic horse sacrifice, the Rigvedic hymns 1,162–
163, we know that the horse was slaughtered by cutting off its head with an ax or a butcher’s
knife (svadhiti-, RV 1,163,9), not by strangling, as in the “classical” Vedic ritual. Hence, it
would appear, the existence of two burials of a headless body of a horse in the BMAC capital
Gonur (chapter 8). The Dadhyañc legend also speaks of providing a human sage with a horse’s
head. Such an operation may have been an integral part of the secret rite of reviving a dead
hero. It is tempting to see this Aśvin-related revival rite revealed in a human skeleton with the
skull of a horse, which was excavated in a unique grave near Samara in the Mid-Volga region
of Russia; it belongs to the Potapovka culture, dated to c. 2100–1700 BCE, and was very
probably the skeleton of a Proto-Indo-Aryan speaker (chapter 7) (Fig. 11.2). The secret
knowledge of revival is, moreover, connected with honey, which the Vedas connect above all
with the Aśvins, the divine healers. In Finnic epic poems of Kalevala, which contain survivals
of Proto-Aryan religious concepts (chapter 7), the mother of a dead hero collects his
dismembered corpse and revives it by means of honey (Uralic *mete from early Proto-Aryan
*medhu) fetched from the highest heaven by a bee.

FIGURE 11.2 A human skeleton with a horse’s skull. Potapovka kurgan burial, grave 1, near Samara on the mid-
Volga in Russia, c. 2100–1700 BCE. The arrow points to the north (C). After Vasil’ev et al. 1994:115, fig. 11. Courtesy
Pavel V. Kuznetsov.

It must be noted, however, that the evidence of the Potapovka grave has been questioned,
because of a suspected mixing of the archaeological layers. Moreover, some molded tablets
from Harappa show anthropomorphic deities with animal heads: a dancer resembling later
Śiva Naṭarāja has the head of the water buffalo (CISI 1:207 H-175), and a long-armed deity
within a fig tree has a ram’s head (Fig. 21.16a) like the later god Naigameṣa (chapter 21).
I assume that the Kāṇvas originally worshipped the Aśvins with a horse sacrifice and with
funeral rituals implying revival of the dead. In the Rigveda, Sage Atri is the person most
closely connected with the gharma rite, the “hot offering” of milk and ghee to the Aśvins,
which the Atri clan apparently developed under a strong Kāṇva influence. Later called
pravargya, this offering was performed twice a day, in the morning and evening, during the
preparatory phase of the soma sacrifice, the principal rite connected with the cult of Indra.
Some of the principal acts in the pravargya rite—all accompanied by mantras—are the
following. A specially prepared clay pot is placed on a separate mound near the gārhapatya
fire, filled and anointed with ghee, surrounded with live coals and fuel, and fanned until the
pot glows red-hot. Then all priests and the sacrificer stand up and reverently watch the pot.
The adhvaryu and pratiprasthātar priests (identified with the Aśvins among the gods) milk a
cow and a goat and pour the milk into the heated pot, which is full of boiling ghee. A pillar of
fire issues from the pot. When the pot has cooled down a little, it is brought to the āhavanīya
fire, and offerings to Indra and the Aśvins are poured into the fire. The pot is filled with curds
or milk, which overflows into the āhavanīya. The remainder of the curds is partly used for an
agnihotra offering and partly eaten by the priests and the sacrificer. The gharma pot and all
accessory implements are placed upon a black antelope skin on a throne. On the last day the
gharma implements are taken to the place of their disposal and laid down on the ground in the
shape of a man, with the gharma pot as the head. This displacement of the gharma
implements is paralleled by the placement of a Vedic sacrificer’s ritual implements upon his
corpse in the funeral. In the fractional burials of the Gandhāra Grave culture, the skull is
always placed in the uppermost position, as mentioned in chapter 8.
The concept of “head” is very much associated with the gharma vessel, which, as we have
seen, is equated with the head of a decapitated divine hero, called mahāvīra, “great hero,” and
“Makha’s head.” The gharma pot may be compared with the vessel most characteristic of the
Gandhāra Grave culture, namely the “face urn,” the funerary vessel made to look like a human
head. Such an urn functions as an ossuary for selected bones collected after cremation
burials. As already observed in chapter 8, cremations constitute about one-third of the
Gandhāra Graves and predominate in the Ghālegay V Period. The face urns are characteristic
of the Ghālegay V Period; some have just holes for the “eyes” and the “mouth,” others have
both these holes and a protruding “nose” (Fig. 11.3a).

FIGURE 11.3 Funerary “face urns” of the Gandhāra Grave culture. (a) An urn from Zarif Karuna, near Peshawar,
with a human-like face including a big nose. Photo Asko Parpola. (b) A lid with a horse-shaped handle from Loebanr,
Swat. Photo Asko Parpola.

Attempts have been made to trace parallels with this distinctive artifact in other
archaeological cultures of Eurasia. Giorgio Stacul (1971) suggested a connection with the
anthropomorphically inspired face urns with cremation remains found in the final phase of the
Middle Danubian culture of Baden-Pécel (c. 2000 BCE). In the Balkans and the Middle Danube,
anthropomorphic containers go back to the Late Neolithic period, where they were not as yet
related to burial rites. In spite of the several other parallels noted by Stacul between the
Gandhāra Graves and the cultures of the Great Hungarian Plain, it has not been possible to
confirm any historical connection. The distance in space, and to some extent in time, is
considerable. The Gandhāran face urns seem to be an independent local development.
The symbolic parallels between the gharma vessel in the cult of the Aśvins—the
“possessors of horses”—and the face urn of the Gandhāra Graves appear not to be just
coincidental, for the horse-shaped handle of the lid of a face urn (Fig. 11.3b) suggests that
these urns had a cultic relationship with the horse and therefore probably with the Aśvins.
Equally significant is the fact that the conspicuous three-dimensional “nose” of the later face
urns has a counterpart in the gharma vessel. The Śatapatha-Brāhmaṇa (14,1,2,17) describes
the preparation of the gharma vessel thus:

He then takes a lump of clay and makes the Mahāvīra (pot) with [the mantra], “For
Makha thee! for Makha’s head thee!” … a span high, for the head is, as it were, a span
high;—contracted in the middle, for the head is, as it were, contracted in the middle. At
the top he then draws it out [unnayati] (so as to form) a spout [mukham “mouth”] of
three thumb’s breadths (high): he thereby makes a nose [nāsikām] to this (Mahāvīra, or
Pravargya).

It is true that this description does not exactly match the making of the face urns of the
Gandhāra Graves, but it is remarkable that the gharma pot, alone of all the vessels described
in the Vedic literature, is expressly told to have a “nose.” It is also true that the Śatapatha-
Brāhmaṇa is the only Vedic text to mention this “nose,” and although its present redactions
are younger than the texts of the Black Yajurveda, it goes back to an earlier version and its
contents in some respects differ entirely from all the other texts. That it could well have
preserved ancient traditions prevalent among the Kāṇvas is quite likely on the basis that one
of the two redactions belongs to the Kāṇvas. The “nose” of the face urn and the gharma pot
should have some specific function, and indeed it is meaningful when considered in the
context of the Aśvin cult.
Comparing the gharma vessel with the cinerary urn of the Gandhāra Grave culture implies
that there should be a close relationship between the gharma rite and the funeral. I think such
a relationship does exist. According to the Rigveda, the gharma rite was instituted by Atri and
offered by his descendants out of gratitude to the Aśvins, because the Aśvins had saved Atri
from the distress of the ṛbīsa pit. It was the Aśvins who had first given the hot gharma drink to
Atri while saving him, so the gharma rite imitates the service rendered by the Aśvins. I hereby
argue that the Atri legend reflects the Atri clan’s cremation burial and its association with the
adopted cult of the Aśvins as funeral deities, who revive the dead by means of their drink. This
funeral cult the Atris took over from the Kāṇvas, with whom they had established friendly
relations while settling in Gandhāra. In RV 6,50,10, the Aśvins “delivered Atri from great
darkness,” in RV 7,71,5 “from a narrow place and darkness.” These expressions usually refer
to death and to the state of an embryo. Atri’s case is similar to that of Vandana and many
persons saved by the Aśvins when they seem to be already dead or in a death-like state. As
“healers” and “saviors” the Aśvins were thus essentially revivers of the dead, in other words,
funeral gods, though this function has not been generally recognized.
What, then, is the meaning of the “nose” of the gharma pot and the face urn of the
Gandhāra Graves? In PIE, the verbal root *nes- meant “to come home safely,” and, as noted
above, from this are derived Gothic ganisan, “to be saved,” and Modern German genesen, “to
be healed.” As a result of the sound change that turned PIE *e into Indo-Iranian *a, that root
became homonymous with the word nas-, “nose.” In the Vedas, the Nāsatyas were associated
with the nose. The word nas-, “nose,” occurs only twice in the Rigveda. In RV 2,39,6 the two
Aśvins are compared to the nose with its two nostrils (nāsā in the dual)—the nose of one
Gandhāran face urn has the detail of two nostrils. Secondly, in RV 5,61,2c, in a hymn
addressed to the Maruts, the gods of storm and battle, who in verses 12–13 are represented as
driving a chariot, but in verse 2 apparently are riders: “(where is) the saddle (literally, seat) on
the back (of the horse), (where) the rein (literally, the restrainer) of the two nostrils (of the
horse)?” (pṛṣṭhé sádo nasór yámaḥ). This verse shows that the Sanskrit words yama- (meaning
“twin” as well as “rein”) and nas- (meaning “nose” as well as “to save”)—both associated with
the Aśvins in the Rigvedic language—had connotations definitely linked with a key equestrian
skill, the control of a horse through its nostrils. (This is one of the two certain references to
riding in the Rigveda.)
The Aśvins are invoked as deities of generation, and in Vedic rites of human fertility, the
nose plays a central role. In RV 10, 184, 2–3, a prayer for the birth of a son, the two Aśvins,
the lotus-garlanded gods, are asked to place an embryo in the wife. The embryo is equated
with the fire, the embryo of the waters, who is hidden in the aśvattha wood used as the fire-
drill: “The embryo, whom the Aśvins rub out of the two golden kindling woods, him we call
hither for you, so that you may bear him in the tenth month.” Fire is generated by rotating a
vertical male kindling-stick of aśvattha wood in a hole of a horizontal female plank of śamī
wood. In the Atharvaveda (AVŚ 6,1), this act is conceived of as sexual intercourse, and Fire as
the son of these two pieces of wood. According to the Kauśika-Sūtra (35, 6–10) of the
Atharvaveda, the life-cycle rite of puṁsavana aiming at the birth of a son is performed as
follows. After fire has been produced with the fire-drill, it is ground into powder, which is
mixed with butter coming from a cow that has born a male calf. The mixture is put with the
thumb of the right hand into the right nostril of the wife.
The myth about the “nose-birth” of the Aśvins is told briefly in a Rigvedic Yama hymn
(10,17,1–2) and more extensively in the Nirukta (12,10), in the Bṛhaddevatā (6,127–7,6 ed.
Tokunaga 1997) and elsewhere. Saraṇyū, the mother of the divine twins, had assumed the
shape of a mare. She became pregnant with the Aśvins when sniffing the seed of her husband,
the bright sun god, Vivasvant, who had assumed the shape of a stallion but in haste had
emitted his seed on the ground.
This “nose-birth” of the Nāsatyas is ritually included also in the pravargya rite. When
collecting the clay out of which the gharma pot is to be fashioned, at the moment when a goat
is milked so that its milk flows over the clay, a horse is made to sniff at the clay while the
adhvaryu priest recites: “Grant life, grant prāṇa, apāna, and vyāna, sight, hearing, mind, voice,
body, strength, mass—grant me all this!” (ĀpŚS 15,2,2–3). The milk symbolizes the generating
seed, as does the reviving gharma drink that the Aśvins gave to Atri. According to the
Aitareya-Brāhmaṇa (1,22), the gharma vessel is the male member, and the milk is the seed.
The seed is poured into the fire, the womb of the gods, as generation, so one who sacrifices
with the pravargya rite is reborn as an immortal god.
Thus there are many reasons to connect the funeral face-urn of Gandhāra, with its
prominent “nose,” to the gharma pot of the Aśvin cult, which also has a “nose,” representing
the head of a great hero called Makha. Here at last we have a very concrete link between
archaeological evidence and Vedic culture, a link that many people have been missing.
The gharma/pravargya ritual is the only context in the Rigveda in which sāmans (ritual
songs) are mentioned by name. The name of the most important sāman, rathantara, refers to
the chariot (ratha-), the vehicle of the Aśvins. The name occurs in the Rigveda just twice: in RV
1,164,25, immediately before the verses 26–29, which speak of the gharma/pravargya rite of
the Aśvins; and in RV 1,180,1 (along with the also important bṛhat sāman mentioned in verse
2), immediately before the gharma ritual (mentioned in verse 3). The gharma rite of the Aśvins
is the only ritual, besides the soma sacrifice, that is mentioned a number of times in the
Rigveda. The hymns addressed to soma were collected from all the “family” books of the
Rigveda in order to make Book 9—the basis for the Sāmavedic ritual. In view of the almost
complete absence of sāman names, and the omission of chanter priests from the Rigvedic
enumerations of priests, it would seem that the Sāmavedic liturgy for the soma sacrifice was
created out of the tradition of sāman singing in Aśvin rituals.
The sāmans in the gharma rite are much simpler than those sung in praise of soma; they
are solo songs chanted by just one person, although all participants sing the finale. In the
soma sacrifice, the sāmans are chanted by a trio of priests, consisting of the prastotar
(literally, “pre-lauder”), who sings the first portion, the pratihartar (“the receiver”), who sings
the second portion, the udgātar (“high-chanter”), who sings the most important portion, the
udgītha (“high chant”), while all three priests sing the finale, joined by other participants in
the ritual on certain occasions. This scheme seems to me to be an elaboration of the solo
sāmans, devised when the Sāmavedic ritual was created for the soma sacrifice.
The relatively simple preparation of the haoma juice in the Zoroastrian yasna ritual
includes filtering, but without the praise songs that in the Vedic soma sacrifice are addressed
to Soma Pavamāna, “Soma that is being purified,” by being strained through sheep’s wool. I
suspect that Soma’s praise songs may have been inspired by an earlier liturgy that praised the
surā, the beer made of barley that is sacred to Varuṇa. We know from the sautrāmaṇī ritual
that surā was also purified, using a horse-hair (vāla- / vāra-) filter. In this ritual, the king, after
first becoming a brāhmaṇa and a soma drinker for the royal consecration (rājasūya), is
transformed back into a warrior (kṣatriya) and therewith a surā drinker in this
“reconsecration” (punar-abhiṣeka). Perhaps the original drink of pavamāna hymns was surā,
but the laud-singing was transferred to soma when the Rigvedic community expanded to
include Kāṇva and Āṅgirasa poets, who were initially worshippers of the Aśvins and Mitra-
Varuṇa but later converted to the Indra cult after the Aśvins became accepted as soma
drinkers.
It is the Kāṇva and Āṅgirasa clans who must have created the Sāmaveda, for in addition to
Book 9, the Sāmaveda-Saṁhitā mainly consists of their hymns in Books 8 and 1. The Aṅgirases
are mentioned as composers of sāmans in the latest books of the Rigveda, in RV 1,107,2
(áṅgirasāṃ sā´mabhiḥ stūyámānāḥ) and RV 10,78,5 (áṅgiraso ná sā´mabhiḥ). The Sāmaveda
favors the meters (gāyatrī and pragātha) and the strophic structure (tristichs) characteristic of
the poetry of the Kāṇva and Āṅgirasa clans, which in this respect differs from the poetry of the
“family” books (Oldenberg 1884). The same applies to some extent to the language: thus the
root gā(y)-, “to sing, chant,” which is fundamental to Sāmavedic terminology, is seldom
attested in the “family” books, mostly in gāyatrī or pragātha stanzas in hymns with tristich
structure (Hopkins 1896). The words gāyatrá- and gāyatrī΄- are found in Books 2–7 only in the
admittedly late hymn RV 2,43, which is the last of Book 2.
The Sāmavedic priest udgātár- is mentioned in the Rigveda just once, in this late three-
versed hymn 2,43, which also speaks of singing a sāman in verse 2a (udgātéva . . . sā´ma
gāyasi) and of “a sāman singer who speaks in two ways, mastering the songs in gāyatrī and
triṣṭubh meters.” Though the Sāmaveda appears to have been created in South Asia, Aṅgiras,
the ancestor of the Āṅgirasa clan, already figures in the “family” books’, so the Rigvedic Indo-
Aryans must have met Aṅgirases already in southern Central Asia. The Old Iranian Avesta has
two different poetic traditions: Zarathuštra’s gāthās resemble the “family” books of the
Rigveda, while the later Yašt hymns are comparable to the Kāṇva and Āṅgirasa poetry.
Probably, the Yašt hymns were collected and recorded later than the gāthās for reasons
similar to the reasons why the Atharvavedic hymns were collected later than the hymns of the
Rigveda.
12
The Atharvaveda and the Vrātyas

The hymns of the Atharvaveda form the second-oldest text collection of the Vedic literature
after the Rigveda. The Atharvaveda of the Śaunaka school (AVŚ)—the most commonly used
recension—contains 730 hymns divided into twenty books. By contrast, the recension of the
Paippalāda school (AVP) has 923 hymns also divided into twenty books, but arranged in an
order different from that of the Śaunaka school. The AVP first became known from a single
corrupt manuscript discovered in Kashmir in 1873, but remained largely incomprehensible
until a living Paippalāda tradition with many manuscripts was discovered in Orissa about 1950
by Durgamohan Bhattacharyya. Its publication and exegesis are one of the most exciting
current projects of Vedic philology.
The hymn collections of the Rigveda, Sāmaveda, and Yajurveda mainly relate to the soma
sacrifice or other śrauta rituals and have therefore been considered to represent the “hieratic
religion” of the Vedas, in contrast to the “popular religion” of the Atharvaveda, with its “white
and black magic” and domestic (gṛhya) rituals such as initiation into learning (upanayana),
marriage, and funerals. These labels represent an attempt to understand the “Triple Veda”
and the Atharvaveda as two sides of a single integral religion, rather than a single phase in
the acculturation of two different religions: “Rigvedic” and “Atharvavedic.” Actually, “popular”
is not a good fit as a label, given that royal rituals and the duties of the king’s court priest (for
whom “magic” was important) are among the central topics of the Atharvaveda and its ritual
manual, Kauśika-Sūtra.
The “Atharvavedic” religion had already started to be accepted into the fold of the Rigvedic
religion in Books 1 and 10 of the Rigveda, which contain many hymns of clearly Atharvavedic
character, including the marriage and funeral rituals and the (not so “popular”) cosmogonical,
speculative, and riddle hymns. “Most western scholars are … agreed that the Rigveda contains
a number of hymns which by their special characteristics in subject, metre and language differ
from the main body of the poems and associate themselves instead with the Atharvaveda.
They are generally described as ‘popular’ and regarded as later additions …” (Gonda
1975:28).
By the time the Atharvavedic hymns were collected, their poets had accepted the Rigvedic
tradition, and vice versa. Many Atharvavedic hymns copy and imitate Rigvedic hymns, yet
without handling the poetical meters nearly as skillfully as in the Rigveda. This partial
adoption of Rigvedic traditions cannot, however, hide the fact that the Atharvaveda goes back
to a separate, and in many respects more archaic, Indo-Aryan tradition. The language of the
Atharvaveda, continued in the Brāhmaṇas and other later Vedic texts, as well as in epic and
classical Sanskrit, cannot be directly derived from the Rigvedic language. This is clear above
all from the fact that the “Atharvavedic” language has preserved the Proto-Indo-European
distinction between the sounds l and r in many words. The original Rigvedic language, on the
other hand, had completely lost the PIE *l as a result of so-called rhotacism, which merged
every *l with *r; and this change must have happened in southern Central Asia, before the
migrants entered South Asia, because the *l/*r distinction was also lost in Mitanni Indo-Aryan
(chapter 8) as well as in Avestan and Old Persian (Mayrhofer 2002 [2004]). Words with PIE *l
start to infiltrate the Rigveda in Books 1 and 10, in which l occurs eight times more frequently
than in the “family” books, while in the AV l occurs seven times more frequently than in the
RV. In the “family” books, l occurs mainly in foreign loanwords and proper names (see chapter
9), but no verbs preserve PIE *l (“to hear” is śru- against Greek klúō, Latin cluēre, Russian
slušat’) and only a few nouns (uloka-/loka-, śloka-, -miśla-), which have crept in when the text
of the Rigveda was finally fixed around 700 BCE (Pinault 1989:36–37).
The dialectal difference between the Rigvedic language on the one hand and the
Atharvavedic language and classical Sanskrit on the other was proposed by Maurice
Bloomfield (1899:46–47) and seconded by M. B. Emeneau (1966) and a number of other
scholars. The loss or preservation of the *r/*l distinction is paralleled in other linguistic traits.
In the declination of the nouns with a-stems, the Rigvedic dialect has innovated in the
masculine plural nominative by adding a second plural nominative marker -as (used in
consonantal stems) to the ending -ās, which continues the original PIE ending; the resulting
double ending -āsas is parallelled in Avestan (-āηhō), Old Persian (-āha), and Pāli (-āse).
Besides this innovated double ending, the original -ās is also found in the Rigvedic language
(though -ās is often to be read as -āsas to mend defective meter) as it is found in Avestan (-ā)
and Old Persian (-ā). While -ās is about twice as frequent as -āsas in the Rigveda, in the
Atharvaveda it is twenty-four times as frequent, and classical Sanskrit knows only -ās.
Similarly, in the instrumental plural of the a-stems, the Rigvedic language has the innovative
ending -ebhis (paralleled by Old Persian -aibiš) besides the slightly more common inherited
ending -ais (Avestan -āiš); while Rigvedic -ebhis is continued in Pāli -ehi and Prakrit -ehiṃ, in
the Atharvaveda -ais is five times as common as -ebhis, and classical Sanskrit knows only -ais.
In the verbal conjugation, the innovative ending -masi of the first-person plural present active
(paralleled by Old Iranian -mahi) is about five times as common as the inherited ending -mas,
while in the Atharvaveda -mas is commoner than -masi, which is no longer used in classical
Sanskrit. There are many further differences, not least in the lexicon.
In chapter 11 we saw that some poets of the Kāṇva and Āṅgirasa families early on joined
the Rigvedic culture, composing hymns in Books 1, 8, and 9, which form the basis of the
Sāmaveda that they apparently created. Evidently other Kāṇvas and Āṅgirasas remained more
faithful to their own traditions and composed the Atharvaveda, the oldest name of which is
atharvāṅgirasaḥ. This name divides the contents of this fourth Veda into two parts and relates
them to these two clans: the Atharvans are associated with the peaceful (śānta) and auspicious
(pauṣṭika) rituals of medication (bheṣajāni), while the Āṅgirasas are connected with the terrible
(ghora) rituals of sorcery (ābhicārika).
Comparing the Śaunaka and Paippalāda versions of the Atharvaveda, Stanley Insler has
studied the methods by which their composers have pieced their hymns together. “This is
made possible because both known recensions of the Atharvan Saṃhitā often present versions
of the same hymn that differ not only in the choice of forms, words, phrases or whole lines
(pāda), but that also differ in the sequence and number of stanzas constituting a recognized
sūkta [‘hymn’] in the respective collections.” Insler concludes that the Kāṇva and Āṅgirasa
poets of the Rigveda belong to the same tradition as the poets of the Atharvaveda:

This proposal can be supported linguistically, which I shall undertake on another

occasion. If the Kaṇva and Āṅgirasa bards were in fact the poets of the Atharvan
compositions, an interesting question arises: Do the Dharma texts state that one should
stop studying the Veda if the singing of sāmans is overheard because the Kaṇvas and
Āṅgirases, the sāman-singers par excellence, were also yātudhānas [i.e., sorcerers]?
How else is their bad reputation to be explained, a matter so clearly unpleasant that
the singing of sāmans is mentioned along with the noise of dogs, jackals, asses and
owls? Support for this idea stems from the fact that the black spells among Atharvan
hyms (ghora, yātu, ābhicārika) are called āṅgirasa. (Insler 1998:17)

Káṇva- (whence the patronym Kāṇvá-) is among those words of the Rigveda that exhibit a
sound change characteristic of the later development of the Old Indo-Aryan language
(Sanskrit) into Middle Indo-Aryan (Prakrit), in this case ṛ >a. (Such Prakritisms one would
expect to find in the “Atharvavedic” language, which had supposedly been subject to Dravidian
substratum influence for centuries in the Indus Valley [chapter 14]. In this connection one can
also point to the unusually high number of non-IE retroflex phonemes in Book 8 of the
Rigveda, belonging to the Kāṇvas and Āṅgirasas.) Karl Hoffmann (1940) convincingly derived
káṇva- from *kṛṇva-, “acting (magically).” The name appears to mean “bewitcher, sorcerer.”
The root kṛ-, “to do,” is itself used in the meaning “to bewitch” in the Atharvaveda (AVŚ 5,14,9
yáś cakā´ra tám íj jahi) and has such derivatives as kártra- and kṛtyā-, “sorcery.” Moreover, in
the hymn AVŚ 2,25, káṇva- denotes demoniac beings who eat embryos (verse 3) and obstruct
life (verses 4 and 5), while AVP 13,10,4 speaks of “a káṇva- associated with a sorcerer
(yātumāvān).”
As we shall now see, key sorcery terms are more or less totally absent from the “family”
books of the Rigveda and occur mainly in Books 1 and 10, whereas they occur frequently in
the Atharvaveda (here AVŚ). The word abhicārá-, “sorcery, witchcraft,” is not found anywhere
in the Rigveda, but it is implied in the only Rigvedic verse (10,34,14b) where the verb abhi +
car- is attested: mā´ no ghoréṇa (scil. abhicāréṇa) caratābhí dhṛṣṇú, “do ye not mightily
bewitch us with terrible (sorcery)!” In the AVŚ, on the other hand, abhicārá- occurs four times;
abhicārín-, “one commanding sorcery,” once; pratyabhicáraṇa-, “counter-incantation” once;
while the verb abhi + car-, “to bewitch,” is used three times. Another key sorcery term, kṛtyā-,
“witchcraft,” occurs only twice in the Rigveda, in 10,85,28–29, that is, in the marriage hymn of
Book 10. In the AVŚ, on the other hand, it is attested as many as fifty-eight times, plus twenty-
six times in the compound kṛtyā-kṛt-, “practicing witchcraft.”
The sorcery rites are usually directed against a rival, and favor the words bhrā´tṛvya- and
sapátna-. The first of these originally meant “brother’s son” or “cousin,” a relative who for a
king or prince is a dangerous rival as a potential seizer of royal power. In the Rigveda bhrā
´tṛvya- is attested just once (with a privative prefix) in RV 8,21,13, used by a Kāṇva poet, while
in the Atharvaveda, it occurs fifteen times, sometimes in prose passages, including three
instances of compound words that denote “destroying,” “killing,” or “driving away” the rival.
The second term, sapátna-, “rival,” is derived from the corresponding feminine gender term
sa-pátnī-, “co-wife, wife having the same husband,” which is attested twice in the old “family”
books of the Rigveda, but without originally implying rivalry, since both hymns (RV 3,1,10 and
3,6,4) speak of the two world halves (ródasī), heaven and earth, as co-wives of the fire god
Agni. Female rivals are involved, however, in two hymns belonging to the late books of the
Rigveda: RV 1,105,8 (the hymn is ascribed to Kutsa Āṅgirasa) and 10,145,1–5. The masculine
sapátna-, “rival,” is attested four times in the Rigveda, all in Book 10, along with the
compound sapatna-hán-, “killing the rival.” In the Atharvaveda, on the other hand, sapátna-
occurs sixty-seven times, along with various compounds meaning “killing the rival” or the like:
sapatna-hán- (twenty-two times), sapatna-kṣáyaṇa- (five times), and sapatna-kárśana-, sapatna-
cā´tana-, and sapatna-dámbhana- (once each).
I return to Insler’s question about the Sāmavedins as sorcerers. The Sāmavedic texts are
the first to prescribe some soma sacrifices specifically for witchcraft purposes. But the
Sāmavedic singers can also fulfil wishes by other, less drastic means. The following list of
options is not exhaustive. One may choose between several different stomas, that is, how
many verses are to be chanted in a laud; between several different viṣṭutis, that is, ways of
counting the verses repeated to fulfil the stoma; between several alternative verses on which
the sāmans are chanted; between several alternative sāmans and within individual sāmans,
between alternative syllable numbers in the parts of the sāmans; and between several
different finales (nidhana). We may exemplify the use of the song’s finale by quoting
Pañcaviṁśa-Brāhmaṇa 7,1,11, which gives the following options for replacement of the regular
finale ā of the gāyatra sāman: “As the finale he should take (the word) iḍā (literally
‘refreshment’, but often equated with cattle) for one who desires cattle; svaḥ (‘heaven’) for one
who wishes (to attain) heaven; yaśaḥ (‘fame’) for one who desires spiritual luster; āyuḥ (‘long
life’) for one who suffers from disease; haṁsi (‘you kill’) for one who practices sorcery.” Similar
options exist for all the means enumerated above; they usually include witchcraft.
Before continuing with the Sāmaveda, the mention of the options to fulfil specific wishes
leads me to comment briefly on the beginnings of the Yajurveda. The principal Yajurvedic
priests are the Adhvaryu and the Pratiprasthātar, human representatives of the Aśvins, whom
they worship as the main priests of the pravargya ritual (see chapter 11). The earliest
Yajurvedic texts, the Maitrāyaṇī and Kāṭhaka Saṁhitās, introduce to the Vedic śrauta ritual
many kāmya-iṣṭi rites, devised to fulfil all kinds of wishes, including the desire to harm rivals
and enemies (Caland 1908). The principal ritual manual of the Atharvavedins, the Kauśika-
Sūtra, consists of two parts; while the latter part describes domestic (gṛhya) rites, the first
fifty-two chapters are devoted to kāmya-iṣṭis of the same kind (Caland 1900). This suggests an
ultimately “Atharvavedic” background for the iṣṭi rites, where the sacrificial oblations usually
consist of cakes baked of rice or barley (grains sacred to Varuṇa). These sacrifices differ from
the original Rigvedic soma sacrifice (yajña), which probably resembled the yasna of the
Zoroastrians. In chapter 11, I compared the milk offerings to the Aśvins of the pravargya to
the daily agnihotra offerings of milk to sun and fire (equated with Mitra and Varuṇa,
respectively), which, with the slightly more complicated iṣṭis of the new- and full-moon days,
constitute the very core of the “classical” Vedic śrauta ritual; the iṣṭis function also as models
for the simpler domestic sacrifices (which first emerge in the late hymns of the Rigveda and
the Atharvaveda). The origins of the Yajurveda thus also seem to lie in the “Atharvavedic”
worship of the Aśvins and their doubles Mitra-Varu ṇa.
The earliest sources on soma sacrifices connected with sorcery belong to the Sāmaveda,
which is the source for most texts dealing with these rites in the other Vedas. The original
Tāṇḍya-Brāhmaṇa seems to have been purged of most “barbaric” material, including the rites
of sorcery, which are missing in the twenty-five chapters of the Pañcaviṁśa-Brāhmaṇa (in
contrast to the Jaiminīya-Brāhmaṇa) but appear in its supplementary twenty-sixth chapter, the
Ṣaḍviṁśa-Brāhmaṇa (Caland 1931:xx–xxi). ṢB 3,8–11 contains the sorcery rites called śyena
(“falcon”), iṣu (“arrow”), saṃdaṁśa (“tongs”), and vajra (“mace, thunderbolt”). The śyena rite
is described as follows in the Sāmavedic Lāṭyāyana-Śrautasūtra (8,5):

As officiating priests of the “falcon” rite one should choose learned sons of warriors
(yaudha-) who form (raiding) gangs (vrātīna-); (learned sons) of worthy persons (arhat-
), according to Śāṇḍilya. One should churn butter for sacrificial ghee from the milk of
cows which are among sick cows but are not yet sick. After cutting down all the many
plant stems and trees that have grown in a given place they should make the sacrificial
area there; or on a burning hot ground. The words śavanabhye adhiṣavaṇe (in the
Ṣaḍviṁśa-Brāhmaṇa) mean: they should make the two planks on which the soma is
pressed from the nave(-plank)s of a cart with which the corpse (śava), i.e., dead person,
is carried away (to the place of funeral). The sacrificial stake should not have the
(usual) caṣāla-ring (at top) but instead a wooden sword as its tip. The officiating priests
should perform their duties wearing red turbans, dressed in red clothes, having their
sacred thread hung from the neck (as is done when one defecates), carrying quivers
and bows with strings strung. … [instructions for sāmans as deadly weapons omitted]
… (The prescription about the sacrificial fees given in the Ṣaḍviṁśa-Brāhmaṇa means
that) in groups of nine (animals), one should give as many as possible cows that are
blind in one eye, lame, without horns, and without tails. Moreover, these (defective
animals) they should make bleed at the time the animals are given as priestly fees.

The term vrātīna- connects these performers of the “falcon” and other sorcery rites with
people termed vrātya- in Vedic texts. The Vrātyas are known especially from rites called
vrātyastoma, the oldest sources on which belong to the Sāmaveda: Jaiminīya-Brāhmaṇa 2,221–
227, and Pañcaviṁśa-Brāhmaṇa 17,1–4; the Lāṭyāyana-Śrautasūtra (8,6) gives divergent
explanations of ritual authorities for the obsolete terms used in PB 17,1. The vrātyastomasand
related rites are centrally important for understanding the prehistory of the Vedic ritual, for
they have many features and components not normally found in the śrauta rites. Actually it is
nowadays widely agreed that they are a kind of fossil remains of an earlier, more archaic stage
of ritual development. Discussing them without giving the reader a chance to form an idea of
them from the source texts seems futile. I therefore start by quoting the Pañcaviṁśa-
Brāhmaṇa’s exposition of the first, basic vrātyastoma, omitting much irrelevant data, but
adding some clarifications in brackets:

The Gods, forsooth, went to the world of heaven; of them the adherents of “the God”
[i.e., Rudra] were left behind (on earth), leading a Vrātya-life. . . . Those who lead the
life of a joined group [vrāta-] are destitute, left behind. For they neither practise the
study of the Veda nor do they plough or trade. … The joined group is unequal [among
its members], as it were. … Swallowing poison are those, who eat foreign food as
Brahman’s food, who call good words bad, who use a stick to strike the guiltless [JB
2,225 is more explicit by stating that the Vrātyas violate a Brahmin who should not be
violated], who, though being not initiated, speak the speech of the initiated [according
to JB 2,222 the Vrātyas speak with their voice what is impure and unsuited for
sacrifice; in front of elder people they speak obscene things, āhanasyaṃ vadanti]. The
guilt of these may be removed by the sixteen-versed stoma …
A turban, a goad, a bow without arrow, a board-covered rough vehicle [vipathaḥ
phalakāstīrṇaḥ; according to LŚS 8,6,9–11, vipatha- is an uncovered chariot of the
people living in the east (i.e., Magadha), prācyaratho nāstīrṇaḥ; it is yoked with an agile
pair of a horse and a mule, according to Śāṇḍilya, or with any pair of horses or mules,
according to Dhānaṃjayya], a garment with black fringes, two goat-skins: one white,
one black, a silver ornament (worn around the neck), (all) that is (the equipment) of the
Gṛhapati [i.e., the leader]. The other (Vrātyas) have (upper garments) with red borders
and corded fringes, with strings at each side; each of them has a pair of shoes and
doubly-joined goat’s hides. This is the possession of the Vrātyas; on him, to whom they
bestow (this possession), they transfer (their) guilt (or unworthiness, so that henceforth
they are qualified to take part in the sacrifice of the Āryas). [LŚS 8,6,28–29 specifies
that those who give up roaming around as Vrātyas should give their equipment to
Vrātyas (who practice this way of life), or to a religious student from the country of
Magadha, brahmabandhu- māgadhadeśīya-; after having performed the vrātyastoma
rites they should live following the three Vedas.] Each of them brings to their Gṛhapati
thirty-three (cows). For thirty-three adherents of “the God” had come (through this
vrātyastoma) to prospering. (So this rite serves) for prospering. (PB 17,1, translated
Caland 1931:454–458)

Besides Sāmavedic texts, the earliest source on the Vrātyas is the Atharvaveda, which devotes
a whole book with eighteen chapters in Brāhmaṇa prose to this topic (AVŚ 15, AVP 18,27–43);
the description of the vrātyastoma in the old Yajurvedic Baudhāyana-Śrautasūtra (BŚS) 18,24–
26 clearly follows the Atharvavedic tradition and provides important specifications.
AVŚ 15,1: The Vrātya-Book of the Atharvaveda begins with a cosmogony. In the beginning
this universe was the Vrātya, going around making requests (īyamāna-); he stirred up the
creator god Prajāpati who started to create. This resulted in the birth of the Great God
(mahādeva), the Lord (īśāna), who became the “solitary” Eka-Vrātya (this expression may
refer to the sun as the protypal solitary wanderer). He took to himself a bow, which became
the rainbow: its belly is dark blue (nīla), its back red (lohita); with the dark blue he envelops a
hostile cousin or rival (bhrātṛvya-), with the red he pierces one hating him. The quarters of
space are connected with the Vrātya’s archers (iṣvāsa-), called Bhava, Śarva, Paśupati, Ugro
devaḥ, Rudra, Mahādeva, and Īśāna (these same plus Vajra as the eighth are the eight names
of Rudra in Kauṣītaki-Brāhmaṇa 6,1–9).
AVŚ 15,2: The Vrātya is described as moving through all the directions of space, starting
from the east. He (BŚS: the agreed leader, sthapati, of the Vrātyas, who performs the vows,
vrata-, on behalf on the whole group: he sleeps on the ground, does not eat meat, and does not
have sex) is dressed in a garment (vāsas, BŚS: it is black with black fringes, this being the
garment of one who has been initiated), wears a turban (uṣṇīṣa, BŚS: it is black, as he is
initiated), two round ornaments (pravarta-, BŚS: radiant disks, rukma-, of gold and silver: the
BŚS expressly connects them with the golden and silvery disk placed on both sides of the
heated gharma vessel which symbolizes the head of the Great Hero in the pravargya rite of
the Aśvins; cf. also the golden disk, rukma-, hung from his neck by the sacrificer when he is
about to carry the fire-pot ukhā in the agnicayana), a jewel-necklace (maṇi). (BŚS: A bow with
three arrows in a quiver made of leather serves as his stick of the initiated; he speaks the
speech of the Vrātyas, which serves as the speech of the initiated. He wears hoofed sandals,
khuryāv upānahau, because as an initiated one, he should not tread on anything impure.)
He drives around in a “rough vehicle” (vipatha-, “able to move in roadless terrain”) pulled
by two draught-animals (vipatha-vāha-) and driven by a charioteer (sārathi-) with a whip
(pratoda-). (BŚS 18,24 here speaks of the leader as having a war chariot, ratha-, and
correlates the chariot’s parts with the components of the sacrificial area of the Vedic śrauta
ritual: the whip’s handle and lash, the chariot front, the two shafts of the forked pole, the two
yokes, the two bow-stands, the two sides of the chariot, the railing, the stands of the servants,
the seat of the charioteer, the chariot-lap, i.e., its floor, the standing-place of the chariot. BŚS
18,25 however mentions an inauspicious old chariot yoked with old draught-animals.) Running
on each side of the Vrātya’s carriage are two footmen (pariṣkanda), while in front of it run two
other men (puraḥ-sara). A prostitute (puṁścalī) and a bard (māgadha) accompany the Vrātya.
(BŚS 18,25 enigmatically states that among the Vrātyas is “a celibate man who is not a
māgadha but is called a māgadha,” brahmabandhur amāgadho māgadhavākyaḥ, and “a
celibate nonprostitute who is called a prostitute,” brahmabandhur apuṁścalū puṁścalūvākyā.)
AVŚ 15,10: When the Vrātya comes to royal houses as a guest (atithi-), the king should
esteem him higher than himself; “thus he will not offend against the royal rule (rāṣṭra-).” From
this arose spiritual power (brahman-) and mundane power (kṣatra-), which entered Bṛhaspati
(the divine purohita) and Indra (the divine king) respectively.
AVŚ 15,3: After standing for a year erect, the Vrātya asked for a throne-seat (āsandī-) to be
brought. The four seasons (spring, summer, rains, autumn) are represented in its four feet, the
sāman songs rathantara, Bṛhad, vāmadevya, and yajñāyajñīya in its two lengthwise and two
crosswise frames, the Rigvedic verses in its forward cords and the Yajurvedic formulae in its
cross-cords, the Vedas in its cushion, and the Brahman in its pillow, the sāman in its seat and
the udgītha in its rest. On this seat the Vrātya ascended. (The principal chanter priest Udgātar
ascends a throne seat with almost identical ritual at the mahāvrata, see LŚS 3,12 below.)
AVŚ 15,2: The sāmans move with the Vrātya, starting with rathantara and bṛhad in the
east, vāmadevya and yajñāyajñīya in the south, vairāja and vairūpa in the west, and naudhasa
and śyaita in the north.
The Rigveda mentions by name very few sāmans, all in connection with the
gharma/pravargya rite of the Aśvins, where they are sung as solo songs (chapter 11).
In the Atharvaveda, sāmans mentioned by name occur mostly in this vrātya-book (other
references, almost all of them just to rathantara and bṛhat—probably connected with the
gharma rite, as in AVŚ 9,10,3, which repeats RV 1,64,25—occur in 8,9,4, 8,10,13–17, 11,3,16,
and 13,3,11–12). All those mentioned in the vrātya-book are important sāmans, mostly used at
the pṛṣṭha lauds sung at the midday service of a soma sacrifice in the Sāmavedic liturgy. This
particular combination of sāmans in AVŚ 15,2, however, suggests that the ritual context
probably is the mahāvrata laud chanted at the midday service of the mahāvrata day, at the end
of a ritual year, when solo sāmans are sung around the completed fire altar as a part of the
mahāvrata laud (PB 5,1–4; LŚS 1,5,11–22, 3,9–4,4). Other reasons, too, require a closer
examination of the mahāvrata day, as we shall now conduct using the Lāṭyāyana-Śrautasūtra.
LŚS 3,9: The Sāmavedic Lāṭyāyana-Śrautasūtra begins its description of the mahāvrata day
with the solo sāmans around the fire altar: when, by whom and where exactly they should be
chanted; the finale parts are sung in chorus.
LŚS 3,10: Next, the king and the warriors who follow him in a minimum of two chariots are
made to put on armor and hand protection, then to string the bow, equip themselves with
three arrows, and board the chariot. Two pieces of antelope skin are tied up for targets on the
north side of the fire-kindler’s hearth, one to the east, the other to the west. The king is
instructed to drive sunwise around the sacrificial ground three times, on each round piercing
the eastern target skin, so that each arrow pierces the skin higher than the previous arrow,
and without the arrow falling to the ground. The other charioteers should in the same way
pierce the western skin with three arrows. After shooting the third arrow the king should
drive northwards and shoot a fourth arrow in any direction he likes, stopping his drive after he
has seen cows. There his armor will be taken off.
LŚS 3,11: A hole has been dug in the ground west of the fire-kindler’s hearth, so that it is
half inside and half outside the sacrificial area. A bull’s hide with its hair side upwards is fixed
over the hole so that it is fully covered. They hit (ā-han-) this earth-drum (bhūmi-dundubhi)
with the tail (of the bull hide), uttering the mantra: “You are (Goddess) Vāc! Announce, O
drum, success … for us! Contradict (i.e., ward off by sound) him who hates (us), contradict
(him by emitting) a terrible sound (ghorāṃ vācam)! Then announce noble fame for us, O drum,
(by emitting) a sound that is kind and friendly (for us)! …”—The verb ā-han-, here used for the
“beating” of the drum, has two connotations: the “slaying” of the bull and the bull’s “pushing
his penis” (symbolized by the bull’s tail with which the drum is beaten) into the vagina of the
Goddess. From this verb is derived āhanasya-, “obscenity,” which JB 2,222 says the Vrātyas
impudently speak in front of elders, and which denotes the obscene dialogue taking place
during the sexual union of the horse sacrifice.
LŚS 3,12: Then a throne seat (āsandī-) made of the wood of the udumbara fig tree and
cords of woven muñja grass is brought to the north side of the pillar made of udumbara wood
in the sitting hall (sadas). The Udgātar touches the four feet of the throne, which he calls the
rathantara and bṛhat, naudhasa and śyaita (sāmans), and the four horizontal beams, which he
calls vairāja and vairūpa, and raivata and śākvara (sāmans). Then he touches the crosswise
and lengthwise woven strings, which he verbally equates with Rigvedic verses and Yajurvedic
formulae, he touches the seat, which he equates with Sāmavedic sāmans, and so on. The
Udgātar then ascends the seat, according to the mantra he thereby utters, for becoming
Samrāj, “the Universal King” (on this term, see chapter 10). (The Vrātya ascends a throne with
an almost identical ritual according to AVŚ 15,3, see above.)
LŚS 4,1: A harp (vāṇa-) has been made of udumbara wood, with its soundbox covered with
the hide of a red bull, hair side out. Ten holes are made in the rear part of the harp, in each of
which ten strings of muñja grass are fixed (alternatively, there are three holes, with thirty-
three, thirty-four, and thirty-three strings each). The Udgātar priest now plays this hundred-
stringed harp by stroking it with a reed plectrum, pronouncing a mantra, linking its sound
with goddess Vāc, who overpowers the enemies and grants life for a hundred years.
LŚS 4,2: Now different stringed instruments (alābuvīṇā, vakrā, kapiśīrṣṇī, mahāvīṇā, and
apiśīlavīṇā) are played by professional musicians at the eastern and western doors of the
sitting hall. Choristers accompany them with mantras. To the west of them, the wives of the
participants play the kāṇḍavīṇā (with a plectrum) and the picchorā (upon the mouth)
alternately in pairs.
LŚS 4,3: A Brahmanical praiser (abhigara-) sits at the eastern door of the hall facing
westwards and praises the accomplishment of the participants in the sacrificial session. A
commoner reviler (apagara-) sits at the western door facing eastwards and blames the
participants for their failure.
A man belonging to an Aryan social class stands south of the mārjālīya hearth (the
fireplace for washing utensils) inside the sacrificial area facing southwards. Opposite him,
outside the sacrificial area, a man of the Śūdra class faces northwards. These two men fight
for a round white skin by pulling it alternately toward themselves, starting with the Śūdra;
eventually the Aryan man wins, the Śūdra drops the skin and runs away, while the Aryan man
follows him, beating him with the skin.
East of the fire-kindler’s hearth a celibate student (brahmacārin-) stands inside the
sacrificial area facing north, while a prostitute (puṁścalī-) stands outside the sacrificial area,
facing southwards. They scold each other, she blaming him for bad conduct and emitting
semen, while he blames her for being a wretched prostitute, the cleanser of the host, and the
washer of every man’s penis. West of the fire-kindler’s hearth outside the sacrificial area in an
enclosed space couples (of men and women) should copulate, without caring about their social
class. At least five, maximally 500, but ideally twenty-five slave girls (dāsī-) of the Gṛhapati
should go sunwise around the mārjālīya hearth, carrying new vessels full of water, shouting
“Haimahā, this is honey!” in an increasingly higher voice. Drums tied in all corners (of the
sitting hall or the sacrificial area) are beaten. All sounds should become increasingly louder
and continue as long as the mahāvrata laud is chanted by the Sāmavedic priests.
The vrātya-book was the first sample of the Atharvaveda to appear in a Western language;
it was translated into German by Theodor Aufrecht in 1849. At the same time, Albrecht Weber
(1849) published a pioneering survey of the Sāmaveda literature, in which he drew attention
to passages in the Pañcaviṁśa-Brāhmana (PB 17,1–4) and the Lāṭyāyana-Śrautasūtra (LŚS 8,6)
that deal with the vrātyastomas. Weber saw them as rituals by means of which Aryan people
living in a non-brahmanical manner could enter the Vedic community. He also connected the
vrātyas with the sorcery rites of the vrātīnas, noting that the terms used for the warrior
(yaudha) and the priestly class (arhat) differ from those normally used in Vedic literature but
have a parallel in the Buddhist title arhat. Weber also noted another pointer to eastern India in
the phrase brahmabandhu māgadhadeśīya, which refers to the country of Magadha. Jarl
Charpentier (1911), on the basis of AVŚ 15, saw the Vrātyas as early worshippers of Rudra-
Śiva. J. W. Hauer (1927) on the basis of the Vrātyas’ association with a prostitute and a
brahmabandhu māgadha recognized the “popular” celebrations on the mahāvrata day as a
Vrātya rite. Hauer also pointed to the similarity between the dialogue of the mahāvrata couple
with the highly erotic kuntāpa verses of AVŚ 20,136 applied during the “sacred marriage” of
the stallion and the chief queen at the culmination of the royal horse sacrifice.
In two papers, “Vrātya and sacrifice” (1962) and “The case of the severed head” (1967),
Jan Heesterman laid the foundations for the hypothesis that he and others (especially Harry
Falk 1986) have developed, that the vrātyastomas were rites for purifying military sodalities
after raiding expeditions, and that they, as well as the mahāvrata and some other rites, in
particular the horse sacrifice (aśvamedha), the royal consecration (rājasūya), the vājapeya,
and the building of the fire altar (agnicayana), are “fossils” preserving earlier, sexually more
explicit “preclassical” forms of Vedic ritual that also included more violence and competition
between two antagonistic parties.
The vrātyastoma texts speak of “beating with a stick people who should not be beaten,” or
“violating a Brahmin who should not be harmed.” What is meant is clarified by Śatapatha-
Brāhmaṇa 13,4,2,17, where the following advice is given to the guardians of the sacrificial
horse when they start their year-long expedition: “Whenever ye meet with any kind of
Brāhmaṇas, ask ye them, ‘O Brāhmaṇas, how much know ye of the Aśvamedha?’ and those
who know naught of the Aśvamedha, know naught of anything, he is not a Brāhmaṇa, and as
such liable to be despoiled.” It is for the sake of such ritualistic and theological disputations
that the vrātīnas should select the most learned among them as their leaders. Ignorance was
used as a pretext for robbing and killing, and such ancient practice seems to be reflected in
the learned disputations of the Upaniṣads, where we sometimes come across threats like the
following: “He went off to that sacrifice which had already been begun. There … he said to the
Prastotar priest: ‘Prastotar, if you shall sing the prastāva (the initial part of the laud) without
knowing the divinity which is connected with the prastāva, your head will fall off’ ” (ChU
1,10,7–9).
The military escort of the sacrificial horse had to prevent the horse from mounting any
mare during its year-long expedition and to guard it from all sorts of dangers, including
enemies of the sacrificing king who might wish to obstruct the sacrifice. This army consisted
of four socially different groups close to the royal household, associated with the numbers four
to one, expressed by the respective number of chariot horses. The most detailed description of
them is given in the Vādhūla-Śrautasūtra (11,10): four lots of 300 chariot teams and four lots
of 300 footmen; each group of 300 is divided into four batches, seventy-five on either side of
the horse and seventy-five behind and in front of it. The four different kinds of chariot teams
are the following: (1) 300 royal princes armed for battle, clad in bronze mail on both sides,
with charioteers armed for battle, driving chariots covered with shields and yoked with four
horses; (2) 300 nonroyal warriors armed for battle, with charioteers not armed for battle,
driving chariots covered with shields and yoked with three horses; (3) 300 heralds and
headmen not armed for battle, driving chariots not covered with shields and yoked with two
horses; (4) 300 meat carvers and charioteers, who drive on off-track carts (yoked with one
horse).
The reading vipṛthu- in Śāṅkhāyana-Śrautasūtra 14,72,3 for the last-mentioned type of
vehicle may throw some further light on the vipathá-, “off-track cart,” which according to the
vrātyastoma texts and the vrātya-Book of the Atharvaveda is the vehicle of the Vrātya leader.
Vipṛthu- might be just a corruption of vipatha-, but it might also represent the original, non-
Prakritized form of the word, meaning “not broad” = “narrow” = one-man chariot.
The chief queen of the sacrificing king, who is to lie with the main victim, the horse (or the
male chief victim of the otherwise identical human sacrifice, puruṣamedha, to be discussed in
more detail shortly, and also in chapters 16 and Chapter 19), and the three other queens, are
escorted by the corresponding number of young maidens from the same four social classes as
the male guardians of the horse. All these women, including the junior queens, go around the
couple of the “sacred marriage,” smiting their thighs and fanning the couple with their
clothing. The sexually loaded mantras uttered on this occasion mention three “mothers”
(ambā, ambikā, ambālikā) that seem to connect the chief victim with Rudra, one of whose
epithets is Tryambaka, “having three mothers.” Actually the victim of the horse/human
sacrifice is at this culmination of the rite addressed as “the host-leader of the hosts” in the
mantra gaṇānāṃ tvā gaṇapatim . . . (ŚB 13,2,8,4; ĀpŚS 20,18,1). He would indeed have led for
one year the armed host accompanying him, thus resembling the Vrātya leader, who is armed
with a bow and directly identified with Rudra in no uncertain terms: in AVŚ 15,1,4–5 he is the
Great God (mahādeva) and the Lord (īśāna); in AVŚ 15,5 he is escorted by the archers Bhava,
Śarva, Paśupati, Ugra Deva, Rudra, Mahādeva, and Īśāna, which are all names of Rudra. From
the mantras of the śatarudriya section of the Vājasaneyi-Saṁhitā (Book 16), which accompany
the 425 oblations to Rudra at the completion of the fire altar (supposed to coincide with the
completion of a ritual year, just like the horse/human sacrifice), it is plain that Rudra was
feared as “leader of hosts,” “lord of robbers,” a hero mailed and armored, whose bow, arrows,
and wrath were much feared. Rudra’s Hindu successor, Skanda, is the divine army leader
(senāpati), who is escorted by a host of “mothers.” In this respect he is like the Hindu god
Gaṇapati, alias Ganeśa, who is also escorted by “mothers” and whose title gaṇapati, “host-
leader,” can be understood also as a military title. Gaṇapati therefore appears to be a
duplication of Skanda and the leader of Śiva’s gaṇas in later mythology.
Gaṇa- is a synonym of vrāta-, “group, troop,” and is also used of a vrātya-gaṇa- in
Kātyāyana-Śrautasūtra 22,4,3. But Śiva’s gaṇas in classical Hinduism consist also of singers
and dancers like the kinnaras and gandharvas. Actually, gaṇānāṃ tvā gaṇapatiṃ in Rigveda
2,23,1—words addressed to the sacrificial horse—were originally addressed to
Bṛhaspati/Brahmaṇaspati, singer of magic songs. In the Atharvavedic Kauśika-Sūtra (135,9)
Bṛhaspati is an Āṅgirasa, and in the Mahābhārata Bṛhaspati is frequently called “the best of
Aṅgirases” (aṅgirasāṃ śreṣṭhaḥ). As the divine purohita and charioteer, Bṛhaspati was a sūta-,
“charioteer” and “bard,” whose task it was to advise the chariot’s warrior and to encourage
him by singing of the brave deeds of his ancestors. During the year-long expedition of the
sacrificial horse, there are “revolving” (pāriplava) oral performances of all sorts of literary
genres for the entertainment of the sacrificing king and the general public. On this occasion,
also, “master musicians accompanied by hosts (gaṇa) of harp players” (vīṇāgaṇakinaḥ) are
present and the Adhvaryu asks them to sing about the sacrificing king and about the righteous
kings of former times. A Brahmin harp-player-singer (vīṇāgāthī) sings three strophes
composed by himself on such topics as “Such a sacrifice he performed, Such jpgts he gave”
(ŚB 13,4,2,8). Likewise, a rājanya harp-player and singer sings three strophes about “Such
war he waged, Such battle he won” (ŚB 13,4,3,5; TB 3,9,14,1–3). These tristichs agree with
the tradition of the Kāṇvas and Āṅgirasas, poets of the Rigvedic Proto-Sāmaveda and the
Atharvaveda, who favored tristich composition. The Brāhmaṇas and Upaniṣads have preserved
a number of gāthās and ślokas, including those of the Śunaḥśepa legend recited at the royal
consecration, which on the one hand are connected with the vrātya tradition, and on the other
feed into the epic tradition (Horsch 1966).
The Sāmavedins of the “classical” Vedic ritual do not play harp, but at the culmination of
the mahāvrata the Udgātar ascends a throne (āsandī) and plays a hundred-stringed harp with
a plectrum. And just before the horse is killed in the aśvamedha, the horse is chosen to be the
Udgātar priest and made to “sing” the udgītha of the out-of-doors laud by neighing, which is
deliberately provoked by the proximity of a mare (cf. ĀpŚS 20,13,5–8; LŚS 9,9,18–23); in other
words, the horse personifies the principal chanter priest Udgātar. According to the Vādhūla-
Śrautasūtra (11,12,2–3), “after having decorated him, they bring to the place the son of the
noblest bard (sūtaśreṣṭhasya putraṃ), a virginal youth (kumāram asiktaretasam) to be the
cutter of the horse, lamenting him as if he was to die (rudanto yathā mariṣyantam evam). For
they say that formerly the head of him would fall off severed who was the first to make a cut.”
The next sūtras (4–12) tell how this custom came to be changed and the youth to survive. I
have suggested that this youth (kumāra) personified god Rudra, whose name the Brāhmaṇa
texts derive from the root rud-, “to cry,” and who is also called Kumāra. Rudra was thus
represented by a bard—in the Rigveda (1,43,4) Rudra is called gāthapati-, “lord of the song”—
and his gaṇas were also groups of singers, comparable not only to the kinnaras of classical
Hinduism but also, for example, to the Bāul singers of today’s Bengal. In later Hinduism,
Rudra’s successor, Śiva, is said to be fond of the Sāmaveda; in south Indian iconography one
of Śiva’s manifestations is “lute-holder,” Vīṇādharamūrti, and the classical Indian musical
instruments include the Rudra-vīṇā.
The fourth vrātyastoma in Pañcaviṁśa-Brāhmaṇa 17,4 is “for those Vrātyas whose penis is
quiet and hangs down (śama-nīca-meḍhrāḥ),” which the Lāṭyāyana-Śrautasūtra (8,6,4) explains
to be “those who by old age are precluded from sexual intercourse.” From this it is clear that
by early Vedic times Rudra’s wandering troops included old people or perhaps even men like
today’s naked ascetics who assemble at the Kumbha Melā and have literally destroyed their
virility by self-castration. Both gaṇa- and saṅgha- are used as synonyms of vrāta- in the case of
the vrātyas. They are also the terms used by the early ascetic communities of Jains and
Buddhists.
The Jaiminīya-Brāhmaṇa (2,69–70) describes a sacrificial contest between the creator god
Prajāpati (predecessor of the Hindu god Brahmā and the divine representative of the
Brahmins in the Brāhmaṇa texts) and Mṛtyu, the god of death. This contest continued for a
long time, many years, as neither could triumph. Eventually Prajāpati “saw” the ritualistic
technique of establishing equations or identifications between the components of the
microcosm of the ritual and the macrocosm of the universe, so that cosmic events could be
controlled by means of ritualistic symbols, especially numerical matches between entities
(manipulating for instance the number of syllables in a poetic meter). Thereby he conquered
Mṛtyu, whose sacrifice gradually vanished, leaving Prajāpati’s sacrifice as the only existing
one. Prajāpati’s army is said to have consisted of the Sāmavedic lauds, Rigvedic śastras, and
Yajurvedic operations of the soma sacrifice, while Mṛtyu’s army consisted of singing to the
accompaniment of the harp, dancing, and frivolous activities. Particular attention is paid to
Mṛtyu’s harp (vīṇā), which clearly played a vital role in Mṛtyu’s ritual, as its various parts—
such as the stick (daṇḍa) of the harp, its body (sūnā) and its seven strings—are enumerated
and compared with elements of the classical sacrifice. There is no doubt that rites such as the
mahāvrata are hereby meant (at the mahāvrata, the Udgātar seated on a throne plays a harp,
slave girls perform a dance around the mārjālīya hearth, slapping their thighs, and sexual
intercourse is performed). These archaic practices are described in the Vedic texts, some of
which however declare them as obsolete (utsanna) and not to be done. Thus Prajāpati’s
victorious sacrifice represents the reformed “classical” Vedic ritual and Mṛtyu’s the “pre-
classical” ritual of the vrātya rites.
From the description of Mṛtyu’s sacrifice it is clear that Mṛtyu stands for Rudra, the chief
god of the Vrātyas as raiding gangs, also known as bhūtapati-, “lord of ghosts or evil spirits of
the dead.” The name Mṛtyu of course means “Death.” The building of the funeral monument
(for a someone who during his lifetime built a fire altar) and the burial of the vessel with the
ashes or bones of the deceased is accompanied with ritual acts that are clearly connected with
the Vrātya type of sacrifice, such as music (with harps, conches, pipes, flutes, and a metal
vessel beaten with an old shoe), songs and dances by male and female relatives, as well as
professional dancing women, and feasting. Sexual intercourse also takes place, at least
symbolically. A brahmabandhu or a Śūdra asks the seniormost widow for sexual intercourse on
behalf of her deceased husband; she denies him twice but then agrees for one night. The
vessel with the ashes or bones is placed under a tripod holding a pot with a hundred holes,
and sour milk with curds is poured into it. This symbolic seed-laying is accompanied by a
“fanning” (dhuvanam) similar to that around the senior wife lying in “sacred marriage” with
the chief victim in the horse sacrifice (ŚB 13,2,8,4). These archaic practices were soon
regarded as obsolete, since they were declared optional in the Pitṛmedhasūtras (Caland 1896).
Only a man who had built a fire altar (agni-citi) was entitled to a funeral monument. The
close connection between these two kinds of monuments is underlined by the fact that it is
Yama, the god of death, who recommends the building of a fire altar as a means to overcome
death in the Kaṭha-Upaniṣad. Significantly, Yama tempts Naciketas not to ask for the ultimate
secret by offering him “these lovely maidens, with chariots and harps” (imā rāmāḥ sarathāḥ
satūryāḥ), which Naciketas rejects as follows: “thine be the vehicles, thine the songs and
dances!” (tavaiva vāhās tava nṛtyagīte) (KU 1,25–26).
In the Kāfir pantheon of northeastern Afghanistan, the high god is Imra, from *Yama-rāja,
“King Yama.” He is also called Māra, “Killer, Death.” Māra had seven daughters, worshipped
as goddesses of fertility who protect agriculture. In the Buddhist legend, Māra, ruler of the
phenomenal world, in vain tries to prevent Buddha Śākyamuni from attaining enlightenment
by sending his terrible armies to threaten him, followed by his three appealing daughters to
seduce him. Mṛtyu corresponds to Maccu in Pāli texts, and some early Buddhist texts speak of
this deity as a king, Maccu-rājā and Mṛtyu-rāj, providing thus a counterpart to the Kāfir name
Imra and to Yama-rājan- in RV 10,16,9, as well as to King Yima in the Avesta. The “evil Death”
(māro pāpimā) of the Buddhist texts corresponds to the synonymous deity called pāpmā
Mṛtyuḥ in the Atharvaveda and Middle Vedic texts, where he is equated with Yama and Śarva
(= Rudra). There are many correspondences between the Vrātya rituals and the Kāfir religion,
in which raiding expeditions, circular dances, and heroic songs sung to the accompaniment of
the “Kāfir harp” (similar to the classical—not modern—Indian vīṇā) play a very central role, so
the concordance between Kāfir Māra and Atharvavedic/Vrātya Mṛtyu is significant.
Yama, the son of Vivasvat, as the king of the dead appears in the Rigveda mainly in Books
1 and 10. Yama as the first mortal or first man duplicates Manu (“man”), the son of Vivasvat,
the first sacrificer and the ancestor of the Rigvedic Indo-Aryans. Avestan Yima, the son of
Vīvaŋhvat, as the first sacrificer of haoma and the first king of the Aryans/Iranians, plays an
important role in Old Iranian religion, while in the Vedic religion Yama seems to duplicate also
Varuṇa, who has an ancient Indo-Aryan background; according to RV 10,14,7–8, the pious
dead arriving in the highest heaven see there the two kings, Yama and Varuṇa. Yama seems to
me to have come to the Vedic pantheon from the religion of the Dāsas who were Iranians (see
chapter 9) and whose religion seems to have survived in the Kāfir religion of Nuristan (see
chapter 20). In Rigveda 10,10,13, Yama’s twin sister Yamī exclaims “bato batāsi,” when Yama
declines her advances. Hereafter the word bata is attested in Indo-Aryan only as an
interjection of reproach or frustration. It is found only in this Rigvedic verse as a noun, in the
nominative and vocative cases, probably denoting “weakling.” Is it a mere coincidence that
Khotanese Saka has an adjective bata meaning “small, weak, bad,” with cognates in
Zoroastrian Persian and other Iranian languages?
Yama’s vehicle in Hinduism is the water buffalo, but there is no record of this in the Vedas.
In the single context of a buffalo sacrifice in the Brāhmaṇa and Sūtra texts, the water buffalo
is sacred to Varuṇa. The water buffalo is an animal native to South Asia, not originally
connected with Aryan deities. On the other hand, it is connected with Harappan “Proto-Śiva,”
who wears the horns of a water buffalo. So the buffalo’s connection with Yama, the god of
death, suggests something about Proto-Śiva, too: Proto-Śiva may have been a god of death,
but most probably he was also a divine king, like Varuṇa, who guarded the cosmic order (ṛta-),
as well as Yama, known as the king of righteousness, dharma-rāja. Yama of the Hindu period
and Vedic Varuṇa also share the attribute of a “noose” (pāśa-), with which to catch and bind
the sinners who are to die. “Indeed, Varuṇa is Death” (mṛtyur vai varuṇaḥ, KS 13,2).
13
The Megalithic Culture and the Great Epics

In the preceding chapters, I have tried to throw new light on the religion of the early Aryan-
speaking peoples chiefly by reference to archaeological evidence. In this chapter, archaeology
continues to make a key contribution. I shall consider three major archaeological complexes of
the first millennium BCE: those of the Painted Grey Ware (PGW), the Northern Black Polished
Ware (NBPW) (Fig. 13.1), and the Black-and-Red Ware (BRW). The last of these is largely
associated with Megalithic graves, which seem to represent a further Aryan immigration to
South Asia. Their archaeological data provide reasons to suggest a wholly new interpretation
of the Mahābhārata war. In addition to archaeology, there are textual sources: Vedic literature
for the period before about 400 BCE, and thereafter Pāṇini and the beginnings of the
Mahābhārata and Rāmāyaṇa, although we know them only from later versions. This chapter
focuses on the central narratives of these great epics, which involve the transition from Vedic
to Hindu religion in the field of Viṣṇuism.

FIGURE 13.1 Distribution of the Painted Grey Ware and the Northern Black Polished Ware. After Thapar 1985:122.
Courtesy ASI.

This is contested scholarly territory, with an unavoidably imprecise chronology until we

reach the inscriptions of Asoka in the third century BCE. To make it easier for the reader to
follow my arguments, I shall sum up at the outset the main theses of this chapter, while noting
that they are not likely to be accepted by all scholars working on these texts. In my view, the
first phase of the PGW reflects the stabilization of the Vedic/Brahmanical culture in the
powerful Kuru kingdom, while its second phase temporally and geographically coincides with
the Mahābhārata war. The Pāṇḍava heroes of the Mahābhārata appear to have been originally
“pale” (pāṇḍu) Iranian horsemen; they entered western India from about 800 BCE and during
the following centuries spread widely to peninsular India and Sri Lanka. The Pāṇḍava takeover
of the Kuru kingdom is the main theme of the Mahābhārata, while the Pāṇḍava takeover of Sri
Lanka appears to be reflected in the Rāmāyaṇa. The latter event is associated with the
introduction of Aryan rulers from north India (the Mathurā region and Gujarat) into Tamil
Nadu. In Hinduism these events are connected with the formation and spread of the Vaiṣṇava
or Bhāgavata religion, in the form of worship of Bala-Rāma and Kṛṣṇa. To my mind, these
divine brothers are continuations partly of the divine Aśvin twins in Vedic worship—this again
is a new idea—and partly of the earlier religion of the Harappan age founded in agriculture
and pastoralism.
The Rigvedic and Atharvavedic traditions had fused by the time the earliest texts of the
Yajurveda were created. These codified the complex Vedic śrauta ritual in more or less its final
form. Around this time—perhaps 1000 BCE—the existing hymns of the Rigveda and the
Atharvaveda were assembled and arranged systematically into text collections that from now
on were memorized and handed down as faithfully as possible. This unification of the Vedic
community took place in the Kuru or Kaurava kingdom. The first hint of its emergence is the
name of King Kuruśravaṇa, mentioned as a descendant of King Trasadasyu in two hymns of
the late tenth book of the Rigveda. Instead of about thirty tribes referred to in the Rigveda,
the Brāhmaṇa texts mostly speak of just the Kurus, or of the Kurus and the Pañcālas, who
clearly were closely allied. The geographical distribution of the Vedic culture is defined by the
place-names occurring in the texts, and seems to offer a close match with the iron-, horse-,
and rice-using culture of the PGW culture (c. 1100–350 BCE) discovered by archaeologists in
the northwestern plains (Fig. 13.1). The early phase of this culture (c. 1100–700 BC), with few
and small towns, appears to represent the Middle Vedic culture and its Kuru kingdom.
Archaic Vedic rituals such as the horse sacrifice and the royal consecration included
“proto-epic” bardic performances. They included both praise songs by priestly singers,
praising the ritual achievements and liberality of their sponsors, as well as praise songs by
warrior bards who lauded the king’s military feats. In addition, there were narratives for the
entertainment of the crowds in the audience. According to the Śatapatha-Brāhmaṇa (13,4,3,2–
14), the pāriplava narratives of the horse sacrifice included also purāṇas and itihāsas, ancient
myths and legends, anticipating those of the epic and Purāṇa texts. The “proto-epic” Vedic
verse preserved in the Brāhmaṇas (AB 8,21,3, and ŚB 13,5,4,2) glorifies the horse sacrifice of
King Janamejaya, a descendant of the Kuru king Parikṣit known from the Atharvaveda. A
mythical snake sacrifice is described in Vedic texts (PB 25,15 and BŚS 17,18), in which royal
snakes take human form and officiate as priests; they include Kings Janamejaya and
Dhṛtarāṣṭra. The second of these names is also the name of the blind Kaurava king against
whom the Pāṇḍava brothers fight in the Mahābhārata. Indeed, the epic states in its first book
that it was first recited at Janamejaya’s snake sacrifice.
The war between the Kauravas and the Pāṇḍavas takes place in Kurukṣetra, the “Kuru
plain,” located in the Ganges-Yamuna Doab. The PGW is found in the lowest strata of all major
sites associated with the main story of the Mahābhārata. The late phase of the PGW culture (c.
700–350 BCE) may represent the period of the Mahābhārata war, which has been difficult to
date; the oldest parts of the epic were in existence in 400–350 BCE, when the grammarian
Pāṇini mentions the Mahābhārata by name (6,2,38), along with some of its places
(Hāstinapura in 6,2,101) and personalities (Vāsudeva and Arjuna 4,3,98; Yudhiṣṭhira 8,3,95;
the Vṛṣṇi and Andhaka tribes 6,2,34). On the other hand, King Pāṇḍu and the five Pāṇḍava
brothers are not mentioned even once in any older Vedic text; the Mahābhārata and its reciter
at Janamejaya’s snake sacrifice, Vaiśampāyana, are mentioned only in the Āśvalāyana- and
Kauṣītaki-Gṛhyasūtras, which belong to the latest phase of Vedic literature. The Pāṇḍavas,
therefore, arrived on the scene only after the completion of most of the Vedic literature. They
could crush the Kurus by making a marriage alliance with the Kurus’ eastern neighbors, the
Pañcāla tribe. To consolidate their rule, the victorious Pāṇḍavas grafted themselves on to the
Kuru genealogy as cousins of their former foes, the defeated Kauravas. In this regard, the
latest version of the Mahābhārata was intended as a form of political propaganda.
This “political” distortion may account for the contradictions in the Mahābhārata, which
some scholars argue originally celebrated the Kauravas rather than the Pāṇḍavas. In his
celebrated History of Indian literature, Moriz Winternitz comments:

While the poem in its present form absolutely takes the part of the Pāṇḍavas, and
describes the Pāṇḍavas as not only brave beyond measure, but also as noble and good,
and on the other hand represents the Kauravas as treacherous and mischievous,—the
poem, in remarkable self-contradiction, relates that all the heroes of the Kauravas fall
through treachery or in unfair fight. It is still more striking that all the treachery
emanates from Kṛṣṇa, that he is always the instigator of all the deceit and defends the
conduct of the Pāṇḍavas. And this is the same Kṛṣṇa who … is praised and glorified as
an incarnation of Viṣṇu, the highest god, and as the ideal and prototype of every virtue.
How can these remarkable contradictions be explained? Upon this there can only be
conjectures. First, there is probably justification for the supposition, although we have
only the authority of the Mahābhārata itself for it, that a change of dynasty did actually
once take place in the Northwest of India as the result of a great war, and that these
quasi-historical events form the foundation of the epic itself. Starting out from this, we
can well imagine that the original heroic songs … were sung among the bards who
were still near … the house of the Kauravas, but that, in the course of time, as the rule
of the victorious Pāṇḍavas was more and more established, these songs were
transmitted to bards who were in the employ of the new ruling race. In the mouths of
these bards those alterations were then undertaken which made the Pāṇḍavas appear in
a favourable light and the Kauravas in an unfavourable one, without its being possible
to eradicate completely the original tendency of the songs. In our Mahābhārata the
nucleus of the epic, the descriptions of the great battle, is placed in the mouth of
Sañjaya, the charioteer of Dhṛtarāṣṭra, that is, in the mouth of the bard of the Kauravas.
It is precisely in these battle-scenes that the Kauravas appear in the most favourable
light. (Winternitz 1927:I,454–456)

Where did the Pāṇḍavas come from? Apart from their total absence from Vedic texts, other
indications point to their having a foreign, and specifically an Iranian, origin.
For example, the polyandric marriage of the five Pāṇḍava brothers to Draupadī shocked
their audience at court: “Drupada said: ‘It is laid down that one man may have many queens,
… but never that one woman may have many men! … you may not perpetrate such a breach of
the Law that runs counter to the Veda and the world … !’ … Vyāsa said: ‘On this Law, which is
mocked and runs counter to the Veda and the world, I wish to hear the view of each of you.
…’ ” (Mahābhārata 1,187–188, translated van Buitenen 1973:I,367–368, modified). This
Pāṇḍava custom may be compared to that of the Iranian tribe of the Massagetae who were
formidable warriors, according to Herodotus: “The following are some of their customs: Each
man has but one wife, yet all the wives are held in common. . . . When a man desires a woman
he hangs his quiver in front of her wagon and has intercourse with her unhindered”
(Herodotus 1,216, translated Rawlinson [1860] 1942:114). The name of the Massagetae is
thought to contain Proto-Iranian *masya- from Proto-Aryan *matsya-, “fish,” and the tribe has
been compared to that of the Matsyas, among whom the Pāṇḍavas were hiding in
Virāṭanagara.
A foreign, northerly origin is suggested by the Pāṇḍavas’ pale skin. The Mahābhārata
(1,100,17–18) notes that their father, King Pāṇḍu, was destined to be born pale because his
mother was pale with fright at the sight of his horrid-looking father during the conception of
Pāṇḍu. In Sanskrit, pāṇḍu-, pāṇḍura-, pāṇḍara-, “white, whitish, yellowish, pale,” attested since
about 800 BCE, are loanwords from a Proto-Dravidian root *pal- / *paṇḍ-. This implies that the
appellation of the foreign immigrants as “pale” may have originated among the local
population in an area with a Dravidian linguistic substratum. The Pāṇḍavas’ hiding in
Virāṭanagara (Bairāṭ near Jaipur), their alliance with Kṛṣṇa Vāsudeva of Mathurā, and the
location of their first kingdom in the wooded southern half of Kurukṣetra, all suggest that
these foreigners of probably West Iranian descent entered the subcontinent from the west, via
Sindh, Gujarat, and Rajasthan. The Mahābhārata (2,23–29) and early northern Buddhist texts
speak of the Pāṇḍavas as marauders over wide areas.
Is there any counterpart in the archaeological record for the Pāṇḍavas as foreigners of
Iranian affinity coming to India c. 800–400 BCE? In my view, a good match is the “Megalithic”
culture, first attested around 800 BCE at sites such as Mahurjhari and Khapa in Vidarbha in
northeastern Maharashtra. These graves are simple stone circles, in which people were buried
with weapons and horses; the horse equipment resembles that found in Central Asia, the
Caucasus, and western Iran. The circular huts with wooden posts and a fireplace are similar to
the yurts used by the nomads of the Central and Inner Asian steppes. The Megalithic
immigrants could have belonged to the West Iranian speakers, who were assumed (see
chapter 7) to have initiated the Yaz II culture (c.1000–500 BCE) of southern Central Asia and to
have continued also westward along the southern shore of the Caspian Sea to emerge in
history as the Medians and Persians. About 518 BCE, the Achaemenid king Darius I annexed
the Indus Valley to the Persian empire. The Persians might be mentioned in Baudhāyana-
Śrautasūtra 18,44, which lists the Gāndhāris, Parśus (emended from Sparśus), and Araṭṭas as
western peoples, while according to 18,13 the countries of Āraṭṭa (see chapter 17) and
Gāndhāra were non-Brahmanical.
After their arrival in western India, the carriers of the Megalithic culture adopted the
Black-and-Red Ware pottery (of local Chalcolithic origin) and during the following several
centuries spread over wide areas, mainly southwards to the Deccan, south India, and Sri
Lanka. In many regions, folklore associates the megaliths with the Pāṇḍavas; this is very
understandable, especially in Tamil Nadu, where the megaliths are called paṇḍu-kal, “old
stones” (Leshnik 1974). (Tamil pala / paṇḍu, “old,” goes back to the Proto-Dravidian root *pal- /
*paṇḍ-, “to be old.” With the sense “to become old” it can be linked, I suggest, with the
homonymous root meaning, “[fruit] to become ripe, be[come] yellowish, be pale,” which is
attested in Rigvedic Sanskrit phalam, “fruit,” from Proto-Dravidian *palam, “ripe fruit.”)
Numerous iron tridents suggest a Śaiva religion. Martial traditions of Megalithic origin still
continue in the Deccan, where horsemen accompanied by dogs worship Śaiva deities such as
Bhairava, Khaṇḍobā, and Birobā with tridents in yurt-like shrines (Sontheimer 1989:26–27). In
Tamil Nadu the Megalithic culture continued till the second century CE and is reflected in the
Old Tamil heroic poetry.
While the focus of the Mahābhārata is the Upper Ganges Valley, the Rāmāyaṇa situates
itself in the middle Ganges Valley. Its oldest portions are estimated to date from c. 750–300
BCE, and correlate with the early Northern Black Polished Ware culture (c. 550–400 BCE),
found at sites identified with various places mentioned in the Rāmāyaṇa, including Ayodhyā
(Fig. 13.1). Thus the two epics reflect a gradual eastwards movement of the Brahmanical
culture. The distribution of the NBPW appears to coincide with the formation of the kingdom
of Magadha and the expansion of its power. Magadhas and Aṅgas are mentioned in a
medicinal spell of the Atharvaveda as a far-off and undesirable people of the east, to whom the
disease of fever is banished.
The possible BMAC origin of the Gangetic copper hoards (mentioned in chapter 8)
suggests that Indo-Aryan speakers moved to the middle Ganges Valley so early that they did
not become part of the Vedic culture. Such early Indo-Aryan speakers of the “Atharvavedic”
wave of immigration seem to be meant when Vedic texts refer to “easterners” as asura
worshippers and speakers of a language resembling eastern Middle Indo-Aryan, the Māgadhī
Prakrit. According to the Śatapatha Brāhmaṇa (13,8,1,5), Vedic people who worship devas,
make their funeral monument four-cornered; but their enemies, who worship asuras, the
“easterners” and others, make their funeral monument round. The Buddhist stūpa must surely
have developed from these round funeral monuments of eastern India, which preserve the
round layout of most of the kurgans (tumuli) of the Eurasian steppes. It is significant that the
Atharvavedins of the Śaunaka school prescribe either a round or a rectangular shape for the
funeral monument, giving preference to the circular form (Kauśika-Sūtra 85,8). In the
Brāhmaṇa period, however, the Vedic culture expanded to the middle Ganges Valley, which led
to an upsurge of ideas and practices in the Upaniṣads that were novel from the Vedic point of
view but that had been developing for a longer time in eastern India, as is evident from their
heterodox religions.
Christian Lassen (1847:I,535) early on proposed that Vālmīki’s Rāmāyaṇa “contains the
legend of the first attempt of the Āryans to extend their power southwards by warring
expeditions.” Albrecht Weber (1870 [1871]) was inclined to accept this view, although he
believed that the epic was composed in north India and that its author did not have any exact
knowledge of the southern parts of the subcontinent. Yet it is clear that “the poet knew of an
island kingdom, whether real or mythical, said to lie some distance off the coast of the Indian
mainland” (Goldman 1984:I,28). Indeed, as early as the second or third century CE, an Old
Tamil poem (Akanānūru 70) refers to Kōṭi (= Dhanuṣkōṭi, the tip of the mainland, opposite to
Adam’s Bridge in Sri Lanka) as the place from which the victorious Rāma crossed over to
Laṅkā.
The archaeology of early historical Sri Lanka, so far largely ignored in this connection, has
only recently become clearer (Coningham & Allchin 1995). The oldest, “Mesolithic,” period is
evidenced by locally manufactured stone tools. In the second, “Iron Age,” period (c. 600–450
BCE) the habitation area of Anurādhapura was about 18 hectares with circular huts indicated
by postholes. People had “typical Black-and-Red burnished ware,” iron, and cattle. In the
“Early Historic 1” period (c. 450–350 BCE), the site and the circular huts are larger, and there
are strong similarities with south Indian Megalithic burials. The pottery is still dominated by
Black-and-Red burnished ware. Horse bones are found, and indications of a major expansion
of trade in conch shells, iron ore, amethyst and quartz. In the “Early Historic 2” period (c.
350–275 BCE), the site was more than 66 hectares and surrounded by a defensive wall. Finds
include mother of pearl, cowrie and conch shells, lapis lazuli from Afghanistan, and carnelian
from Gujarat, five Brahmi inscriptions on potsherds, and, toward the end, coins stamped with
a single arched hill or caitya. The “Early Historic 3 and 4” periods (c. 275–225 and 225–150
BCE) have yielded also typically Hellenistic objects. It must be noted that the dates given here,
though said to be radiocarbon-based, have been very seriously questioned, together with the
claim of a pre-Aśokan date for Brahmi script in Sri Lanka.
Widespread evidence covering the entire island suggests that Sri Lanka was inhabited only
by tribes of Mesolithic hunter-gatherers until c. 800–600 BCE, when agriculture and cattle-
raising were introduced by an Iron Age culture characterized by “Megalithic” burials and
Black-and-Red Ware. It is so similar to the Iron Age Megalithic culture of the Indian mainland
that its spread must be ascribed to actual movements of people. But where exactly did these
settlers come from? It is sensible to seek an answer from the legends in the chronicles of Sri
Lanka.
The legend of the colonization of Sri Lanka is related in the Dīpavaṁsa (chapters 9–11) and
with slight variation in the Mahāvaṁsa (chapters 6–10), written in about 400 and 500 CE,
respectively, but based on older records. This legend derives the Simhalas from Gujarat, which
is most reasonable on the basis of linguistic evidence, which generally classifies Sinhalese
with Gujarati and Marathi. Pāli, the Middle Indo-Aryan language of the Theravāda Buddhism
of Sri Lanka, is closest to Aśoka’s inscriptions at Gīrnār in Gujarat, and is generally considered
nowadays to have originated in western India. Gujarat and Maharashtra are also precisely the
areas where the Megalithic culture seems to have spread first.
According to legend, at first 700 Simhalas led by Prince Vijaya came to Sri Lanka from
Sīhapura (Siṁhapura) in Lāḷa (Lāṭa in southern Gujarat). “Prince Vijaya was daring and
uneducated; he committed most wicked and fearful things, plundering the people.”He was
therefore expelled from Sīhapura by his father, King Sīhabāhu. Vijaya and his men sailed down
the west coast of the subcontinent, stopping at the cities of Bhārukaccha (Broach in Gujarat)
and Suppāra (Śūrpāraka = Sopāra near Mumbai). In both places they were offered hospitality
and honors, but they exasperated the inhabitants with their “cruel, savage, terrible, and most
dreadful deeds,” which included “drinking, theft, adultery, falsehood, and slander.” Finally
they arrived at the island of Laṅkā—by coincidence at the time when the Buddha reached the
complete extinction (parinirvāṇa). In nine months Vijaya and his men exterminated the island’s
indigenous population of yakkhas (Sanskrit yakṣas), who may denote the Mesolithic hunter-
gatherer population of Sri Lanka, ancestors of the Veḍḍa aboriginals. Vijaya founded
Tambapaṇṇi, the first town in the island of Laṅkā. After ruling for thirty-eight years, Vijaya sent
a message to his brother Sumitta in Sīhapura, asking that a relative succeed him in Laṅkā
after his death.
Vijaya is usually dated to 1–38 after the Buddha’s parinirvāṇa, that is (according to the
traditional chronology) c. 486–448 BCE; his successor, Paṇḍu-Vāsudeva, to 38–39 (448–447
BCE), and so on. However, Laṅkā is said to have been kingless for one year, and Paṇḍu-
Vāsudeva to have come from Siṁhapura on a separate mission. The Vijaya story may be just an
attempt to fill the earlier history with a vague memory of the first immigration: it seems to me
that regular dynastic record was started only with the arrival of Paṇḍu-Vāsudeva, whereafter it
was continuous. Indeed, Lassen (1852:II,96) suggested that “Vijaya” did not refer to any
specific person but rather to an event: the “conquest” of Sri Lanka. The “cruel and savage”
Vijaya, like the demon king Rāvaṇa of the Rāmāyaṇa, may simply symbolize the early rulers of
the island.
The Purāṇas associate Rāvaṇa and his brother Kubera (the god of riches) with the
Himalayas. When people migrate, they often transfer the name of their old domicile to their
new habitat. Siṁhapura, Vijaya’s home town in Gujarat, has a namesake, Siṁhapura, in the
Indus Valley, conquered by the Pāṇḍavas (Mahābhārata 2,24,19); according to the Chinese
pilgrim Xuan-Zang/Hsüan Tsang, this Siṁhapura was some 200 km southeast of Taxila. In the
next verse (2,24,20), the Mahābhārata mentions the Cola as a people crushed by the Pāṇḍavas.
Moreover, Vijaya’s brother Sumitta, King of Siṁhapura, married a princess of the Madra
country, which is situated in the Upper Indus Valley. Cola is otherwise known only as the name
of another dynasty of ancient Tamil kings; like the Pāṇḍyas, this royal house is also likely to
have come to Tamil Nadu from north India.
The second Siṁhala king was called Paṇḍu-Vāsudeva. The word Pa ṇḍu(ka) figures in the
names of other ancient Sinhalese kings as well, and associates them with the Pāṇḍavas of the
Mahābhārata, whose father Pāṇḍu is called Paṇḍu or Paṇḍu-rājā in Buddhist texts written in
Pāli. Paṇḍu-Vāsudeva’s father-in-law, who ruled in a kingdom on the Ganges River, was
likewise called Paṇḍu. He belonged to the Śākya clan, being a relative of the Buddha. Śākya is
derived from Śaka, one of the principal names of Iranian steppe nomads. Its association with
the name Paṇḍu is an additional hint of the Iranian origin of the Pāṇḍavas.
The beginning of the second phase (c. 450–350 BCE) of the Megalithic culture of Sri Lanka
coincides almost exactly with the traditional dates for Paṇḍu-Vāsudeva’s rule. The phase is said
to resemble much the Megalithic culture of South India. These archaeological parallels are
mirrored in the chronicles. According to the Mahāvaṁsa (chapter 7), a fierce demoness
(yakkhinī) called Kuveṇī or Kuvaṇṇā had fallen in love with Prince Vijaya and helped the
invader to kill the yakkhas, who lived in their cities of Laṅkāpura and Sirīsavatthu. She bore
him children. But when his companions proposed to perform the Vedic sacrifice necessary to
consecrate King Vijaya, he accepted only if they procured him a queen of high rank. So they
sent a delegation with jewels and other presents to southern Madhurā (dakkhiṇa-madhurā). Its
ruler, called Paṇḍu/Paṇḍava, decided to send his daughter Vijayā in marriage to Prince Vijaya
along with 700 daughters of his nobility for Vijaya’s retinue of 700 men. After marrying
Paṇḍava’s daughter, Vijaya rejected Kuveṇī, sending her away from his house but promising to
maintain her as a goddess with a thousand bloody bali offerings.
Southern Madhurā is modern Madurai in Tamil Nadu, the capital of the Pāṇḍya kings. The
dynastic name Pāṇḍya is derived from Pāṇḍu, according to Patañjali’s commentary on Pāṇini’s
grammar (4,1,168). Sri Lankan kings kept contact with the city of southern Madhurā later on,
as well. The Greek ambassador to the Mauryan kingdom, Megasthenes, writing about 300 BCE,
refers to the Pāṇḍya country when speaking of an Indian “Heracles”:

this Heracles … had only one daughter. Her name was Pandaea [Pandaíē], and the
country in which she was born, the government of which Heracles entrusted to her, was
called Pandaea after the girl. . . . Some other Indians tell of Heracles that, after he had
traversed every land and sea, and purged them of all evil monsters, he found in the sea
a new form of womanly ornament … the sea margarita [pearl] as it is called in the
Indian tongue. Heracles was in fact so taken with the beauty of the ornament that he
collected this pearl from every sea and brought it to India to adorn his daughter …
among the Indians too the pearl is worth three times its weight in refined gold. (Arrian,
Indica 8,6–13, translated Brunt 1983:II,329–331)

The Arthaśāstra (2,11), ascribed to Megasthenes’s contemporary, the Maurya minister

Kauṭilya (although it actually dates from a somewhat later time), mentions as sources of pearls
several places on the coasts of southernmost India and northern Sri Lanka, among them
Pāṇḍya-kavāṭa and Tāmraparṇī. Tāmraparṇī is the name of the chief river of the southernmost
(Tirunelveli) district of Tamil Nadu, at the mouth of which was the Pāṇḍya port town of Korkai,
famed in Old Tamil literature for its pearl fishery. Tāmraparṇī is also the name of the first
Sinhalese capital on the north coast of Sri Lanka, called Tambapaṇṇi in Mahāvaṁsa 7,38–42
and Taprobane by Onesicritus, the admiral of Alexander the Great, who learned it as the name
of the whole island in 325 BCE in the Indus Valley. Vijaya’s contacts would have been with
Korkai, before the Pāṇḍya capital was moved to Madhurā inland.
The Pāṇḍya capital is called Southern Madhurā to distinguish it from the northern
Madhurā, that is, Mathurā, the famed domicile of Kṛṣṇa Vāsudeva, after which the Pāṇḍya
Madhurā obviously was named. This is suggested also by the name of the second Siṁhala king
coming from Gujarat, Paṇḍu-Vāsudeva. It seems to me that it was this second wave of Pa ṇḍu
princes coming by sea to Sri Lanka and Tamil Nadu that brought the Vaiṣṇava religion to the
south. This is suggested further by the legend of the God Uppalavaṇṇa (Sanskrit Utpalavarṇa,
“having the color of blue lotus”) being appointed by the Buddha as the guardian deity of the
island and taking the immigrants under his protection, even if the Mahāvaṁsa (chapter 7)
associates this with Vijaya. The earliest form of Vaiṣṇava religion in south India is supposed to
be the Pañcavīra cult, that is, the worship of the five Vṛṣṇi or Yādava heroes, in particular
Kṛṣṇa Vāsudeva and his elder brother Bala-Rāma, worshipped both independently and together
in Tamil Nadu in the early centuries CE. Such a migration of the Yādavas seems to be hinted
at in the northern Sanskrit sources too: Kṛṣṇa Vāsudeva moved from Mathurā to Gujarat,
where he founded the coastal city of Dvārakā or Dvāravatī. Sanskrit dvāra, “door,”
corresponds to Tamil kavāṭam/kapāṭam, “fold of a door,” found in the names Pāṇḍya-kavāṭa
(one of the pearl sources in Arthaśāstra 2,11,2) and Kapāṭapuram (the legendary seat of one of
the ancient Tamil literary academies). According Old Tamil tradition, Sage Agastya brought
the eighteen Vēḷir chiefs and the rulers of the Aruvāḷa country from Dvārakā. The Āy rulers of
the eighth- to ninth-century south Travancore likewise traced their descent from the Yādavas.
Who was the Indian Heracles? “This Heracles is chiefly honoured by the Surasenians, an
Indian tribe, with two great cities, Methora and Clisobora [Kleisóbora]; the navigable river
Iomanes flows through their territory. Megasthenes says that the garb this Heracles wore was
like that of the Theban Heracles by the account of the Indians themselves; he also had a great
many sons in this country, for this Heracles too wedded many wives, but he had only one
daughter. Her name was Pandaea” (Arrian, Indica 8,5–7, translated Brunt 1983:II,327–329).
Practically all scholars have identified the Indian Heracles with Kṛṣṇa worshipped by
Śūrasenas in Mathurā on the Yamunā river. An exception was James Tod, who in 1835
identified Heracles with Bala-Deva, the god of strength (bala). Strength is of course a
characteristic of the Greek Heracles; and there is other evidence suggesting that Tod was
right. Cicero, while discussing the six appearances of Heracles, states that in India he was
called Belus (Cicero, De natura deorum 3,42). Bala-Rāma’s strength is mentioned in the Old
Tamil poems of the first centuries CE: “In Pur[anānūru] 56, Krishna is invoked for his fame,
Balarāma for his strength. Krishna is described as having a body like blue sapphire, having a
bird (presumably the garuḍa) on his flag, and being accompanied by Balarāma, who has a
body the colour of a conch, a plow for his weapon, and a palmyra for his banner” (Hart
1975:57)
Textual and iconographic evidence from about 400 BCE onwards shows that Bala-Rāma was
in early Viṣṇuism a very important deity, especially in the Mathurā area. Mathurā is called
Madhurā not only in Pāli sources but also by Patañjali in his Mahābhāṣya, around 150 BCE. The
form Madhurā figures in the Mahābhārata, too, which derives the name from the demon
Madhu, who lived in the Madhu-vana forest on the Yamunā river but was slain by Kṛṣṇa, who is
therefore called Madhusūdana, “the killer of Madhu.” The “demoniac” god earlier worshipped
at Madhurā seems to have been a snake deity connected with plowing and identified with Śiva
(Śiva’s names listed in the Mahābhārata include Madhu) and who was addicted to drinking
wine (madhu). Madhu’s cult was then absorbed into that of Kṛṣṇa Vāsudeva by transferring all
the attributes of this earlier local god to Kṛṣṇa’s “elder brother” Bala-Rāma, who, among other
things, is a great wine drinker. Demon Madhu (with the demon Kaiṭabha, whose name may be
compared with Pāli keṭubha, “Brahmin ritualist”) is said to have stolen the Vedas from the god
Brahmā, which were then regained by Viṣṇu; this myth seems to imply that a Vedic tradition
prevailed at Mathurā before Viṣṇuism.
In the Vedas, madhu is specifically associated with the Aśvins. As discussed in chapter 10,
these divine charioteers, twin sons of the Sky (divo napātā), probably represent, among other
things, the (white) day and the (black) night. In Rigveda 3,55,11, day and night are spoken of
as twin sisters (yamyā) who have assumed different colors, one shining bright (tayor anyad
rocate), the other black (kṛṣṇam anyat); the Aśvins, too, are twins and are identified with day
and night (MS 3,4,4: ahorātre vā aśvinā). Some verses of the Rigveda suggest that the sun is
the chariot of the Aśvins. According to RV 6,9,1, “the white day and the black day—(the pair
of) light and darkness—manifestly turn around.” The color terms here used of day and night,
árjuna-, “white,” and kṛṣṇá-, “black,” are connected with the two members of the chariot team
in the Mahābhārata. The mightiest warrior of the Pāṇḍavas is Arjuna, whose name means
“white.” Arjuna’s charioteer is the wise and crafty Kṛṣṇa, whose name means “black.”
Originally Kṛṣṇa’s teammate was undoubtedly his elder brother, the strong Balarāma, who is
white in color. When the Vedic king at the royal consecration as a warrior (kṣatriya) ascends
his chariot, he calls himself Arjuna in the mantra that he thereby utters (ŚB 5,4,3,7, where
Arjuna is said to be Indra’s secret name).
The Aśvins drive around the world in a triple chariot accompanied by the fair goddess of
Dawn (Uṣas), daughter of the Sky or Sun (Sūre/Sūro duhitā, Sūryā), their sister and wife. This
trio has a counterpart in the divine horsemen of the Greeks, Kastor and Poludeukes (originally
Poluleúkēs, “much shining”)—who are sons of the sky god Zeus and brothers of Helen—as well
as in the Lithuanian twin gods expressly identified with the morning and evening star wooing
the daughter of the Sun. Many of the Aśvin hymns of the Rigveda belong to the Kāṇva family of
poets that was associated with the early Vedic tribe of Yadu, from which Kṛṣṇa’s Yādava tribe
is descended. It therefore appears likely to me that the trio of Aśvins and their sister/wife is
the model of the early Vaiṣṇava trio consisting of two brothers connected with the colors white
and black and their sister/wife. Chariotry is one further characteristic in common. The Aśvins
are chariot gods, and in the Mahābhārata, the “black” Vāsudeva is the charioteer of the
“white” chariot-warrior Arjuna. The two brothers Kṛṣṇa-Vāsudeva and Bala-Rāma and their
sister were a popular trio in early Vaiṣṇava iconography and are still worshipped as such at
Puri in Orissa.
Megasthenes, in the above-quoted passage, gave Pandaíē as the name of Heracles’s
daughter and of the country called after her, the source of pearls (i.e., the Pāṇḍya country of
south India). This name may correspond to Sanskrit Pāṇḍeyā, “daughter of Pāṇḍu.” In
Megasthenes’s account, Heracles is both the father and husband of Pandaíē:

In this country where Heracles’ daughter was queen, the girls are marriageable at
seven years, and the men do not live longer than forty years. There is a story about this
among the Indians, that Heracles, whose daughter was born to him late in life, realizing
that his own end was near, and having no man of his own worth to whom he might give
his daughter, copulated with her himself when she was seven, so that their progeny
might be left behind as Indian kings. Thus Heracles made her marriageable, and
thenceforward the whole of this line which began with Pandaea inherited this very
same privilege from Heracles. (Arrian, Indica 9,1–3, translated Brunt 1983:II,331)

I quote this account, because it parallels a well-known Sāvitrī legend, one of the many
side-episodes included in the Mahābhārata, and one of importance for the topic of chapter 19.
Princess Sāvitrī’s father, King Aśvapati of Madra (in the northern Indus Valley), fails to marry
off his daughter in time, and therefore sends her to search for and choose a husband on her
own. The texts do not directly indicate that the king had an incestuous relationship with
Princess Sāvitrī, but they do quote in this context a law text stating that if a girl experiences
her first menses in her father’s house, the father incurs a great sin. According to the
Mahābhārata 3,277,32, Aśvapati asks Sāvitrī to find a husband “equal to herself” (sadṛśam
ātmanaḥ) as no wooer is forthcoming, but according to the Skanda-Purāṇa (7,166,16), Aśvapati
says that however much he looks, he cannot find for his daughter a bridegroom who in worth
is equal to himself (vicārayan na paśyāmi varaṃ tulyam ihātmanaḥ).
In the Sāvitrī legend, the human couple (Princess Sāvitrī and Prince Satyavat) corresponds
to the divine couple (Goddess Sāvitrī and God Brahmā). It was through the grace of Goddess
Sāvitrī and her husband that the princess was born, and both the human and the divine Sāvitrī
along with their husbands are to be worshipped in the vow and ritual of vaṭa-sāvitrī-vrata that
is associated with the legend. Even the fate of the human couple has its counterpart at the
divine level. In accordance with the prophesy of Sage Nārada, the husband (Satyavat, alias
Citrāśva, the young “alter ego” of Sāvitrī’s father Aśvapati) dies after one year has passed
from his wedding, with his head on the lap of Princess Sāvitrī. Sāvitrī, as a faithful wife, Satī,
follows her husband to death when Yama comes to fetch him, and with her loyalty gains his
life back. The vaṭa-sāvitrī-vrata is still today centrally important for married Hindu women
whose husband is alive.
Parallel to the Sāvitrī legend, the Skanda-Purāṇa (3,1,40) tells how the creator god
Brahmā, alias Prajāpati, has sex with his own daughter Vāc and is therefore killed by Śiva, but
Brahmā’s wives Sarasvatī and Gāyatrī pacify Śiva and make him join Brahmā’s severed head
with the body. This myth is directly based on a Vedic myth most explicitly told in Aitareya-
Brāhmaṇa 3,33: Prajāpati is guilty of incest with his daughter Vāc and is killed by Rudra in
punishment. Vāc (“speech, voice, sound”) is another name of Goddess Sāvitrī, known best as
the holiest stanza of the Veda composed in the gāyatrī meter: its recitation at sunrise and
sunset, and (later) at noon are considered to manifest the goddesses Gāyatrī, Sāvitrī, and
Sarasvatī.
Prajāpati thus had an incestuous relationship with his daughter Vāc, who is explicitly
identified with the goddess of Dawn (Uṣas or Sūryā or Sāvitrī), and had to die, in punishment
for this sin. Pandaíē’s incestuous father Heracles also died soon after the copulation. Pāṇḍu,
the father of the Pāṇḍava heroes (and of Pandaíē = Pāṇḍeyā), after he had killed a mating deer,
was cursed to die if he ever copulated again, which came to pass when he had intercourse
with his wife Mādrī. Mādrī was a princess of the Madra country, and ascended the funeral
pyre of Pāṇḍu, resolute as the goddess Dhṛti. In both respects Mādrī resembles another
princess of the Madra country, namely Sāvitrī, who is the prototype of a Satī, and the human
counterpart of goddess Sāvitrī, the wife-daughter of Brahmā/Prajāpati. We have seen that the
female member of the early Vaiṣṇava trio (Kṛṣṇa’s sister Subhadrā, Rāma’s wife Sītā) seems to
continue the goddess of dawn (Sūryā/Sāvitrī) in the trio that she forms with the two Aśvins.
Not only Sāvitrī but this entire earlier trio appears to have been worshipped in the Madra
country, because Nakula (clever like Kṛṣṇa) and Saha-Deva (whose name is a synonym of Bala-
Deva as “the god of strength”), the two Pāṇḍavas sired by the Aśvins, had Mādrī as their
mother. Mādrī’s brother Śalya, King of Madra, had Goddess Sītā in his banner, and the Vedic
Taittirīya-Brāhmaṇa 2,3,10 mentions Sītā Sāvitrī as the daughter of Prajāpati. All this suggests
that Pandaíē, Uṣas/Sūryā/Sāvitrī, and Sītā are each other’s aliases.
I agree with Weber (1850), who considered Rāma’s spouse Sītā to be at least partly
mythical. An agricultural goddess, Sītā, the personified furrow, is known from the Rigveda
(4,57,6–7), and her worship is described in detail in Pāraskara-Gṛhyasūtra 2,17; according to
the Gobhila-Gṛhyasūtra (4,4,27–29), she was to be worshipped at plowing. It makes sense that
the husband of “Furrow” is the god of plowing. Weber therefore thought that the hero of the
Rāmāyaṇa had developed from Rāma Halabhṛt (“carrier of the plow,” i.e., Bala-Rāma), and that
he too personified an agricultural divinity, like Sītā. Bala-Rāma’s distinctive iconographic
emblems, the plow (lāṅgala, hala, and phāla) and pestle for pounding grain (muṣala) definitely
mark him as primarily an agrarian deity. The agricultural connection is also plain from his
alternative name, Saṃkarṣaṇa, which is derived from the activity of plowing (kṛṣi). Weber’s
hypothesis that the Rāma of the Rāmāyaṇa is actually Bala-Rāma is supported by the fact that
the name Bala-Rāma is not found in the Mahābhārata, where Bala-Rāma is called simply Rāma
(143 times).
The plow is instrumental in placing the seed in the womb of the earth, and plowing thus
symbolizes sexual intercourse. But the plow also creates the furrow, thus representing its
generator. In Rāmāyaṇa 1,66,14–15, Sītā emerges out of the furrow when Janaka, the king of
Mithilā, is plowing a field, and is given the name Sītā and raised as his daughter by Janaka. In
Rāmāyaṇa 7,88,9–14, Sītā finally returns to her mother Earth: the goddess comes to fetch her
and the two disappear underground. Janaka’s name denotes “progenitor, father.” It is one of
the names used in the Purāṇas of the Hindu creator god Brahmā, and Brahmā directly
continues Vedic Prajāpati, whom Taittirīya-Brāhmaṇa 2,3,10 mentions as the father of Sītā
Sāvitrī. On the other hand, as noted above, the plow and the field plowed (or the furrow) form
a couple, so that Prajāpati is also Sītā Sāvitrī’s husband through incest. In the Rāmāyaṇa, a
plow-god seems to be both Sītā’s father (Janaka) and husband (Rāma = Bala-Rāma).
Queen Pandaíē of Megasthenes has been compared with the guardian Goddess of the
Pāṇḍya capital Madurai, Mīnākṣī. In the local Tamil Tiruviḷaiyāṭar-Purāṇam (twelfth to sixteenth
centuries CE), Mīnākṣī is the daughter of a Pāṇḍya king of Madurai and his queen, who was the
daughter of a Cōla king called Śūrasena. Childless, they performed a Vedic sacrifice to obtain
a son, but received from the sacrificial fire a girl. (The birth of Princess Sāvitrī to King
Aśvapati in the Indus Valley was similar.) The girl had three breasts, and a voice from heaven
told that she should be educated in the military arts like a prince, and that she would conquer
the whole world. The third breast would disappear when she met her future husband. All this
happened, and finally when fighting at Mount Kailasa, she met God Śiva and the third breast
disappeared. After their marriage, Śiva ruled Madurai as King Sundara-Pāṇḍyan.
Here the spouse of Mīnākṣī is called Sundareśvara, “Beautiful Lord,” and considered to be
Śiva. However, there is in Madurai a local form of Viṣṇu whose Tamil name is Alakar,
“Beautiful Lord.” Alakar is the brother of Mīnākṣī who gives the bride away to the groom. The
Alakar-Malai temple near Madurai, with a standing form of Viṣṇu, dates to pre-Pallavan times,
and is one of the oldest in Tamil Nadu. Especially in a city called Madhurā, Alakar could have
been both the brother and the husband of the goddess in ancient times, as was the case with
Rāma and Sītā according to the Dasaratha-Jātaka. Both Sundara and Alakar might render
Sanskrit Rāma, which in classical Sanskrit means “pleasing, charming, handsome, lovely,
beautiful.” Some iconographic manuals prescribe that Rāma is to be depicted as beautiful
(sundara), others that Rāma and Lakṣmaṇa are to be exceedingly handsome. According to
Bhāgavata-Purāṇa 10,2,13, Bala-Rāma was called Rāma because he charmed people (rāmeti
lokaramaṇād) with his beauty.
Vijaya’s Sri Lankan yakkhinī wife Kuveṇī or Kuvaṇṇā likewise had three breasts, and she
had also been told that one of them would vanish when she would see her future husband,
which happened when she saw Vijaya. As David Shulman (1980) has pointed out, the Tamil
word kaṇ in Mīnākṣī’s Tamil name, Aṅ-kayar-kaṇṇ-ammaiyār, “lady of the beautiful carp-eyes,”
means both “eye” and “breast-nipple.” In the Śrīvidyārṇava-Tantra, Sītā is three-eyed and
wears the crescent of the moon on her head; she has four arms holding a noose, a goad, a bow,
and an arrow. Sītā Sāvitrī is an aspect of the warrior goddess Durgā, as is sometimes made
explicit in texts. In the case of Mīnākṣī, this relationship with Durgā is clear from her local
legend. This legend must be old, for in the Mahāvaṁsa (c. 500 CE), the daughter of King Paṇḍu
of the southern Madhurā is called Vijayā, which designates her as the goddess of victory.
The legend of a three-breasted princess recurs even at Nāgapaṭṭinam in Tamil Nadu: here
this “lady of the long dark eyes” (Karun-taṭaṅ-kaṇṇi) is the daughter of Ādi-Śeṣa, king of the
snakes, an ardent worshipper of Śiva. Of her, too, it was prophesied that the third breast
would disappear as soon as she saw the king who would wed her, in some variants a Nāgarāja.
Shulman (1980) has discussed her relationship with Mīnākṣī and with Kaṇṇaki, the heroine of
the Tamil epic Cilappatikāram, who destroys the city of Madurai with one of her breasts: all
multiforms of the three-eyed warrior goddess Durgā-Kālī. At Madurai, too, the bridegroom
appears to have been the local Śiva-related snake god, called in Tamil Āla-vāy (Sanskrit
hālāsya, “having poison in its mouth”).
Bala-Rāma incarnates a snake deity connected with fertility and the subterranean regions,
called Śeṣa, “remainder” (the name seems to refer to the seed grain left over for next sowing),
or Ananta, “endless.” Serpent Śeṣa drinks palm-wine, and has the palmyra palm and the wine
cup as his iconographic attributes. In this regard he is like Bala-Rāma, who in turn has the
triple-bended (tri-bhaṅga) pose associated with snake deities. Buddhist Sanskrit texts know
Pāṇḍuka, Pāṇḍuraka, Paṇḍulaka, and Paṇḍaraka as names of a nāga king, one of the guardians
of the great treasures.
The Mathurā region is considered to be the stronghold of Saṃkarṣaṇa-Baladeva worship.
The identity of Bala-Rāma is likely to have been pasted on the earlier local divinity there. The
myth of Kṛṣṇa’s subduing the snake Kāliya living in the Yamunā River and driving him away
from his home has been explained to symbolize the replacement of a snake cult earlier
prevalent at Mathurā with the cult of Kṛṣṇa. Excavations at Sonkh near Mathurā have
confirmed that snake worship prevailed to a remarkable degree at Mathurā still around the
beginning of the Christian era. The only major shrine discovered is a nāga temple. The
associated finds comprise images and panels representing serpent deities and inscriptions
referring to their cult. Nāga, Nāga Bhūmo, and Nāgarāja Dadhikarṇṇa are mentioned by name.

Although Saṅkarṣaṇa appears as a Vaiṣṇavite divinity in the Mahābhārata and the

Purāṇas, there are traces of his close connection with the cult of Rudra-Śiva also. The
Pañcarātra Saṃhitās often identify Saṅkarṣaṇa with Rudra-Śiva. The Brahmāṇḍa Purāṇa
states that Rudra was known as Halāyudha [“having the plow as his weapon”]. The
Viṣṇu Purāṇa speaks of Saṅkarṣaṇa-Rudra, who comes out of the mouth of the serpent
Śeṣa at the end of every aeon. . . . Śiva also is intimately associated with the nāgas.
(Jaiswal 1981:54)

In Bengal, Śiva is worshipped as Lāṅgaleśvara, “lord of the plow,” and the most important
phallic god of Hinduism could really be expected to be the god of plowing and generation.
Megasthenes’s account (c. 300 BCE) of the so-called Indian “Dionysus” almost certainly
describes the cult of Śiva, underlining his connection with agriculture and the plow: “The
Indians, he [Megasthenes] says, were originally nomads … until Dionysus reached India. But
when he arrived and became master of India, he founded cities, gave them laws, bestowed
wine on the Indians as on the Greeks, and taught them to sow their land, giving them seed. …
Dionysus first yoked oxen to the plough and made most of the Indians agriculturalists instead
of nomads, and equipped them also with the arms of warfare” (Arrian, Indica 7,2–7, translated
Brunt 1983:II,325–327.)
All these sources preserve important information about the earlier, pre-Aśvin background
of the Vaiṣṇava/Bhāgavata religion, connecting it with the Vedic and Hindu myths of the
incestuous father-and-daughter couple and the young prince-husband (an alter ego of the
father) resurrected from death, to be discussed in chapter 19. The agricultural divinities
involved here are very likely to have their origin in the indigenous, Harappan religion.
PART II
The Indus Civilization
14
The Language of the Indus Civilization

The identity of the Indus language is perhaps the most important puzzle about the Indus
civilization. At the present time, a number of languages are spoken in the Indus Valley. Indo-
Aryan languages include Sindhi in Sindh and Punjabi in the Punjab (with its dialects Siraiki
and Lahnda) and a number of languages of the Dardic subgroup (Shina, Khowar, Kohistani,
and others) with relatively few speakers in the northernmost part of the valley. Other Indo-
Aryan languages border the Indus Valley in the east, chiefly Hindi, with its dialect Marwari in
Rajasthan, and in the south, Gujarati, in Gujarat. Iranian languages are spoken immediately
west of the Indus Valley: Baluchi in Baluchistan, Dari and Pashto in Afghanistan (Pashto
extending also to the northern Indus Valley), plus Wakhi and some others with relatively few
speakers in the Hindu Kush. The Nuristani languages, spoken in northeastern Afghanistan,
have both ancient Iranian and ancient Indo-Aryan features. Outside the group of Indo-Iranian
languages is the Dravidian language Brahui, spoken in Baluchistan and Sindh, and the isolated
Burushaski, spoken in the northernmost part of Pakistan, close to the Chinese border.
In the third millennium BCE, it is quite likely that many dialects, and perhaps even
languages, were spoken in the Indus Valley. But we may be sure that just one language was
used in writing the Indus script, because the sign sequences of its inscriptions are repeated
throughout the Indus realm. Only in the West Asian Indus seals do we find clearly different
sign sequences testifying to the use of another language (probably Sumerian or Akkadian in
Mesopotamia, an early West Semitic language in the Arabian or Persian Gulf).
Some of the West Asian Indus seals are of the typically square-stamp Harappan type, and
their inscriptions, too, agree with South Asian Indus texts. Others, however, have common
Indus signs in sequences dissimilar to seals from the Indus Valley (Fig. 14.1). It appears that
the owners of these latter West Asian seals had adopted local, Mesopotamian names along
with the local language(s) they needed to know for their business, just as some of these
Harappan trade agents in Mesopotamia adopted the local cylinder type of seal. That the
language of Meluhha (the probable Sumerian name for the Greater Indus Valley, see chapter
17) was not understood in Mesopotamia is clear from an inscription found on an Akkadian
cylinder seal, which reads: “Šu-ilišu, interpreter of the Meluhhan language” (Fig. 17.4). This,
plus the fact that hardly a single object of clearly West Asian origin has been excavated in the
Indus realm, makes it very unlikely that the language spoken by the Harappans was any of the
West Asian languages.

FIGURE 14.1 A round Gulf-type stamp seal (see chapter 17) of glazed white steatite in the British Museum (Reg. No.
1883,116.1, BM 120228). A Harappan-type bison bull feeds from a manger beneath an inscription engraved with
five frequently occurring Indus script signs, which, however, are arranged in a unique order. Image 272276001 ©
The Trustees of the British Museum.

What about the Indo-Aryan languages? As demonstrated in Part I, there is an unbroken

chain of archaeologically connected cultures via which the Indo-Iranian languages passed
from their Proto-Indo-European homeland in the Ukraine to their earliest recorded areas of
speech. (Even today, the distribution of the Indo-Iranian languages covers almost the entire
distance from the Ukraine to India; the steppes north of the Black Sea, in Ukraine and
southern Russia, are now void of Indo-Iranian, but the Scythian tribes, who lived there as
nomads, spoke an Old Iranian language in the fifth century BCE, as we know from Herodotus.)
Since the Indo-Iranian languages started arriving in the South Asia only in the Late Harappan
period, around the twentieth century BCE, they cannot account for the Harappan language.
Archaeology supports this conclusion. The domesticated horse is totally absent from the art
and faunal finds of the Indus civilization. (The published claims to the contrary are based on
ambiguous bones, which may have belonged to the wild ass; Meadow 1991; Meadow & Patel
1997; Parpola & Janhunen 2011.) This fact is in drastic contrast with the important role played
by the horse in the culture and religion of the early Indo-Iranian speakers. No less striking is
the difference between the primitive wooden huts of the Vedic people and the Harappan brick-
built and preplanned cities and towns with their elaborate water engineering.
Speakers of the language of the Bactria and Margiana Archaeological Complex (BMAC),
too, are likely to have arrived in greater numbers to the Indus Valley only around the
twentieth century BCE, accompanying the first wave of Indo-Aryan speakers. As discussed in
chapter 8, the original non-Indo-European BMAC language appears to have given a large
number of loanwords to Old Indo-Aryan. Indeed, the majority of the many (more than 300)
foreign words of unknown origin in the Rigveda seem to have come from this source.
Burushaski, a language isolate in northernmost Pakistan, must have come from the north
and never spread further, as there are no linguistic traces of Burushaski south of its present
area of speech. The Tibeto-Burman languages (which belong to the great Sinitic family of
China, Tibet, and the Southeast Asian mainland) are spoken on the fringes of the Himalaya.
They probably came from the north prior to Burushaski; they may have been spoken in the
Northern Neolithic culture (2800–1500 BCE) of Kashmir and the Swat Valley. Characteristic
features of this culture, such as semisubterranean dwellings and dog burials, are paralleled in
China. The Northern Neolithic was in some contact with the Early Harappan Kot Diji culture
(Figs. 4.1 and 15.2b), but not demonstrably with the Mature Harappan culture, which did not
spread to the northernmost Indus Valley.
The Munda languages spoken in eastern India and belonging to the Austro-Asiatic family
have linguistic relatives in Southeast Asia; their westernmost language, Korku in central India,
is evidently an offshoot from eastern India. Toshiki Osada (2006), following a remark made by
Norman Zide (1969:420), another leading authority on the Munda languages, has shown the
untenable nature of many of the most frequently quoted Proto-Munda or Austro-Asiatic
etymologies proposed for Sanskrit words (including lāṅgala-, “plow,” kadala- / kadalī-,
“banana,” marīca-, “pepper,” and karpāsa-, “cotton”). Osada argues that the role of the
Munda/Austro-Asiatic languages in the linguistic history of South Asia has been overestimated
(cf. also Anderson 2008). Yet Michael Witzel (2005:176–180), followed by Franklin Southworth
(2005:64–69,325), considers “Para-Munda” as the most likely language of the Indus
civilization. All the linguistic evidence adduced in support, however, is questionable: among
the “clearest cases” of a Munda prefix cited from the Rigveda is the word vi-bhindu- (Witzel
2005:177), which however is a -ú- derivative from the present stem of the Indo-Aryan root
bhi(n)d-, preceded by the Indo-Aryan prefix vi- (thus Wackernagel & Debrunner
1954:II.2:473). Among the words allegedly formed with the Austro-Asiatic prefixes ka- and ki-
(no Austro-Asiatic cognates are given) are kaparda- and kiyāmbu-, which can be explained as
loanwords from Dravidian (see below and chapter 18).
Burushaski, the Tibeto-Burman languages, and the Austro-Asiatic languages are all spoken
by very small linguistic minorities in South Asia and have been restricted to the marginal
geographical areas, so far as is known. The Austro-Asiatic-speakers, who form the largest
group, amount to just 1.13 percent of India’s population (as of the 1991 census). None of these
minority languages has contributed generally recognized loanwords to the Rigvedic
vocabulary. They cannot be considered to be serious candidates for the Harappan language,
spoken by about one million people, a remarkably large population in the Bronze Age.
Only one group of South Asian languages remains as a candidate for the Indus language:
the Dravidian language family, by far the second largest after the Indo-Iranian family, spoken
by 22.53 percent of India’s population in 1991. As early as 1816, Francis Whyte Ellis noted
that Tamil, Malayalam, Kannada, Kodagu, Telugu, and Malto “form a distinct family of
languages” with which “the Sanscrit has, in later times, especially, intermixed, but with which
it has no radical connection.” In 1856, Robert Caldwell published the first edition of his
pioneering Comparative grammar of the Dravidian or South-Indian family of languages, in
which he named the language family “Dravidian,” taking it from the Sanskrit word draviḍa-,
used of the Tamil language and its speakers and more vaguely of other south Indian peoples.
Caldwell described just twelve Dravidian languages, adding Brahui in his second edition
published in 1875. Today, some twenty-six Dravidian languages are recognized, which early on
branched into three or four subgroups: North, Central, and South Dravidian, the last of which
split into two (Fig. 14.2). Four are major languages, Tamil, Malayalam, Kannada, and Telugu,
with roughly fifty-three, thirty, thirty-three, and sixty-six million speakers, respectively (in the
1991 census), and long literary traditions. Linguistically and culturally Tamil is the most
important, as it has remained highly conservative in structure and has a rich literature going
back some 2000 years, and even earlier in the case of some short inscriptions. Most of the
other Dravidian languages have relatively few speakers, many being tribal languages with
only an oral literature.

FIGURE 14.2 The geographical distribution of the Dravidian languages in South Asia: (1) Tamil (Tamil); (2)
Malayalam (Malayāḷam); (3) Irula (Iruḷa); (4) Kurumba (Kurumba); (5) Kodagu (Koḍagu, Kuḍagu, Coorg); (6) Toda
(Tuda); (7) Kota (Kōta); (8) Badaga (Baḍaga); (9) Kannada (Kannaḍa, Kanarese, Canarese); (10) Koraga; (11) Tulu
(Tuḷu); (12 Telugu (Teluṅgu, Teliṅga, Tenugu, Tenuṅgu); (13) Gondi (Gōṇḍī); (14) Konda (Koṇḍa, Kūbi); (15) Kui
(Kūi); (16) Kuvi (Kuwi, Kūvi); (17) Pengo; (18) Manda (Manḍa); (19) Kolami (Kōlāmī); (20a) Naikri (Naikṛi); (20b)
Naiki of Chanda; (21) Parji; (22) Ollari; (23) (Koṇḍēkōr) Gadaba (Gaḍba); (24) Kurukh (Kuṛukh, Kuṛux, Orāōn, Urāōn,
Kisan); (25) Malto (Maltō); (26) Brahui (Brāhūī). After Krishnamurti 2003:18, with the caption enlarged to include
alternative spellings of the language names and their synonyms. Copyright © 2003 Bhadriraju Krishnamurti.
Reprinted with the permission of Cambridge University Press.

Most Dravidian speakers inhabit south and central India. The exceptions are the speakers
of Brahui, Kurukh, and Malto. Brahui-speakers live close to the Indus Valley in Baluchistan,
now mainly in Pakistan, but previously also in Iranian Baluchistan. Though Dravidian, Brahui
is very heavily influenced by Baluchi, the West Iranian language that arrived in Baluchistan
from the northwest in several waves between 700 and 1200 CE (chapter 7). Brahui was
probably spoken in Baluchistan before this time, for its speakers had higher social prestige
than Baluchi-speakers until 1750, when the situation was reversed. Until quite recently, all
speakers of Brahui knew Baluchi, and vice versa, and families were often bilingual in Brahui
and Baluchi. The two other members of the North Dravidian subbranch, Kurukh and Malto,
which are closely related, are mainly spoken in mountainous regions of Bihar. It is possible
that the two languages branched off from Brahui in the first millennium CE, perhaps because
Gurjara or Muslim invasions prevented some Brahuis from returning to Baluchistan from
Sindh, where many of them had spent the winter season in an age-old transhumance from the
mountains to the plains and back.
The Dravidian languages appear to have spread to central and southern India from the
area of the Indus Valley. Copper Age cultures of the Deccan, which derive from the Early and
Mature Harappan cultures of the Greater Indus Valley, spread farming and animal husbandry
to central and southern India, in place of hunter-gathering—an archaeological configuration
that matches the current distribution of the Dravidian languages in central and southern India
(Fig. 4.4; Fig. 14.2). The Indus civilization itself covered Gujarat and the northern part of
Maharashtra (Fig. 4.2), where a Late Harappan seal with Indus script was excavated at
Daimabad (Fig. 14.3). Gujarati and especially Marathi, despite being Indo-Aryan languages,
are acknowledged to contain a considerable number of loanwords of Dravidian origin, while
the evidence of place names also suggests a Dravidian substratum. Particularly common are
the suffixes -ul, -uli, -ol, -oli, -vli, -vali, -valli, -vāli, -vāl, -palli, -pāli, and -pāl, all of which are
supposed to go back (via North and Central Dravidian *palli) to Proto-Dravidian *paḷḷi, “small
village, hamlet” (Southworth 2005).

FIGURE 14.3 A Late Harappan stamp seal (Dmb-1) from Daimabad, Maharashtra, period II, c. 1800 BCE. (a) The
obverse; (b) a side view. After CISI 1:352. Photo EL, courtesy ASI.

Not only loanwords reveal a Dravidian substratum; so do kinship relations. In theory,

Dravidian and Indo-Aryan-speakers adhere to two fundamentally different kinship systems.
Dravidian kinship terminology is based on the cross-cousin marriage, where a man may marry
his mother’s brother’s daughter or his father’s sister’s daughter (both of whom are cross
cousins), but not his mother’s sister’s daughter or his father’s brother’s daughter (who are
both parallel cousins, because they are the children of same-sex siblings). By contrast, in the
Indo-Aryan marriage system exogamy of consanguines prevails, although the exact rules vary.
For instance, among the Rajputs one may not marry any member of the gotras (extended
patrilineages) of one’s four grandparents, and according to the even more restrictive sapiṇḍa
rules of the Sanskrit law texts, marriage is prohibited with all near-relatives by birth, not just
patrilineal relatives. In practice, however, some Indo-Aryan speakers follow Dravidian kinship
practices.

Although for the most part Dravidian speakers are also Dravidian in kinship and Indo-
Aryan speakers are Indo-Aryan in kinship, that is by no means wholly the case. In
particular, the kinship map serves to establish a Dravidian presence right up the
western coast of India, well to the north of the limit of Dravidian speech, through
Maharashtra and into Saurashtra in Gujarat. This is a matter of great historical
significance … there are a number of communities that, although they speak Indo-
Aryan languages, nevertheless have a Dravidian system of kinship. (Trautmann
1981:112)

If the Harappan population numbered about one million in the last quarter of the third
millennium BCE, its language is bound to have left some trace in the Vedic language that was
spoken in the Indus Valley a thousand years later. Indeed, even structural features in the Vedic
language not found in the closely related Old Iranian have been ascribed to Dravidian
substratum influence. These involve both phonology and syntax, including the retroflex
consonants as phonemes; an increasing use of the gerund; the quotative construction with the
particle iti following direct speech; and onomatopoeic expressions with echo-duplication (e.g.,
Rigvedic budbuda-, “bubbling” ~ Telugu buḍa-buḍa, “with a bubbling noise”; Atharvavedic
kurkura-, “dog” ~ Kannada kure kure and Malto kur kur, “call to a dog”; Rāmāyaṇa
kaṭakaṭāpayati, “gnashes the teeth” ~ Kannada kaṭakaṭa kaḍi, “to grind one’s teeth”). From
later Indo-Aryan one may mention the simplification of the syllable structure and the use of
dative-subject constructions as evidence of Dravidian influence. Moreover, it is widely
recognized that the following seven Rigvedic words, at least, are derived from Proto-Dravidian
words:

phalam, “fruit” < *pa lam, “ripe fruit”

mukham, “mouth, face, beginning” < *mukam, “mouth, face, beginning”
khala-, “threshing floor” < *kaḷam, “threshing floor”
lāṅgalam, “plow” < *ñāṅgal, “plow” (in my opinion, ñāṅgal is originally a compound of Proto-
Dravidian *ñāl, “earth, ground” + *kal, “to dig, excavate”)
ukham, “hip” < *ukkam, “hip”
kāṇa-, “blind in one eye” < *kāṇ-, “to see” + negative marker *-aH-, “to see not”
kaṭuka-, “pungent, bitter” < *kaṭu “pungent, bitter”

Although the Rigveda is the earliest recorded text collection in Indo-Aryan, it was composed
by the later wave of immigrants, who may be assumed to have met fewer Dravidian speakers
in the Indus Valley than the Indo-Aryan speakers who came earlier to the Indus Valley. The
number of Dravidian loanwords in the later Vedic texts and the epics that continue the
Atharvavedic tradition is well over 150; and there are many more if still later Indo-Aryan is
taken into consideration. For instance, the following two words central in cultic Hinduism,
discussed in the next chapter, are very likely to have Dravidian origins: pūjā and ghāṭ. I now
consider six further early Vedic words that in my personal view offer fairly convincing
examples of Dravidian loanwords. The first two concern the flora and fauna of the Indus
Valley; the remaining four are my own etymologies.

1. kiyāmbu, a kind of aquatic plant

This word is known only from variants of a single Vedic verse. The Rigvedic phrase
(10,16,13c) kiyāmbu atra rohatu, “may kiyāmbu grow here,” appears as kyāmbūr atra rohatu
in the Atharvaveda (18,3,6c). There is general agreement that kiyāmbu denotes a kind of
aquatic plant, because the context of the verse concerns cooling the place of corpse cremation
with plenty of water. No cognates have been recorded from later forms of Indo-Aryan, and the
etymology has been considered unclear. However, Nikita Gurov (2000) has proposed an
excellent etymology from Proto-Dravidian *kiyampu, which through regular sound changes
became Tamil cēmpu, a flowering plant known variously as taro, aroid, and Colacasia. The
lengthened -ā- in the Rigveda and Atharvaveda may result from the idea that kiyāmbu is a
compound with Sanskrit ambu-, “water,” as the last member (although ambu- is not attested
before the Upaniṣads). Incidentally, this etymology is important for the botanical history of
South Asia, as it suggests that taro was present in the Indus Valley in 1000 BCE; in a recent
survey by a leading researcher, taro is supposed to have come to south India from Orissa or
eastern India (D. Q. Fuller 2006:41,46–47).

2. gardabha, “donkey” (sometimes also “wild ass”)

The wild ass is the only native equid of South Asia, its habitat being restricted mainly to
the extensive saline soils and salt marshes of Sindh and Gujarat. Its bones have been found at
many Harappan sites and the animal itself is depicted on one Indus seal from Kanmer in
Kutch, as well as in a sign of the Indus script, and in several terracotta figurines. After the
similar-looking donkey was introduced (besides the domesticated horse and the camel) from
Central Asia to the Indus Valley by Indo-Aryan speakers, the native Dravidian word for “wild
ass” was also used for “donkey.” Proto-Dravidian *kaẓutay, which is preserved in Tamil kalutai,
“donkey,” seems originally to have meant “the kicking animal of the salt marsh” (from *kaẓ-,
“saline soil, salt marsh” + *utay, “to kick”). How Proto-Dravidian *kaẓ(u)ta(y) became garda-
in Sanskrit may be easier to understand if we note that it has become kalte in Kannada and
gāṛde in Kuvi. Adding the ancient Indo-Aryan suffix -bha- to animal names has parallels even in
later Indo-Aryan (Parpola & Janhunen 2011).

3. oṁ
The sacred syllable is not attested in the form oṁ in the Rigveda, but is thought to be
implied by the expression akṣara “syllable” in three verses of the riddle hymn RV 1,164. It is,
however, much used in the later Vedic liturgy. When the Rigvedic, Sāmavedic, and Yajurvedic
priests ask the Brahman priest for permission to carry out an action (e.g., brahman apaḥ
praṇeṣyāmi, “O Brahman, I am about to carry forwards the water”), oṁ introduces the
permission (anujñā), followed by an imperative of the appropriate verb (e.g., oṁ, praṇaya,
“yes, do carry it forward!”). The meaning of oṁ is explicitly stated in the Chāndogya-Upaniṣad
(1,1,8): “This syllable is one of permission; for when one permits anything, he says oṁ.” When
the Śunaḥśepa legend is related during the royal consecration, “oṁ is the response to a
Rigvedic verse (ṛc), tathā (‘be it so’) to a profane verse (gāthā); oṁ is divine, tathā is human”
(AB 7,18). Distinction is made here between the hieratic speech of the sacrifice, where oṁ is
used to express agreement, and ordinary mundane conversation, where tathā or tathāstu is
used. But occasionally oṁ and na are used for “yes” and “no” in ordinary prose dialogue as
replies to questions (e.g., in BĀU 6,2,1). I therefore propose that oṁ (with a long ō, as always
in Sanskrit) derives from Proto-Dravidian *ām, “yes,” a contracted variant of *ākum, the non-
past (habitual present-future without personal endings) regularly formed with the suffix *-um
of the very basic auxiliary verb *ā / *āku, meaning “it is, it is so, yes.” This usage of ām to
mean “yes” is still very common in Dravidian languages, for instance Tamil, and in the Jaffna
dialect of Tamil in Sri Lanka, where ām has become ōm, “yes,” clearly repeating the same
phonetic process (vocalic anticipation of m) that resulted in Vedic ōm. Hindi hāṁ, “yes,” owes
its initial h- to North Dravidian, where all initial vowels are introduced with a glottal stop
(Parpola 1981).

4. śakaṭī-, śakaṭam, “ox-cart”

This word is attested once in the latest book of the Rigveda (10,156,3), where it has the
feminine gender (śakaṭī-), apparently because it is connected with the goddess of the forest.
Thereafter it mostly appears in the neuter gender as śakaṭam, which in later Vedic texts
replaces the inherited Indo-Aryan word anas-, “ox-cart” (cognate with Latin onus, “load”). This
suggests that the local, Indus Valley, name of the much-used ox-cart broke through to Vedic
usage, given that the population of the Indus Valley had an ox-cart from the Early Harappan
period, around 3000 BCE. The cognates of śakaṭam in the various modern Indo-Aryan
languages presuppose variant forms *śakkaṭam and *śaggaṭam (Turner 1966:709; Mayrhofer
1996:II,601). I therefore suggest derivation, with metathetical sound transposition, from
Proto-Dravidian *caṭṭakam, “wooden frame or platform of a cart” (also used of the “frame of a
door, bedstead, bier”). The Harappan terracotta models of carts consist of a platform-like
frame with holes for vertical pins, and separate wheels; ox-carts of the same type are still used
in Sindh (Parpola, in press d).

5. bhekuri/bekuri, “calendrical asterism, mistress of the moon”

The oldest Yajurvedic texts in one context refer to the nakṣatras—that is, the calendrical
asterisms (lunar mansions of Hindu astrology)—by the name bhekuri or bekuri. They treat the
nakṣatras as heavenly mistresses of the moon. Śatapatha-Brāhmaṇa (9,4,1,9) compares the
moon joining the nakṣatras to a heavenly playboy (gandharva) having sexual intercourse with
dancing girls (apsaras); and gives a folk-etymological explanation of bhekuri by stating that
the nakṣatras “make light” (in Sanskrit, bhāṃ kurvanti). A more plausible etymology, in my
view, is to trace b(h)ekuri back to what seems to be the original Dravidian appellation of the
nakṣatras, preserved in several Dravidian languages in words denoting “morning star,”
particularly Old Tamil vaikuru mīn, derived from the Proto-Dravidian roots *vayku, “to stay
overnight, to outlast the night, to begin to grow light, to dawn,” and *u ru, “to be joined, to be
close together, touch, have sexual intercourse.” The alternations v-/b- and ai/e are very
common in Dravidian languages, for example in the representation of the Proto-Dravidian root
*vayku in Tulu, baiku, Telugu, vēgu, Kuvi, vēy- and Malto, bije (Parpola 1994a:206–207).

6. kiṃpuruṣa and kinnara, mythical musicians

In epic and classical Sanskrit literature, kiṃpuruṣa and kinnara are imagined to be half-
human, half-animal (usually a horse). According to Māgha’s Śiśupālavadha (4,38), the
kiṃpuruṣa has a horse’s head and a human body, while the kinnara has a human head and an
equine body. Both words can be analyzed as compounds consisting of Sanskrit words meaning
“what?” (kim) and “man” (puruṣa and nara), yielding the overall sense, “what kind of man?”
Of the two words, only kiṃpuruṣa occurs in the Vedic literature, in Brāhmaṇa texts. These
specify that when a fire altar is built of bricks, five victims are sacrificed and their heads
deposited in the ukhā pot in the center of the lowest layer of the altar. These five victims
(including a man) are spoken of as “animals of the village”: man (puruṣa), horse (aśva), cow,
sheep, and goat. They are correlated with five “animals of the forest,” which have been kept
imprisoned but are then released in order to remove any evil created by the distress and pain
suffered by the victims of the sacrifice. The five “wild” animals are: kiṃpuruṣa, wild ass
(gaura), bison, urial or camel, and markhor or ibex. “Mock-man,” “depraved man,” “deformed
man,” “dwarf,” “savage,” and “ape” have been suggested as possible meanings for kiṃpuruṣa.
Because kiṃpuruṣa occurs in the Vedic literature but kinnara does not, kinnara has
generally been considered to be a later transformation of kiṃpuruṣa. In my view, however,
kinnara predates kiṃpuruṣa: it is a Dravidian word, which was misunderstood to be Sanskrit,
so that nara-, which in Sanskrit means “man,” was rendered by Sanskrit puruṣa-, “man.”
Indeed, kinnara(m) is probably the very earliest Dravidian word on record. The Harappans
may have brought the word to Mesopotamia, because it seems to be written in cuneiform
sources. It appears first with the cuneiform logogram for “harp.” Then, from the eighteenth
century BCE onwards, it is written phonetically in Old Babylonian cuneiform as GIŠki-in-na-
ratim and in the Ugaritic cuneiform and alphabet as GIŠki-na-rum, knr. These words definitely
denote a “stringed musical instrument,” and in this meaning cognates are known from a wide
range of languages used at various times in the region: New Kingdom Egyptian k[]-nù-rú and
k.nn.r, Syriac kennārā, Arabic kinnāra, kinnīra, and kinārum, Hebrew kinnōr, Greek kinúrā
and kinnúrā. In Hittite, kinirtalla denotes “a musician (zither player?).” Assyriologists have
suspected that the stringed musical instrument denoted by these words is an import from
India to West Asia; the words are considered decidedly non-Semitic in structure.
I regard kinnaram as a compound word, where the first part is the Proto-Dravidian root
*kil, “to sound, resound,” used in Old Tamil to express the sound of a harp; its final *-l has
been assimilated to the initial *n- of the second member of the compound. This second
member is Proto-Dravidian *naram < *ñaram, “nerve, tendon, sinew, vein,” in Old Tamil
“string of the harp” (made of tendon). The original meaning of kinnaram is therefore “a
resounding stringed musical instrument,” and in this meaning the compound word is attested
in nine Dravidian languages: Tamil and Malayalam kinnaram, kinnari, Kannada and Tulu
kinnari, Telugu kinnara, kinnera, Parji kindri, Kuvi kinēri, Kurukh kendrā, and Malto kendre.
Moreover, three Indo-Aryan languages have words with the same meaning: Sanskrit kiṃnarā
(attested only in a native Sanskrit lexicon), kiṃnarī (in Hemacandra’s lexicon and the
Kathāsaritsāgara), Sindhi kīniro, kīnaro, and Old Marathi kinarī. According to Hemacandra,
kiṃnarī is “a lute of the tribals, caṇḍālas” (Parpola 1983).
Based on textual evidence (partly discussed in chapters 12 and Chapter 19), I propose that
the human victim in Vedic sacrifices was a harp-playing bard, who was called (like any
member of a modern orchestra) after his musical instrument. The sacrifice included a
“revival” ritual that involved swapping the severed heads of the human victim and the animal
(horse) victim, resulting in two “mythical” creatures of the kind described above by Māgha
(see chapter 11).
Thus kiṃpuruṣa, an old problem relating to human sacrifice in the early layer of Vedic
rituals, turns out to be a folk-etymological translation loan from Dravidian kinnara, which may
be the earliest recorded word from any South Asian language.
Sanskrit śakaṭam, derived from Proto-Dravidian *caṭṭakam, denotes the ox-cart, the most
important transport vehicle of South Asia from Early Harappan to modern times. The sacred
syllable oṁ, nowadays one of the primary symbols for Hinduism, also has a Dravidian
etymology. These new etymologies increase the probability that Proto-Dravidian was the
language of the Indus civilization—a conclusion reinforced by the evidence of the Indus script,
as we shall see in chapter 21.
15
Fertility Cults in Folk Religion

Excavation in the Neolithic villages of Baluchistan, such as Mehrgarh, has brought to light not
only evidence for the cultivation of crops and the rearing of animals for food but also what
appear to be religious images: terracotta figurines of human females and bulls as the principal
cultic artifacts. They suggest that the earliest village religion involved prayers for the fertility
of plants, animals, and people. The female figurines are thought to represent the goddess
Earth, the progenitor of plants from seeds sown into her womb and the nurturer of
worshippers, like a mother who gives birth to children and takes care of them. The bull
figurines relate to animal husbandry. They represent the thundering and pouring Sky, which is
conceived in many archaic religions as a roaring bull, a powerful male god whose semen
fertilizes his spouse, the Earth. This interpretation is endorsed by West Asian parallels, to be
discussed in chapter 19.
Numerous terracotta figurines of human females and a variety of male animals show that
this fertility cult continued into the Early and Mature Harappan periods dated several
thousand years later, and indeed even up to our times in South Asian villages. Male human
figurines make their appearance in the Early Harappan period, first at Mehrgarh, where they
usually are turbaned and sometimes have a child in their arms. Of the human figurines
excavated at Harappa, 1,143 have been identified as female (some also holding infants) and
407 as male. Most of the 500 or so human and animal figurines excavated at Mohenjo-daro
were found in rooms with just one door, that is, in a private part of the house. Thus the cult
images were granted a special place comparable to the pūjā room or alcove in Hindu homes.
The Indo-Iranians—and the Vedic priests who followed them—did not worship their deities
in image form. They invited the invisible gods to come and sit on the grass next to the
sacrificial fire, into which they poured food and drink for the gods, while praising them with
newly composed hymns. The Hindus, by contrast, perform their ritual of worship directly and
concretely upon the divinity, be it a statue, statuette, an icon, or a living being, either human
(such as bride and groom during the wedding, when they personify Śiva and Pārvatī, or a
guru) or animal (such as cattle during the yearly festival, snakes on the nāgapañcamī day), the
banyan tree (while performing the vaṭasāvitrīvrata) or a lifeless object (such as a book or a
weapon during the navarātri festival, or wooden sandals of saintly persons).
The Hindu image worship, called in Sanskrit pūjā, according to the learned manuals
(written by Brahmins) normally consists of sixteen “services,” but this number can vary from 1
to 108. Some of the earliest (post-Vedic) texts list these as follows: (1) invocation, (2) seat, (3)
water for washing the feet, (4) arghya water offered to an honored guest, (5) water for
sipping, (6) bath, (7) garments, (8) sacred thread, (9) anointing with unguents or sandalwood
paste, (10) flowers, (11) incense, (12) lamp, (13) food, (14) mouth perfume, (15) (praise hymn
and) prostration, and (16) circumambulation and jpgt or dismissal. One can immediately see
that items (2) to (5) come from the Vedic ritual for receiving an honored guest, and (8) the
sacred thread likewise belongs to the Brahmanical tradition.
The original non-Aryan mode of treating an honored guest, human or divine, seems to be
first recorded in our sources in some rituals described in late Vedic texts. One is the sarpa-
bali, offerings made without sacrificial fire to the snakes, another the ancestor rituals. The
main thing seems to be bathing the object of worship, anointing the person’s, animal’s, or
image’s body with unguents, and giving garments and various adornments and items of
toiletry (eye-liner, comb, mirror) and flower garlands. In temple worship, the image may be
bathed with many different kinds of liquid (milk, curds, ghee, honey, sugar-cane juice …). It is
also important for the worshipper to drink part of the bathing water and eat the remnants of
the god’s meal, which are charged with holy power from the divine touch.
If one turns from textual sources and temple ritual, which have been in the care of
Brahmins, to the actual worship of simple villagers, Jarl Charpentier (1926) has collected
several pages of references from all over India from the nineteenth century, where the
worship consists of anointing the icon—which may be nothing but a peculiar stone under a
sacred tree—with oil and vermilion paste. The red ointment, the red kunkuma powder, and the
yellow turmeric powder, which all are so often used in smearing Hindu objects of worship, are
supposed to stand for blood, which is poured on the icon of the goddess in animal sacrifices in
south India and Bengal. In the Vedic ritual this has one singular parallel, which is undoubtedly
of non-Aryan origin. In the horse sacrifice the blood of this chief victim is poured on the head
of a bald old man with projecting teeth, who, up to his head in water, personifies the god
Varuṇa. The description of the man and his being in water reminds one of the crocodile as
Varuṇa’s mount, discussed in the sequel. Chapter 19 will argue that the Vedic horse sacrifice
has replaced a buffalo sacrifice of Harappan origin.
The Sanskrit root pūj-, “to worship, honor,” with its derivates, is not used in Vedic texts
before the Gṛhyasūtras, where the object of worship can, in Hindu fashion, be an animal such
as the bull (GGS 3,6,11). Its etymology from the Proto-Dravidian root *pūcu, “to smear,
anoint,” suggested by Hermann Gundert (1869), agrees well with the central act of the
simplest form of Hindu pūjā. Dravidian pūcu corresponds in its meaning to the Sanskrit verbs
lip- and añj-, “to smear, anoint,” both of which occur in the descriptions of the sarpa-bali
ritual.
The emphasis on bathing in Hindu rituals may possibly be traced back to the Indus
civilization, as first suggested by Marshall. Mortimer Wheeler agreed, noting as follows:

the importance … of water in the life of the Harappans is stressed by the Great Bath on
the citadel of Mohenjo-daro and by the almost extravagant provision for bathing and
drainage throughout the city, and may provide yet another link with the later Hinduism.
The universal use of ‘tanks’ in modern Indian ritual, and the practice of bathing at the
beginning of the day and before principal meals, may well derive ultimately from a
usage of the pre-Āryan era as represented in the Indus civilization. (Wheeler 1968:110)

In my view Hindu bathing places, such as the ghats at Varanasi, may have existed from the
time of the Indus civilization. It is supposed that a canal or a branch of the Indus flowed next
to the lower city of Mohenjo-daro, which appears to have been surrounded by revetments
functioning as flood defences. Next to what Ernest Mackay took to be “a small fort on the city
wall,” he found a “ghat-like staircase” that “led down at least as far as the present water-
level” (Wheeler 1968:47). In any case the derivation of Sanskrit ghaṭṭa- from Dravidian *kaṭṭa,
“river bank, embankment, dam,” is fairly generally accepted for the Hindi word ghāṭ, “bathing
place,” and its cognates in many modern Indo-Aryan languages; and it is surprising that the
corresponding Sanskrit word ghaṭṭa- is not attested in Vedic, Epic, or Purāṇic literature, but
only in the works of lexicographers. All the nuances of the equivalent Indo-Aryan term tīrtha,
“ford, sacred bathing place,” are supplemented by a near homophone of Dravidian *kaṭṭa,
namely *kaṭa, “to cross over, ford, transgress” (Parpola 2003).
Early Harappans used the plow and the ox-cart; indeed all their basic crafts were well
developed. In practical respects, South Asian villages have changed little until very recently.
We can probably expect the same to be true of religion, since folk religion in all cultures is
notably conservative. Speaking of Gujarat, a nineteenth-century scholar noted: “Every village
has its own special guardian mother, called Mātā or Ambā”—some 140 different “mothers” in
all. “Generally there is also a male deity, who protects like the female from all adverse and
demoniacal influences. But the mother is the favourite object of adoration” (Monier-Williams
1885:222). The same held true in India at large, not least in the Dravidian-speaking south
India. In every village the mother goddess personified the place and its soil, out of which her
cult images were made by the local potter. She was prayed to for blessings such as children,
good health, and animals, but also feared, for she could bring about terrible calamities such as
a pestilence, if she were neglected or angered.
Early Harappan cultures started moving toward the east and south in about 3000 BCE, and
later waves of influence in these same directions came from the Indus civilization (Figs.4.1–2).
That the Harappan water-buffalo cult (Figs. 15.2, 16.6) had reached peninsular India by the
Late Harappan or Chalcolithic times is suggested by the large bronze sculpture of water
buffalo (Fig. 15.1) discovered in 1974 in a hoard at Daimabad, the southernmost Indus site in
Maharashtra (Figs. 4.4, 14.3). Throughout south India, until relatively recently, village
goddesses have been worshipped through water-buffalo sacrifices. The goddesses have been
associated with a male deity called the “buffalo king,” represented by a wooden post or a
pillar made of stone, or by the pipal tree (Biardeau 2004).
FIGURE 15.1 A water buffalo, 31 cm high and 25 cm long, standing on a platform attached to four solid wheels. One
of four bronze sculptures weighing together over 60 kg, found in a hoard at Daimabad, Maharashtra, the
southernmost Indus site, and ascribed to the Late Harappan or Chalcolithic period. Photo Asko Parpola.

In the form of worship at Kannapuram, a south Indian village in Tamil Nadu studied by
Brenda Beck (1981), the tree-trunk (known as kampam in Tamil, from Sanskrit skambha,
“pillar”) in front of which the sacrificial buffaloes are decapitated, is said to be the husband of
the goddess. At the end of the annual marriage rite, after the last victim has been slaughtered,
this trunk is uprooted and the goddess divested of her ornaments like a widow. The pillar and
its uprooting correspond to Śiva’s phallus and its castration. The sacrificial post used to be
burned after the divine marriage festival.
In a story about the origin of the tree-trunk, the goddess is said to have implanted herself
in the womb of a Brahmin woman and been born as a Brahmin girl. An untouchable Paraiya
boy fell in love with her and obtained her parents’ consent to marry her. His untouchable
status was discovered only later, to the anger of the goddess. With a scorching look she
burned her husband to ashes and from them caused a margosa tree to grow, to stand forever
outside her house. This is a typical version of the myths explaining the origin of the trees and
posts symbolizing the male divinity who is the low-caste guardian or servant of the goddess.
The myth of Śiva and Kāma, known all over India, is quite parallel. With an angry look from
his third eye, Śiva incinerated Kāma, the god of love and sexual lust, in revenge for Kāma’s
having shot Śiva with arrows of desire that made him fall for goddess Pārvatī. The ashes
covering Śiva’s ascetic body are said to be those of Kāma. According to some variants of this
myth, Kāma became a tree. In contemporary Sri Lanka, a Tamil festival enacts Śiva’s burning
of Kāma, in which Kāma is impersonated by a post, which is also burnt. In central and south
India, the spring festival of Holi (at the vernal equinox) commemorates the death and
resurrection of Kāma with a big bonfire. Human figures representing Kāma and his wife Rati
(“sexual pleasure”) are made and the male figure thrown into the fire. In Maharashtra, people
walk around the fire shouting words for female genitals, and dances imitating copulation are
performed. In Kumaon, Holi is celebrated by planting a pole in ground, dancing around it,
singing erotic songs, and burning the pole on the last day of the festival.
In the Vedic ritual, the animal victim was tied to a wooden sacrificial stake (yūpa). This
symbolized the primeval cosmic tree growing from the navel of the earth, upholding the sky
and thus leading to heaven. “It is through the sacrificial stake that the offerings go to the
heavenly world,” says the Maitrāyaṇī Saṃhitā (4,8,8). After a Vedic sacrifice the post was
either left standing or thrown into the sacred fire. The burning of the stake is explained by
saying that the fireplace is the womb of the gods, and the sacrificer will be born in heaven
with a body of gold (AB 2,3).
That the sacrificial post also had a phallic connotation, even in the Vedas, is evident from
the Vedic rite of puṁsavana, “causing the birth of a male child.” Here various symbols of the
male generative organ are used, for example, two beans (for testicles) together with a barley
corn (for the phallus), a shoot of the banyan tree with a fruit on either side and an aerial root
of the banyan. The banyan shoots and air roots are pounded with millstones (an activity
symbolizing sexual intercourse) and the husband inserts the resulting paste—expressly
equated with sperm—into the right nostril of his wife. Among the other things that may be
mixed with this potent paste is “a splinter of a sacrificial post … exposed to the fire” (HGS
2,1,2,6).
The deity prayed to in the puṁsavana ritual is Prajāpati, “the lord of offspring”; but the
means used in it, barley and beans (KS 36,6; MS 1,10,12), as well as the banyan-tree (GGS
4,7,24), are all said to belong to Varuṇa. According to the Śunaḥśepa legend (discussed in the
next chapter), Varuṇa is the deity to approach in order to conceive a son. This fact underlines
Varuṇa’s phallic nature and his relationship with Prajāpati, Śiva, and liṅga worship, which has
become such a vital part of Hinduism.
In descriptions of the south Indian folk religion the guardian goddess of the village and the
bloody sacrifices to her indeed occupy a central position:

On the night before the day appointed for the offering of animal sacrifices by the
villagers, a male buffalo, called Devara Potu, i.e. devoted to the deity, is sacrificed on
behalf of the whole village. First, the buffalo is washed with water, smeared with yellow
turmeric and red kunkuma, and then garlanded with flowers and the leaves of the
sacred margosa tree. It is brought before the image; and a Mādigā cuts off its head, if
possible at one blow, over a heap of boiled rice, which becomes soaked with the blood.
The right foreleg is then cut off and placed crosswise in its mouth, according to the
widespread custom prevailing in South India, the fat of the entrails is smeared over the
eyes and the forehead, and the head is placed in front of the image. A lighted lamp is
placed, not as in other villages on the head itself, but on the heap of rice soaked with
blood. This rice is then put into a basket; and a Mādigā, the village vetty or sweeper,
carries it round the site of the village, sprinkling it on the ground as he goes. The whole
village goes with him, but there is no music or tom-toms. The people shout out as they
go “Poli! Poli!” i.e. “Food! Food!” and clap their hands and wave their sticks above their
heads to keep off the evil spirits. The rice offered to the goddess, but not soaked with
blood, is then distributed to the people. … On the next day … various householders,
even Brāhmans and Bunniahs, bring animals for sacrifice. All are killed by a Mādigā,
and then the heads are all presented and placed in a heap before the goddess.
Sometimes an extraordinary number of animals is sacrificed on occasions of this kind,
as many as a thousand sheep on a single day. (Whitehead 1921:56–57)

If the buffalo sacrifice is common in south Indian village religion, it is absent from north
Indian village religion. The reason is probably the conscious efforts of Vedic Brahmins to
eradicate it. Extravagant buffalo sacrifices were at first adopted by the Rigvedic Aryans in the
Indus Valley (cf. RV 5,29,7–8, etc.), but thereafter this mode of worship is not heard of. In the
later Vedic literature a buffalo sacrifice (to Varuṇa) is mentioned in only a single context, in a
list of hundreds of different animals offered as subsidiary victims in the horse sacrifice. It
appears that generally speaking, the Brahmins have been fighting against bloody sacrifices
from Rigvedic times onwards. In the Rigveda, the cow is called aghnyā, “not to be slain,” and
while the Gṛhyasūtra rules (apparently reflecting the behavioral code of the Atharvavedic
Indo-Aryans) include the slaughter of a cow when a guest of honor is received, a later rule
leaves it to the guest to decide whether this is done or the cow is set free. The Brāhmaṇa and
Śrautasūtra texts record, even mentioning the names of the Brahmins concerned, how in the
Vedic sacrifices human, horse, and other animal victims were successively discontinued, and a
final rule says that the prescribed victims should be made out of rice and barley paste.
In the Indus civilization the water buffalo was connected with an important divinity, for the
famous “Proto-Śiva” on a seal from Mohenjo-daro wears the horns of a water buffalo and this
animal itself is also depicted on the seal (Hiltebeitel 1978) (Fig. 16.6). The water buffalo is
prominent among the motifs on painted pottery in the Early Harappan Kot Diji culture; it is
attested in different corners of the Indus Valley: at Kot Diji in Sindh, Gumla in the northwest
and even in a vessel imported to the Northern Neolithic site of Burzahom in Kashmir (Fig.
15.2).
FIGURE 15.2 Water buffalo on painted pots of the Early Harappan Kot Diji culture. (a) A pot excavated at Kot Diji.
After Khan 1965:58, fig. 16. Courtesy DAMGP. (b) A pot imported to the Northern Neolithic site of Burzahom in
Kashmir, period II. After Kaw 1989:88, fig. 7. Courtesy ASI.

Even without the buffalo sacrifice, the religion of North Indian villages, too, largely goes
back to traditions established in Early Harappan times. Buddhist and Jain texts describe in
detail early folk cults, in which male and female deities, called in Sanskrit yakṣa and yakṣī or
yakṣiṇī, occupy a central position. These deities are connected with sacred trees. For example:

Near Campā there was a sanctuary (cēïya) named Puṇṇabhadde. It was of ancient
origin … and well known. It had umbrellas, banners, and bells; it had flags … to adorn
it. . . . It had daises (veyaḍḍi) built in it, and was reverentially adorned with a coating of
dry cow-dung, and bore figures of the five-fingered hand painted in … sandal. There
was a great store of ritual pitchers. . . . It smelt pleasantly … with sweet-smelling fine
scents. … It was haunted by actors, dancers, rope-walkers, wrestlers, boxers, jesters,
jumpers, reciters, ballad-singers, story-tellers, pole-dancers, picture-showmen, pipers,
lute-players, and minstrels. … This sanctuary was encompassed round about by a great
wood … . In this wood was a broad mid-space. Therein … was a great and fine Aśoka
tree. It had its roots pure with kuśa and vikuśa grass. … Underneath this fine Aśoka
tree, somewhat close to its trunk was … a large dais of earthen blocks. … It was shaped
like a throne. (Aupapātika-sūtra, translated Barnett 1907; quoted in Coomaraswamy
1928:I,19–21)

As noted by A. K. Coomaraswamy in his important study of the yakṣa cult (1928:I,17):

the essential element of a yakṣa holystead [i.e., shrine] is a stone table or altar (veyaḍḍi,
mañco) placed beneath the tree sacred to the Yakṣa. … It was just such an altar
beneath a sacred tree that served as the Bodhisattva’s seat on the night of the Great
Enlightenment; Sujātā’s maidservant, indeed, mistakes the Bodhisattva for the tree-
spirit himself (Nidānakathā). It is very evident that the sacred tree and altar represent
a combination taken over by Buddhism from older cults, and in the case of the Bodhi-
tree we see the transference actually in process.

The texts also describe how new yakṣa shrines were established:

At Sāvatthī … lived a householder named Great-Wealth. … He was rich …, but at the

same time he was childless. One day, as he was on his way home from bathing at a
ghāṭ, he saw by the roadside a large forest tree with spreading branches. Thought he,
“This tree must be tenanted by a powerful tree-spirit.” So he caused the ground under
the tree to be cleared, the tree itself to be enclosed with a wall, and sand to be spread
within the inclosure. And having decked the tree with flags and banners, he made the
following vow: “Should I obtain a son or a daughter, I will pay you great honour.”
Having so done, he went on his way. (Dhammapada Atthakathā, translated Burlingame
1921:146)

The Mahābhārata refers to holy trees (caitya-vṛkṣa) in villages and towns (6,3,37); these trees
should not be injured as they are the abodes of gods, yakṣas, demonic spirits, and so on
(12,69,39–40). The epics and other texts often mention gandharvas and apsaras together with
the yakṣas and yakṣiṇīs as subjects of Kubera, the god of riches. In Vedic texts the gandharvas
and apsaras are said to reside in different varieties of fig trees, where their cymbals and harps
resound (AVŚ 4,37,4; TS 3,4,8,4).
Worship of sacred trees, especially the figs banyan and pipal, and the neem, continues to
be a very important part of folk religion all over India. Images of deities, including snake
deities (nāgas), are often kept beneath the sacred trees. There cannot be any doubt that this
tree worship, using the mightiest trees native to South Asia, has its origin in the Early and
Mature Harappan culture, where fig trees are important art motifs. Several Indus seals and
tablets depict an anthropomorphic deity standing inside a fig tree, sometimes with a
worshipper kneeling in front of the tree, hands raised up, as if in prayer. Trees with railings
around them are depicted in Indus tablets. A possible tree temple has been identified at
Mohenjo-daro: it has a higher floor, to which lead two flights of steps on opposite sides of the
room, which contains the remains of what may have been a railing for a sacred tree. In the fig
deity seal from Mohenjo-daro we see in front of the sacred tree with its deity a throne-shaped
altar table (Fig. 21.9). One tablet shows a cobra on such a throne beneath a sacred tree. Many
anthropomorphic deities are shown to wear a branch of a fig or some other sacred tree in
their horned headdress (Figs. 18.12, 19.5, 21.6b).
Besides the water buffalo and the fig tree, the fish is among the much repeated motifs of
Early Harappan painted pottery that continue as conspicuous symbols in the Indus civilization.
Fish constituted a significant part of the Harappan diet, but there were certainly other
reasons, too, that made the fish an important symbol. Fish breed rapidly, have a phallic shape,
and are the principal creatures inhabiting water, which is of course vital for life and
vegetation. For all these reasons, the fish has remained an auspicious symbol of fertility in
South Asia.
In Tamil, the Hindu water god Varuṇa is called mīn-ūrti, “having the fish for his mount”;
usually in Hindu iconography, the mythical water monster, the makara, is Varuṇa’s mount.
Kāma, the god of sexual love, likewise rides on a makara and has the emblem of fish in his
flag. As the most important aquatic animal, fish symbolizes water, and fish are offered waters
in the Vedic horse sacrifice (MS 3,14,2 abdhyo matsyān).
On a tablet from Mohenjo-daro, a seated anthropomorphic god with a horned crown is
flanked on either side by a fish and a long-snouted crocodile, as well as a wavy line that might
represent water or a snake (Fig. 15.3). A cylinder seal probably made in Mesopotamia by a
Harappan craftsman seems to represent the same divinity, because here, too, there is on
either side a fish and a horned snake; in addition, a pair of water buffaloes is depicted under
the throne on which this god is seated; this god wears a crown of buffalo horns topped by a
branch of a fig tree (Fig. 18.12). The water buffalo, the fish, and the crocodile are all animals
connected with water, and their presence strongly suggests that this anthropomorphic deity is
the Harappan predecessor of the Hindu water god Varuṇa. (In Vedic texts, Varuṇa is, among
other things, the god of waters, including the heavenly waters, and the master of all kinds of
aquatic animals.) The fish symbol of this Harappan water-god reminds one of Sumerian Enki,
the god of waters, who is depicted with rivers issuing from both sides of his body, with fish
swimming either in those rivers or around the god (Fig. 15.4). Enki is praised as a phallic god,
who fertilizes the soil with flood or rain water considered to be his semen.
FIGURE 15.3 An anthropomorphic deity with a crown of water buffalo horns, seated on a throne with bovine legs,
flanked on either side by a fish, a crocodile, and a wavy line that may represent a snake or water. One side of a
triangular prism from Mohenjo-daro (M-2033 B), in the Department of Eastern Art (Md 013), Ashmolean Museum,
Oxford. After CISI 3.1:404 no. 101. Photo courtesy Ashmolean Museum, Oxford.

FIGURE 15.4 The Sumerian water god Enki with rivers containing fish flowing from his shoulders or loins. (a) An
Akkadian greenstone cylinder seal from Sippar, Mesopotamia, in the British Museum (Reg. No. 1891,0509.2553, BM
89115), inscribed “Adda, scribe” (cf. Collon 1987: no. 761). Image 135126001 © The Trustees of the British
Museum. (b) Portion of an Akkadian cylinder seal from Tell Asmar (As. 32:593). After Frankfort 1955: no. 619.
Courtesy of the Oriental Institute of the University of Chicago.

The Indus seals and tablets often depict the long-snouted fish-eating crocodile of north
Indian rivers (Fig. 15.5). Known as the gharial, this crocodile is in danger of becoming extinct,
because it has been hunted for the swelling protuberance on the tip of the male gharial’s long
snout, which is thought to represent a phallus and is believed to be an aphrodisiac (Fig.
15.6a). An Indus tablet from Harappa appears to represent the gharial in the act of fertilizing
a woman by pushing its snout into her vagina (Fig. 15.6b). According to Marshall (1931:52) in
this tablet “a nude female figure is depicted upside down with legs apart and with a plant
issuing from her womb,” this being a “striking representation of the Earth Goddess with a
plant growing from her womb.” Marshall’s interpretation has been often quoted ever since,
although already Mortimer Wheeler (1968:106) thought that “what has been interpreted as a
plant … may equally well be … even a crocodile.”
FIGURE 15.5 The gharial with fish in its mouth. (a) The Indus seal M-410 (DK 8037, ASI 63.10.191) from Mohenjo-
daro. After CISI 1:98. Photo EL, courtesy ASI. (b) The reverse side of the molded tablet M-482 (E 2500, NMI 69)
from Mohenjo-daro in the collection of the National Museum of India, New Delhi. After CISI 3.1:387. Photo JL,
courtesy NMI. (c) One side (C) of the molded terracotta prism M-1429 (UPM/MD-602, Exc. Branch) from Mohenjo-
daro. After CISI 3.1:386. Photo courtesy © J. M. Kenoyer and DAMGP. (d) The reverse side of the molded tablet H-
172 (XXXII,22, ASI 63.11.144) from Harappa. After CISI 1:207. Photo EL, courtesy ASI.
FIGURE 15.6 The fish-eating river crocodile (gharial) and fertility. (a). The snout of the male gharial. Photo Bo Link
2006, Wikipedia Commons. (b) A gharial depicted as fertilizing a human female on a moulded tablet H-180 (649,
NMI 32) from Harappa in the collection of the National Museum of India, New Delhi. After CISI 3.1:398. Photo EL,
courtesy NMI.

Childless young married women are reported to have fed crocodiles “in the hope of being
blessed with offspring” (Crooke 1906:112). Before the British stopped this cruel practice in
the early nineteenth century, “women in performance of a vow used to throw a first-born son
to the crocodiles at the mouth of the Hooghly [river] in the hope that such an offering would
secure them additional offspring” (Crooke 1926: 377, based on Ward 1811). Symbolic feeding
of a baby to a crocodile continues in Bengali Hinduism. During the gājan festival of Śiva, in
some temples a crocodile is made of clay, its mouth colored red with vermilion, and a baby
made of clay is placed near its mouth (cf. Mahapatra 1972:132–133). A tablet from Dholavira
in Kutch, Gujarat, suggests that the Harappans offered human children in sacrifice to a
crocodile god.
In his essay on the Indus religion, Marshall discussed the present-day Hindu worship of
the various animals depicted on their seals and tablets by the Harappans. Stating that these
parallels “afford, of course, no proof that they were similarly worshipped five thousand years
ago,” yet considering the conservatism of India’s religious cults, he thought that “we are
justified in inferring that much of the zoolatry which characterizes Hinduism and which is
demonstrably non-Aryan, is also derived from the prehistoric age” (Marshall 1931:I,73). Yet
serious doubts have often been expressed with regard to the assumed religious continuity
from the Indus civilization to historical times. It seems possible to demonstrate, however, that
cults of Harappan origin have survived with little change even to the present day.
One such case concerns the crocodile, often depicted on Harappan “sacrificial tablets” and
some seals. A Mature Harappan potsherd from Amri in Sindh shows two crocodiles who,
instead of hind legs, have a 90° sideways projecting extension (Fig. 15.7). An explanation of
this image comes from the unique way in which the crocodile is worshipped in some fifty
villages of three tribes in southern Gujarat, documented by Fischer and Shah (1971). Images
of crocodiles, normally a couple, are made of wood and installed on wooden posts, which goes
through the middle or back part of the body (Fig. 15.8). The installation ceremony celebrates
their wedding, and the images are worshipped by smearing them with red vermilion paste and
with offerings of chicken and goats, milk or alcohol afterwards consumed by the adorants. The
male crocodile can be substituted with a liṅga-like post, or the couple by a single crocodile
with a head at either end. This now rapidly declining crocodile cult is both communal and
individual: people ask the crocodile gods for cows to give milk and calves, for good crops, for
offspring, and for help against illness and sorcery, making a vow to install and worship the god
if the boon is granted (Parpola 2011a).

FIGURE 15.7 Two crocodiles set on poles. A Mature Harappan painted pot from Amri, southern Indus Valley. After
Casal 1964: II, fig. 75 no. 323. Courtesy Mission Archéologique de l’Indus, CNRS/Musée Guimet, Paris.

FIGURE 15.8 The Devlimadi sanctuary of crocodile gods in southern Gujarat. After Fischer and Shah 1971: pl. 2.
Photo courtesy © Eberhard Fischer 1970.

The peacock is yet one important motif of Early, Mature, and Late Harappan painted
pottery. One clue to its meaning for the Indus people is given by the funeral urns of the Late
Harappan Cemetery H at Harappa, where the bird is seen carrying dead people to the stars
(Fig. 15.9). At the funerals of the Dravidian-speaking Maria Gonds of Bastar in central India,
wooden posts with rude representations of peacocks are set up next to the grave. In classical
Hinduism, the peacock is the vehicle of Skanda, the successor of Vedic Rudra and the “son” of
Śiva, who in south India is connected with a snake cult. Śiva is called Nīlakaṇṭha, “blue-
necked,” because he drank the deadly poison created at the churning of the milk ocean, which
threatened to destroy the world. The Sanskrit word nīlakaṇṭha also denotes the peacock,
which is a blue-necked bird. Possibly the color of the peacock’s neck was imagined to result
from the peacock’s habit of eating poisonous snakes.
FIGURE 15.9 Peacocks carrying dead people to the stars. Painting on a funerary urn of the Late Harappan Cemetery
H at Harappa. After Piggott 1950:234, fig. 29.

This chapter has dealt with some important religious conceptions and practices in present-
day folk Hinduism that with great likelihood can be traced back to Harappan times and even
earlier. Most of them are connected with outstanding aspects of South Asian nature, notably
its fauna and flora. They all attest to a remarkable religious continuity over millennia. But
there is no doubt that the Harappan religion also included the priestly elaborations of the
elite.
16
Astronomy, Time-Reckoning, and Cosmology

The calendrical asterisms of the Vedas were not brought to South Asia by the Rigvedic Indo-
Aryans, for they are not mentioned in the family books and have no counterpart in Iranian or
other Indo-European traditions. Some nakṣatras are mentioned in the late Rigvedic “marriage”
hymn 10,85, but not until the Atharvaveda and Yajurveda are all twenty-seven or twenty-eight
nakṣatras listed for the first time. The Yajurvedic texts connect the nakṣatras and the full and
new moon with specific “bricks” laid down in brick-built fire altars (agni-citi), which are
unknown to the Rigveda and belong to the archaic layer of Vedic rituals connected with the
Atharvavedic tradition (chapter 12). The relatively late Vedic Śulvasūtras describe the
elaborate geometry and rules of orientation used in the construction of these altars. This
Śulvasūtra tradition probably goes back to the Indus people, who lived for a millennium in
brick-built cities and needed solar time-reckoning, as did other agriculturally based riverine
and urban early cultures, for example, in Mesopotamia and Egypt, in which astronomy and
astrology invariably form an important part of religion.
The planets cannot have failed to draw the attention of early astronomers, in whatever
civilization, because they differ from other heavenly bodies through their independent
movement and sometimes belong to the brightest phenomena of the night sky. Venus was the
symbol of the Great Goddess as early as the fourth millennium BCE in Mesopotamia; in China
astronomers were observing planetary movements by 1576 BCE; and in Old Tamil texts, there
are native Dravidian names for the planets, some of which can be read in the Indus texts (see
chapter 21). So it is surprising to come across no reference at all to the planets in the
Rigveda. This might be taken as further proof that the Rigvedic Aryans did not create the
Vedic star calendar, although it is at least conceivable that the Vedic Brahmins who descended
from the Rigvedic Aryans actually knew about the planets but chose to avoid referring to them
because they had a significant role in the pre-Rigvedic religion of South Asia.
Indeed, certain proper names in the Vedic literature reveal that the planets were not
unknown to Vedic Aryans. The clearest example connects the planets with the Atharvavedic
tradition of the Vrātyas. According to Pañcaviṁśa-Brāhmaṇa (24,18), the daiva vrātyas (who
were adherents of “the God,” i.e., Rudra), had Budha, the son of Soma, as their leader. Budha
means “wise” and is the later Sanskrit name of the planet Mercury, while Soma denotes not
only the sacred drink of Vedic Aryans but also the moon. One of the best-known astral myths
of classical India, told in many Purāṇas and the Mahābhārata, concerns the birth of the planet
Mercury. Soma the moon abducts Tārā (“the star”), lawful wife of Bṛhaspati, the planet Jupiter,
and produces a splendid son, Budha. Willibald Kirfel (1952) speculates that this myth may
have come from Mesopotamia, and that Tārā might be the bright star Spica in Virgo. (I
propose a different explanation later in this chapter.)
Among the ritualistic fossils of the Veda, also connected with the Vrātyas, is the mahāvrata
day, discussed in chapter 12 (and again in chapter 19). The mahāvrata concludes year-long
rites and celebrates one of the turning-points of the year. A fire altar piled with bricks, some of
which represent the nakṣatras, is a necessary part of the rite; when completed, it is praised
with special songs on the mahāvrata day. The god Rudra, who is otherwise ignored in Vedic
soma sacrifices, is identified with the fire-god Agni and receives on the completed altar an
elaborate burnt offering of hundreds of libations of liquid butter (śatarudriya). The fire altar is
an image of the creator god Prajāpati, whose body is said to consist of both the cosmos (space)
and the year (time). There are various types of fire altar, but one common type has five layers
and 10,800 bricks. The year is reckoned to comprise 360 days-and-nights, each having thirty
muhūrtas (“hours” of forty-eight minutes), so that one year has 10,800 muhūrtas.
The thirty muhūrtas are based on the parallelism of three principal time cycles, each split
into luminous and dark halves in the Veda: (1) a (twenty-four-hour) nychthemeron of thirty
muhūrtas comprising a day of fifteen muhūrtas and a night of fifteen muhūrtas; (2) a month of
thirty days comprising the “white half” of the waxing moon (fifteen days) and the “black half”
of the waning moon (fifteen days); and (3) a year of twelve months (twenty-four half-months),
comprising the auspicious “northern course” of the sun (six months or twelve half-months),
and its ominous “southern course” (again six months or twelve half-months). In Yajurvedic
ritual, it was of the greatest importance to make a sacrifice to accompany every transition
from one half to the other in these three time cycles.
The idea of a year of 360 days, attested in the late riddle hymn of the Rigveda (1,164,11)
arises naturally from twelve solar months of approximately thirty days. Of course, this round
number 360 falls short of the actual length of a solar year, so the difference had to be
regulated with an intercalary thirteenth month (implied in the late Rigvedic hymn 1,25,8
ascribed to Śunaḥśepa, who figures later in this chapter); this extra month was added once
during every five-year cycle. Five times 360 plus 30 equals 1830 days, and 1830 divided by
five yields 366 days, which is less than a day longer than the actual solar year.
Ancient astronomers determined the seasons by observing the position of the sun among
the stars on its heavenly path known as the ecliptic. In Egypt, and in West Asia until about
1100 BCE, heliacal risings of stars were observed just before sunrise, in order to discover
which asterism was closest to the rising sun. However, this method leaves little time for
observation and the star is difficult to spot, on account of the brightness of the sun and
atmospheric disturbances on the horizon. Chinese and Indian astronomers avoided this
difficulty by adopting a different method, based on the fact that the full moon is exactly
opposite to the sun. Asterisms were chosen in pairs with the two members exactly opposite
each other (Fig. 16.1). From the conjunction of the full moon with a given asterism the
astronomer knew that the sun was in conjunction with the opposite asterism. Even relatively
small stars were selected in preference to more luminous ones. The Chinese and Indian
calendars are therefore luni-solar; a purely lunar calendar based on just the observation of the
moon’s phases is out of step with the seasons, as exemplified by the movement of the Muslim
month of Ramadan around the solar year.

FIGURE 16.1 The nakṣatras as pairs of opposing asterisms. A sketch by Jean Filliozat. After Filliozat 1962:350, fig.
3. Courtesy Pierre-Sylvain Filliozat.

The relationship between the Indian, Chinese, and Mesopotamian calendars has been
much debated. Recent archaeological discoveries in China strongly suggest that the Chinese
and Indian calendars developed independently of each other, despite their similarity. It just so
happened that they evolved in much the same way. A Mesopotamian origin for the Chinese
and Indian calendars, suspected by Joseph Needham (1959), is unlikely, because the
opposition of the sun and the full moon was utilized only at a late phase of Mesopotamian
astronomy, around 1000 BCE, and was not the basis of the regular calendar.
The opposite asterisms of the luni-solar calendar are never simultaneously visible;
however, the position of the invisible asterism below the horizon could be figured out from the
position of the pole star and the circumpolar stars, especially the seven bright stars in the Big
Dipper of Ursa Major, which never rise or set. Chinese astronomers are known to have
observed the circumpolar stars systematically. The rotation of Ursa Major functioned as a
celestial clock marking the hours of the night as well as the seasons. The circumpolar stars
play a dominant role in Chinese astronomy and cosmology. In addition to the four “palaces”
defined by the equinoctial and solstitial points, the Chinese distinguished a fifth, central
“palace,” which was the celestial archetype cosmically empowering the Chinese emperor.
Around 500 BCE, Confucius equated the emperor with the pole star, as follows: “The Master
said: To conduct the government by virtue may be compared to the Northern Asterism: it
occupies its place, while all other stars revolve around it.” (Lunyu 2,1, translated Pankenier
2004:288). The Chinese concept of this particular astral-terrestrial correspondence goes back
to the Bronze Age or even the Neolithic. The heavenly prototype (of the center ruling the four
quarters) was used for political legitimation. Ever since the earliest dynastic state Xia (from
2000 BCE), palatial structures and royal tombs are quadrilateral and cardinally oriented.
The usual presumption is that a calendar was created to fulfil the needs of agriculturalists.
“Farmers generally find the signs of nature to be a safer indication of the season than a civil
[official] calendar. The latter is probably an invention that followed upon urbanization, not
least as a tool of social control” (David Pankenier, personal communication 2012).
In the Indus Valley, urbanization began with the Early Harappan Kot Diji culture (3200–
2600 BCE). One of its first towns, Rahman Dheri in the northwest, has streets and buildings
oriented according to the cardinal directions (Fig. 16.2). A grid pattern enables easy access
for ox-carts and an instant allotment of house plots to a large number of settlers arriving
simultaneously. Such a town plan implies a strong regulating authority, as does the use of a
standardized brick for construction in the Indus cities. The preplanning of level streets would
have been necessary for the functioning of the elaborate drainage system of Mohenjo-daro in
the Mature Harappan period (Ratnagar 1991).

FIGURE 16.2 The Early Harappan town of Rahman Dheri from the air. After Durrani 1988:210, pl. 6. Courtesy
Directorate of Archaeology and Museums, Government of Khyber Pakhtunkhwa, Peshawar, Pakistan.

The spread of the Kot Diji culture all over the Indus Valley was accompanied by the spread
of a new type of stamp seal—an important instrument of administration—which continued to
exist in the Harappan period (Uesugi 2011). The basic design of these Kot Diji seals consists of
“concentric circles” (one to three circles centered on a dot): usually four sets placed in the
four corners of a square or cross-shaped seal but with a common variant consisting of a fifth
set of concentric circles in the center of the seal (Fig. 16.3). The seals seem to reflect the
importance of the four cardinal directions and the center in the cosmology and political
ideology of the Early Harappans. “Rays” surrounding the concentric circles in one seal
suggest solar symbolism (Fig. 16.4). In a painted bowl from Mehrgarh VI–VII (c. 3200–2600
BCE) the field is divided into four squares occupied by sun-like circular images surrounded by
“rays” (Fig. 16.5). This probably depicts the annual course of the sun, divided into four
quadrants by the equinoctial and solstitial points, which also define the four cardinal
directions.

FIGURE 16.3 Kot Diji type seals. (a) Trq -2, and (b) Trq-3 from Taraqai Qila. After CISI 2:414. Photo (a) JL and (b) J.
R. Knox, courtesy Directorate of Archaeology and Museums, Government of Khyber Pakhtunkhwa, Peshawar,
Pakistan. (c) H-638 (725, HM 13377) from Harappa. After CISI 2:304. Photo S. M. Ilyas. Courtesy DAMGP. (d) H-
1535 (H00-4495, 9597-0001) from Harappa. After CISI 3.1:211. Photo Richard Meadow, courtesy Harappa
Archaeological Research Project.

FIGURE 16.4 A seal from Rahman Dheri with the motif of “rays around concentric circles.” After Durrani, Ali, &
Erdosy 1994–19951:207. Courtesy Directorate of Archaeology and Museums, Government of Khyber Pakhtunkhwa,
Peshawar, Pakistan.
FIGURE 16.5 “The sun in the four quadrants” painted on a Faiz Mohammad style gray ware bowl from Mehrgarh,
period VI (c. 3200–2900 BCE). After C. Jarrige et al. 1995:160. Courtesy Mission Archéologique de l’Indus,
CNRS/Musée Guimet, Paris.

The Rigvedic and Atharvavedic hymns often refer to four directions of space, which are
undoubtedly the four main directions, for example, RV 7,72,5. The occasionally mentioned fifth
direction is the center. In the Yajurvedic Saṁhitās and Brāhmaṇas, the fifth point is the zenith;
it is consistently associated with Bṛhaspati, the Vedic predecessor of the Hindu god Brahmā,
who in classical Sanskrit texts occupies the center (Wessels-Mevissen 2001). One Vedic ritual
connected with the directions of space is crucially important for understanding their
ideological significance. This is the “mounting of the regions” during the royal consecration.
The king first dons the tārpya garment, the royal robe of the divine king Varuṇa (which is
likely to derive from the trefoil-ornamented “sky garment” of the Harappan “priest-king,”
modeled on Mesopotamian prototypes, as discussed in chapter 18). Then the king makes five
steps, one after the other, in each of the five directions, thereby ascending to the zenith: “from
the quarters he goes to the heaven,” for “the heaven is the quarters of space” (MS 4,4,4). As
the Śatapatha-Brāhmaṇa (5,4,1,8) explains: “It is the seasons, the year, that he [the adhvaryu
priest] thereby makes him [the king] ascend; and having ascended the seasons, the year, he is
high, high above everything here.” In a parallel consecration ritual, in the vājapeya sacrifice,
also known as the Bṛhaspati-sava, the ascent to the zenith is more concrete: the king dons the
tārpya garment and with the help of a ladder ascends the sacrificial post; reaching the top he
declares: “We have reached the sun/heaven, we have become immortal.” He then descends
and seats himself on the throne placed at the foot of the pillar.
In the Mahābhārata, Yudhiṣṭhira, the eldest of the five Pāṇḍava brothers, aspires to perform
the royal consecration (rājasūya). Since only a king of the whole world is entitled to perform
this, Yudhiṣṭhira sends his four younger brothers to conquer the four cardinal directions. In
such a digvijaya, a “conquest of the directions,” the victorious king has the sun as his role
model: it rises in the east, creating light and expelling darkness, and goes through all the
regions (cf. MS 4,14,14; ŚB 10,3,5,3). The sun is accordingly known as “four-cornered”: the
directions are his corners (ŚB 14,3,1,17). The sun defines these four directions through its
daily as well as its annual course through the equinoxes and solstices (cf. JB 2,26). It is the sun
that the gods anoint on the royal throne, and the person who knows this, when he sits on the
royal throne, becomes the sun (cf. JB 2,25–26).
On the basis of the Kot Diji seals, I suggest that the connection of the ruler with the sun
and the center of the four directions defined by the sun’s daily and annual course, important
in Vedic and epic royal ideology, may have originated in the Early Harappan culture. Further
evidence that it prevailed in the Indus civilization comes from the so-called Proto-Śiva seal of
Mohenjo-daro, in which an anthropomorphic figure wearing the horns of a water buffalo is
seated on a throne (Fig. 16.6). The throne is a major symbol of royal authority in Vedic culture
(cf. ŚB 12,8,3,4; TB 1,3,9,2). John Marshall (1931a:52–55), who coined the name “Proto-Śiva,”
suggested that the figure is three-faced, like many later Hindu images of Śiva. I prefer the
alternative idea that the figure has four faces, like Śiva in the shape of caturmukha-liṅga,
symbolizing the axis mundi or the Hindu god Brahmā, whom Indian architectural texts
connect with the center. The fourth face is hidden, because it faces away from the viewer.
FIGURE 16.6 The four-faced “Proto-Śiva” seated on a throne and surrounded by four animals on the seal M-304 (DK
5175, NMI 143) from Mohenjo-daro in the collection of the National Museum of India, New Delhi. After CISI 1:382.
Photo EL, courtesy NMI.

I further suggest that the four faces looking in four directions are in all likelihood related
to the four male animals depicted on either side of “Proto-Śiva,” arranged so as to form a
rectangle: elephant above tiger on one side, rhinoceros above water buffalo on the other. (The
fact that there are definitely four animals, rather than three, supports the hypothesis of a four-
faced, rather than a three-faced, figure.) Marshall himself suggested their connection with the
four directions. Four animals are associated with the four heavenly palaces in China: blue
dragon with east, red bird with south, white tiger with west, and black tortoise with north.
Among the principal ways to find due east is to observe the points where the sun rises
during the year. The same method may be applied with other visible stars. Holger Wanzke
(1987) studied the axes of the buildings and streets in Mohenjo-daro and found that they
diverge one to two degrees clockwise from the cardinal directions, which would match the
setting of the star Aldebaran in the west against the horizon of the Kirthar Mountains.
Although no likely astronomical observatory has been identified in Mohenjo-daro, Mayank
Vahia and Srikumar Menon (2011), two Indian astronomers, have suggested that a peculiar
circular stone structure near the acropolis of Dholavira in Kutch, Gujarat, may have been an
astronomical observatory.
Another way to find out the directions of space is by means of a sundial or gnomon, the
oldest astronomical instrument, comprising a straight peg orthogonal to a level base. By
recording the length and the direction of the peg’s shadow every day of the year one can
define the hours of the day, and from measurements made at noon discover the solstices and
equinoxes. The Kātyāyana-Śulvasūtra (1,2–3) gives the rules for finding the cardinal directions
as follows:

Having fixed a peg on level ground and having drawn a circle around it by means of a
rope that has the same length as the peg (and is attached to it), he fixes two pegs at the
two points where the shadow of the tip of the peg falls on the line (of the circle in the
fore- and afternoon); that (line joining these two new pegs) (is) the east(-west line).
After adding to the rope its length and making two loops (at its either end), he fixes the
loops at the two pegs marking the east-west line. Stretching the rope (draw a circle
around each of the pegs) and fix a peg south and north in the middle (area where these
two large circles meet). That (line joining these new pegs) is the north(-south) line.
In this case, I propose that ten stone pedestals excavated at Mohenjo-daro may have been
made for wooden gnomons, rather than being the bases of liṅgas (cult images of Śiva’s
phallus), as suggested by Ernest Mackay (1938). One stand made of dark red stone is
decorated with the trefoil motif that almost certainly had an astral significance in both the
Indus Valley and Mesopotamia (Fig. 16.7). Great care was taken to make the bottom of the
base flat and to secure the stability and uprightness of the shaft fixed in the round depression
in the middle with two dowels (the dowel holes survive). For a gnomon the ground and the
base must be as level as possible, and the peg as straight, stable, and orthogonal as possible.
The Kātyāyana-Śulvasūtra (7,4) states that nothing is more level than the surface of still water,
and that the peg should be made of the particularly stable core part of old, hard-wooded
acacia tree.

FIGURE 16.7 A finely polished pedestal made of dark red stone and decorated with trefoil figures, probably meant
for a gnomon (sun-stick). Excavated in Mohenjo-daro (DK 4480, cf. Mackay 1938:I,412 and II, pl. 107.35), in the
National Museum of Pakistan, Karachi. After Parpola 1985: fig. 9a. Photo Asko Parpola 1971, courtesy DAMGP.

The Baudhāyana-Śulvasūtra (1,22–28) describes a method to construct an oriented square

and thus to define the cardinal and intermediate directions by using a gnomon and a cord with
marked midpoint. This produces a pattern of “intersecting circles,” which is an important
motif on Early and Mature Harappan painted pottery, and also a favorite motif on Harappan
bathroom floors and “bath tubs.” I suspect this motif relates to the necessity of leveling the
ground for the gnomon by means of water. In later Indian astronomical texts water is
recommended for making ground perfectly level.
It is of the greatest importance to note that the first calendrical asterism in the Vedic
nakṣatra lists is the Pleiades. We can fully appreciate this fact only by noting carefully how the
Vedic texts comment on the position of the Pleiades. As is well known, the priestly Brahmins
established themselves as the highest social class of Vedic society; the Kṣatriyas or Rājanyas,
warriors and nobility, were the second social class; and the Vaiśyas, the common people, were
the third class. The oldest Yajurvedic texts prescribe that a Brahmin should establish his
sacred fires in the spring, the first of the seasons, a Rājanya in the summer, and a Vaiśya in
the autumn. A Brahmin should establish his fires under the Pleiades (kṛttikāḥ), the first among
the calendrical asterisms, for the Pleiades belong to the fire-god Agni, the “foremost” of the
gods (since he is always mentioned first, being the purohita, literally “set in front”), and the
Brahmin, a member of the priestly class, belongs to Agni, who is the priest of the gods. The
Pleiades are said to be seven, and since there are seven “breaths” in the head, the Pleiades
are regarded as the head of the creator god Prajāpati, while Agni, who eats the offerings, is
his mouth (mukham, “mouth, beginning”). The Śatapatha-Brāhmaṇa (2,1,2,1–5) adds that
originally the Pleiades were the wives of the Seven Sages, but are now precluded from
intercourse with their husbands, as the Seven Sages rise in the north, the Pleiades in the east.
So now the Pleiades have Agni as their mate. In the early Vedic texts, Agni is the god of the
east, the direction of the rising sun.
In this context, the Śatapatha-Brāhmaṇa (2,1,2,3) states that the Pleiades “do not move
away from the eastern quarter, whilst the other asterisms do move from the eastern quarter.”
Jean Filliozat (1962) interpreted this to mean that while the east is defined by the sun’s rising
above the horizon, the east par excellence is defined by the sun’s rising in the east at the
vernal equinox, since at this point of time the sun rises into the northern hemisphere, too,
beginning its so-called “northern course.”
In the same context, the Śatapatha-Brāhmaṇa (2,1,2,4) also states: “the Seven Sages
(ṛṣayaḥ) were in former times called Bears (ṛkṣāḥ).” Sanskrit ṛkṣa- means both “bear” and
“star”; it occurs in Rigveda 1,24,10, which asks: “In the daytime, where have those bears/stars
gone that are seen in the night sky as fixed high up?” Since this verse belongs to the
Śunaḥśepa hymn (to be discussed below), it is fairly certain that the stars of Ursa Major are
intended. In Greek antiquity the stars of Ursa Major were conceived as a mother bear (Greek
árktos) followed by three bear cubs.
From the time of the Atharvaveda, however, Ursa Major was almost exclusively called
sapta rṣayaḥ, “Seven Sages.” The choice of the word ṛṣi-, “sage, seer, holy man, rishi” (a word
the Indo-Aryans probably adopted from the original language of the BMAC, see chapter 8) was
influenced by the inherited Graeco-Aryan term ṛkṣa-, “bear.” But the semantic content of the
new name undoubtedly goes back to conceptions that prevailed in pre-Vedic India. Harappan
parentage, ultimately going back to Mesopotamia, may be involved here, for the Sumerians
also spoke of “Seven Sages” (Sumerian abgal). They were supposed to be kings who ruled
before the Flood, and servants of Enki, the god of water and wisdom, represented in Sumerian
art as half-man half-fish. The Indian “Seven Sages” are also connected with a Flood, for
according to the Mahābhārata (3,185,1–54), they were on an ark with Manu, the first man and
ancestor of the human race. This conception may be due to the fact that the seven stars of
Ursa Major being near the pole of heaven, never go beneath the horizon. Homer says that
Ursa Major “circles ever in her place, and alone has no part in the baths of Ocean” (Iliad
18,489).
Late Harappan people were much more numerous than the first Indo-Aryan-speaking
immigrants. The fusion of these two population groups probably started right away with mixed
marriages and growing bilingualism. The newcomers needed middlemen to deal with the local
population, and this gave the local leaders and priests a chance to obtain important positions
in the emerging new social order.
The Seven Sages are considered to be the ancestors of the Brahmanical clans since the
Rigveda; they are probably the same as the “seven sacrificers” of yore (sapta hotrāḥ), with
whom Manu performed the first sacrifice to the gods (cf. RV 10,63,7). The Mahābhārata (3,43)
says the ancient saints shine with a light of their own, acquired by their merits, standing
ablaze on their own hearths (dhiṣṇya): these lights are seen as the stars looking from the earth
below, like tiny oil-lamp flames because of the distance. This conception can be traced to early
Vedic texts. While dealing with the building of the fire altar, the Taittirīya-Saṁhitā (5,4,1,3–4)
states: “He puts down the constellation bricks; these are the lights of the sky; verily he wins
them, the Nakṣatras are the lights of the doers of good deeds.” The Maitrāyaṇī Saṃhitā (1,8,6)
is most explicit in connecting the stars with ancient sacrificers. In the Vedic ritual, the word
dhiṣṇ(i)ya in the strict sense denotes the fireplaces of seven priests officiating in a soma
sacrifice; six of them are built in a row in the sitting hall, one is outside the hall on the border
of the sacrificial area. A very similar row of seven Indus Valley fireplaces has been excavated
on a ceremonial brick platform in the acropolis of Kalibangan.
The Śatapatha-Brāhmaṇa’s account of the divorce of the Seven Sages and the Pleiades is
the earliest version of the famous Hindu myth of the birth of the war-god Skanda, whose
metronym Kārttikeya connects him with the Pleiades (kṛttikāḥ). According to the fuller epic
and Purāṇic versions, Agni or Śiva seduced the Pleiades in the absence of their husbands, the
Seven Sages. In one variant, the Seven Sages cursed Śiva to lose his fiery phallus, which
dropped down and started burning the world, and did not stop until it was placed in the
female member—the yoni base upon which Śiva’s cult image (symbolizing an erect phallus) is
installed. Thus the origins of liṅga worship come from this myth. Śiva’s four-faced liṅga
represents the axis mundi, and in the earliest Vedic texts, the rising sun (representing the
mate of the Pleiades, Agni or Śiva) is described as a pillar which separates heaven and earth
and props up the sky. (On the burning pillar or tree symbolizing the phallus, Kāma, or the
husband of the goddess, and its connection with the spring festival, see chapter 15.)
The Vedic nakṣatra lists start with the asterism kṛttikāḥ because the heliacal rising of the
Pleiades at the vernal equinox started the new year. The Pleiades were closest to the
equinoctial point in about 2240 BCE. Through the equinoctial precession, the position of the
asterism marking the vernal equinox with its heliacal rising slowly changed. Therefore, by 80
CE Indian astronomers had revised the nakṣatra list to make it start with Aśvinī, the
constellation closest to the equinoctial point between 655 BCE and 300 CE. A similar
calendrical adjustment was apparently made by Indus astronomers at the peak of the Mature
Harappan period, for there are good reasons to suspect, with Albrecht Weber (1861
[1862]:275–277), that it was originally Rohiṇī, the second nakṣatra in the old list, that
launched the calendar. Rohiṇī, “the red (female),” is Aldebaran, the large red star Alpha Tauri,
which was closest to the equinoctial point in 3054 BCE.
According to the Brāhmaṇas, the creator god Prajāpati gave his daughters, the nakṣatras,
in marriage to the moon. However, the moon neglected all his other astral wives and
cohabited only with Rohiṇī, his favorite. The other wives in anger returned to their father, who
severely reprimanded his son-in-law. The moon promised to treat all his wives equally but still
continued to cohabit only with Rohiṇī; so he was inflicted with an illness that makes him wane.
Both the Mahābhārata (3,219,10) and the Rāmāyaṇa (5,31,5) speak of the time when “Rohiṇī
was the first of the stars.” In the Atharvaveda (AVŚ 13,1,22), Rohiṇī is actually said to be the
devoted wife of Rohita, the “Red,” that is, the rising sun. The Jaiminīya-Upaniṣad-Brāhmaṇa
(4,27) mentions Savitṛ, the “instigator” god, that is, the rising sun, and his mate, the “solar
maiden” Sāvitrī, as the archetypal couple. But in the marriage hymn of the Rigveda (10,85),
Soma, the moon, is the bride of Sāvitrī. These references suggest to me that it was the heliacal
rising of Rohiṇī, together with the sun, that originally marked the beginning of the new year in
India. The old astral myth, known already from Rigveda 10,109, in which Soma abducts (and
eventually returns) the wife of Bṛhaspati, therefore originally seems to refer to the changeover
from the solar to the luni-solar calendar. Bṛhaspati, who is later the planet Jupiter, is here the
rising sun, and Tārā, the “star” par excellence, was originally Rohiṇī. The Vedic texts have
preserved an obsolete word for “calendrical asterism,” b(h)ekuri; its Dravidian original,
discussed in chapter 14, suggests that it was an asterism observed at dawn, that is, when it
rose heliacally with the sun.
To begin with, there were only twenty-four nakṣatras, corresponding to the twenty-four
half-months of the solar year, with twenty-four separate names. Three more were later added
by doubling three nakṣatras and adding the qualifications “former” and “latter” to the original
three names. This was to provide the moon with twenty-seven mansions, one for each night of
its sidereal revolution (the period it takes the moon to return to a given point in its orbit:
twenty-seven days, seven hours, forty-three minutes). Thus it became possible to refer to each
day of the month: according to Pāṇini (4,2,4), the statement “today is Puṣya” meant that the
moon was in the Puṣya asterism.
In one version of the earliest nakṣatra list (Taittirīya-Saṃhitā 4,4,10) there is another star
with same name, Rohiṇī, exactly opposite to Aldebaran. This is the large red star Alpha
Scorpii. Weber (1861 [1862]) proposed that the two stars were both called Rohiṇī because
originally each of them marked the beginning of a half-year period at the two equinoxes.
Maybe at that time these were the only calendrical stars in use. In most versions of the old
nakṣatra list, the second Rohiṇī (Alpha Scorpii) is called Jyeṣṭhā. This word means “the eldest
(lady),” and the asterism is connected with the goddess of misfortune, the elder sister of
Lakṣmī, the goddess of fortune. However, Jyeṣṭhā is likely to abbreviate jyeṣṭhaghnī, “killer of
the eldest (son),” a word mentioned in the Atharvaveda as the name of an ominous
constellation. This name recalls the sacrifice of a first-born child mentioned in the Śunaḥśepa
legend, the recitation of which was an integral part of the royal consecration, the rite
mentioned above as involving such components of probable Harappan origin as the donning of
the tārpya garment and ascending the regions of space.
The Śunaḥśepa legend tells of King Hariścandra, who had a thousand wives but no son. He
was advised to resort to Varuṇa. The god granted him a son, but only on condition that
Hariścandra would sacrifice the child to Varuṇa. The boy, named Rohita, was demanded by
Varuṇa several times, but each time Hariścandra postponed the sacrifice on various pretexts.
When Rohita turned sixteen—the coming of age of a warrior youth—the father told the son
that he would redeem his promise to Varuṇa. Rohita, however, took his bow and arrows and
disappeared into the forest. His father was punished by Varuṇa with dropsy (an illness
supposed to make the moon swell). Eventually Rohita purchased a surrogate victim, a
Brahmin boy called Śunaḥśepa, whose father was prepared to slaughter his son for money.
Bound to the sacrificial stake, Śunaḥśepa prayed to Varuṇa and other gods, who set him free.
At the same time, Hariścandra was healed of his illness.
In chapter 15 we saw that Varuṇa was the successor to the Harappan god of water and
fertility, which had the crocodile (Varuṇa’s mount in Hinduism) among his symbols. The
crocodile’s long snout with a swelling protuberance at its tip represents a phallus. Crocodiles
are believed to grant babies, and Hindus are known to have sacrificed their first-born child to
crocodiles in the hope of further offspring. Now the Taittirīya-Āraṇyaka (2,19) speaks of a
divyaḥ śākvaraḥ śiśumāraḥ, “a mighty divine crocodile” with a tail (puccha) of four sections.
Because fore and hind legs are also mentioned among this creature’s body parts represented
by various stars, śiśumāra- cannot here have the alternative meaning “dolphin,” but must
denote “crocodile.” Literally, śiśumāra- means “baby-killer.”
The context of this heavenly crocodile is the worship of Brahma, which a Vedic
householder is supposed to perform by uttering a prayer at dusk while facing the region of the
pole star, Dhruva. This prayer begins with “dhruvas tvam asi, dhruvasya kṣitam” (“You are
firm, foundation of the firm”), and ends with “obeisance to Śiśukumāra” (perhaps a corrupt
reading for Śiśumāra). The oldest parts of the cosmographic descriptions of the Purāṇas tell
that the god Viṣṇu appears in the sky in the shape of a crocodile consisting of stars, and that
the pole star is in the tail of the crocodile. Thus the pole star occupies in the sky a position
comparable to that of the pole in the back part of the body of the cultic crocodile images in
Gujarati tribal villages and in the Harappan painted pot from Amri discussed in chapter 15
(Figs. 15.7, 15.8). In the sky the heavenly crocodile turns around with the pole star as its
pivot, in the same way as tribal crocodile images. Moreover, on the Indus tablets depicting a
procession of animals, the crocodile appears not in the lower register with the other animals
but in the upper register corresponding to the sky.
In the Śunaḥśepa legend, Hariścandra’s name means “Yellow Moon,” and his thousand
wives apparently stand for stars. His son Rohita has a name used of the rising sun in the
Atharvaveda. Weber (1853:237) suggested that the name Śunaḥśepa has an astral denotation.
It means literally “dog’s tail,” and corresponds exactly to Greek kunósoura and its Latin
translation canis cauda, “dog’s tail,” which describes the asterism Ursa Minor (actually
kunosourìs árktos, “female bear with a dog’s tail”) in the Greek and Roman tradition.
Śunaḥśepa is thus an astral term of Indo-European origin for a circumpolar constellation,
which moreover has the concept of “tail” in it. Śunaḥśepa is therefore the most natural
inherited term that the Indo-Aryans had for what seems to have been the Harappan
conception of Ursa Minor—the tail of a celestial crocodile, with the pole star in that tail. The
heavenly crocodile itself might be compared to the Draco constellation (Fig. 16.8).

FIGURE 16.8 The circumpolar stars. After Parpola 1994a:243, fig. 14.4, based on Liebert 1968 (1969):168.

The Gṛhyasūtras prescribe that the groom should show the bride the pole star as a
paragon of faithfulness. In the old khila verse interpolated as the last stanza of the marriage
hymn Rigveda 10,85, the groom says: “be faithful!” (dhruvaidhi). In 4000–1900 BCE the pole
star was Thuban (Alpha Draconis). In 2780 BC, Thuban was only 0.6° distant from the heavenly
pole. Thuban is the only star that could really be called the “fixed” (dhruva) center of the
rotating heavens before our own Polaris (Fig. 16.8). Therefore, the conception of the pole star
as “fixed” must date from a much earlier time than the Gṛhyasūtras, that is, from the time of
the Indus civilization.
The Hiraṇyakeśi-Gṛhyasūtra (1,22–23) prescribes that the groom should actually address
the pole star. The contents of the groom’s mantra suggest that it was originally the king who
spoke this mantra, probably in the royal consecration, known as “Varuṇa’s sacrifice”:
I know thee as the nave [hub] of the universe. May I become the nave of this country.
I know thee as the center of the universe. May I become the center of this country.
I know thee as the string that holds the universe. May I become the string that holds this
country.
I know thee as the pillar of the universe. May I become the pillar of this country.
I know thee as the navel of the universe. May I become the navel of this country …

Varuṇa is the divine king, and in the mantra the pole star is a symbol of royalty, just as in
imperial China.
In the Old Tamil epic Cilappatikāram, the groom points out the pole star, known as vaṭa-
mīn, to the bride in the wedding ceremony, and this is still the custom in south India and Sri
Lanka. The phrase actually means “north star” and can be read in the Indus inscriptions, as
shown in chapter 21. A homonym of Dravidian vaṭa, “north,” is vaṭam, “banyan fig,” the mighty
tree with rope-like air-roots from which it takes its name (Dravidian vaṭam means “rope”) (Fig.
16.9). This original Dravidian appellation of the pole star with its homonyms opens up a new
way to comprehend some fundamental conceptions of ancient Indian cosmology and kingship.

FIGURE 16.9 The banyan tree with its hanging air roots. After Parpola 1994a:242, fig. 14.2. Photo Asko Parpola.

Immediately after speaking of the starry crocodile with the pole star in its tail, the Purāṇas
describe the crucial function of the pole star as the hub and upholder of the entire stellar
system: “As the pole star revolves, it causes the moon, sun and other planets to turn also, and
the lunar asterisms follow that revolving (pole star) in the manner of a wheel (turning around
the nave). The sun and the moon, the stars, the lunar asterisms along with the planets are all
in fact bound to the pole star by cords consisting of an array of winds.” (Viṣṇu-Purāṇa 2,9,2–3)
A Rigvedic hymn, in which reference is made to Śunaḥśepa, speaks of a banyan tree in the
sky: “King Varuṇa holds up the crown of the (heavenly banyan) tree in infinite space; high up
is the base of its (aerial roots) which hang down: may these beams of light be fixed on us!” (RV
1,24,7). The Vedic and Hindu texts repeatedly refer to a heavenly fig tree. This conception
seems to be pictured on an Indus tablet (H-179), which depicts an anthropomorphic deity
inside a fig tree: at bottom, the fig tree is flanked on either side by a star, which suggests a
heavenly connection for the tree (Fig. 16.10). All principal varieties of fig trees are associated
with kingship in the Aitareya-Brāhmaṇa (7,27–34); the mightiest of them, the banyan tree,
belongs to Varuṇa (GGS 4,7,24). It is Varuṇa who holds the heavenly banyan tree in the sky.
The “ropes of wind,” which, according to the Purāṇas, bind the stars and planets to the pole
star and prevent them from falling out of the sky, are undoubtedly the imagined air roots of
this cosmic banyan tree. These conceptions follow naturally from the original Dravidian name
of the pole star: vaṭa-mīn, “north star” = “banyan star” = “rope star.”

FIGURE 16.10 The heavenly fig tree with its deity, flanked by a star on either side, on the molded tablet H-179
(2410, NMI 30) from Harappa. After Parpola 1994a:244, fig. 14.5; cf. CISI 1:209. Photo EL, courtesy NMI. The stars
are surrounded by extensions of the tree’s arch curving like the horns of a buffalo; compare the Indus deity with
stars in the loops of the horned headdress in fig 21.6 b.

The most fundamental change in the passage from Vedic to later Indian cosmology, shared
by the Hindu, Buddhist, and Jaina traditions, is the addition of a golden mountain which rises
from the center of the earth to the zenith. On the top of the mountain is the city of Brahmā or
Indra, and directly above it is the pole star. The slopes of the mountain in the different
directions have different colors. Willibald Kirfel (1920:2*–3*, 14*–19*) connects this central
mountain with the astronomical conceptions associated with the pole star. Since the pole-star-
related cosmological ideas seem to be much more ancient than has been thought, the central
mountain and its name, Meru, attested in all three traditions, could also go back to the Indus
civilization. The etymology of Mēru has remained unclear; I propose deriving it from Proto-
Dravidian *mēl / *mēlu, basically “what is over or above,” hence also “the upper part of
anything, high place, top.” While Old Tamil has mēl-ulakam, “the celestial world,” a modern
Tamil lexicon uses the word mēl in its explanation of ucci, “zenith”: talaikku nēr mēlāka
irukkum vānattinpakuti, “the part of the sky that is directly above the head” (Kriyā 2008:177).
Semantically mēl(u) resembles ucci, “crown of the head, top, summit, zenith,” which is
supposed to be of Dravidian origin (DEDR 579), but actually can hardly be separated from the
Sanskrit adverbs uccā, uccais, “high, above,” derived from the Sanskrit verbal and nominal
prefix ud, “up.”
From Rigveda 1,24,7, it appears that the air roots of the heavenly banyan tree were
conceived as beams of light that supply human beings with life energy. A parallel idea
connected with the sun and its rays is found in the Purāṇic cosmology (Brahmāṇḍa-Purāṇa
24,64–72; Liṅga-Purāṇa 60,18–26; Vāyu-Purāṇa 53,44–50): the sun is said to feed the other
planets and stars with its thousands of rays (raśmi). The seven principal rays nourish the
moon, the calendrical stars (nakṣatra/ṛkṣa), and the planets Mercury (budha), Venus (śukra),
Mars (lohita), Jupiter (bṛhaspati), and Saturn (śanaiścara). The solar ray that brings
nourishment to the moon is called suṣumṇa-, “very benevolent.”
The Buddhist tradition ascribes different colors to the rays of the sun. Thus according to
the Avadānaśataka (ed. Speyer 1906:I,4–6), when the Buddha smiles, “a myriad band of light
with a thousand colours comes out of his mouth and illuminates the ten directions like a rising
sun”; the text specifies that “whenever the Blessed Buddhas produce a smile, rays of blue,
yellow, red and white light issue from their mouths.” David Fiordalis (2013) has compared this
with the Pravargya-Brāhmaṇa of the Taittirīya-Āraṇyaka (5,1,2), where brilliant energy (tejas)
escaped from Makha when he smiled. In chapter 11 we saw that Makha’s severed head
became the sun.
The solar ray suṣumṇa (suṣumṇaḥ sūrya-raśmiḥ) is mentioned in a mantra occurring in the
Yajurvedic Saṃhitās (MS 2,12,2; KS 18,14; VS 18,40; and as suṣumnaḥ in TS 3,4,7,1) and the
Śatapatha-Brāhmaṇa (9,4,1,9). In this mantra, which is used in the fire-altar ritual, the stars
are said to have the name bekuri / bhekuri-, which probably comes from Dravidian (see
chapter 14), and they are connected with the gandharvas and apsarases, whose abode is the
banyan and other fig trees (TS 3,4,8,4).
In haṭha-yoga or kuṇḍalinī-yoga, which is supposed to have emerged around 1000 CE and
best known from such texts as the Gorakṣaśataka (c. 1400 CE) and Haṭhayogapradīpikā (c.
1450 CE), suṣumṇā is the name of the central and most important one of the 72,000 veins or
arteries (nāḍī-) of the human body, connecting the nerve center of the mūlādhāra at the
bottom of the spinal cord with the brahma-randhra and its cakra of 1000-petaled lotus (which
represents the 1000-rayed sun). This tradition is very late, but its elements can be traced back
to the Upaniṣads. The Bṛhad-Āraṇyaka-Upaniṣad (2,1,19) mentions 72,000 arteries, and the
Maitrī-Upaniṣad (6,21) names one of these as suṣumṇā. Suṣumṇā is meant also in one of the
early ślokas (associated with the vrātya tradition, cf. Horsch 1966) quoted in the Chāndogya-
Upaniṣad (8,6,6) and the Kaṭha-Upaniṣad (6,16):

One hundred and one, the veins of the heart,

Of them one runs up to the crown of the head.
Going up by it, he reaches the immortal.
The rest, in their ascent, spread out in all directions. (Translated Olivelle 1998:279)

The foregoing passage of the Chāndogya-Upaniṣad (8,6,1–5) explains this verse:

1. Now, these veins [nāḍī-] of the heart consist of the finest essence of orange [piṅgala-],
white [śukla-], blue [nīla-], yellow [pīta-], and red [lohita-]. The sun up there, likewise, is
orange, white, blue, yellow, and red. 2. Just as a long highway traverses both the
villages, the one near by and the one far away, so also these rays of the sun traverse
both the worlds, the one down here and the one up above. Extending out from the sun
there, they slip into these veins here, and extending out from these veins here, they slip
into the sun up there. 3. So, when someone is sound asleep here, totally collected and
serene, and sees no dreams, he has then slipped into these veins. No evil thing can
touch him, for he is then linked with radiance. 4. Now, when someone here has become
extremely infirm, people sit around him and ask: “Do you recognize me?” “Do you
recognize me?” As long as he has not departed from the body, he would recognize
them. But when he is departing from his body, he rises up along those same rays. He
goes up with the sound “Oṃ.” No sooner does he think of it than he reaches the sun. It
is the door to the farther world, open to those who have the knowledge but closed to
those who do not. (Translated Olivelle 1998:279)

Here only five rays of different colors are mentioned, apparently those that feed the five
planets, for according to the earliest detailed descriptions of the planets, in texts such as
Yājñavalkya-Smṛti 1,290–303 and Jaiminīya-Gṛhyasūtra 2,9, Jupiter is golden, Venus white,
Saturn dark-blue or black, Mercury greenish-yellow, and Mars red. The passage makes clear
that the many veins bringing nourishment (blood) to all parts of the human body are thought
to be connected with the rays of the sun, understood to be similar channels bringing
nourishment from the sun to all beings of the universe. The human body with its center in the
heart and the universe with its center in the sun are two networks that correspond to, and are
connected with, each other.
The parallelism between the microcosm of the human body and the macrocosm of the
universe is indeed one of the very basic conceptions of the haṭha- or kuṇḍalinī-yoga. The spinal
column is called in Sanskrit meru-daṇḍa, a word that equates the column with the central
mountain of the universe, Mount Meru. At the top of the human body is the head, and at the
top of the head the brahma-randhra, “Brahma’s crevice,” the suture or aperture in the crown
that is still open in a new-born child, and through which the soul is said to escape at death. In
the macrocosm, the brahma-randhra would correspond to the sun in the zenith, which the
Upaniṣads present as the gate to the Brahma-loka beyond the universe.
The conception that the universe has the shape of a human body—also prevalent in later
Jaina cosmology—emerges in one of the latest hymns of the Rigveda (which are of
Atharvavedic affinity). According to RV 10,90, the primeval cosmic “man” (puruṣa), who had a
thousand heads, a thousand eyes, and a thousand feet (these descriptions refer to the sun,
which is often said to be thousand-rayed), rose up and expanded to fill the entire universe.
Verse 4 uses the verb vi+kram-, “to stride through,” which older hymns of the Rigveda use to
describe the three steps with which the solar god Viṣṇu traversed the universe. This primeval
man was the victim of the first-ever sacrifice performed by the gods, the model of the Vedic
human sacrifice (puruṣa-medha) and other sacrificial rituals. The manifested cosmos arose
from the dismembered body parts of this primeval cosmic man: the moon from his mind, the
sun from his eyes, the wind from his breath, the atmosphere from his navel, the sky from his
head, the earth from his feet, and the directions of space from his ears.
Herman Tull (1989) has pointed out that this idea of an originally anthropomorphic cosmos
that becomes dismembered differs fundamentally from the earlier Rigvedic cosmogonies.
While the idea that the universe resulted from the union and separation of Father Heaven and
Mother Earth goes back to Proto-Indo-European conceptions, “the chief cosmogonic myth of
the early Ṛgveda is the slaying of Vṛtra by Indra and Indra’s subsequent separation of the
heavens and the earth through propping up the sky with the cosmic tree or pillar” (Tull
1989:54). Tull concentrates on showing in the ideology of Brāhmaṇa texts, on the other hand,
that the concept of cosmic Puruṣa and his dismemberment remains central, particularly in the
fire-altar ritual (which is unknown to the Rigveda).
In the Brāhmaṇa texts, Puruṣa becomes the creator god Prajāpati, as is explicitly stated in
the Śatapatha-Brāhmaṇa (6,1,1,5). In the sequel (ŚB 6,1,2,12–26) this text notes that creating
the universe out of himself, Prajāpati exhausted himself and came to fall apart, that is, to die.
He asked his son Agni, “Fire,” to restore his body by building up the fire altar (agni-citi).
Prajāpati is here equated with the year and the cosmos, and his disjointed body parts are
equated with the seasons and the directions, as well as the layers of brick in the fire altar. The
fire laid on the completed altar is identified with the sun, and Prajāpati, the father of Agni, is
also said to be his own son Agni, while Agni as the restorer of Prajāpati is said to be
Prajāpati’s father. (This somewhat paradoxical notion may be explained thus: when the sun
has completed its yearly round and “brought out” all the seasons, it “dies” as the old or past
year, but is restored in the new year when the sun is “reborn” as the successor and is thus the
“son” of his own self.)
In miniature the same solar process takes place in the sun’s diurnal cycle. During the day,
the sun in the zenith represents the ruler of the universe. In the night this central position is
held by the pole star, to which all planets, stars, and beings are bound with invisible “ropes of
wind” (vāta-raśmi)—a conception I have endeavored to show goes back to the Indus
civilization: Dravidian vaṭa-mīn, “north star,” also means “rope star” and “banyan star.” The
pole star is thus the nightly counterpart of the sun, and indeed RV 1,24,7 refers to the
(invisible) air ropes of Varuṇa’s heavenly banyan tree as “beams of light” (ketávaḥ). According
to an astronomical conception current in Vedic times, which is probably of Harappan origin,
the sun actually never rises or sets but is in the sky all the time. The sun has a bright and a
dark side. In the evening it simply turns its dark side toward the earth and its bright side
toward the sky, returns invisible from west to east, and turns around again in the morning.
(This idea is most explicitly stated in AB 3,44, less clearly in MS 4,6,3; KS 27,8; and TS
6,4,10,2–3.)
The pole star is the pivot on which the whole sky turns. In this respect it is like the tortoise
(Sanskrit kūrma), which in the epic and Purāṇic myth of the churning of the milk ocean serves
as the pivot and firm foundation of the floating earth and its cosmic mountain Meru (alias
Mandara), used as a churning stick by gods and demons (cf. Mahābhārata 1,15–17; Matsya-
Purāṇa 249–251). The etymology of Sanskrit kūrma has not been settled, but as the word is
not attested in the Rigveda and lacks clear Indo-European cognates, it is likely to be a
loanword from a South Asian language. Moreover, it is a local South Asian animal, and one
that is represented in the molded tablets from Harappa and among the terracotta figurines
from Mohenjo-daro; further, one sign of the Indus script (no. 350 in Parpola 1994a:77) may
depict the tortoise. A possible Dravidian etymology for kūrma is suggested by Tamil kuruḷai,
“tortoise,” with cognates in other Dravidian languages denoting “shell,” such as Kannada
guruḷe; these words are probably derived from the root kur-uḷ, “to curl,” which at least in Tulu
has the further meaning “to contract, shrink” (cf. further the root kur-aṅku, “to bend, curve,
crouch, contract, shrink”). A potential Proto-Dravidian homonym of kūrma is provided by the
root *kūr, “to be sharp, pointed or tapering,” from which Tamil derives kūr, “pivot on which
door swings,” and kūrmai, “sharpness, pointedness.” In Vedic texts, the lower shell of the
tortoise is said to be the earth “which is firmly fixed (pratiṣṭhitam),” and the upper shell is the
sky, “which has its ends bent down” (ŚB 7,5,1,2).
The tortoise is the shape assumed by the creator god Prajāpati when he created the living
beings (ŚB 7,5,1,3). The tortoise is the life-sap and the life-breath (prāṇa), and a living tortoise
is placed at the bottom of the brick-built fire altar, a Vedic cosmogram and restructured image
of Prajāpati, in order to provide it with life-sap and breath (ŚB 7,5,1,1.3.7). “Now this tortoise
is the same as yonder sun. … (It is placed beneath the altar, wrapped in water plants, with the
mantras VS 13,30–32 beginning) ‘Seat thee in the depth of waters …’, for that indeed is the
deepest (place) of the (heavenly) waters where yonder (sun) burns; ‘the lord of waters, the
bull of bricks’ ” (ŚB 7,5,1,6–9). As observed by John Brockington (1998:279), the tortoise as
the lord of waters is here “a representative of Varuṇa.” The heavenly banyan tree (i.e., the pole
star) also belongs to Varuṇa.
The tortoise as a representative of the sun can be connected particularly with the setting
sun or night sun. The most common tortoise of South Asia, from Sindh to Kerala, is the starred
tortoise (Fig. 16.11). As we have seen, the Vedic texts equate the tortoise with the sun, and
state that during the night the black side of the sun is turned towards the earth. The black
shell of the starred tortoise is ornamented by several star-like figures, so that it looks like the
black night sky studded with stars.

FIGURE 16.11 The starred tortoise (Geochelone elegans). After Daniel 1983: opposite p. 30. Photo courtesy © Isaac
D. Kehimkar.

When setting, the sun draws in its rays, as the tortoise draws in its limbs. Three different
Dravidian words for “tortoise” are derived from this characteristic: besides the above-
mentioned Tamil kuruḷai, North Dravidian ekkā, eke, “tortoise,” from Proto-Dravidian *ekku,
“to contract the abdominal muscles, to draw in the stomach,” and Tamil oṭunki, “tortoise (as
that which contracts).” The Proto-Dravidian verb *oṭuṅku, “to contract, shrink, become
reduced, become hidden” is also used in Tamil in the sense of “to be restrained, as the senses
or desires, to become tranquil,” and the transitive form (used when the verb requires a direct
object) oṭukku in the meaning “to restrain the senses.” Patañjali, in his Yogasūtra (2,54),
advises the yogi to withdraw the senses from their objects to attain complete mastery of the
sense organs. The Mahābhārata often compares this yogic sense-retraction (pratyāhāra) to the
tortoise’s drawing in its limbs (thus Bhagavadgītā 2,58). This comparison is also made in
Tirukkuraḷ 126, where the Proto-Dravidian verb *aṭaṅku / *aṭakku is used. Besides meanings
similar to those of *oṭuṅku / *oṭukku, this verb is also attested in Tamil in the sense of “to
disappear, set as a heavenly body.”
In post-Vedic worship of planetary gods, Saturn may be represented as riding a tortoise,
and the Indus script suggests that Saturn was understood to have the tortoise as his vehicle
already in Harappan times (chapter 21). According to the epics and Purāṇas, Saturn is the son
of the sun; and as his color is, moreover, black or dark blue and the main deity presiding over
Saturn is Yama (the Righteous King judging the dead and ruling over them), Saturn seems to
represent the nightly aspect of the sun. In Babylonian astrology, too, Saturn is the “black star,”
representing the sun, and functions as the judge of the dead (Jeremias 1929:176).
We have seen that the tortoise placed at the bottom of the fire altar in water plants is said
to be the sun in the deepest place of heavenly waters (ŚB 7,5,1,6–9). This suggests its
connection with the night sun in the region of the pole star, which seems to correspond to the
skull suture (brahma-randhra) of the macrocosmic man. The Śatapatha-Brāhmaṇa (7,5,1,35)
indeed says that “the tortoise (is) the head (of the fire altar), and the vital airs in the tortoise
are the vital airs in the head.” The heads of the five principal victims are placed in the middle
of the lowest brick layer of the fire altar, the human head in the center, surrounded by those of
the horse, the cow, the ram, and the sheep, and the sacrificer is required to put into each head
seven chips of gold, one into each of the “seven vital airs of the head”: the two eyes, the two
ears, the two nostrils, and the mouth (ŚB 7,5,2,8–12). One reason why the tortoise is
connected with the head is likely to be its shell, for words denoting “shell” often stand also for
the “shell of the head”: English skull, for instance, has such a derivation, and Dravidian ōṭu,
“shell,” has the meanings “tortoise shell” and “skull” in several languages.
According to one of best-known late Rigvedic cosmogonic speculations, there was neither
asat, “nonexistent,” nor sat, “existent,” in the beginning (RV 10,129,1). Śatapatha-Brāhmaṇa
(6,1,1,1–7) modifies this by saying that in the beginning there was asat, and this asat was the
Seven Sages, who were the seven vital airs. The creation could start only after these seven
puruṣas were made into one Puruṣa, who became Prajāpati; the life-sap of those seven puruṣas
is concentrated in Prajāpati’s head. The seven circumpolar stars of Ursa Major are thus in
Vedic texts evidently connected with the head of the cosmic man (puruṣa) as its seven
openings (see Ehlers 2007). Being close to the pole star, to which other stars are bound with
“ropes of wind” (vāta-raśmi), the “Seven Sages” of Ursa Major may be meant when the
Rigveda (10,136) speaks of the long-haired, orange-clothed, and wind-girdled (vāta-raśmi)
ascetics (muni) who, having drunk poison from Rudra’s cup, fly through the air carrying fire.
The above passage, speaking of the seven openings of the head and their connection with
the Seven Sages, mentions seven vital airs. Normally only five kinds of air in the body are
spoken of, are enumerated and identified with the five directions. Just two are important,
however, namely the inhalation (prāṇa) and exhalation (apāna), which the Seven Sages are
said to observe in a śloka verse and its commentary contained in Jaiminīya-Brāhmaṇa 2,27.
This looks like a reference to the regulation of breathing (prāṇāyāma), which is yet another
important component of Patañjali’s yoga. The words prāṇa and apāna do not occur in the
Rigveda at all, while both words occur many times in the Atharvaveda; in the Vrātya book, the
breaths of the Vrātya are identified with different elements of the universe (AVŚ 15,15–17), as
are his body parts (AVŚ 15,18), rather like the body parts of the primeval Puruṣa in RV 10,90.
Ever since its discovery, the Indus civilization has been regarded as a possible source for
the traditions of yoga. So far, the principal evidence for this hypothesis has consisted of the
postures of seated Indus divinities on the seals, especially the “Proto-Śiva” (Fig. 16.6), whom
John Marshall (1931a) compared with Śiva as Yogeśvara, “master of yoga.” However, this
posture seems to be an artistic convention taken from Proto-Elamite art (see Fig. 17.2). While
this does not prevent it from also being a prototype of a particular yogic āsana, it somewhat
weakens its power as evidence for the Harappan beginnings of the yoga. On the other hand,
the human body plays an important role in the Tantric ritual, and the meditative practices of
yoga undoubtedly do feel rather un-Aryan and native to South Asia, given the latter’s strong
ascetic traditions evident from the dawn of the historical period. This chapter has linked some
central concepts and practices of the yogic tradition with a cosmology that can with good
reason be linked to the Indus civilization, and thus it provides new evidence for the view that
yoga, too, is a part of the Harappan cultural heritage.
17
Dilmun, Magan, and Meluhha

There were overland and then maritime contacts between West Asia and the Indus Valley
during the third millennium BCE (Fig. 17.1). The eastward expansion of the Proto-Elamite
civilization started around 3200 BCE and reached Seistan, Kerman, and Makran in eastern Iran
by about 2900 BCE (Lamberg-Karlovsky 1986). Typically Proto-Elamite bevel-rimmed vessels
(along with compartmented copper seals known widely from the Iranian plateau) have been
discovered at Mir-i Qalat in Pakistani Baluchistan (end of period IIIa, c. 2900–2800 BCE). The
layer above this one (IIIb, c. 2800–2600 BCE) has various kinds of pottery with motifs such as
zebu bulls, leopards, and fish that are very similar to pottery motifs in the Early Harappan
cultures of the Indo-Iranian borderlands. Still younger layers demonstrate the transition
period (IIIc, 2600–2500 BCE) and the mature period (IV, c. 2500–2000 BCE) of the Indus
civilization, respectively (Besenval 2011; Didier 2013).

FIGURE 17.1 Map illustrating interaction between West, Central, and South Asia, c. 2600–1900 BCE. © Anna
Kurvinen and Asko Parpola. Country names in bold versals are those of cuneiform sources. The approximate location
of archaeological sites is indicated with numbers: (1) Akkad; Al Hiba, see Lagash; (2) Altyn Depe; (3) Amri; (4)
Anshan (Tall-i Malyan); (5) Babylon; (6) Bala-kot; (7) Bampur; (8) Chanhu-daro; (9) Dholavira; Djokha, see Umma;
(10) Eshnunna (Tell Asmar); (11) Faiz Mohammad; (12) Girsu (Tello); (13) Godin Tepe; (14) Gonur; (15) Harappa;
(16) Hili; (17) Jiroft; (18) Khafajeh; (19) Kish; (20) Kulli; (21) Lagash (Al Hiba); (22) Larsa; (23) Lothal; (24) Maysar;
(25) Mehrgarh; (26) Mir-i Qalat; (27) Mohenjo-Daro; (28) Mundigak; (29) Nal; (30) Namazga; (31) Nindowari; (32)
Nippur; (33) Qala’at al-Bahrain; (34) Quetta; (35) Ra’s al-Hadd; (36) Ra’s al-Jinz/al-Junayz; (37) Sar; (38) Shahdad;
(39) Shah-i Tump; (40) Shahr-i Sokhta; (41) Shah Tepe; (42) Shortughai; (43) Sotka-koh; (44) Susa; (45) Sutkagen-
dor; Tall-i Malyan, see Anshan; (46) Tell Abraq; Tell Asmar, see Eshnunna; (47) Tell es-Sulema; Tello, see Girsu; (48)
Tepe Gawra; (49) Tepe Giyan; (50) Tepe Hissar; (51) Tepe Sialk; (52) Tepe Yahya; (53) Tureng Tepe; (54) Umma
(Djokha); (55) Umm an-Nar; (56) Ur; (57) Uruk.

More debatable evidence for overland contact is the so-called yoga posture of the
Harappan deity dubbed “Proto-Śiva” (Fig. 16.6), which seems to derive from the way in which
seated bulls are represented in Proto-Elamite art (Fig. 17.2) (Parpola 1984a). On the other
hand, the motif may have been borrowed c. 2500–2400 BCE from the so-called “Trans-Elamite”
art, which flourished in the Jiroft region of Kerman and in the BMAC of southern Central Asia;
it may even have come from Mesopotamia itself in Early Dynastic IIIAB times (2600–2334 BCE)
via maritime contacts between Indus Valley merchants and Sumerian rulers.
FIGURE 17.2 Sitting bulls on two Proto-Elamite seal impressions from Susa. After Amiet 1980: nos. 581, 582.
Courtesy Pierre Amiet. Compare the sitting posture of the “Proto-Śiva” (Fig. 16.6).

In the second half of the third millennium BCE, the Indus people began to travel in the
opposite direction, westwards to Mesopotamia. During the transition from the Early to Mature
Harappan phase (2600–2500 BCE), the Indus people built boats and ships that enabled them to
navigate the Indus River and other waterways with large and heavy loads. Harappan seals and
tablets depict ships with a hut-like cabin and a steersman at the helm with an oar. These ships
resemble quite closely the present-day houseboats of the Mohanas, fishermen of the Lower
Indus and the Manchar Lake. In the early Mature Harappan period (2500–2200 BCE), there
was a dynamic cultural expansion. Enterprising seafaring merchants from the Indus Valley
founded outposts on the Makran coast (Sotka-koh, Sutkagen-dor), in Oman (Ra’s al-Junayz,
Ra’s al-Hadd, Maysar, Hili), along the Persian Gulf, and in Mesopotamia itself.
Among the items they traded was Harappan jewelry of high value, which has been found in
Mesopotamia. Exceptionally long (more than 5 cm) barrel-cylinder-shaped carnelian beads
involved the difficult drilling of long holes in this very hard stone. Other carnelian beads were
stained with white designs by first drawing a pattern on the stone with soda and then heating
it until the carnelian absorbed the alkali, making the pattern permanent. Such “etched”
carnelian has survived as an industry in Sindh until our times. These two kinds of distinctive
bead—excavated at Kish and in the Royal Graves at Ur, and dated to the Early Dynastic (ED)
IIIA period—are the earliest objects of Harappan manufacture discovered in Mesopotamia.
According to the conventional chronology followed in this book, their date is 2600–2500 BCE,
but according to a more recent chronology gathering support, their actual date may be as late
as 2400–2300 BCE (Reade 2008). It is possible that the jewelry reached Mesopotamia via
middle men, either overland (through Iran) or by sea. But it appears more likely that the
Harappans were already present in Sumer, on the basis of the iconographic motifs of ED III
date borrowed into Harappan art, to be discussed in chapter 18.
Some Indus seals excavated in Mesopotamia in 1923–1925 provided the very first
intimation of the high age of the Indus civilization. About twenty seals or seal impressions
connected in one way or another with the Indus civilization come from different sites in
Mesopotamia (Ur, Telloh/Girsu near ancient Lagash, Djokha/Umma, Babylon, Nippur, Kish, Tell
Asmar/Eshnunna, and Tell es-Sulema) and Elam (Susa); in addition, about a further ten seals
are unprovenanced but probably come from Mesopotamia or Elam. Unfortunately, only a few
of these seals can be dated with fair certainty.
The earliest of the seals are attributed to the Akkadian period (2334–2154 BCE). They have
the typically Mesopotamian cylindrical shape but, given their Harappan style and motifs
(including a frieze of an elephant, a rhinoceros, and a crocodile), they were almost certainly
carved by Harappan craftsmen residing in Mesopotamia. Two come from stratified contexts in
the northerly sites Tell es-Sulema and Tell Asmar, and lack an inscription. An iconographically
important cylinder seal without provenance is very probably contemporary with these first two
(Fig. 18.12). A clumsily carved cylinder seal with an inscription excavated at Susa but in an
unknown context may also be contemporary with them.
A corroded square copper seal and an Akkadian-style cylinder seal of lapis lazuli (without
an inscription) attached to a metal cloak pin were excavated next to the arm of a skeleton
buried in Private Grave no. 489 at the cemetery of Ur (Reade 1995). In an X-ray image of this
copper seal shown to me in 1992, an Indus inscription with sign sequences common in the
Indus Valley was clearly visible. A closely similar square copper seal was excavated at Ra’s al
Jinz (Ra’s al-Junayz) in Oman, an important source of copper for the Indus people and the
Mesopotamians (see Parpola 1994b: nos. 21 and 2). Only a single copper seal is known from
the Greater Indus Valley, discovered in Lothal, a port town in Gujarat that served as an
international entrepôt. Steatite square seals with native Harappan sign sequences have been
excavated at Nippur and Kish, but cannot be dated with certainty; they were probably
imported from the Indus Valley, like the single rectangular stamp seal with an inscription
found at Telloh, dating from the time of Gudea (c. 2141–2122 BCE).
The round stone seals found in West Asia fall into two categories (Laursen 2010). The first
category is known as “Gulf type.” It depicts a Harappan-type bison (with or without a
“manger”), with or without signs of the Indus script, and has a Harappan-type knob (with a
single groove) on the reverse. A total of twenty-eight Gulf-type seals with an inscription are
known. Twelve are from Mesopotamia (half being from Ur), three are from Iran (one from
Susa, another of unknown source with an inscription in Linear Elamite), nine are from the Gulf
(seven from Bahrain, two from Failaka), and five are from the Lower Indus Valley (four from
Mohenjo-daro, one from Chanhu-daro). On one of these last, Harappan-style animal front parts
are united into a whorl, a motif common in West Asia after 2000 BCE. There are also ninety-
five Gulf-type seals without an inscription, almost all of them (eighty-seven) from Bahrain, the
rest from elsewhere in the Gulf, except for one from Iran (Tepe Yahya in Kerman).
The oldest of the Gulf-type seals, dated c. 2100–2050, come from Mesopotamia and Iran.
After this period, it appears that a community of Harappan merchants moved from
Mesopotamia/Iran to Bahrain and settled there, for the majority of Gulf-type seals, dated c.
2050–2000 BCE, come from Bahrain, and there is evidence of local seal manufacture. Cubic
weights of polished stone typical of the Indus civilization have been found in Bahrain, with
numerous seals in the rooms of an ancient customs house next to the city gate of Qala’at al-
Bahrain (Fig. 17.3). The majority of the seals come from burial mounds typical of the local
culture, which suggests that the Indus people married into the local society. This supposition
is reinforced by the fact that a great number of the Gulf-type seal inscriptions show sequences
not known in the Indus Valley seals (Fig. 14.1). They probably write local proper names, given
by bilingual Harappan merchants to their children now running the business. One seal from
Bahrain was found with a small cuneiform tablet containing three Amorite proper names,
datable on the grounds of the forms and combinations of its signs to c. 2050–1900 BCE. This
suggests that the local language in Bahrain may have belonged to the West Semitic group.

FIGURE 17.3 An ancient customs house in the City II area within the north wall at Qala’at al-Bahrain in the Gulf. To
the right is the cul-de-sac, to the left the city gate, and between them the guard room with its well and staircase to
the ramparts. Find spots of the seals (×) and weights (•) are marked. After Bibby 1972:368. Courtesy Moesgaard
Museum, Højbjerg, Denmark.

The Gulf-type seal was a model for the second category of round seals, the “Dilmun type,”
hundreds of which replaced the Gulf type in Bahrain and Failaka around 2000 BCE and
continued in use for about a century. These later seals have Mesopotamian-influenced motifs,
no inscriptions, and three grooves on a convex reverse, flanked by two dots-in-a-circle on
either side. One Dilmun-type seal was excavated at Lothal in 1963. This is practically the only
object in the Greater Indus Valley that clearly comes from West Asia, apart from some
Mesopotamian-type “barrel weights” from Mohenjo-daro and Harappa that were used in the
Gulf in the early second millennium BCE. The Lothal seal suggests that the maritime trade
between Mesopotamia and the Indus Valley was wholly in the hands of sailors from the valley
and Gulf merchants of Indus descent, and that Mesopotamian people did not come to South
Asia.
Dilmun is a mythical paradise in a Sumerian creation myth, but cuneiform texts mention it
as a real place, too, where Mesopotamian sailors could replenish their stocks of drinking
water. It was an important center of trade to which goods came from a number of places.
Dilmun figured in Mesopotamian maritime trade from the late fourth millennium but
especially in the late ED III and Akkadian periods (c. 2500–2200 BCE). It is scarcely mentioned
between 2200 and 2000 BCE, but becomes prominent again thereafter. Both archaeological
and textual evidence suggests that Dilmun stretched from the island of Failaka in the north of
the Persian Gulf to the island of Bahrain in the south, which acted as the center of Dilmun.
Bahrain, with its thousands of tombs and its flourishing date palms, fits the image of paradise
described in the Sumerian myth.
Another key location in cuneiform texts, Magan, was an important source of copper for the
Sumerians. Archaeological excavations and surveys have identified Magan with Oman
(especially Umm an-Nar) during the third millennium. Magan may, however, have included the
Makran coast of Iran on the opposite side of the Gulf from Oman. For one thing, the name
Magan may be related to that of Makran, which goes back to Maka, one of the satrapies of the
Persian empire. Besides, “the meš wood of Magan,” mentioned in the trilingual inscriptions of
Darius the Great, was imported from northwest India (Gandhāra) and southeastern Iran
(Kerman). The wood is called yakā- in Old Persian; its modern Iranian cognate, ag, denotes
sissoo wood from a large tree growing in the mountains of Oman, the Makran coast of
southern Iran, and Pakistan.
While Dilmun and Magan are mentioned in early cuneiform texts, the third foreign country
involved in the Mesopotamian sea trade, Meluḥḥa, appears for the first time in an Akkadian
inscription of Sargon the Great (ruled 2334–2279 BCE). Sargon boasts that ships from
Meluhha, Magan, and Dilmun docked in his new capital Akkad. To impress his readers, Sargon
has reversed the order in which these countries were usually listed, putting the most distant
place first. Several administrative texts mention food rations for “men of Meluhha” boarding
“ships of Meluhha.” The identification of Meluhha with the Greater Indus Valley is now almost
universally accepted.
How did the Mesopotamians communicate with the Meluhhans? An Akkadian cylinder seal
is inscribed “Šu-ilišu, interpreter of the Meluhhan language” (Fig. 17.4). This indicates that
the Meluhhan language was not widely understood in Mesopotamia and that some Meluhhan
speakers took local names, in this case a typical Semitic name.

FIGURE 17.4 Impression of an Akkadian cylinder seal that, according to its inscription, belonged to “Šu-ilišu,
interpreter of the Meluhhan language” (for Dominique Collon’s comments on this seal, see the caption of fig. 8.4 in
Parpola 1994:132). Collection De Clercq, Département des Antiquités Orientales (AO 22310), Musée du Louvre.
Photo courtesy Musée du Louvre/Béatrice André-Salvini.

The Akkadian kings endeavored to expand their power to the east, but never reached
Meluhha. Apparently the easternmost country they conquered, first under Sargon himself, was
called Marḥaši (in Sumerian) or Paraḥšum (in Akkadian). A historical text relates how other
Iranian countries and Meluhha allied themselves with Marhaši against Sargon’s son and
successor Rimuš (ruled 2278–2270 BCE), who reconquered Parahšum: “[Rimuš, king of the
world, in battle] was victorious over Abalgamaš, King of Parahšum. Zahar, Elam, Gupin and
Meluhha assembled in Parahšum for battle, but [. . .].”
An Old Babylonian text, “Gula-AN and the seventeen kings against Narām-Sīn,” recounts
how almost the whole world rebelled against Narām-Sīn (ruled 2254–2218 BCE), the grandson
of Sargon. This text, unfortunately mutilated, lists eighteen countries or peoples with their
mutinous kings, most of whom lack the first part of their names. In eleventh place comes: “[. .
.]-ib-ra, man of Meluhha.” This tantalizing reference is the single historical name we have for
a person from Meluhha, that is, from the Indus civilization. The eighteen countries or peoples
are, in order: Gutium, Kakmum, Lullubum, Hahhum, Turukkum, Kaniš, Amurru, Dēr, Arrites,
Kassites, Meluhha, Aratta, Marhaši, the confederacy of Elam, Land of Apum/the Canebrake,
Land of Fifty, Land of Armanum, Land of Hana.
Aratta, the country listed between Meluhha and Marhaši, is chiefly known in
Mesopotamian sources from the late Sumerian epic, “Enmerkar and the Lord of Aratta,” in
which King Enmerkar of Uruk negotiates the importation of lapis lazuli with the king of Aratta,
a legendary country situated far away to the east. This is presumably the same country as that
mentioned under the non-Aryan name Araṭṭa/Āraṭṭa/Arāṭṭa/Ārāṭṭa in Indian sources, such as the
Vedic Baudhāyana-Śrautasūtra (18,13 and 18,44) and the Mahābhārata, where Āraṭṭa
comprises the “immoral” regions of the northwest: Sindh, Gandhāra, and the western Punjab.
The Mahābhārata includes within Āraṭṭa a region called Vāhīka or Bāhīka, which frequently
occurs as a variant spelling of Bāhlīka/Bālhīka/Bahlika/Balhika/Balhīka. Vāhīka or Bāhīka was
apparently created by ancient grammarians (who derived it from Sanskrit bahiṣ, “outside”) in
an attempt to understand these Sanskrit counterparts (Bāhlīka, etc.) of the Avestan name
Bākhdī and Bākhtrī, that is, Bactria, modern Bālkh in northern Afghanistan—which is of
course “outside” the Indian plains. A Harappan colony in Bactria, at Shortugai, controlled the
nearby lapis lazuli mines of Badakhshan.
Narām-Sīn’s son and successor, Šarkališarri (ruled 2217–2193 BCE), married a princess
from Marhaši. He seems to have visited the country and received a delegation from it at
Nippur. During the Third Dynasty of Ur, Marhaši had good diplomatic relations with Sumer
and close economic relations with Magan. As tribute, Marhaši presented to Ibbi-Sīn, the last
king of Ur III (whose rule ended in 2004 BCE), a “speckled dog of Meluhha”—possibly a
leopard (cheetah), although this is a cat rather than dog, or the fierce Indian wild dog known
as the dhole (Cuon alpinus), although this animal is red rather than speckled. (A hyena would
perhaps suit the description best, were it not for the fact that hyenas were already known in
West Asia.) It even appears that Marhaši could export Indian animals to Mesopotamia, for in
the Old Babylonian text “Curse over Akkad,” holy Inanna arranges that: “foreigners would
cruise about [in Akkad] like unusual birds in the sky; that even Marhaši would be reentered in
the tribute rolls; that monkeys, mighty elephants, water buffaloes, exotic animals, as well as
thoroughbred dogs, lions, mountain ibexes, and alum sheep with long wool would jostle each
other in the public squares.”
Marhaši/Parahšum has been identified with great probability as the above-mentioned Jiroft
region of Kerman, where dramatic archaeological finds were recently made, around 2000.
Hundreds, if not thousands, of vases, cups, boxes, plaques, and handled “weights” with
beautiful iconography carved in soft stone (chlorite, serpentine, steatite) were illegally
excavated from five vast graveyards. Part of this trove was confiscated by police and published
by Iranian archaeologists, who have since excavated the related habitation site Konar Sandal
B near Jiroft. Important evidence supporting Marhaši’s identification with Kerman comes from
the cuneiform inscription “Rimuš, King of Kiš, the slayer of Elam and Parahšum,” carved on
two Jiroft-like chlorite bowls, one from Ur and another of unknown provenance. They probably
belonged to the booty consisting of various kinds of stones that Rimuš states he brought from
Parahšum to Mesopotamia (Steinkeller 1982).
Sumerian literary texts composed in the Old Babylonian period (c. 1800–1600 BCE) reflect
Mesopotamian conceptions of foreign countries in the third millennium BCE. A composition
entitled “Enki and the World Order” contains interesting information about Meluhha. It
describes how the wise water god Enki moves around organizing the world and decreeing the
fate of different countries. Lines 124–130 speak of foreign partners in maritime trade: “Let the
lands of Meluhha, Magan and Dilmun look upon me, upon Enki. Let the Dilmun boats be
loaded with timber. Let the Magan boats be loaded with treasure. Let the magilum boats of
Meluhha transport gold and silver and bring them to Nibru for Enlil, the king of all lands”
(translated Maekawa & Mori 2011:262).
Sumerian magilum (gišma2-gi4-lum) is a word exclusively used of the boats of Meluhha. In
my opinion, it might well render Proto-Dravidian *maṅki (*maṅgi), a form that can be
reconstructed as the ancestor of South Dravidian *mañci (*mañji), as an affrication of *k/g due
to the following i may be assumed. The meaning of this latter word, attested in Tamil,
Malayalam, Kannada, and Tulu, is as follows: “single-masted large cargo boat with a raised
platform used in coasting trade, holding 10–40 tons”—which matches remarkably well the
contextual meaning of magilum in the above Sumerian passage. The above Proto-Dravidian
reconstruction is supported by the following Sanskrit words recorded only by the twelfth-
century Gujarati lexicographer Hemacandra and some other lexica (including one for Hindi):
maṅginī-, “ship, boat,” and maṅga-, “forepart of a ship, mast.”
Lines 219–234 of “Enki and the World Order” deal with Meluhha itself:

Then he proceeded to the land (kur) of Meluhha. Enki, Lord of Abzu (the Sweet-Water
Ocean), decreed its fate: “Black land/mountain (kur), may your trees be great trees,
may your forests be forests of highland (kur) meš trees! Chairs made from them will
grace royal palaces! May your reeds be great reeds [. . .]! Heroes shall [. . .] them as
weapons in the battlefields! May your bulls be great bulls, may they be bulls of the
mountains/highland (kur)! May their roar be the roar of wild bulls of the
mountains/highland (kur)! The great powers (me) of the gods shall be made perfect for
you! May the francolins (dar) of the mountains/highland (kur) wear carnelian beards!
May your birds all be peacocks (haia)! May their cries grace royal palaces! May all your
silver be gold! May your copper be tin-bronze! Land (kur), may everything you possess
be plentiful! May your people be [. . .]!” (Translated Maekawa & Mori 2011:262,
modified)

The reference to “reeds” and “battlefields” relates to the fact that arrows have traditionally
been made of reeds in India: Sanskrit śara, a principal word for “arrow,” has as its primary
meaning “reed” (Saccharum sara). Cuneiform texts inform us that reeds from Magan were
used to make spears or arrow shafts. The reference to peacocks calls for a comparison with
the much later Buddhist Bāveru-Jātaka. A story that may date from the Achaemenid times
(sixth to fourth centuries BCE) speaks of Indian merchants who “came to the kingdom of
Bāveru (i.e., Babylon, called in Babylonian Bābilu and in Old Persian Bābiru) [and] brought a
royal peacock which they had trained to scream at the snapping of the fingers and to dance at
the clapping of the hands”; the Babylonians bought it for a thousand gold coins, and paid the
royal peacock high honor. Evidently the beautiful Indian bird has always been a desired luxury
item of export.
Sumerian kur, “mountains, highland,” which occurs many times in the above passage,
usually refers to the Iranian plateau and to the foreign countries east of Mesopotamia. “Black
land” could mean either “country of dark-skinned people” or “country of dark soil.” I opt for
the second alternative. The south Indian state Karnataka derives its name from Dravidian
karu-nāṭu “black land,” on account of its black loamy soil. In the Indus Valley, “very
characteristic of Sind is the ‘kalar,’ soil which contains an excessive proportion of salts. Its
composition has been known to include nearly forty per cent of sodium sulphate. In such
ground, almost black in colour, and glutinous, no vegetation can subsist, but in the darkish
brown medium kalar several species of salvadora … flourish, particularly in lower Sind.”
(Lambrick 1964:16). The Sindhi word kalar goes back to Proto-Dravidian *ka lu / *ka lar /
*kaḷar “saline soil.”
“Black land” recurs as an attribute of Meluhha in the “Curse over Akkad,” which describes
events in the times of Narām-Sīn. In lines 43–50, people from many foreign countries bring
goods to the city of Akkad:

At the city gate, like the Tigris flowing toward the sea, holy Inanna opened the portal of
its gate. Sumer brings its own possessions upstream by boat. The highland Martu,
people ignorant of agriculture, brought befitting cattle and kids for her. The
Meluhhans, people of the black country, brought exotic wares to her. Elam and Subir
loaded pack-asses with goods for her. (Translated Maekawa & Mori 2011:260,
modified)

“Black country,” denoting Meluhha, also occurs in an interesting inscription (B, lines 218–220)
of King Šulgi of Ur from the Neo-Sumerian period (ruled 2094–2047 BCE). The king boasts that
he can answer in five languages, being those of Elam, Sumer, the “black country” (Meluhha),
Martu, and Subir. From precisely this reign, there are six Gulf-type Indus seals found at Ur.
During a period of at least thirty-five years (2105–2071 BCE), a village called Meluhha
existed near the city of Lagash during the Third Dynasty of Ur. Meluhha is also the proper
name of some of those who resided there, while its other residents have purely Sumerian
names (S. Parpola, A. Parpola, & Brunswig 1977). One Ur III cuneiform tablet connects the
“Meluhha village” with Guabba. This does not mean that the village was identical with the
well-known large town called Guabba, as Vermaak (2008) has understood, but simply its
location in the Guabba (“Sea-shore”) province south of Lagash (Simo Parpola, personal
communication). The dates of this Meluhha village agree closely with those of the early Gulf-
type seals of Mesopotamia, such as the one from Telloh/Girsu, near Lagash. A little bit earlier,
Gudea, the Sumerian king of Lagash (ruled 2141–2122 BCE), states in his inscriptions that “the
Meluhhans came from their country” and supplied ušū wood, gold dust, carnelian, and other
luxury items for the construction of the main temple in Gudea’s capital.
The Old Babylonian text “Enki and Ninhursag” tells of the countries from which various
goods came to Mesopotamia:

May Dilmun become a storehouse on the quay for the land. May the land of Tukriš hand
over to you gold of Harali, lapis lazuli and [. . .]. May the land of Meluhha load precious
desirable carnelian, meš wood of Magan and the best abba wood into large ships for
you. May the land of Marhaši yield you precious stones, dušia stone. May the land of
Magan offer you strong, powerful copper, diorite, ú stone, and šumin stone. May the
Sea-land offer you its own ebony wood, [. . .] of a king. May the “tent”-lands [i.e., the
country of shepherds?] offer you fine multicoloured wools. May the land of Elam hand
over to you selected wools, its tribute. (Translated Maekawa & Mori 2011:261)

The abba (“sea”) wood imported from Meluhha for making chairs and dagger sheaths during
the Third Dynasty of Ur was in all likelihood the very hard wood of the mangrove tree growing
on the coast of east Baluchistan and Sindh. The ú stone was used to drill carnelian; the
Harappans employed for this purpose “a type of indurated tonstein flint clay that has been
deliberately heated to produce or enhance properties that made it a highly effective material
for drilling hard stone beads” (Law 2011:83).
Ivory, though unmentioned in “Enki and Ninhursag,” was another product of Meluhha that
was imported into Mesopotamia via the Gulf (the texts mention it as coming from Magan and
Dilmun). But after the end of the Isin-Larsa dynasty about 1800 BC, it ceases to be mentioned
in cuneiform texts. Indeed, the very name Meluhha disappears altogether until the fifteenth
century BCE, when it reappears referring to a new source of ivory: Nubia or Ethiopia, places
that had been unknown to Mesopotamians in the third millennium BCE. Now the texts clearly
speak of “black men of Meluhha,” instead of the earlier “black land of Meluhha.”
18
Royal Symbols from West Asia

The Harappan people, although technologically advanced in some respects—in water

engineering, textile weaving, and bead making—were undoubtedly greatly impressed by the
achievements of the Mesopotamian civilization. The elements they borrowed from
Mesopotamia were, however, adapted to their own art traditions and environment by
Harappan craftsmen; there is no evidence for Mesopotamian craftsmen working in the Indus
Valley. Of these borrowed elements the most important were several royal symbols that
embrace the so-called contest motif, which constitutes the principal iconography of the
Sumerian and Akkadian royal seals, coupled with its associated details: the hairstyle of the
royal hero and the “victory pose” of his foot. A further stylistic borrowing is the royal robe
known as a “sky garment.” This chapter explores the origin and significance of the “contest”
motif and the above-noted details in Mesopotamia.
Five Indus seals from Mohenjo-daro and six molded tablets from Harappa depict a
standing male hero who holds back two rearing tigers with his bare hands. Ernest Mackay
(1938:I,337) noted that although the seals were manufactured in the Indus Valley they copied
a Mesopotamian motif. This shows a male figure with two rearing lions, reminiscent of the
epic of Gilgamesh: an example of a “contest scene” in West Asian iconography, which could
involve combat with a range of animals. The replacement of the Sumerian lions with tigers, in
the scene’s Harappan equivalent, seems to be a natural adaptation of the “contest” motif to
the ecology of the Indus Valley. (The lion is not represented anywhere in Indus iconography,
but it is depicted in the seals of the BMAC, see Fig. 20.2, and also occurs in the Rigveda as
siṁha-, “lion,” a foreign loanword in Sanskrit.) However, as Heinz Mode (1944) pointed out,
the Harappan representation has a closer parallel in the Nile Valley than on the royal seals of
Mesopotamia, notably on an ancient Egyptian flint knife from Gebel el-Arak, which will be
examined below (Fig. 18.1). Moreover, the man fighting with lions in Sumerian art (and also in
later Assyrian and Persian glyptics) is often the king—a figure for whom the evidence from the
Indus Valley is dubious.
FIGURE 18.1 A Late Uruk style priest-king with two lions on the ivory handle of a royal flint knife from Gebel el-
Arak, Egypt, c. 3300–3200 BCE. Musée du Louvre (AE 11517). Photo courtesy Musée du Louvre/Béatrice André-
Salvini.

The “contest” motif has an interesting history, which provides clues to understanding its
symbolism. Its background is the “master of animals” or “divine hunter” of the upland tribes
of southwestern Iran, who lived in symbiosis with the people of Susa in the adjacent lowlands
of Elam. He was originally an anthropomorphic figure surrounded on either side by a snake,
or sometimes also by a game animal, an eagle, or a panther. But around 4500–4000 BCE, in the
stamp seals of Archaic Luristan and the style-A seals of Susa, the “master of animals” is
depicted with the head of a wild goat. Later, in a style-B stamp seal from Susa (4000–3500
BCE), while retaining the horns of a wild goat, the figure holds back two rearing lions, one on
either side. When the Mesopotamians of the Uruk culture expanded their power to Elam
shortly before 3500 BCE, this “contest” motif became part of their iconography. Seal
impressions from Uruk and Susa II depict an apparently naked hero holding back two lions.
The next manifestations of the “contest” motif are from the subsequent Late Uruk period; but
they come, somewhat surprisingly, from Upper Egypt.
The Late Uruk period, 3500–3100 BCE, was one of the most dynamic periods in
Mesopotamian history, which witnessed the birth of the world’s oldest writing system around
3400 BCE in Uruk. Before this, in the fifth millennium, southern Mesopotamia had already
exerted influence, but now, during the so-called Uruk expansion, large-scale Late Uruk
colonies—Habuba Kabira, Jebel Aruda, and Sheikh Hassan—were established on the Upper
Euphrates (3500–3000 BCE) (Rothman 2001). They are nowadays reckoned to have played a
significant role in the relations of the Uruk culture with Egypt, by linking Upper Egypt with
West Asia via Lower Egypt and the Levant, rather than via the Persian Gulf and the Red Sea,
as was previously thought. Egypt undoubtedly had trading contacts with West Asia from the
Middle Uruk period onwards. Lapis lazuli, originally from Badakshan in Afghanistan, has been
found in the Naqada region of Upper Egypt, dating from as early as 3800–3600 BCE, and at
several Upper Egyptian sites in great quantities, dating from 3500–3300 BCE. Cylinder seals
are another item providing clear proof of early Western Asiatic influence in Egypt. The
cylinder seal was invented in Mesopotamia around 3750–3550 BCE as an administrative tool
that would be more efficient than the existing stamp seal, because much larger areas of wet
clay could be quickly and continuously marked by rolling the cylinder. The five earliest
cylinder seals in Egypt come from the richest graves in the Naqada region and date from
3500–3300 BCE; and on the basis of their material (limestone), iconography, and style, they
were all imported from Mesopotamia (Boehmer 1974; Honoré 2007).
In this period, the Egyptians were still behind the Uruk people in cultural development, for
instance in their knowledge of metallurgy, and not nearly so well organized. It should have
been possible for a well-armed group of well-traveled Uruk traders—already familiar with the
local conditions, people, and language (at least via interpreters)—to take over Upper Egypt. I
would like to resurrect the suggestion that the first pharaohs were Sumerian or Elamite
invaders, proposed by the eminent Egyptologist Flinders Petrie (1853–1942) and strongly
supported by many of his colleagues until about 1960. It was then generally rejected, because
of the undoubted untenability of its associated ideas: a “Dynastic race” and a sudden invasion
of Egypt via a sea route from Mesopotamia through the Persian Gulf around the Arabian
peninsula and into the Red Sea.
The creation of the Egyptian state in Upper Egypt demanded above all strong leadership,
and Uruk trader-invaders could have introduced Uruk’s effective methods of administration:
cylinder seals and record-keeping with logosyllabic protowriting. It is true that the Egyptian
hieroglyphic script used its own, non-Mesopotamian, signs and language from its beginning at
the end of the fourth millennium; but this may be explained by postulating that the Uruk
rulers left such matters to preexisting local officials, as did the Indo-Aryan kings of Mitanni.
The palaces and royal tombs of Pre-Dynastic Egypt resemble the so-called recess-buttress
architecture typical of Late Uruk culture, which spread widely in West Asia with the Uruk
expansion (Sievertsen 1999). The king’s name was written within the façade of such a
buttressed palace, as depicted in a so-called serekh frame; this frame may have been the
origin of the later cartouche of the classical hieroglyphic script.
We might envision a parallel between prehistoric Uruk traders sailing up the Nile and
taking over power in administratively weak Upper Egypt and the birth of the Russian state in
historic time. By 750 CE, if not before, armed Viking traders were sailing from Scandinavia to
the Black Sea and the Caspian Sea along the rivers of Russia, Belorussia, and the Ukraine.
The Old Russian Nestor’s Chronicle (the work of a monk called Nestor) records that sometime
after 862 the Russians, who lacked laws and a firm government, invited the Vikings to be their
rulers. From this same source we know that these chiefs of Viking origin had Scandinavian
names in the first two generations, but thereafter Russian names. Likewise, adventurers of
Uruk origin in Upper Egypt might have become Egyptianized within a couple of generations.
At least this is what happened to Uruk art objects in Egypt. The superb flint knife from
Gebel el-Arak is without clear provenance but can be dated to 3300–3200 BCE on the basis of
other similar luxury knives with ivory handles found in context in Upper Egypt, all locally
made but decorated with Late Uruk motifs (Sievertsen 1992). The blade made of yellowish
Egyptian chert is of extremely high quality, being considered the most accomplished example
of the silex tool-making technique. The Gebel el-Arak knife is clearly a royal object, made in
Upper Egypt for the king of Upper Egypt. The handle is exquisitely carved in pure Late Uruk
style, by a native Uruk craftsman working in Egypt. It contains both Egyptian and Sumerian
motifs, depicting two kinds of boats, and a combat between almost nude men wearing penis
sheaths. Uppermost on one side of the handle is a bearded man holding back two lions, which
stand up on their hind legs on either side (Fig. 18.1). He wears the long frock and head-band
of the priest-king of the Late Uruk culture, matching exactly the royal dress on a number of
Late Uruk artifacts from Mesopotamia, such as a stone stela from Uruk, where the king is
shown hunting lions (Fig. 18.3). The craftsmanship of the Gebel el-Arak knife handle attests to
the real presence of Sumerians in Upper Egypt, working for the king of Upper Egypt; it is
logical to conclude that the Late Uruk priest-king depicted here is a Pre-Dynastic pharaoh.
Thus this particular “contest” scene—royal motif par excellence—is of extraordinary
importance for understanding the Egyptian state formation and the extent of the Uruk
expansion.
FIGURE 18.3 The Sumerian priest-king hunting lions on a Late Uruk–style basalt stone stela from the Eanna temple
in Uruk, Mesopotamia, c. 3200 BCE. Height 78 cm. The Iraq Museum (IM 23477), Baghdad. After Strommenger
1962: pl. 18. Photo Max Hirmer, © Hirmer Fotoarchiv, Munich.

The Egyptian slate palettes of Late Naqada III times (3200–3100 BCE) are already
Egyptianized in style, but still have Late Uruk motifs. In the famous stone palette of Narmer,
the monsters with intertwined snake-like necks closely parallel a Late Uruk seal. At the
bottom of one side of Narmer’s palette the king is depicted as a wild bull, goring his enemy,
the chief of Lower Egypt; the other side depicts conflict with an enemy, but here the king is
shown in human shape. In another palette, from Abydos, the king is represented as a bull, and
in yet another, the “battle” palette in the British Museum, the king is shown in the shape of a
lion. So the two most feared beasts, the wild bull and the lion, both manifest the royal power.
While the already mentioned (Fig. 18.3) stone stela from Uruk shows the king shooting
lions, a Mesopotamian Late Uruk cylinder seal depicts the king hunting wild bulls (Fig. 18.2).
These depictions convey the message that the king is more powerful than the lion and the bull,
or that he is a match for these kings of the beasts. But the Late Uruk king was not only a
warrior; he was also the high priest, who is shown performing priestly duties.

FIGURE 18.2 The Sumerian priest-king hunting wild bulls on a Late Uruk–style cylinder seal. After Amiet 1980: no.
1614. Courtesy Pierre Amiet.
The Late Uruk depiction of a king holding back two lions continued in Mesopotamia until
at least about 2600 BCE (the Early Dynastic II period). It appears in a plaque from the Inanna
temple of Nippur dating from this period; the two bulls confronting each other at the bottom
are similar to some Harappan representations (Fig. 18.4a). In the Early Dynastic III period the
Mesopotamian “contest” scene continues in the old fashion (Fig. 18.4b), but also changes (Fig.
18.4c), as a result of the influence of Proto-Elamite seals.

FIGURE 18.4 “Contest” with two lions in the Near East of the Early Dynastic II–III periods, c. 2750–2334 BCE. (a) An
ED II (c. 2750–2600 BCE) plaque from Nippur. After Parpola 1994:248, fig. 14.10. Drawn after Pritchard 1969:356
no. 646. (b) An ED III cylinder seal from Susa, southwestern Iran. After Legrain 1921:252. Courtesy Musée du
Louvre/Béatrice André-Salvini. (c) An ED II–III cylinder seal from Fara. After Amiet 1980: no. 893. Courtesy Pierre
Amiet.

In Proto-Elamite art, 3200–2600 BCE, human activities are, as a rule, expressed through
animal actors. The lion and the bull dominate, often shown in the “pose of power.” In a variant
of the man-plus-two-lions “contest” motif, the lion and the bull are shown as adversaries.
Particularly interesting and important are seals in which these two animals are alternately
represented as victorious or defeated, the killer or the killed. On one seal, a standing lion
shoots a seated bull with a bow and arrow, and a standing bull clubs a seated lion (Fig. 19.8).
On two others, an enormous bull masters two small lions and, conversely, an equally
disproportionately big lion dominates two small bulls (Fig. 19.2). Edith Porada (1950) and
Pierre Amiet (1956) have plausibly suggested that the lion and bull are personifications of
cosmic forces, whose alternation between strength and weakness preserves equilibrium in
nature, as manifested in the cycle of day, night, and the seasons. Indeed, the golden-colored
lion may stand for the sun, day, light, heat, and fire; the darker bull with crescent-shaped
horns for the moon, night, darkness, cold, and water. Other dualities, too, seem to have been
ascribed to these two forces: male versus female, heaven versus earth, life versus death; these
will be discussed in the next chapter.
In one Proto-Elamite seal impression, a large bull dominating two small lions is balanced
by a second scene where a man lifts a bull (Fig. 18.5). Amiet (1980:109) has suggested that in
this composition, the bull is the prototype of the “bull-man” Enkidu, who in later
Mesopotamian seals fights lions alongside the naked human-shaped hero Gilgamesh, who in
turn kills the sky-god An’s “Bull of Heaven.” To Amiet’s observation I add the fact that the
human figure (rarely seen in Proto-Elamite seals) has here been substituted for the expected
lion. In Sumerian seals, the human hero has six locks of hair, arranged in three hair curls on
either side of his head, reminding one of the way the lion’s mane is represented on some seals.
Thus, this particular Proto-Elamite seal impression links the Proto-Elamite “contest” between
animal adversaries with the most important motif of the Mesopotamian royal seals of the Early
Dynastic III and Old Akkadian periods.

FIGURE 18.5 A Proto-Elamite “contest” seal with a man lifting a bull. After Amiet 1980: no. 586. Courtesy Pierre
Amiet.

The Harappan seals with the “contest” motif suggest that the Late Uruk–style “contest”
motif continued in Mesopotamia until Early Dynastic III times (2600–2334 BCE) (Fig. 18.4). In
the same period, there are six dots over the head of the hero who holds back two tigers in a
seal from Mohenjo-daro (Fig. 18.6). The dots correspond to the six curling locks of hair in the
“contest” scenes in Early Dynastic III seals, including those of the Fara style (Fig. 18.4c), and
in the Old Akkadian seals of Sargonic times. In the Old Akkadian period (2334–2200 BCE), this
hairstyle was replaced by a conical cap, but toward Neo-Sumerian times (2200–2000 BCE) the
six locks come back.
FIGURE 18.6 “Contest” of a six-locked hero with two tigers on the seal stamp M-308 (DK 11794, ASI 63.10.388)
from Mohenjo-daro. After CISI 1:384. Photo EL, courtesy ASI.

The plaited double-bun or chignon of the Mesopotamian hero is an important indicator of

the period and place of origin of West Asian influence on Mature Harappan glyptics.
Mesopotamian kings wore their hair in this fashion in the Late Early Dynastic period. The
golden helmet of the Sumerian king Meskalamdug from the Royal Cemetery of Ur comes from
the Early Dynastic IIIA period (2600–2500 BCE). The helmet shows “a large encircling braid
holding the hair in place. Perhaps in the Early Dynastic period this braid replaced the fillet …
that binds the hair of earlier rulers in Uruk and Jamdat Nasr representations” (Hansen
2003a:34–35). The same plaited chignon is worn by Eannatum, the Sumerian king of Lagash
(but not by ordinary soldiers) in the battle scene carved on the ED IIIA “Stele of Vultures”
(Fig. 18.7a), and in other depictions of rulers, for example that of Ishqi-Mari, who ruled Mari
in Syria during the ED IIIB period (2500–2334 BC). Sargon the Great (c. 2334–2279 BCE) wears
his hair in this fashion in a stela found at Susa, but by the time of Sargon’s Old Akkadian
successors, the plaited chignon had already changed into something more elaborate.
Indus people visiting Mesopotamia from their native Meluhha must have been deeply
impressed by this royal symbol, given its prominence in their own art and iconography. This
was first noted by Elisabeth C. L. During Caspers (1979:133–135), who refers to a number of
human portrayals on Indus seals and sculptures: “In each case, the hair is dressed in a bun,
which is then secured horizontally, by means of a ribbon or hair-slide of some sort, resulting in
the division of hair into two protuberances one above the other.” Other examples, additional to
those of During Caspers, are as follows. In a cylinder seal from Kalibangan depicting a military
context corresponding to the Sumerian helmets, two warriors, both with the divided chignon
at the back of the head, fight each other with spears while the goddess of war (who is also
shown riding a tiger in this scene) holds them by the hand (Fig. 19.5). In the “fig-deity” seal
from Mohenjo-daro, a human head placed on an offering table in front of the deity is probably
is a warrior’s head since it appears to have the two-parted chignon (Fig. 21.9). Among three
sculptures of human heads from Mohenjo-daro, the male head illustrated here has its hair
gathered at the nape of the neck into a bun, which is divided into two (now-broken) parts by a
horizontal pin connected at both ends with the ribbon encircling the head (Fig. 18.7b–c).
There are examples of the Harappan hero of the “contest” scene also wearing the divided
chignon.
FIGURE 18.7 The “double-bun” headdress. (a) Eannatum, the Sumerian king of Lagash, leading his soldiers on the
Stele of the Vultures (EDIIIA, c. 2600–2500 BCE) from Tell Loh in the Musée du Louvre. After Strommenger 1962:
pl. 66. Photo Max Hirmer, © Hirmer Fotoarchiv, München, courtesy Musée du Louvre/Béatrice André-Salvini. (b–c)
A sandstone sculpture of a male head from Mohenjo-daro in the Mohenjo-daro Museum. After NHK 2000:83, no.
320. Courtesy NHK & NHK Promotions and DAMGP.

In the Sumerian language, this hairstyle was called gú.bar, whence Akkadian gubāru /
gupāru (gú denotes “the nape of the neck”). It is possible that it survived as a warrior’s
hairstyle to the Vedic Vrātya rites, for the war god Rudra wears a hairstyle known as kaparda.
This Sanskrit word is likely to be derived from the Proto-Dravidian root *kavar, “to bifurcate,
be divided into two,” and the Proto-Dravidian noun *kavaram / *kavari, “braided hair,” which
became a loanword in Sanskrit, kabara- / kabarī-, “plaited hair” (first attested in Pāṇini’s
grammar).
In the Old Akkadian royal seals, the bull of the “contest” motif is suddenly replaced by the
water buffalo during the last third of the sixty-year-long rule of Sargon the Great (Fig. 18.8a).
At the end of the Old Akkadian period, the water buffalo disappears from the Mesopotamian
iconography. The water buffalo did not exist in Mesopotamia before this, nor was it
reintroduced to Mesopotamia before the Sasanid period some 2500 years later. The water
buffalo is so well carved in the Akkadian seals that the craftsmen must have seen the animal
alive. In ancient times exotic animals were royal jpgts. It has been plausibly suggested that
Harappans brought some water buffaloes from the Indus Valley by ship, given that Sargon
boasts that ships came from far-off Meluhha to his new capital Akkad. These water buffaloes
would have been kept in his royal park, where they apparently soon died (Boehmer 1975). It is
noteworthy that the Akkadian artists connect the water-buffalo with the water god Enki/Ea:
his servants slake the thirst of the buffalo (Fig. 18.8b).
FIGURE 18.8 The water buffalo in Old Akkadian seals. (a) A hero grappling with a water buffalo victoriously puts his
foot on the head of the beast in a “contest” scene on a seal from Mari in the Damascus Museum. After Parrot 1960:
186 no. 224. Courtesy Musée du Louvre/Béatrice André-Salvini. (b) Enki’s servants (nude heroes with hair in six
locks) slake the thirst of water buffaloes on an Akkadian cylinder seal in the De Clercq collection of the Musée du
Louvre (AO 22303). Inscribed: “Šarkališarri, King of Akkad: Ibnišarrum, the scribe, (is) your servant.” After Collon
1987:124 no. 529. Courtesy Dominique Collon and Musée du Louvre/Béatrice André-Salvini.

FIGURE 18.9 A hero places his foot on the horns of a water buffalo while spearing it in front of an erect cobra. One
side (C) of the molded tablet M-492 (DK 8120, NMI 151) from Mohenjo-daro in the collection of the National
Museum of India, New Delhi. After CISI 3.1:396. Photo EL, courtesy NMI.

A notable new iconographic detail is associated with Akkadian seals depicting water
buffaloes: the hero places his foot upon the conquered enemy (Fig. 18.8a). From now on, this
is the conventional symbol of victory in Mesopotamian art. There are several representations
of buffalo spearing in Harappan seals and tablets, and in each case the hero who kills the
buffalo raises his foot on the head of the beast (Fig. 18.9). In reality, it would be impossible to
put one’s foot on a buffalo’s head while spearing it. So this pose is simply a conventional
symbol of “domination,” like the restraining of two lions or tigers with bare hands. In
Akkadian glyptics the pose is a natural development, less so in some Harappan examples,
which have the appearance of rather clumsy imitations. It is just one addition to the several
motifs already enumerated that the Harappans adopted from West Asia during the period from
Late Early Dynastic to Sargonic times.
In Mesopotamia, the lion and bull had astral counterparts in the zodiacal signs Leo and
Taurus. Sumerian sculptures dating from 3100–2900 and 2100–1900 BCE show a reclining bull
covered with inlaid trefoils (Fig. 18.10a). These “Bulls of Heaven” may be compared with the
cow-shaped Egyptian “Lady of Heaven,” since both the bull and the cow are covered in
trefoils. The trefoil motif resembles the Sumerian pictogram mul, “constellation,” consisting of
three stars placed in a triangle. From Mohenjo-daro comes a fragmentary steatite sculpture of
a bull covered with trefoils (Fig. 18.10b).

FIGURE 18.10 The Bull of Heaven decorated with trefoils and other astral motifs. (a) A Neo-Sumerian steatite bowl
from Ur (U.239), bearing also symbols of the sun and the moon. (b) A fragmentary steatite statuette from Mohenjo-
daro. After Ardeleanu-Jansen 1989:205, fig. 19 and 196, fig. 1. Courtesy © Michael Jansen.
FIGURE 18.11 The “priest-king” statuette from Mohenjo-daro, with a robe decorated with trefoils. After Marshall
1931: III, pl. 98.3. Courtesy ASI.

West Asian gods and divine kings had festive cloths with golden stars, rosettes, and so
forth sewn on to them, which were known in Sumerian and Akkadian as “sky garments.” What
this term meant was apparently stated in the fifth century CE, when the poet Nonnos, writing
in Greek, described a Syrian deity as “clad in a patterned robe like the sky” and “called
Starclad, since by night starry mantles illuminate the sky.” The cloak of the “priest-king”
statue from Mohenjo-daro is clearly a form of “sky garment” decorated with trefoils, originally
filled with red paste (Fig. 18.11). This Indus garment seems to have been transmitted to the
Vedic royal consecration, in which the king is invested with a so-called tārpya garment. The
Vedic texts ascribe this garment to the divine king Varuṇa, who is associated with the night
sky, and state that it had images of sacrificial fireplaces (dhiṣṇya) sewn onto it. In the
Mahābhārata (3,43), the stars are described as heavenly fireplaces (dhiṣṇya) of pious ancient
sages. As discussed in chapter 16, this conception can be traced to early Vedic texts.
What does the Harappan borrowing of the “contest” motif—which contained centrally
important symbols of kingship in the Late Uruk, Proto-Elamite and Old Akkadian cultures—tell
us about kingship in the Indus Valley? When the Harappans adopted this motif, and the royal
“double-bun” headdress and “sky garment” of Mesopotamian rulers, did they retain their royal
associations? In my view, the answer is yes: the Indus people were ruled by kings.
As discussed in chapters 15 and Chapter 16, the Vedic god Varuṇa, the divine king, seems
to have succeeded the Harappan water god, symbolized by the fish, crocodile, and water
buffalo. Buffalo horns connect this water god with the four-faced Proto-Śiva, who seems to be
connected with the sun as the lord of the four directions in the center, and is seated on a
throne (Fig. 16.6). This apparently royal divinity is represented on a Harappan-style cylinder
seal, where he is balanced by a standing hero who holds back tigers (Fig. 18.12). Beneath the
god seated on a throne with bovine legs is a pair of water buffaloes, and he is surrounded by
fish and horned snakes. Beneath the standing god is the markhor goat of the Hindu Kush. In
later Indian mythology the markhor goat is an animal form of the fire god Agni, who in Vedic
texts is equated with Rudra, the divine robber, predecessor of Skanda, the divine warrior in
Hinduism. Exceptionally in Harappan seal iconography, there is also a bird, probably the
vulture, beneath this standing god. In the agnihotra ritual of Vedic India, the sacrificer must
stand up while pouring the offering libation for the rising sun in the morning, but in the
evening, while pouring the offering for the setting sun, he must be seated. If standing and
sitting in the cylinder seal discussed refer to the rising and setting sun, or the day sun and the
night sun (or the moon), the antithetical pair would represent youth and old age, fire and
water, and war/destruction and peace/prosperity. The entire composition reminds one of ideas
connected with kingship in later Indian tradition (cf. Heesterman 1957; 1989): the old king
resides in the capital in the center of his realm and looks after the welfare of his subjects in all
peace, while the crown prince, called yuvarāja, “young king,” or kumāra, “prince,” is in charge
of war expeditions.

FIGURE 18.12 Impression of an unprovenanced Harappan-style cylinder seal in the Département des Antiquités
Orientales, Musée du Louvre (Collection De Clercq 1.26). Photo courtesy Musée du Louvre/Béatrice André-Salvini.

This cylinder seal is to some extent paralleled by an Early Dynastic III chlorite box from
Khafajeh in Mesopotamia, where we see the lion-bull antithesis. A divine hero stands on two
lionesses and masters two snakes (Fig. 18.13a). On another side of the same box, a deity sits
upon two bulls and holds streams of water in his hands (Fig. 18.13b). Both men are marked as
divine by a rosette that appears to function like the determinative for divine names in the
cuneiform script, the star pictogram. Beside the head of the god sitting on the bulls is also the
sickle of the moon. Above the bulls are ears of corn. The sickle of the moon connects the man
upon the bulls with night, which is cool and pleasant like the fertilizing waters. The snakes
conquered by the man standing upon the lionesses are nightly creatures. The day-night
opposition is present also in the standing versus seated posture of the two deities.
FIGURE 18.13 Two sides of a chlorite box carved in the “intercultural style” of eastern Iran, Early Dynastic III
period (c. 2600–2334 BCE), from Khafajeh, Mesopotamia, in the British Museum (Reg. No. 1936,1217.2, BM 12887).
(a) A male deity (marked as such with the star or rosette symbol) standing upright on two lionesses and holding a
snake in both of his hands. (b) A male deity (marked as such with the star or rosette symbol) seated upon two zebu
bulls and holding streams of water in both of his hands. He is surrounded by the sickle of moon and ears of corn.
Photos AN 319934001 and AN 6044191001 © The Trustees of the British Museum.

Chapter 16 argued that the sun was a symbol of the king and the ruling power in the Early
Harappan culture, and that the conceptions associated with the pole star go back to the Indus
civilization: the Vedic texts connect the pole star with the king of the country, with the divine
king Varuṇa and his banyan tree, yet another symbol of royalty. In the present chapter we have
seen that the Indus elite adopted iconographic symbols of royalty from Mesopotamia. The next
chapter argues that the most important royal rituals of West Asia had a counterpart in the
Indus civilization, and that they were continued with some modification in the Vedic rites of
kingship.
19
The Goddess and the Buffalo

In chapter 15 we saw that the principal evidence of the folk religion in the Greater Indus
Valley from 7000 BCE onwards consists of statuettes of human females and bulls. It is fairly
safe to assume that they represent Mother Earth and Father Sky, an archetypal couple who
unite sexually when the roaring sky-bull inseminates the earth with rain. Such ideas appear
throughout the texts of West Asia that throw light on the religion of early agricultural
communities.
A typical example is the religion of Syria in the fourteenth century BCE. Here Ba’al, the god
of thunder and rain, was symbolized by the bull, and Anat, the virgin earth goddess, who
personified nature’s power of reproduction, was symbolized by the cow, the lioness, the
furrow, and the fruit tree. Ba’al—the name means “husband, owner, lord”—was understood to
own the cultivated land, just as a husband owned his wife, for whom he had paid the bride-
price. The cult culminated in their sacred marriage. Ba’al ruled the season of inseminating
rains, fresh vegetation, and growth, whereafter he had to abdicate and descend to the
netherworld. After his death, Anat gave birth to a male calf, the young Ba’al, with whom she
was to unite the next year. Ba’al was succeeded to the throne by Mot, the god of drying wind
and drought, which made everything fade and die, but also ripened the corn. Eventually Anat
furiously attacked Mot and cut off his head as if she were harvesting ripe corn. Ba’al’s return
to the throne coincided with the new-year festival and a further sacred marriage with Anat,
whose virginity was renewed each year.
In all these early religions, the earth goddess gave birth to new life, but also took back the
life she had given: the dead were buried in the earth, her womb. Indeed, the production of
new corn presupposed the sacrifice and burial of seed corn, so birth demanded death.
Correspondingly, living beings, especially seed-laying males, whether human or animal, were
sacrificed to the earth goddess, but these victims were also supposed to be resurrected. The
worship of the earth goddess developed into mystery cults in which not only the goddess but
also her devotees partook of the life-containing blood and flesh of the victim.
Around 6000 BCE the earth goddess is represented at Çatal Höyük in Turkey as a fat
woman who gives birth while seated on a throne flanked by felines (Fig. 19.1). Bulls are
sacrificed in her shrines. Known to the Hittites as Kubaba and to the Greeks and Romans as
Kybele, this ancient Anatolian goddess was in historical times worshipped with an orgiastic
cult that involved the castration of her young lover Attis and also her priests. An important
part of Kybele’s mystic cult was the taurobolium, a sacramental ceremony of cleansing the
worshipper of sins with the blood of a sacrificed bull and eating its testicles.
FIGURE 19.1 Clay figurine of the Great Goddess seated on a throne flanked by felines. Neolithic Çatal Höyük in
Anatolia, c. 6000 BCE. The figurine is on permanent display at the Anatolian Civilization Museum in Ankara. Photo
courtesy © Çatalhöyük Research Project, Stanford University.

The earth goddess also came to preside over killing and bloodshed that took place in war.
From early Neolithic times all over West Asia and the Aegean she was represented by a lion or
some other feline, who safeguards her worshippers as a lioness fiercely guards her cubs.
Lionesses guard the gates of many ancient cities, such as the Hittite capital Boğazköy in
Anatolia, Zincirli and Carchemish in Syria, and Mycenae in Greece. The protective function of
the earth goddess manifests itself, too, in the concept of a “mural crown,” that is, a crown
depicting the (often crenelated) walls of a city. Kybele—who defended Rome against Hannibal
—wears such a crown, while the goddess of the city of Mari in Syria was known as “mistress of
the city walls.”
In chapter 18 we have seen that in Proto-Elamite art (c. 3200–2600 BCE) the lion and the
bull are represented as each other’s adversaries. Thus on one seal from Susa, an enormous
bull masters two small lions, and a giant lion dominates two small bulls (Fig. 19.2). They have
been interpreted as representatives of opposite and complementary cosmic forces, which
appear to have included heaven and earth, male and female. Next to the lion there is a
triangular pictogram similar to the Sumerian sign KI, which denotes “earth.” The pictogram
occurs on other seals, where a lioness is represented, Atlas-like, upholding with her paws tree-
clad mountains, apparently as a chthonic deity (Fig. 19.3). According to the Babylonian “Epic
of Creation,” after the separation of heaven and earth, the sky-god An was the father of the
natural deities, the earth-goddess Ki their mother.
FIGURE 19.2 A Proto-Elamite seal impression from Susa, with lions and wild bulls dominating each other in turn.
The script sign KI flanks the lion on either side. After Amiet 1980: no. 585. Courtesy Pierre Amiet.

FIGURE 19.3 Impression of a Proto-Elamite seal, showing lionesses upholding mountains and trees, and the script
sign KI. After Amiet 1980: no. 577. Courtesy Pierre Amiet.

The principal Mesopotamian goddess, called in Sumerian Inanna and in Akkadian Ištar,
was symbolized by both the lion and the planet Venus. As the evening star she presided over
sexual love and prostitutes. As the morning star, she was a bloodthirsty warrior about to begin
the battle, standing by her princely favorites. The fertility of vegetation, animals, and humans
was believed to depend upon the union of Inanna and Dumuzi (Akkadian Tammuz), her young
shepherd lover. Dumuzi’s death was ritually lamented. (According to one variant of the myth,
when Inanna visited the underworld ruled by her sister, she could not return to the earth
without a substitute, so demons kidnapped Dumuzi to replace Inanna. However, Dumuzi was
allowed to return to the earth for six months every year.) Dumuzi was identified with the king,
but the sacred marriage was enacted by priestly personnel. This was the most important
festival, celebrated on the new year’s day, with processions, dance, music, and feasting.
Turning now to the Indus civilization, we find evidence for the presence of key elements of
a West Asian kind of cult of the earth goddess. A seal from Chanhu-daro depicts a “sacred
marriage”: a bison bull with exposed long penis is about to mount a prostrate human female
with an elaborate headdress whose legs are apart and vulva is visible (Fig. 19.4). Importantly,
the groom is an animal (bull) and the bride a woman (priestess?), as with the stallion and chief
queen in the Vedic horse sacrifice discussed in chapter 12, where the early Indo-Aryans (of the
Atharvavedic wave) seem to have substituted their principal sacrificial animal (the horse) for
the victim of the Harappan tradition (chiefly the water-buffalo).
FIGURE 19.4 Intercourse of a bison bull and an anthropomorphic priestess. Impression of a lost Indus seal
excavated at Chanhu-daro in Sindh (cf. Mackay 1943: pl. 51, no. 13). Exposure 3148 of the old print B7392 in the
Museum of Fine Arts, Boston. Courtesy Museum of Fine Arts, Boston.

The Harappans also had a tiger-riding goddess of war, depicted on a cylinder seal from
Kalibangan (Fig. 19.5). Mesopotamian influence is visible in the type of the seal and in the
hairstyle of the seal’s warriors, which sports a band forming a “double bun,” as worn by
princes in West Asia around 2400 BCE. But the goddess herself looks native: she has a long
skirt and long plaited hair, and holds the hands of two warriors who are spearing each other.
Next to this group she is depicted again, now with the body of a tiger and wearing an
elaborate crown with the horns of the markhor goat topped by a branch of acacia. In this
latter form she also appears on a Harappan-type square seal, found in the same house as the
cylinder seal. The inscriptions on these two seals are very similar, but not identical.

FIGURE 19.5 A tiger-riding goddess wearing the horns of a mountain goat under an acacia tree. She is also shown,
without her mount, as holding by hand two warriors spearing each other. The warriors wear their hair in a two-
parted chignon at the nape of the neck. There are two signs of the Indus script, one representing the numeral three.
Modern impression of the cylinder seal K-65 (KLB2-9734, ASI 68.1.77) from Kalibangan. After CISI 1:311. Photo EL,
courtesy ASI.
The West Asian sacrifice of the bull is also paralleled in the Indus civilization. The sacrifice
of a water buffalo bull by spearing is represented on an Indus seal and on several Harappan
tablets. We have seen in chapter 18 that the Harappan water buffalo, imported from the Indus
Valley, is the counterpart of the West Asian urus bull in the Sargonid “contest” seals, and that
the “victor’s pose,” one foot placed on the head of the buffalo, is replicated in Mesopotamia
and in the Indus Valley (Figs. 18.8a, 18.9). The buffalo sacrifice also takes place in front of an
erect cobra (Fig. 18.9); according to Vedic texts, “the Queen of the snakes (sārparājñī) is the
earth” (PB 4,9,6).
In the ancient West Asia and Mediterranean, notably in Minoan Crete, acrobatic bull-
baiting has been an integral part of the cult; the sport continues in Spain today. Some Indus
seals depict people (in one case clearly women) jumping over a water buffalo (Fig. 19.6). Two
bulls facing each other as if in fight are also depicted (H-1997 B). Some tablets from Harappa
show men grappling with bison or buffalo bulls, holding them from horns (H-2023 C to H-2030
C). According to seven Old Tamil poems describing life in the pastoral landscape, Kalittokai
101–107, herding communities regularly arranged festive contests of bull grappling (ēru-kōḷ or
ēru-taluvu), where wild bulls with sharpened horns were let loose in the arena, and young men
showed their physical prowess in order to win girls in marriage; the poems describe how they
could be gored or trampled to death. Bull baiting is still popular in some parts of south India,
especially Tamil Nadu, where it is currently known as erutu-k-kaṭṭu or calli-k-kaṭṭu.

FIGURE 19.6 Jumping over the water buffalo in Indus seals. (a) The seal M-312 (DK 8321, NMI 147) from Mohenjo-
daro in the collection of the National Museum of India, New Delhi. After CISI 1:385. Photo EL, courtesy NMI. (b) An
Indus seal from Banawali. Photo courtesy ASI.

Hindu India reveres a fight between the lion or tiger and the buffalo in the battle between
the goddess Durgā and the buffalo demon, Mahiṣa Asura, a ferocious warrior. The beautiful
granite panel of the Pallava dynasty (c. 700 CE) at Māmallapuram in Tamil Nadu depicts the
lion-riding goddess Durgā shooting her opponent with bow and arrows, while the buffalo
demon Mahiṣa Asura has the club as his weapon (Fig. 19.7). This composition offers a striking
parallel to the Proto-Elamite seal where the lion kills the bull by shooting it with bow and
arrows, while the bull clubs the lion with a mace (Fig. 19.8).
FIGURE 19.7 The lion-riding goddess Durgā shoots the buffalo-headed Mahiṣa Asura who wields a club. Rock panel
at Māmallapuram, Tamil Nadu, South India, eighth century CE. After Kramrisch 1954: pl. 86. Research inconclusive:
copyright Philadelphia Museum of Art may apply.

FIGURE 19.8 A Proto-Elamite seal in which an upright-standing lion shoots a seated bull and an upright-standing
bull clubs a seated lion. After Amiet 1980: no. 591. Courtesy Pierre Amiet.

But is this comparison really warranted? It is generally thought that the goddess Durgā
and her cult did not exist in South Asia until the Iranian-speaking Kuṣāṇas introduced it from
Afghanistan in the first century CE. In my view, however, as I shall now discuss, although the
name Durgā is indeed not attested in early sources, the Vedic texts do refer to a similar
goddess of war and victory, associated with the earth, the lioness, and the buffalo: the goddess
Vāc, who figures prominently in the Vrātya rite of mahāvrata (chapter 12). Moreover, I argue
for a Vrātya origin of the cult of Durgā and therewith of the worship of the Goddess, with
bloody sacrifices, even of human victims, and sexual intercourse in which all taboos might be
broken: the so-called left-hand Śākta Tantrism. (In “right-hand” worship these practices are
replaced with harmless symbols, for example, cutting a water-melon instead of a head, using
colored water instead of blood.)
The earliest images of Durgā as Mahiṣāsuramardinī, “crusher of the buffalo demon,” date
from early Kuṣāṇa times, around 100 CE. The earliest, and most important, full account of the
Goddess’s victory over the demon is that of chapters 2–3 in the Devī-Māhātmya, which
appears in the Mārkaṇḍeya-Purāṇa. Here the beautiful virgin goddess is created from the
united fiery energy representing the fury of the main gods Brahmā, Viṣṇu, and Śiva, while the
other gods provided her with a range of weapons for the annihilation of Mahiṣa; she also
drinks wine to increase her fury for the fight. During the battle, the demon changes himself
from a buffalo first into a lion, then into a man (a warrior), then an elephant, then back into a
buffalo, and a man. In the Devī-Bhāgavata-Purāṇa (5,18,2–70), the demon first has a human
form, then successively changes into a lion, an elephant, a śarabha, and a buffalo. In the Śiva-
Purāṇa (5,46,48–55), the forms assumed by the demon are those of a buffalo, a lion, an
elephant, and a man.
Comparing the available stone iconography with the descriptions in these texts, Heinrich
von Stietencron (1983) has proposed that none of the textual accounts can predate the fifth
century CE. The earliest icons of the Kuṣāṇa period, where the Goddess embraces the buffalo
and breaks its back, have no textual counterpart. The Devī-Māhātmya states that the Goddess
jumped onto the back of the buffalo demon, which corresponds to the images at Elephanta,
Ellora, and Aihole dated to the sixth and seventh centuries CE; but another detail of the fight
mentioned in the Devī-Māhātmya (and the Vāmana-Purāṇa)—the demon’s final emergence
from the buffalo’s mouth in human form—is not known from iconography before the eighth
century.
Von Stietencron sees a development of the myth in the texts, conditioned by the growing
bhakti movement of the first millennium CE. In the Devī-Māhātmya (and in the Śiva- and
Vāmana-Purāṇa), the buffalo demon is only a straightforward enemy of the Goddess. In later
Purāṇas, though, the demon is driven to battle not by enmity but by passion for the Goddess.
The Goddess, however, is unable to accept the lowly demon as her husband. In the latest
versions of this textual history Mahiṣa’s love is so great that he chooses death at the hand of
his beloved rather than life without her. The demon’s bhakti evokes pity and grace in the
Goddess. She offers him friendship if he will return to the netherworld, but this does not
satisfy the demon. He is slain by the Goddess, and thereby attains liberation.
In the Devī-Purāṇa, Mahiṣa had been in love with the Goddess in a previous incarnation in
the form of the demon called Dundubhi (“Drum”), who had been burnt to ashes by the jealous
Śiva. According to an Orissan tradition, at his death Mahisa could kiss the female organ of the
Goddess, becoming liberated through this intimate contact with her. In the Kālikā-Purāṇa (c.
1000 CE), the love is finally legitimized, as Mahiṣa is identified with Śiva. Now the motive for
Mahiṣa’s death must also change, as he is no longer striving for something forbidden; instead,
Śiva is said to have been cursed by an enraged sage.
Von Stietencron’s hypothesis explains the evolution of the Goddess and Mahiṣa myths of
the Purāṇa texts well, conforming to the history of Indian religions during the corresponding
period. Even so, it does not necessarily follow that the theme of love between the Goddess and
the buffalo demon is as recent as the sixth century CE; it may have been latent in the earlier
tradition. Indeed, David Shulman (1980) has argued for an opposite development in the Tamil
texts of South India: that the myth of divine marriage is central to the Tamil religion, and the
murder of the male spouse by the virginal bride—for the sake of the god’s rebirth—is its most
essential element. Brahmanical ideology later imported from North India has tampered with
the texts and changed the myths, in particular eliminating the idea of the god’s death and
projecting it onto the Goddess’s fight with demons.
Old Tamil literature speaks of a goddess of war, Kotti or Kottavai, who gives victory
(kottam), lives in a great forest, and dances the tuṇaṅkai dance. She is generally considered to
be a native Dravidian divinity, although the Tamil epic Cilappatikāram (probably composed
between 450 and 750 CE), identifies her with the north Indian goddess Durgā
Mahiṣāsuramardinī. Her name Kotti is homonymous with South Dravidian kotti, “cat”; both
words are probably derived from Proto-Dravidian *kol, “to kill.” Kotti’s connection with the
tuṇaṅkai dance appears to me very significant. Tuṇaṅkai is danced on the battlefield by men,
women, and ghosts, who beat their sides with their arms bent at the elbows. The derivation of
tuṇaṅkai from the Proto-Dravidian root *tuḷaṅku, “to shake, sway, be agitated, move violently to
and fro or up and down,” proposed in the Tamil lexicon, seems to suit the situation in the
battlefield as an expression of “grief, sorrow” (Tamil tuḷakku) for the dead. Several rituals of
Vrātya affinity (discussed in chapter 12), including an archaic funeral and the horse sacrifice,
include a dance around the deceased, in which the dancers beat their thighs.
The forms of the demon slain by the Goddess consist of the males of various powerful
animals. In Old Tamil poems, the tiger, elephant, and wild buffalo are used as similes to
portray the anger and courage of heroes; a special theme of Old Tamil poetics is erumai-
maram, comparing the ferocious attack of a hero to the charging of a wild buffalo. The buffalo,
lion, śarabha, and man, mentioned as metamorphoses of Mahiṣa Asura in his fight with the
Goddess, figure high in the list of sacrificial victims in Kālikā-Purāṇa 71,6–18. Here it is
specified which particular animal will give pleasure to the Goddess for which length of time—
which is also the length of time that she will bless the sacrificer of this particular animal
(Table 19.1). The identity of the śarabha, which occupies a prominent place in this list, is
problematic and will be discussed in chapter 20.
TABLE 19.1

victim time
fish, tortoise 1 month
crocodile 2 months
nine species of wild animals 9 months
wild bull, guana 1 year
black antelope, wild boar 12 years
śarabha 25 years
buffalo, rhinoceros, tiger 100 years
lion, śarabha, man 1000 years
three men 100,000 years

The different sacrificial victims offered to the Goddess represent the metamorphoses of
the valiant male, who on the one hand has been killed by the Goddess and at the same time is
her late husband or suitor. The victims of the Vedic Vrātya rituals similarly represent the
metamorphoses of the creator god Prajāpati, who is the father and husband of Vāc, in my
opinion Durgā’s predecessor as the goddess of victory and fertility. He is killed on account of
his illicit pairing with Vāc. Prajāpati is expressly said to be the “lord/husband of (the goddess)
speech” (vācaspati) (ŚB 5,1,1,16). In Kaṭha-Saṁhitā 12,5 the creation is described in the
following manner: “In the beginning Prajāpati (alone) was this (whole universe). His Voice
(vāc-, fem.) became a second (being). He paired with her. She conceived an embryo. She went
away from him. She delivered these creatures.” Rigveda 5,42,13 speaks of the creator god
Tvaṣṭar, “who, wanton in his own daughter’s womb, changing forms made this (form) of ours.”
A more extensive version is in Bṛhad-Āraṇyaka-Upaniṣad 1,4:

1. In the beginning this world was self (ātman) alone in the form of a man (puruṣa) … 3.
. . . He desired a second. He was, indeed, as large as a woman and a man closely
embraced. He caused that self to fall (pat) into two pieces. Therefrom arose a husband
(pati) and a wife (patnī). . . . He copulated with her. Therefrom human beings were
produced. 4. And she then bethought herself: “How now does he copulate with me after
he has produced me just from himself? Come, let me hide myself.” She became a cow.
He became a bull. With her he did indeed copulate. Then cattle were born. She became
a mare, he a stallion. She became a female ass, he a male ass; with her he copulated.
Thence were born solid-hoofed animals. She became a she-goat, he a he-goat; she a
ewe, he a ram. With her he did verily copulate. Therefrom were born goats and sheep.
Thus, indeed, he created all, whatever pairs there are, even down to the ants.
(Translated Hume 1931:81, slightly modified)

Here the males and females of all the animal species are different manifestations of just one
and the same primeval pair, Prajāpati and Vāc. This idea seems to be reflected in the
iconography of the Indus civilization, too, for a number of Harappan seals and amulets portray
fantastic animals, whose body parts belong to different species. Thus, a composite animal has
the horns of a zebu bull, a human face, the trunk and tusks of an elephant, the forequarters of
a ram, the hindquarters of a tiger, and a snake for the tail. These mythical creatures may
express the idea that all the species represented by the different parts belong to just one and
the same being, who can assume different appearances.
In the Vādhula-Anvākhyāna (4,16) the primeval couple is Prajāpati and Vāc, but Prajāpati
also has the sacrifice (yajña-), that is, the sacrificial victim, as his alter ego. “Prajāpati created
(emitted, poured out) the sacrifice (yajña-, masc.); after it had been created, it went away from
him.” (The expression parāṅ ait, “went away,” usually refers to dying; in ŚB 13,2,5,1, it is used
of the pouring out of the sap of life—apparently blood—of the sacrificial horse just
slaughtered.)

He (the sacrifice, the male victim) wished to pair with … the (goddess) Speech (vāc-,
fem.) (in the form of) the anuṣṭubh verse … he united with her. She became a mare
(aśvā-), he a stallion (aśva-); she became a wild ass mare (gaurī-), he a wild ass stallion
(gaura-); she a bison cow (gavayī-), he a bison bull (gavaya-); she a female camel (uṣṭrī-
), he a male camel (uṣṭra-); she a she-goat (ajā-), he a he-goat (basta-); she a female
śarabha (śarabhī-), he a male śarabha (śarabha-); she a she-antelope (eṇī-), he a black
buck (kṛṣṇa-); she a red female nilgau (rohit-), he a male nilgau (ṛśya-). Thus they went
through becoming all the beings, both of them taking one form after another.

Then Vāc went into the man (puruṣa) and “made defence walls (puraḥ) (against him),
namely the teeth.” Then Prajāpati took the form of the porridge (odana) sacred to Vāc, which
he knew Vāc desired. To enhance her lust, he decorated himself with melted butter. Vāc could
not resist him; she removed the defense and swallowed him, thus finally becoming united with
him.
In this variant, Vāc possesses defensive walls (puraḥ) like Durgā as the goddess of the
fortress (durga), and the West Asian goddesses like Kybele wearing a mural crown. Moreover,
it is stated that Prajāpati and Vāc copulated in the form of all beings (sarvāni bhūtāni).
Aitareya-Āraṇyaka 5,1,5 speaks of the mating of many animal couples (bhūtānāṃ ca
maithunam) at the mahāvrata. Mating of animal couples takes place in popular festivals in
India even today: thus a bull and cow that copulate may represent the marriage of Rāma and
Sītā. From the Vādhūla-Śrautasūtra (11,19,25–30) it appears that at the Vedic horse sacrifice
there was full sexual license for all the human participants, as in the Śābarotsava of Durgā’s
“tenth day of victory.”
In the variant of the Vādhūla-Anvākhyāna, the enumeration of the animals stops at the pair
of the nilgai antelope, rohit and ṛśya. (The “unicorn” bull, which is so prominent in the
Harappan iconography, seems to be a cross between the wild urus bull—borrowed as
“unicorn” from Mesopotamia, but not present in the Indus Valley fauna—and the native nilgai
antelope, which represents the creator god in these Vedic myths: Figs. 19.9, 19.10.) This is the
only pair mentioned in a number of other, clearly abbreviated, versions dealing with
Prajāpati’s incest with his own daughter, among them Aitareya-Brāhmaṇa 3,33, which states
that Prajāpati was killed in punishment:

Prajāpati desired his own daughter. . . . Having become a nilgai bull he approached her
who had become a nilgai cow. The gods saw him and said: “Prajāpati is doing a deed
that is not done.” They sought someone to punish him; they did not find such a person
among themselves. They united their most terrible manifestations. Those became This
Deity [a euphemism for the dreaded god Rudra]. . . . The gods said to him: “This
Prajāpati here has done a deed that is not done; pierce him (with an arrow).” “Be it so,”
he (Rudra) replied. … He aimed and pierced him …

FIGURE 19.9 Nilgai, “the blue bull” (ṛśya in Sanskrit). After Roberts 1977:176, ill. 52. Courtesy F. Roberts and Ruth
W. A. Sloan.
FIGURE 19.10 An antelope-like “unicorn” on an Indus seal Nd-1 (Exc. Branch 1187) from Nindowari. After CISI
2:419. Photo AV, courtesy DAMGP.

Rudra’s birth from the most terrible manifestations of the various gods in order to kill
Prajāpati is closely paralleled in the Devi-Māhātmya by the birth of Devī from the combined
forces of the gods so as to dethrone the buffalo demon on the tenth day. While in the
mythology of the Goddess it is the virgin (kumārī) goddess herself in her warrior manifestation
who kills the demon that has dared to approach her, in the Vedic myth it is her male
counterpart, the youthful (kumāra) war god Rudra, born at the instance of the incestuous
union of the sinful father and his daughter. There is an epic version of the Mahiṣa myth, in
which the buffalo demon is killed by the war god Skanda, the Hindu successor of Vedic Rudra.
Skanda’s birth myth and metronym Kārttikeya connect him with the autumn, when the buffalo
is sacrificed in Durgāpūjā.
In the Purāṇas, a lion- or tiger-riding young goddess kills a demon in the shape of a male
buffalo. In Kerala, there is a myth of a tiger-riding young male god (Ayyappan) who kills a
demoness having the form of a she-buffalo (mahiṣī). This Keralan myth underlines the capacity
of a divinity to assume either a male or a female appearance. The lion or tiger and the buffalo
are both symbols for the fearless and destructive warrior, who personifies war, as does the
goddess of victory. Among different Semitic tribes, the war god was either male, Aštar, or
female, Aštart. Based on the myth variants discussed in the preceding paragraph, I would like
to argue that a similar relationship prevails between the male war god Rudra/Skanda, and the
female goddess of war, Durgā, and her doubles, such as Sarasvatī. In Hinduism, these deities
share a number of common attributes (such as the peacock and the harp), epithets, and myths.
They are youthful and virginal, the male one being called Kumāra, the female Kumārī; these
epithets/names imply also sexual chastity (brahmacaryā) required in learning wisdom that also
characterizes these deities. The Goddess rides a tiger, and in the Veda it is Rudra who is called
“the lord of the wild animals” (araṇyānāṃ pati, TS 4,5,3e), an epithet otherwise given to the
tiger only (cf. AB 8,6 kṣatraṃ vā etad araṇyānāṃ paśūnāṃ yad vyāghraḥ).
From the Brāhmaṇa texts it appears not only that Prajāpati assumed the forms of the
various animals in creating them but also that he was sacrificed in these forms. Thus the
vegetarian offering is explained in Śatapatha-Brāhmaṇa 2,1,3,6–9 to be a substitute for an
original bloody sacrifice, starting with the human sacrifice:

6. At first … the gods offered up a man as the victim. When he was offered up, the
sacrificial essence went out of him. It entered into the horse. They offered up the horse.
When it was offered up, the sacrificial essence went out of it. It entered into the ox.
They offered up the ox. When it was offered up, the sacrificial essence went out of it. It
entered into the sheep. They offered up the sheep. When it was offered up, the
sacrificial essence went out of it. It entered into the goat. They offered up the goat.
When it was offered up, the sacrificial essence went out of it. 7. It entered into this
earth. They searched for it, by digging. They found it (in the shape of) those two
(substances), the rice and barley. . . . 9. The man (puruṣa) whom they had offered up
became a bard (kiṃpuruṣa). These two, the horse and the ox, which they had sacrificed,
became a wild ass and a bison respectively. The sheep which they had sacrificed,
became a camel. The goat which they had sacrificed, became a śarabha. For this reason
one should not eat (the flesh) of these animals, for these animals are deprived of the
sacrificial essence (are impure). (Translated Eggeling 1882:I,50–52, modified)

Here two series of animals are correlated, the first being pure animals, which may be
sacrificed, and the second impure animals, which are unfit for sacrifice. (For an interpretion of
kiṃpuruṣaas a “bard”, see chapter 14.) The five pure animals are those whose heads are
placed in the lowest layer of the fire altar (agniciti). The mantras that are employed in placing
these heads also mention the impure animals, here characterized as “belonging to the forest”
(araṇya), that is, wild, while the pure animals by implication belong to the village, that is, are
tame (Table 19.2).
TABLE 19.2

tame wild
1. puruṣa 1. kiṃpuruṣa/mayu
2. aśva 2. gaura
3. go/vṛṣabha 3. gavaya
4. avi/vṛṣṇi 4. uṣṭra/meṣa
5. aja/basta 5. śarabha

The worship of Durgā (durgāpūjā) takes place especially in the autumn, during Dussehra,
from Hindi daśahrā (< *daśāharaka), “festival lasting ten days.” It consists of the “nine-day
festival” (nava-rātra or nava-rātri) followed by “the tenth day of victory” (vijaya-daśamī), which
celebrates Durgā’s victory over Mahiṣa. Everything culminates in the festival celebrated in the
manner of Śabaras (understood to denote “wild tribals,” but another explanation is offered in
chapter 20) on the tenth day, when the Goddess is “dismissed” and the king lustrates his army
before marching to battle. Hindu kings have traditionally started their military operations
when the roads have dried up after the rainy season and there is abundant fodder and grain
available.
Another festival, the now less-known vasanta-navarātri, also connected with the worship of
Durgā, is celebrated around the vernal equinox. In Nepal the goddesses of war called “Nine
Durgās” are worshipped from the autumnal dasaīṃ (< daśamī) festival to the end of May, but
not during the rainy season. The two Durgā festivals of the year may originally have had the
same function as the Vedic vrātyastoma rites. One vrātyastoma was celebrated when the
Vrātyas set out on a raiding expedition, confirming their mutual covenant; another
vrātyastoma was performed on their return, to purify them of the violent deeds they had
committed during the raids. If this was the case, the year was divided into two halves of
violence and peace: the inauspicious dakṣiṇāyana (the sun’s southern course) of raiding, living
outside the village, in the wilderness; and the auspicous uttarāyaṇa of agriculture (the sun’s
northern course), living in the village.
This hypothesis would explain the two series of sacrificial animals, tame and wild, used
respectively during the two festivals celebrating the two diametrically opposed turning points
of the year. When the fire altar is completed (at mahāvrata) and the tame animals are
sacrificed, the wild ones are released. At the horse sacrifice, however, all possible kinds of
wild animals are sacrificed as subsidiary victims. For example, Maitrāyaṇī Saṃhitā 3,14,10 and
Vājasaneyi-Saṁhitā 24,28 state:

To the Lord (Īśāna, i.e., Rudra) he sacrifices rhinoceroses (parasvatāḥ),

to Mitra wild asses (gaurān),
to Varuṇa water buffaloes (mahiṣān),
to Bṛhaspati bisons (gavayān),
to Tvaṣṭar camels (uṣṭrān).

At the autumnal equinox the sacred couple was father (Prajāpati) and daughter (Vāc), who
seem to represent the dying “old year” and the virginal “new year”: their union coincides with
the death of the old and birth of the new year/sun. At the vernal equinox the “sacred
marriage” would also be incestuous but with an inversion of the roles: son and mother. The
dying groom would be the “young sun” born out of the preceding marriage, that is, the leader
of the vrātya sodality who had been raiding during the past six months, that is, the Vedic god
Rudra personified by a war bard (kiṃpuruṣa, māgadha), and the bride would be the chief
(seniormost) queen (mahiṣī, “buffalo cow,” symbolizing Mother Earth). This reconstruction fits
the reasonable assumption (presented in chapter 16) that the earliest Indian calendar
consisted of the opposing asterisms rohiṇī and jyeṣṭhā, associated respectively with a young
virgin girl just coming of age, and “the eldest lady,” representing respectively the goddess of
welfare and goddess of misfortune.
I suggest that the two Durgā festivals correspond to the two great festivals of the Vedic
sacrificial year, the mahāvrata and the viṣuvat. Mahāvrata is always preceded by a ten-day rite
(daśarātra)—equal to Dussehra—and in the year-long sacrificial session this combination
occurs only once, at the end of the yearly cycle. It may be that originally the mahāvrata
coincided with the tenth day, for the tenth day is sacred to the creator god Prajāpati; riddles
about the creation of the world are followed by an abuse of the evil deeds of Prajāpati, that is,
his incest with his daughter resulting in his death. The Vedic texts contain statements
indicating that a human victim was once sacrificed on this day and that his flesh was eaten by
the participants (Falk 1986:37–42)—as was done on the “tenth day of victory” until the British
period in Karnataka—by men leaving for raids (Silva 1955).
I argue that the mahāvrata and viṣuvat originally had the same function as the Durgā
festivals, and that they too were celebrated at the autumnal and vernal equinox, respectively.
Most scholars have connected the mahāvrata with the winter or summer solstice, and indeed
in several Vedic texts the mahāvrata and the middle day of the year, viṣuvat, clearly denote the
solstices. However, some facts suggest that, earlier, the mahāvrata celebrated the autumnal
equinox. For one thing, the Pleiades constitute the first asterism of the oldest nakṣatra lists,
and hence their conjunction with the full moon marked the beginning of the year, and several
of the individual star names of this asterism relate to rain. Secondly, one characteristic action
of the mahāvrata day is the sounding of different musical instruments (chapter 12), so that “all
(manner of) voices (i.e., music) resound” (sarvā vāco vadanti, TS 7,5,9,3; PB 5,5,20).
According to Taittirīya-Brāhmaṇa 1,8,4,2 this takes place in the rainy season (prāvṛṣi sarvā
vāco vadanti), and the rainy season lasts from about the middle of July to the middle of
September. The Taittirīya-Brāhmaṇa passage speaks of rites connected with the regular
plundering tours of the Vedic Kuru and Pañcāla tribes. While the mahāvrata concludes the
year-long sacrifice, the mid-point of the year is the viṣuvat day, and viṣuvat or viṣu denotes the
“vernal equinox” in many Indian languages and is celebrated today as such in many parts of
India. Finally, thus the yearly cycle of the sun comes to correspond with the daily cycle
(sunrise–noon–sunset : new year at vernal equinox–summer solstice–autumnal equinox, see
chapter 16).
The Kālikā-Purāṇa describes the celebration of the śābarotsava in chapters 60–61 (in
another version 62–63). I quote the text at length to enable comparison with the celebration of
the Vrātya rite mahāvrata described in chapter 12, with Goddess Vāc as its central divinity,
and with the spring festival of Holī celebrating the death and resurrection of the love god
Kāma, which in chapter 15 was found to be related to the south Indian buffalo sacrifices:

31–32. Afterwards the Goddess was dismissed with Śabara-festivals, on the tenth day …
42–43. A king should hold a lustration of the army …; a performance must be made
with charming women … ; 43–48. One should worship the Supporteress of the worlds
with dance, song, play, festivity and benediction; with sweetmeat, cakes, drinks, foods
… with parched grain . . . and various kinds of meat, spiritous liquors … with various
sorts of wild animals which belong to the class of offerings … with flesh, blood … 50–52
. . . . The eight beautiful Yoginis … are also to be worshipped, and also the sixty-four
Yoginīs, and the ten-million Yoginīs; the beautiful nine Durgās, … and Jayantī, etc., …
because they are the Goddess’s shapes. 52–53. One should always worship all weapons
of the goddess Mahiṣāsuramardinī and in the same way her ornaments … ; and also the
lion, her mount, with a view to obtaining success. (Translated van Kooij 1972:109–111)

17–23. On the tenth day in Śravaṇa one should dismiss her [the Goddess] with Śābara-
festivals. … People should be engaged in amorous plays with single women, young
girls, courtesans and dancers, amidst the sounds of horns and instruments, and with
drums and kettle-drums, with flags and various sorts of cloths, covered with a
miscellany of parched grain and flowers; by throwing dust and mud; with auspicious
ceremonies for fun; by mentioning the female and male organs, with songs on the male
and female organs, and with words for the female and male organs, until they have
enough of it. If one is not derided by others, if one does deride others, the Goddess will
be angry with him and utter a very dreadful curse. (Translated van Kooij 1972:121)

The śābarotsava includes feasting with meat of sacrificial victims and with alcohol, obscene
abuses, and sexual intercourse. I am convinced that the notorious “circular worship” (cakra-
pūjā) of left-hand Śākta Tantrism is derived from a military festival such as the śābarotsava
(see chapter 20). In the cakra-pūjā, the goddess in the center of a ritual circle may be
represented by a nude virgin surrounded by couples including “heroes” (vīra) and their mates,
who may belong to any caste or even be their own mothers or daughters: they indulge in
eating meat, fish, and cereals, drinking alcohol, and sexual intercourse: these are the famous
“five m’s” (pañca makārāḥ), denoted with five Sanskrit words beginning with m- (madya,
“intoxicant”; māṃsa, “flesh”; matsya, “fish”; mudrā, “seal-impressed wafer”; and maithuna,
“sexual intercourse”). The Tantric cult also mentions circles with an ithyphallic male god in
the center surrounded by “Mothers,” such as the harp-playing god Tumburu (identified with
Bhairava, Rudra Tryambaka, Śiva, etc.) surrounded by four goddesses connected with
“victory” through their names, Jayā, Vijayā, Ajitā, and Aparājitā; the “mothers” work
themselves up to the state of the utmost revelry, dancing, singing, laughing, mocking people,
drinking, and devouring quantities of raw flesh.
Even the nude virgin of the Śākta Tantric cakrapūjā has a counterpart in the Vedic horse
sacrifice. When the four wives of the sacrificing king are led to the victim who is being slain, a
“virgin” (kumārī) follows them as the fifth (ŚB 13,5,2,1). The chief queen lies with the horse in
the center, while the other wives and their companions go around them, beating their thighs.
The archaic Vādhūla-Śrautasūtra (10,20) details the role of the virgin:

Then he (the adhvaryu priest) gives her (the chief queen, mahiṣī) a blade of grass
broken at both ends; she wipes with it that (part of her body) where the male member
of the stallion was placed. They lead to the place this virgin (ānayanty etāṃ kumārīm). .
. . She (the queen) should throw that blade of grass upon this girl, saying: “I pierce you
with the burning of the sexual union” (mithunasya tvā śucā vidhyāmīti). She is called
Sāhā (“mighty, vanquishing”) and can from now on leap forth into the (men’s) hall
(sāhety āhur īśvarā sabhāṃ praskand[it]or) which according to the commentary means
that she becomes a prostitute (sabhāṃ praveṣṭuṃ samarthā, puṃścalī syād iti matam).

The prostitute who accompanies the vrātyas on their raids and performs the sexual
intercourse with a māgadha bard or a student (brahmacārī) evidently represents goddess Vāc,
whose double is goddess Sarasvatī, also called Vāg-īśvarī, the patron goddess of the
courtesans and the fine arts in later Hinduism. In Aitareya-Brāhmaṇa 1,27, the goddess Vāc is
“a great naked lady” (mahānagnī) with whom the gods purchase Soma from the Gandharvas
“desirous of women.”
The Vedic texts speak of goddess Vāc as an “invincible” goddess of victory associated with
the lion. Vāc is represented by the chief wife (mahiṣī) of the sacrificing king, and both are
identified with the earth (ŚB 6,5,3,1–4; 7,4,2,32–39). The queen (and earth) is fertilized by the
male partner in the “sacred marriage” of the horse sacrifice. Thus the sacrificial victim is the
“husband” of the goddess Vāc. This is clear also from Śatapatha-Brāhmaṇa 3,2,1,18–26, where
the Brahmanical author confesses that originally the goddess Vāc belonged to the Asuras, that
is, “demon” worshippers speaking an Indo-Aryan dialect similar to the Māgadhī Prakrit of
eastern India; however, Vāc was taken away from them by the gods, that is, Vedic Aryans
worshipping Devas. The last section of this passage runs:

25. That Yajña (sacrifice) lusted after Vāc (speech), thinking, “May I pair with her!” He
united with her. 26. Indra then thought within himself, “Surely a great monster will
spring from this union of Yajña and Vāc: (I must take care) lest it should get the better
of me.” Indra himself then became an embryo and entered into that union. (Translated
Eggeling 1885:II,30–32).

This suggests that Indra, the Vedic war god, was afraid of the birth of the war god of the Asura
mythology, that is, Rudra, who according to the Aitareya-Brāhmaṇa (3,33) was born at the
moment Prajāpati united with Vāc.
In the Purāṇic ritual, the Goddess is offered the sacrificial victim’s blood and head
immediately after the decapitation. According to Śatapatha-Brāhmaṇa 3,5,1,22–23 and
3,5,2,9ff., Vāc, as the goddess of victory, went over to the side of the gods on the condition
that the offering should reach her before it reached the sacrificial fire: therefore they pour
ghee on the high altar (uttaravedi), which represents Vāc, with the mantra “Thou art a lioness,
Hail!” (VS 5,12). Taittirīya-Saṁhitā 6,2,7,3, in dealing with the construction of the uttara-vedi,
states: “ ‘Thou art a lioness; thou art a buffalo cow’ (simhīr asi mahiṣīr asi), he says, for it (i.e.,
the high altar) taking the form of a lioness went away and remained between the two parties.”
The two parties referred to are those of the gods (deva) and the demons (asura), who were
contending with each other. In Taittirīya-Saṁhitā 6,2,8,1 this is continued: “The high altar
said, ‘Through me ye shall obtain all your desires’. The gods desired, ‘Let us overcome the
Asuras our foes.’ They sacrificed (with the words), ‘Thou art a lioness, overcoming rivals; hail!
(simhīr asi sapatnasāhī svāhā)’ They overcame the Asuras, their foes” (translated Keith
1914:II,509; the two formulae are also given in the mantra collection in TS 1,2,12,2 as e and
k).
The buffalo sacrifice to the Goddess is the cultic counterpart of the Mahiṣāsuramardinī
myth, and really central to the Durgā cult. “The sacrifice of the buffalo in India is offered
exclusively to the Goddess, either at the feast of her temple in the beginning of the spring, or
during the Navarātri,” says Madeleine Biardeau (1981:215, translated). In Nepal, the royal
cult of the living virginal goddess Kumārī has until quite recently involved the sacrifice of
hundreds of water buffaloes. The water buffalo is native to India and was sacrificed in the
Indus civilization, while the horse, first imported to South Asia by the Indo-Aryans, was their
most prestigious animal. It is understandable that in Vedic India the horse sacrifice came to be
the king of the sacrifices.
According to Rigveda 1,162,10–15, the flesh of the sacrificial horse was cooked and
roasted, and eaten with great pleasure. But we are surprised to read in RV 6,17,11 that
“(Agni) cooked a hundred buffaloes” to Indra; RV 5,29,7–8 states that Agni cooked 300
buffaloes for his friend Indra upon his wish; “and after Indra had eaten the flesh of 300
buffaloes and drunk 3 oceans full of soma, all the gods shouted ‘Victory!’ to him … because he
had killed Vrtra.” In RV 8,12,8 Indra gains his real strength only after he has eaten a thousand
buffaloes. Indra’s feasting on buffalo meat increased his warring powers: this sacrificial feast
was therefore performed just before leaving for a warring expedition. It seems evident that for
a short while after their entrance to South Asia, some Rigvedic Aryans adopted the local way
of celebrating the military feast.
However, every trace of those enormous buffalo sacrifices to Indra has been wiped out
from the Brāhmaṇa texts, and with the sole exception of the subsidiary buffalo sacrifice to
Varuṇa in connection with the horse sacrifice, there is no later Vedic buffalo sacrifice. This
suggests a deliberate discrimination against the buffalo by the conservative circles of the
Vedic priesthood. There is some concrete evidence that sides were taken for and against the
horse and the buffalo as sacrificial victims. The Brahmanical discrimination of the buffalo is
matched by the express prohibition of ever offering horses to the Goddess in the Kālikā-Purāṇa
(71,46), which prescribes (71,57–58) the following mantra to be addressed to the buffalo when
it is decapitated: “As you hate a horse and as you carry Caṇḍikā, so kill my enemies and bring
happiness, O buffalo …” (translated Kane 1955:V.1,167).
In the Rigveda, Indra has adopted even other elements of the earlier local religion, among
them the magic ability to assume different shapes, including shapes of different animals
(usually this māyā of the Asuras belongs to Varuṇa). Most significant among Indra’s
metamorphoses is that referred to in RV 1,121,2b: “as a water buffalo (mahiṣa) he desired the
lusty female born from himself.” This replicates Prajāpati’s incestuous approach to his own
daughter (AB 3,33) and reveals that early on Prajāpati was conceived to have the shape of a
water buffalo.
The Vedic texts prescribe a human sacrifice (puruṣamedha), identical with the aśvamedha,
save the chief victim. This is thought to be theoretical priestly extension, but Vādhūla-
Śrautasūtra 11,21 mentions as the carver of the sacrificed horse a virginal youth (kumāra),
son of a celebrated bard (sūta), who was garlanded and lamented as if he were to die, because
in former times he himself was to be cut up. This makes a striking parallel with the
lamentation of the dying bridegroom of the Goddess in the West Asian mystery cults. This
sacrificed youth seems to have personified the divine Kumāra, whose name Rudra the
Brāhmaṇa texts connect with the verb rud-, “to cry, lament”; originally the name is likely to
have been Rudh(i)ra, “red,” referring both to the rising sun and to the planet Mars, as do
many names of the Old Tamil war god Murugan “youth,” such as Cēy, “the red one.” The
human victim represented the Primeval Man (puruṣa) out of whose severed limbs the world
came into being according to RV 10,90,6–16. Sacrifice of a human warrior to the Goddess is
implied by the Mahiṣa Asura myth, and indeed south Indian heroes are supposed even to have
cut off their own heads in order to secure the Goddess’s favor and victory to their king. The
severed head of a warrior, recognizable from the “double-bun” hair-braid at the nape of the
neck (chapter 18), is placed on an offering-table in front of a deity inside a fig tree, whose
worship is depicted on a famous Indus seal; one sign of its inscription can be interpreted as
the Dravidian name of the rohiṇī nakṣatra, the very earliest new-year star connected with the
goddess of victory (chapter 21, Fig. 21.9).
20
Early Iranians and “Left-Hand” Tantrism

The previous chapter argued that the Harappans worshipped the earth goddess using
symbols, rituals, and conceptions shared with West Asia: the Goddess presides over fertility,
death, and war, is represented by a feline, and undergoes a sacred marriage at the new-year
festival, involving the death and rebirth of the bridegroom, represented by a bull or some
other male, including a human. I also presented Vedic evidence for the continuity of this
Harappan Goddess worship in the Vrātya rituals of the Indo-Aryan speakers of the early
Atharvavedic wave, coming from the BMAC culture of southern Central Asia around 2000–
1700 BCE. A further suggestion of mine is that the Vrātya rituals largely form the basis of the
Śākta Tantric worship of the Goddess that begins to emerge in Hindu and Buddhist sources
around the middle of the first millennium CE, at the same time as the earliest Purāṇic accounts
of the myth of the warrior goddess Durgā and her victory over the buffalo demon.
These texts were written later than the first images of the goddess Mahiṣāsuramardinī,
which date from first centuries CE and come from the Kuṣāṇa-dominated region of Mathurā.
The Kuṣāṇas conquered Bactria (northern Afghanistan) around 150 BCE and adopted there the
local Middle Iranian language called Bactrian for administrative purposes. Their own, little-
known language was different, however. The Kuṣāṇas descended from the nomadic tribe called
Yuezhi in Chinese sources, whose original homeland was the eastern end of the Takla Makan
desert in the Tarim basin of Xinjiang; tribes called Wusun and Hsiungnu pushed them to the
steppes of Central Asia, from where they came to Bactria after a few decades. In the late first
century BCE the Kuṣāṇas extended their core area to Gandhāra, which extends from the Kabul
Valley to the Swat Valley. From this base they ruled the northern plains of South Asia up to
Kashmir and the Ganges-Yamuna doab for the first three centuries CE, at times even as far as
Magadha. The Kuṣāṇa kings favored Buddhism but also honored Greek and Iranian deities.
Their coins were first in Greek and then in Bactrian; Sanskrit was used in inscriptions, but
their own texts are written in Greek, in Bactrian, and in an “unknown” script and language in
one case (Vima Takhtu in Dasht-e Nawur) (Harry Falk, personal communication 2014).
In Afghanistan the Kuṣāṇas adopted, among other local divinities, the goddess called Nana
and Nanaya, who is depicted as riding the lion and wearing the mural crown (Fig. 20.1). On
the basis of her iconography and name, many scholars have derived Nana from West Asia. A
goddess called Na-na-a(-a) is mentioned in cuneiform documents as a goddess whose
character is very similar to that of Inanna, being described as follows in cuneiform tablets
found in the Temple of Marduk at Babylon: “Lady of ladies, goddess of goddesses, directress
of mankind, mistress of the spirits of heaven, possessor of sovereign power; the light of
heaven and earth, daughter of the Moon God, ruler of weapons, arbitress of battles; goddess
of love; the power over princes and over the sceptre of kings” (translated L. W. King quoted in
Rosenfield 1967:85).
FIGURE 20.1 The goddess Nanaya wearing a mural crown and seated on a feline, holding the sun and the moon in
her back pair of hands. A Khwarezmian seventh-century silver bowl in the British Museum, diameter 12.7 cm. After
Azarpay 1976:540, fig. 6. Courtesy Guitty Azarpay.

Nana first appears in the Mesopotamian lists of gods in the Ur III period (c. 2100–2000
BCE), classified in the same section as Inanna. Daniel Potts (2001) plausibly suggests that
Nana came to Bactria with the lapis lazuli trade by the end of the third millennium BCE,
probably via Susa, where this goddess apparently had a temple around 2300 BCE and where
her worship continued until Hellenistic times. Besides, many BMAC objects have been found
at Susa. BMAC seals from Afghanistan with a winged goddess escorted by lions very probably
depict Nana, nearly two millennia before the Kuṣāṇas (Fig. 20.2).

FIGURE 20.2 A lion-escorted and winged goddess with a Mesopotamian-type kaunakes skirt on a compartmented
BMAC seal made of gold, c. 1900 BCE. (a) The compartmented stamp side on the obverse. (b) The reverse. After
Ligabue & Salvatori [1989]:202, pl. 58–59. Courtesy Centro Studi Ricerche Ligabue, Venice.

Almost two millennia after the Kuṣāṇas, local people worship “Bībī Nānī” at Hinglāj, in Las
Bela district of Baluchistan. The name comes from Sanskrit hiṅgulāja, “vermilion, cinnabar,”
the red paste applied on the body and on images by Hindus. Hinglāj is the westernmost of the
Śākta pīṭhas, famous temples of the Hindu Goddess, where she is worshipped as the feline-
faced manifestation of Durgā, known as Carcā or Carcikā, a name that might derive from
Proto-Dravidian *car-, “tear, rend, cut with teeth.” The Kuṣāṇas seem to have taken the
worship of Nana to the Kulu Valley in Himachal Pradesh and to Nainital in the Kumaon
foothills of Uttarakhand: both places have a famous temple of Goddess Nainā.
Nana’s mural crown connects her with Kybele and Mesopotamian Ištar, known at Mari in
Syria as “mistress of the city wall.” Durgā’s name is supposed to come from the name of a
demon slain by her called Durga; the texts call her durga-ghnā, “remover of adversity or
danger.” As an adjective, dur-ga- means “difficult to go to, unapproachable,” and as a noun
“inaccessible place, stronghold, citadel, fortress” (as first attested in Rigveda 5,34,7), as well
as “adversity, danger.” “Fortress” seems the preferable meaning because Durgā’s
manifestations have several other names associated with “fortress.”
Most notable among Durgā’s “fortress” names is Tripurā, “goddess of the triple fort,” or
Tripura-Sundarī, “the beauty of Tripura.” Hindu mythology understands Tripura as “three
cities” of the asuras (demons) vanquished by Śiva, but the concept seems to go back to the
BMAC “temple-forts” such as that of Dashly-3 in northern Afghanistan, from around 1900 BCE,
with three concentric circular walls within a square fort surrounded by a moat (Fig. 20.3a).
The tradition of building such temple-forts continued in Afghanistan until Achaemenid times
(Fig. 20.3b), thus embracing the time when the Rigvedic Aryans had confrontations with the
Dāsas. In his study of the Rigvedic passages mentioning Dāsa citadels (pur-), Wilhelm Rau
concluded that “the evidence does not fit the cities of the Indus civilization. It rather suggests
the existence of numerous, frequently concentric, mud or stone ramparts of round or oval
ground plan” (Rau 1976:52). This type of fort seems to have come to the BMAC from the
Sintashta culture of southern Urals, which had fortified settlements with concentric circular
walls (see Fig. 7.4 and chapters 7 and Chapter 8).

FIGURE 20.3 Fortress with triple concentric walls in Afghanistan. (a) The BMAC “temple-fortress” of Dashly-3 in
Bactria, northern Afghanistan, c. 1900 BCE. After Sarianidi 1986:59. Courtesy Margiana Archaeological Expedition.
(b) The fortress of Kutlug Depe, Achaemenid period (559–330 BCE). After Sarianidi 1986:73. Courtesy Margiana
Archaeological Expedition.

The Dāsa fortresses are several times described as “autumnal” in the Rigveda (RV
1,131,4b and 1,174,2b = 6,20,10c puraḥ . . . śāradīḥ). The exact meaning of this epithet has
remained unclear. My own suggestion is that the fortresses were venues for the Goddess’s
autumn festival. In Kashmir, Durgā is for this very reason called Śāradā (“autumnal”), and one
of the principal sites of her worship is a fortified place called Śār(a)dī. In present-day Nepal,
where the Goddess is the guardian of the country, the navarātri festival is always celebrated in
a fortress.
The historical connection of the BMAC tripura with later Tantrism is strongly endorsed by
the layout of the “palace” of Dashly-3 (Fig. 20.4). This is practically identical with the later
Tantric maṇḍala in its square form that represents the king’s palace or fortified city with a T-
shaped gate in the middle of each side. It symbolizes the residence of the supreme ruler of the
whole world extending to all four directions. This cosmogram, centered on an image of the
divinity, has been widely used as a means of worship and reintegration (Fig. 20.5). It must
have been transmitted through the centuries in the form of ephemeral maṇḍala paintings
made with colored powders for ritual purposes. Ritual powder paintings are found all over
South Asia and spread from there to Tibet. In historical times this particular maṇḍala
cosmogram first reappears in the so-called “TLV-mirrors” of Han-dynasty China (c. 200 BCE),
which, with their T-, L-, and V-shaped projections, replicate even more closely the Dashly-3
“palace” with its external corridors.

FIGURE 20.4 The ground plan of the BMAC “palace” of Dashly-3 in north Afghanistan, c. 1900 BCE. After Sarianidi
1986:53. Courtesy Margiana Archaeological Expedition.
FIGURE 20.5 The Hindu Tantric maṇḍala of Mahā-Kālī, the triangle in the center symbolizing the female organ.
After Preston 1980:65, fig. 2. Courtesy James Preston.

The autumnal worship of the Goddess culminates on the final (tenth) day of the “festival of
Śabaras.” Since the 1920s, śabara- has been connected with Śambara, the proper name of one
of the principal Dāsas, whose fortifications were broken by Indra, as mentioned in the Rigveda
(chapter 9). Both words have generally been considered to be of non-Aryan, probably Austro-
Asiatic, origin, mainly because the ethnic name Sora or Saora of an Austro-Asiatic speaking
tribe in Orissa derives from Sanskrit śabara. It is true that Śabara, when first attested (in AB
7,18), is the name of one of the Dasyu tribes (along with Andhra, Puṇḍra, Pulinda, and Mūtiba)
who live beyond the eastern borders of the Vedic realm. However, determining the ethnic
affinity of the people called Śabara in Vedic times from the fact that Austro-Asiatic speakers
are called Sora or Saora today, is tantamount to claiming that the Mleccha people, who
likewise were met by the Vedic Aryans on their eastern border, were Muslims because the
Muslims were later called Mleccha. Yet the Śatapatha-Brāhmaṇa (3,2,1,18–24) makes it plain
that those early Mlecchas spoke an Indo-Aryan language sharing characteristics with the later
Māgadhī Prakrit of eastern India. It is significant that people called Śabara were labeled
Dasyu (which is another appellation of the Dāsas in the Rigveda), and that they were
encountered by the Vedic Aryans in eastern India; the Mleccha parallel suggests that they
were Aryan speakers.
Śambara as an enemy of Indra is also known from the Mahābhārata, and as the name of a
fierce Tantric Buddhist deity, derived from the Indian Śaiva Tantric tradition. In both cases the
name Śambara has a variant form Saṃvara, which allows a good Aryan etymology, suggesting
a Māgadhī-like sound change s > ś. The first part is the prefix sam-, “together, fully,
completely,” often appearing in the reduced grade *sṃ- > sa-; the second part, vara- means
“enclosing, protecting” as well has “obstructing, hindering” (in RV 1,143,5, with reference to
an enemy army), from the root var-, “to surround, enclose, protect, ward off.” In the neuter
gender, śambara- is used to mean “wall, rampart, fortification.” In this sense Sanskrit
śambara-/saṃ-vara- and (saṃ-)varaṇa- correspond to Avestan vara-, “fortress,” which is the
word used to describe the castle built by Yima, the primeval king of the Iranian tradition. That
vara- also existed in the Iranian Saka language spoken on the Eurasiatic steppes can be seen
from its borrowing into Hungarian, where var means “fort, citadel.” Dāsa Śambara seems to
have been the Dāsa counterpart of Indra as the god of war, close to the Avestan deity of war
Ham-varti-, personification of “manly courage.”
This conclusion warrants a more general consideration of the Dāsas and their proposed
connection with eastern India. It is very likely that when Old Iranian (Saka)-speaking Dāsas
entered Afghanistan around 1500 BCE, as suggested in chapter 9, they also entered the
northern plains of South Asia, like the Kuṣāṇas and Mughals long after them. This is suggested
also by the presence of tribal names cognate with Dasa/Dāsa, known from the Upper Indus
Valley (Sanskrit dāsamīya and dāśamīya, “non-Brahmanical tribe”), the western Indus Valley
(Lahnda ḍahā, “a tribe of Jaṭs,” ḍāhrā, “a division of the Kerār tribe”), Sindh (Sindhi ḍāharīand
ḍāhirī, “name of a Sindhi tribe”), and Marwar (Sanskrit daśeraka and dāśeraka, “name of the
people of Maru”; cf. Dāseraka, “name of a people”). The alternation of the dental and palatal
sibilant in the above tribal names is attested in the name Dāsa- itself: besides Rigvedic dāsa-,
“name of enemy people” and “slave,” there is dāśa-, “name of a mixed caste” (Manu) and
“servant” (VS 30,16). In the meaning “son of a slave girl, bastard,” we have in Sanskrit also
both dāsera- and dāśera-, and dāseya- and dāśeya-.
The sound change s > ś (which may be due to Dravidian substratum influence)
characterized the late Prakrit of northern Sindh called Vrācaḍa (this name is derived from the
Sanskrit word vrātya-) as well as the Gāndhārī and Māgadhī Prakrits. Gāndhārī (spoken in
Gandhāra) and Māgadhī (spoken in Magadha in eastern India) have also some other features
in common, notably the change *az > e (instead of *az > o in Sanskrit), which is attested once
in the Rigveda. The testimonies of the Dāsa language in the Rigveda are very scarce, and it
therefore seems significant that three features connect it with Māgadhī: s > ś, v > b, and the
presence of the phoneme l (which in Proto-Rigvedic and Avestan was replaced by r). This
suggests considerable Dāsa influence upon the earliest Indo-Aryan-speaking immigrants in
eastern India, who may be connected with the Ochre Coloured Pottery (OCP) or Copper
Hoards culture of the Upper Ganges Valley in the early second millennium BCE (chapter 8).
Śākta Tantrism has been particularly strong in eastern India, and it may have been taken
there by Harappan and early Indo-Aryan immigrants. In the Vedic realm the Vrātya religion
was more or less extinguished by orthodox Brahmins. The Brahmins of the “Aryan Midland”
were horrified by the “immoral” customs in marginal areas, such as Gandhāra and the western
Punjab: their festivals resembling the śābarotsava are described with strong disapproval in
the eighth book of the Mahābhārata.
The name Śambara/Saṃvara of the Tantric deity thus seems to be an Iranian (Dāsa)
contribution to Goddess worship in South Asia. He is often called Cakra-Saṃvara from the fact
that his square maṇḍala contains three concentric circles (cakra) (Fig. 20.6). This replicates
the Dashly-3 type tripura, “triple fort” (Fig. 20.3a), which, as argued above, also seems to be a
Dāsa import, reinforcing Śambara’s suggested origin. This terrifying Tantric god has come to
Buddhism from the Śaiva circles, in particular from the “skull-bearing” Kāpālika sect
notorious for its antinomian practices. A Mahiṣa-Saṃvara is worshipped in Nepal, and
Śambara is identified with the fierce god Heruka, whose name may come from the Proto-
Dravidian word *ēru, “bull, male of any animal remarkable for its physical strength, male
buffalo”—in several Sanskrit loanwords from Dravidian the initial h- represents the glottal
stop that in Central and North Dravidian reinforces a word-initial vowel. (The same
phenomenon is found in Hindi: I have heard a Hindi speaker pronouncing English “empty” as
hempty.) Some Tantric texts speak of herukas as demonical beings, who according to Tibetan
and Chinese commentaries are “blood drinkers.”
FIGURE 20.6 The Tantric maṇḍala of the four-faced and twelve-armed god Śambara embraced by his consort Vajra-
Varāhī and surrounded in the four directions by four other fierce goddesses. After Raghu Vira and Lokesh Chandra
1995: fig. 65. Courtesy Prof. Lokesh Chandra.

The Tantric skull cult, at least the use of the skull as a drinking bowl for wine and blood,
may be a contribution of the Iranian Dāsas. The word used for “skull-cup” in Tantric and
Purāṇic texts is Sanskrit kapāla-, attested in the meaning “skull” in the Atharvaveda, and
cognate with Middle and Modern Persian kabāra, kabārag, “bowl”; it is suspected to come
from the lost original language of the BMAC. The Vedic Hiraṇyakeśi-Gṛhyasūtra (2,3,7) speaks
of demons “drinking out of skulls” (kapāla-pa-).
According to Herodotus (4,64), the Scythian warrior always cut off the heads of the
enemies he had slain and carried them to his king, because the heads alone entitled him to a
share of the booty. His esteem also depended upon the number of enemies he had killed, and
this he exhibited by hanging their scalps from his bridle-rein, or making himself a cloak by
sewing many scalps together. Others covered their quivers with the skin flayed from the right
arm of the enemy. Herodotus devotes a full chapter (4,65) to describing how the Scythians
made enemy skulls into drinking cups, which were handed around when they had any worthy
visitors, the host telling its story. Moreover, once a year the ruler of each district

mingles a bowl of wine, of which all Scythians have a right to drink by whom foes have
been slain; while they who have slain no enemy are not allowed to taste of the bowl, but
sit aloof in disgrace. No greater shame than this can happen to them. Such as have
slain a very large number of foes, have two cups instead of one, and drink from both.
(Herodotus 4,66; translated Rawlinson [1860] 1942:316)

The Nuristani languages of northeastern Afghanistan seem to represent the mixed Iranian and
Indo-Aryan language spoken by a group of people who very early departed from the
mainstream and migrated into those valleys where they have in relative isolation preserved
some very archaic linguistic and cultural features. Thus the depalatalized affricates
reconstructed for the Iranian protolanguage are attested in appropriate words only in
Nuristani languages (chapter 9), while some other linguistic features connect these languages
with early Indo-Aryan. The former “Kāfir” (pagan) religion of the Nuristani speakers is likely
to have preserved elements that may go back to around the middle second millennium BCE, as
do those aspects of the Dāsa religion that we may understand from the scanty Rigvedic
references.
Among the Ashkun Kāfirs (on the upper reaches of the Alingar River), a man was worth
nothing until he had killed an enemy. The more people he had killed, the higher his rank,
which was exhibited by various symbols. Specific titles were earned after four, eight, and
twelve kills, as were garments of honor with embroidered ornamentation and bells hanging
from the belt and trousers, and so on. A man with four kills was allowed to erect in the main
gathering place of the village a post in which each kill was recorded with a willow twig put
through a hole, the top being decorated with a red cloth. This social ranking system led to
regular head-hunting expeditions; the trophies, especially severed heads, scalps, or ears, were
brought back to the village in a triumphal procession and exhibited during the victory feast.
We have an eyewitness account of how the Kāfirs celebrated immediately before leaving
for a war expedition:

[feasting and dancing is] kept up with great spirit, until about midnight, when on a
given signal, the lights are suddenly extinguished; the men rush on the women; and
each man seizes the hand of the nearest female, or one whom he may have selected
beforehand, if he can manage to approach her in the scuffle which now ensues. He then
takes her away to some private place and retains her until the morning. On these
occasions it makes very little difference who the fair one is, whether his own wife or
that of another—his own daughter or sister or another’s. (Raverty 1859:353–354)

Sexual promiscuity ensured that as many women of the tribe as possible became pregnant, so
that the tribe produced new men, in case all the elders should perish on the war path. This
should be how the Tantric cakrapūjā discussed in chapter 19 originated.
The most important Kāfir goddess, Disani, resembles the Hindu Durgā, for she is also
“known to excite sexual desire, the prerequisite of procreation. … Disani is at the same time a
deity of death … who takes the deceased home into the house of the Great Mother … the
women pray to her when they fear for the lives of their men who participate in a war-party. …
She is armed when in human form, and allegedly carries a bow and quiver. In another story,
she kills with a dagger” (Jettmar 1986:69–70).
Georg Morgenstierne (1953:164) has suggested that Disani’s name might be
etymologically related to that of the obscure Vedic goddess Dhiṣaṇā, but the origin and exact
meaning of both words are debated. I propose connecting Dizani—the form of the more
archaic Kāmdeshi and Urtsun dialects of Kati—with Old Persian didā- from earlier *dizā-,
“wall, palisade, fortress,” Bactrian dizā, “fortress,” and Middle and Modern Persian diz,
“fortress,” with cognates in Sanskrit dehī-, “wall” (in RV 6,47,2 used of Śambara’s walls), and
Greek teîkhos, “wall.”
This etymology agrees with that proposed above for Durgā’s name. In a Kati song Disani is
actually said to have made a four-cornered golden fortress (kuṭ) with seven doors. According
to the Prasun Kāfirs, again, goddess Disani constructed a tower from which seven lanes lead
outwards. The seven gates in Disani’s fortress can be compared to Indra’s seven daughters,
including Disani in Ashkun Kāfir mythology. Other Kāfirs worshipped seven daughters of the
highest god Imra (whose name comes from Yama-rāja), alias Māra (see chapter 12); Disani
was born out of the creator god Imra’s breast and was the wife of Gish, the fierce god of war.
The seven daughters of Kāfir Māra compare with the daughters of Māra, the god of death,
who tempted the Buddha on the verge of his enlightenment. In Hinduism they have a
counterpart in the “Seven Mothers” of the Hindu war god Skanda, who follow him on war
expeditions; these represent the stars of the Pleiades, mythically the wives of the Seven Sages
(chapter 16).
Disani is also said to assume the appearance of a markhor goat. The markhor is
prominently depicted on Indus seals, and the Harappan goddess of war with tiger vehicle on
the Kalibangan cylinder seal wears the horns of a markhor goat (Fig. 19.5). The important “fig
deity” seal from Mohenjo-daro shows a human-faced markhor behind the priest(ess?) who
kneels and raises hands in prayer to the deity (probably a predecessor of Durgā) inside a fig
tree, while a warrior’s severed head has been placed upon an altar table (Fig. 21.9).
On the basis of the cognates in (Nuristani) Kati, in the Dardic languages of the northwest
and in Sindhi, I have suggested that śarabha originally denoted “markhor” (Capra falconeri),
the wild goat of the northwestern mountains. In the Vedic texts discussed in chapter 19,
śarabha is the wild counterpart of the domestic goat. The meaning “markhor” also fits verse
54 of Kālidāsa’s Meghadūta, where śarabha is a proud animal of the Himalayas, jumping high
into the air at the risk of breaking its limbs. Elsewhere in classical Sanskrit literature śarabha
denotes “a fabulous kind of deer with eight legs, which was supposed to kill elephants and
lions.” Most Sanskrit speakers presumably never set eyes on a real markhor because they
lived in areas outside the animal’s geographical range. Cognates of Sanskrit śarabha have
survived also in the Neo-Indo-Aryan languages of the north (Kashmiri, Panjabi, Pahari,
Kumaoni, Nepali, and dialects of Hindi) as well as in the Sinhalese, but in these words the
meaning has mutated into another jumping creature, a locust or grasshopper (Turner
1966:714–715). And yet the Kālikā-Purāṇa, written as far east as Assam around 1000 CE,
places the śarabha on the same level as the water buffalo in its list of sacrificial victims to the
Goddess, second only to man, as we saw in chapter 19.
In this chapter I have argued among other things that a goddess related to Hindu Durgā
was worshipped in Afghanistan from at least as early as 2000 BCE, then by the Iranian-
speaking Dāsas who arrived there around 1500 BCE, and then again by the Iranian-speaking
Kuṣāṇas who arrived more than a thousand years later. The Iranians seem to have contributed
to the Goddess worship of South Asia by introducing generally abhorrent practices that
nevertheless make sense in a military context: sexual promiscuity before a war expedition,
head-hunting, and use of skulls as bowls for drinking wine and blood. However, similarly gory
and revolting elements characteristic of “left-hand” Tantrism undoubtedly existed in South
Asia even before it received this Iranian reinforcement. The West Asian cults of goddesses of
war and fertility, such as Inanna-Ištar, are full of cruelty and sexuality, which seems to be
replicated in Harappan art. Old Tamil poems, too, describe bloody sacrifices to the goddess of
war on the battlefield, and amorous pleasures of warriors returning from their victorious
expeditions.
21
Religion in the Indus Script

The iconography of Harappan seals and molded and engraved tablets, motifs of painted
pottery, terracotta figurines, stone sculptures, buildings, and city planning are important
sources of Harappan religion. The foregoing chapters have endeavored to interpret their
imagery with the help of better-documented parallels, especially in South Asia and West Asia,
connected with the Indus civilization by cultural continuity and trade relations. Yet to be
considered are the signs of the Indus script. Though an undeciphered system of writing, it is
uniquely important in having the potential to tell us the actual names used by the Harappans
of their gods, and thus to verify assumptions regarding cultural continuity with later times and
religions. For, in the other earliest civilizations, writing served religion and administration,
and names of gods were often mentioned, whether independently or as components of human
proper names.
The decipherment of a logosyllabic script without bilinguals presents a tougher challenge
than that of a syllabic or alphabetic script. Syllabaries and alphabets form closed systems that
include the entire phonology of a language; in favorable circumstances they can be fully
deciphered with the help of just a few external phonetic clues. Logosyllabic scripts have many
more signs and variables than syllabaries and alphabets; the phonetic links between their
signs are weaker. With minimal marking of grammatical forms, and a presumably more
complex syllabic structure, logosyllabic scripts offer no opportunity to build phonetic “grids”
of the kind utilized in the decipherment of the Linear B. In that case, sign groups that were
likely to be proper nouns with three different case endings were identified from their context
in Linear B tablets, and the signs of these hypothesized case endings, probably standing for
open syllables, were tabulated according to their shared consonants or vowels in rows and
columns (the grid). The proper nouns were then guessed to be the names of Cretan towns,
such as Knossos (ko-no-so), which provided the initial syllabic values from which other syllabic
values and readings could be deduced and applied to wholly unfamiliar sign groups. (One
Indus script sign occurs in eleven out of the seventy-four inscriptions excavated at Chanhu-
daro, but nowhere else, five times alone forming a second line. It may therefore express the
Harappan name of the town. For an interpretation connecting it with the town of Pātāla, which
according to Hellenistic sources stood near the Indus estuary, see Parpola 1975b; 2003:558;
Parpola & Janhunen 2011:84–90.)
On the assumption that the Indus script is logosyllabic, a complete phonetic decipherment
is certainly not possible with the presently available inscriptions. We can hope to decipher
only a few Indus signs by assuming that the underlying language is Proto-Dravidian. But to
reach even this limited goal we need valid methods.
Most early script signs were originally pictures denoting the objects or ideas they
represented. But while this representation works well enough for concepts such as “tree” or
“eat,” abstract concepts such as “life” are more difficult to express pictorially. These were
expressed by extending the meaning ascribed to a pictogram or ideogram from the word for
the depicted object to comprise all the word’s homophones, that is, words having a similar
phonetic shape but a different meaning. Thus the English phrase, “to be or not to be,” might
be written using the following pictograms: 2 bee oar knot 2 bee. This phonetic functioning of
pictograms in early logosyllabic scripts is called a rebus (Latin, “by means of things”).
Ideally, the homophony in such punning should be exact. However, there are too few exact
homophones in any language to yield an effective writing system. So, if we were to write
English in rebus form, a picture of a “sun” might conceivably be used to represent the concept
“son,” a near-homophone of “sun.” In the Egyptian and Semitic writing systems, vowel
differences are not written, making near-homophony easier to use. But this is not so in
Dravidian scripts, which do write vowels. In deciphering the Indus script we should therefore
demand near-exact consonant and vowel homophony, generally allowing only those differences
that occur commonly within Dravidian etyma, such as a difference in the length of a given
consonant or vowel, for example, kaṇ and kaṇṇu, “eye,” and kāṇ, “to see.”
In the Sumerian script, the problem of representing “life” was solved by drawing a picture
of an arrow. This sign denotes not only “arrow” but also “life” and “rib,” because the same
sound, ti, has three different, homophonous, meanings in the Sumerian language. Puns such
as this must surely have been employed in oral folklore long before they appeared in Sumerian
writing. But because they are language-specific, their point is lost when they are translated
into another language. The biblical myth of Eve’s creation out of Adam’s “rib” makes sense
only in the light of its Sumerian origin. In the Sumerian paradise myth, the rib of the sick and
dying water god Enki is healed by the goddess Nin-ti, the “Mistress of life.” The punning
connection in Sumerian between ti (“rib”) and ti (“life”) does not exist in Hebrew between
ṣelāc (“rib”) and Eve’s name Ḥawwā (although the latter word is still connected with the
concept of “life,” according to Genesis 3,20: “Adam called his wife’s name Eve; because she
was the mother of all living.”).
This means that rebuses may help to identify the language underlying a script and to
decipher some signs of it, if we can guess the meaning of a sign from its pictorial form and/or
its context. But its pictorial meaning may not always coincide with its contextual meaning. Yet,
even if the pictorial and contextual meanings differ, they may still be connected by
homophony. If the pictorial and contextual meanings of a particular sign can be independently
determined, and turn out to be the same, the coincidence strengthens the assumed shared
meaning, though without yielding a phonetic reading.
As an example, consider an Indus text on a tablet from Mohenjo-daro (Fig. 21.1). It shows
the common U- or V-shaped sign, in the middle of the inscription and near the right-hand end.
In the first occurrence, the sign’s pictorial meaning might include a pot or vessel, though
other interpretations are certainly possible. But in the second occurrence, a kneeling human
figure holds the very same U- or V-shaped sign in its hands and offers it to a sacred tree—an
iconographic motif. From this contextual meaning, it appears that the sign must represent an
offering vessel. The contextual meaning of the sign supports its pictorial meaning, without any
postulation of an underlying language. Nonetheless, without postulating a language, we have
no idea of the word for “offering vessel” in the Indus language, nor of the phonetic value of
the sign in the first occurrence.

FIGURE 21.1 The U-shaped sign in an Indus inscription and a kneeling worshipper extending a U-shaped vessel
toward a sacred tree. Obverse of a molded tablet M-478 (DK 10237, IM 10387) from Mohenjo-daro. After CISI
3.1:403. Photo EL, courtesy Indian Museum, Kolkata.

A phonetic decipherment of a logosyllabic script demands that four conditions are

simultaneously fulfilled:

(1) the object depicted in a pictogram can be recognized;

(2) the pictogram has been used as a rebus in a particular context;
(3) the contextual meaning of the rebus can be deduced;
(4) a linguistically satisfactory homophony between these two meanings exists in a historically
likely language.

Deduction can begin at any stage; it does not have to follow the order 1, 2, 3, 4.
Unfortunately, the pictorial meaning of most Indus signs is not clear, or at least not
unambiguous. Furthermore, it is not always possible to identify every instance of a particular
sign, because sometimes it may have been simplified. In most scripts, including the Egyptian
hieroglyphic and Mesopotamian cuneiform scripts, the demand for fluency led to a radical
simplification of sign shapes. This variation in the graphemic shape of a sign is important, for
some variants (known as allographs) may render the pictorial meaning better than others;
also, the proposed pictorial interpretation should fit all of the allographs.
Another kind of contextual clue comes from the function of inscribed artifacts. The vast
majority of Indus texts are seal stamps and seal impressions in clay. In the warehouse of the
Harappan harbor town at Lothal, which burned down, nearly one hundred clay tags carrying
seal impressions were baked in the fire and have therefore survived. The tags were once
attached to bales of goods, for there are traces of cloth, strings, and other packing material on
the reverse of the impressions. The Harappans must have used these seals to control
commercial transactions, to prevent their merchandise from being tampered with. Some of the
clay tags carry seal impressions made with different seals, indicating the need for several
witnesses.
A clay tag found in Mesopotamia is stamped with a normal Indus-type square seal bearing
the script and the “unicorn” motif on the obverse and a cloth impression on the reverse (Fig.
21.2). This usage ties in with similar usage of seals among the Sumerians and Akkadians. The
inscriptions on such Sumerian and Akkadian seals comprise chiefly the proper names of
persons, with or without their occupational or official titles and descent (Edzard 1968). The
Harappans’ long-time presence in West Asia makes it highly probable that the same content
may be expected in the Indus seal inscriptions, whether they are found in Mesopotamia or in
the Indus Valley.

FIGURE 21.2 A clay tag from Umma, Mesopotamia, in the Department of the Ancient Near East (1931.120),
Ashmolean Museum, Oxford. (a) The obverse with impression of a square Indus seal. (b) The reverse with a textile
impression. After Parpola 1994:113, fig. 7.16. Photo courtesy Ashmolean Museum, Oxford.

In Mesopotamian and classical Indian onomastics, the names of gods were used to form
personal names. We can therefore expect to have theophoric components in both proper
names and priestly titles in some fairly large and uniformly distributed sign groups in the
Indus seals. Might this practice apply to the “fish” sign, both in its plain form and when it is
modified with various diacritical additions, which occurs so frequently on Indus seals that
almost every tenth sign is either a plain “fish” or a modified “fish” (Fig. 21.3)?
FIGURE 21.3 Two Indus seals with fish signs from Mohenjo-daro. (a) Impression of the seal M-314 (HR 3005, ASI
63.10.366) with the longest Indus text on one side of an object. After CISI 1:78. Photo EL, courtesy ASI. (b) The seal
stamp M-236 (DK 10965, NMI 103) in the collection of the National Museum of India, New Delhi. After CISI 3.1:371.
Photo EL, courtesy NMI.

That the Indus pictograms looking like fish really do have this pictorial meaning is
supported by some Indus iconography, in which the pictogram is placed in the mouth of a fish-
eating crocodile (Fig. 15.5). Harappans may have offered fish to sacred crocodiles, as people
have done at the Magar Talāo or “Crocodile Pool” near Karachi until recently, despite the
millennial presence of Islam in the Indus Valley. That fish indeed were offered, either to sacred
crocodiles or to sacred trees, as in later Hinduism, and that the “fish” sign in some contexts
means “fish,” is suggested by some Indus tablets (not seals) that seem to mention offerings of
“four pots of fish” or “four fish-pots” (Fig. 21.4). Here the simple “fish” occurs before or after
the U- or V-shaped sign, while the three-sign text begins with number 4, represented by four
vertical strokes. The addition of the fish sign in these tablets is exceptional. The U- or V-
shaped sign occupies the reverse side of most of such tablets either alone or preceded by just
one, two, three, or four vertical strokes representing the numbers 1, 2, 3, and 4 (in some rare
cases there are no strokes but the U sign is repeated: UUU, “three pots”). Occasionally in such
a context, the U sign is placed in the hand or hands of an adjacent sign depicting a man, either
standing or kneeling. This connects the U sign with the previously discussed iconographic
scene in which the U sign is extended toward to a sacred tree by a kneeling worshipper (Fig.
21.1). The occasional addition of the “worshipper” suggests that on these tablets the U signs
stand for pots of offerings.

FIGURE 21.4 Reverse sides of two incised tablets from Harappa, probably meaning: (a) “Four pots of fish” (H-1191),
and (b) “Four fish-pots” (H-1192). After CISI 3.1:164. Photo Punjab 41:4395, courtesy ASI.

In contemporary Mesopotamia fish offerings were made in temples. Yet, although

Mesopotamian administrative texts in cuneiform tablets often record rations of fish, fish are
never mentioned in Mesopotamian seal inscriptions. This is not surprising, if the seals
contained chiefly proper names. Assuming that the Indus Valley followed Mesopotamian
practice, it is likely that in the Indus seal inscriptions the fish sign denotes something other
than fish and may be used as a rebus.
The most commonly used word for “fish” in more than one subgroup of the Dravidian
languages is *mīn. A homophonous word means “star.” Both meanings, fish and star, are likely
to derive from the verb root *min, “to sparkle, to glitter,” which can be reconstructed for
Proto-Dravidian. It is known that Old Tamil speakers imagined the stars to be fish swimming in
the ocean of the night sky. Might the “fish” sign in the Indus seals denote “star,” rather than
“fish”? Two pieces of evidence independent of the Dravidian hypothesis support this proposal.
An artistic motif showing fishes with stars occurs on Mature Harappan painted pottery (Fig.
21.5). Moreover, the “star” symbol appears alone on Early Harappan pottery. Secondly, there
is no “star” pictogram in the signs of the Indus script (although stars do appear as
iconographic motifs on Indus seals); this is understandable, following the principle of
parsimony, if the “fish” sign could also denote “star.” The meaning “star” suits the expected
role of the “fish” signs as a component of theophoric proper names in seal inscriptions.

FIGURE 21.5 The “fish” and “star” motifs combined on Indus pottery from Amri. After Casal 1964:II, fig. 92 no. 487
and fig. 78 no. 343. Courtesy Mission Archéologique de l’Indus, CNRS/Musée Guimet, Paris.

In Mesopotamian inscriptions, by contrast, the pictogram showing a star is easily

recognizable. Indeed, whenever a god or goddess is mentioned in cuneiform texts, the “star”
pictogram is prefixed to the name as its determinative, so as to indicate that what follows is
divine. In the Sumerian script, the “star” pictogram denotes not only “god” but also “sky.” It is
thought to have originated as an attribute of the sky god An. Since An was the leading divinity
of the Sumerian pantheon, his symbol would have come to mean “god” in general. Astronomy,
including the use of a star calendar, played an important role in ancient Mesopotamia, and
deeply influenced its religion: all the main gods were symbolized by particular stars or
planets. In West Asia, one or two “star” symbols placed near the head distinguished divinities
in pictorial representations (Fig. 21.6a; Fig. 18.13) The practice seems to have been borrowed
by the Indus civilization, for a seal from Mohenjo-daro depicts an Indus deity with a star on
either side of his head enclosed by a pair of curved horns (Fig. 21.6b).
FIGURE 21.6 Star on either side of an anthropomorphic deity. (a) Six-locked nude hero with water streams flowing
from his shoulders (as from the Sumerian water god Enki) on a Syrian seal without an archaeological provenance
(VA 2928), dated c. 1980–1880 BCE (cf. Eder 1996: Dok. 188). © bpk / Staatliche Museen zu Berlin,
Vorderasiatisches Museum / Fotoarchiv. (b) The Indus seal M-305 (DK 3884, ASI 60.10.62) from Mohenjo-daro. After
CISI 1:383. Photo EL, courtesy ASI.

The “fish” sign, with the rebus meaning “star,” could well have been a building block for
Harappan proper names. Certainly, there is a long tradition of Indians having astral names
derived from their birth star or constellation; already in early Buddhist texts astral names are
common. But this tradition does not seem to have Indo-Aryan roots. In Vedic texts astral
names are rare, and they are supposed to be kept secret. Hindu law books recommend that
Aryan men should not marry girls “named after a star or constellation (ṛkṣa), a tree, or a river,
nor one bearing the name of a low caste, or of a mountain, nor one named after a bird, a
snake, or a slave, nor one whose name inspires terror” (Manu 3,9). This suggests that names
derived from stars are of non-Aryan, possibly Harappan, origin.
Thus, the interpretation of the plain “fish” sign as denoting Dravidian mīn in the sense of
both “fish” and “star” seems to be a plausible fit with Harappan, Mesopotamian, and post-
Vedic Indian religious practice. We must now attempt to check it against other evidence, both
internal and external to the script.
Internal checking can be compared to doing a crossword puzzle, in which we guess a word
on the basis of the given clue and the number of slots for letters. Whether we are correct is
initially uncertain. If we can fill in another word interlocking with the first guess our first
guess becomes more probable; the more interlocking words we can fill in, the greater is our
confidence in our first guess. In Indus decipherment, we should try to apply similar basic
assumptions and methods of interpretation to a new sign closely associated with the first sign
we have interpreted—either so that two successive signs make a recurring sequence that
probably stands for a compound word, or so that together the signs form a compound sign,
where one sign may function as an auxiliary phonetic or semantic indicator to the other sign.
External checking requires us to propose an underlying language and see if this
assumption generates credible words in the undeciphered script. If we assume that two
successive signs probably form a compound word in a Dravidian language, for example, we
must see if this compound word is actually attested in Dravidian languages, notably Old Tamil.
A key stage in the phonetic decipherment of the Mayan script shows this process at work. In
the later nineteenth century, Léon de Rosny discovered what appeared to be pictorial
translations of accompanying glyphs in a Mayan manuscript known as the Madrid Codex.

He realized that the glyphs for certain animals, such as dog, turkey, parrot and jaguar,
could be identified by examining the glyphs above the pictures of these creatures. …
He now applied [Bishop de Landa’s putative Spanish-Mayan “alphabet” of 1566] to
what seemed to be the first sign in the glyph for “turkey.” Rosny read the first sign in
the glyph as cu, by comparing it with Landa’s cu. He then hazarded a guess that the
entire glyph might be read cutz(u), since the Yucatec Mayan word for turkey is cutz.
(Robinson 2002:121)

In 1952, Rosny’s surmise was validated by Yuri Knorozov, who

noticed that the first sign in the dog glyph was the same as the second sign in the
turkey glyph. If the first sign in the dog glyph had the phonetic value tzu (as proposed
by Rosny), the second could be assigned the value l(u), on the basis of its resemblance
to Landa’s symbol l. Hence the dog glyph might stand for tzul. Was there a Yucatec
word tzul in the dictionary? There was. It meant “dog.” (Robinson 2002:123–125).

In deciphering sequences of Indus signs it is obviously essential to read them in the correct
direction, as with any writing system. We can be sure that the normal direction is from right to
left in the Indus seal impressions. (In the seal stamps, by contrast, the reading direction is
from left to right; the stamp texts are mirror images of the seal impressions.) One bit of
evidence for the direction of the writing comes from the overlapping of the Indus signs as they
were drawn in wet clay. A second bit comes from seals in which part of an edge remains blank,
for example, a seal carved on only three of its four sides, with a gap left at what must be the
end of the inscription (Fig. 21.7). A third bit comes from lines in which the signs are cramped,
or overspill the line, because the writer must have run out of space. Finally, the frequent
repetition of certain signs and sign sequences allows them to be identified as being typical of
the beginning or end of an inscription. Occasionally, however, the small tablets from Harappa
apparently use a reverse order compared to the normal one, as shown by the fact any
asymmetrical signs are reversed. Unless otherwise noted, the sign sequences in this book are
printed in the order of the impression, not the stamp, and should therefore be read from right
to left.
FIGURE 21.7 The direction of writing in the Indus script, and dividing longer texts into shorter phrases. (a)
Impression of the seal H-103 (2789, ASI 63.11.116) from Harappa, with an inscription carved along three of its
sides. The text runs anticlockwise: the uppermost side was written first and is full, running from the right end to the
left; the line written next starts beneath the first line and continues to the end of the side; the line written last
(upside down in the picture) starts from the side of line 2 and on account of its shortness leaves a gap at the end.
Each of the three lines of H-103 forms a phrase that occurs as a complete text among shorter Indus texts from
Mohenjo-daro. (b) Impression of the seal M-1680 (DK 8856), parallel to the first line of H-103, except that it lacks
the first sign. After CISI 3.1:13. Photo Sind 20:212, courtesy ASI. (c) Impression of the seal M-122 (DK 12523, ASI
63.10.40). After CISI 1:41. Photo courtesy ASI. (d) Impression of the seal M-197 (DK 10924). After CISI 1:53. Photo
Sind 22:446, courtesy ASI.

The numerals are among the few Indus signs of which the function and meaning can be
deduced with fair certainty, for two reasons. First, the signs consist of groups of vertical
strokes, which is how numerals are represented in many ancient scripts. Second, they are
mutually interchangeable before specific signs, including the simple “fish” sign. Reading the
sequence “six vertical strokes” + “fish” in Dravidian yields the Old Tamil name of the Pleiades,
aru-mīn, “six stars.” “Seven vertical strokes” + “fish” yields the Old Tamil name of Ursa Major,
alias the Big Dipper, elu-mīn, “seven stars.” The latter sequence forms the entire inscription
on one big seal from Harappa (which might be compared to the Mesopotamian dedicatory
seals sometimes presented to divinities) (Fig. 21.8). In India, since Vedic times, the stars of
Ursa Major have been identified with the “Seven Sages,” the mythical ancestors of priestly
clans. The Pleiades and Ursa Major play a very important role not only in early Indian
mythology—including the mythical origin of Śiva’s liṅga cult—but also in the early history of
Indian calendrical astronomy, which is probably of Harappan origin, as discussed in chapter
16.
FIGURE 21.8 The sign sequence “7” + “fish” on the seal H-9 (115, ASI 80.2.4) from Harappa. Depicted here is the
seal stamp, where the text is in mirror image and reads from left to right instead of the normal writing direction
right to left. After CISI 1:166. Photo EL, courtesy ASI.

Even the modified attributes of the “fish” sign can be interpreted using similar premises, if
more tentatively than the numerals or constellations. One diacritical mark over the “fish” sign
looks like a “roof” (see the third sign from the right in the top row of Fig. 21.3a). The most
widespread lexeme for “roof” in Dravidian languages is *vēy / *mēy. Using phonological
variations reconstructed for Proto-Dravidian, the early form of the word for “roof” was
homophonous with *may, “black.” The modified “roof-fish” sign can thus be read as *mēy-mīn,
functioning as a rebus for *may-mīn, “black star,” a compound word actually attested in Old
Tamil mai-m-mīn as the name of Saturn. Saturn is indeed a dim planet, connected with the
color black in Sanskrit sources. In Sanskrit, Saturn is called Śani or Śanaiścara, “slowly
moving,” on account of the planet’s slow pace. In Buddhist and Jaina iconography, the god
Saturn rides a proverbially slow tortoise. Conceivably, the Indus “roof-fish” pictogram
symbolizes Saturn not only phonetically but even pictorially through his vehicle—for the
tortoise, as an aquatic animal, might be regarded as a kind of “fish,” while its shell is a kind of
roof!
Another diacritic modifying the “fish” sign is a straight line drawn directly or obliquely
across the body of the “fish” pictogram, halving the fish (see the second sign from the left in
Fig. 21.3b). Dividing a fish into two equal parts for its catchers is a narrative motif in Indian
folklore. The Proto-Dravidian root *pacu, “to halve, divide into two,” is homophonous with
*pacu, “green.” The resulting compound word pacu-mīn is attested in Old Tamil with the
meaning “green fish,” while paccai, literally “greenness,” is a Tamil name for the planet
Mercury. Paccai is also a Tamil name for Viṣṇu as green-hued, and one of Viṣṇu’s Sanskrit
names is Hari, “green.” All this agrees with the Indian astrological tradition, according to
which Mercury is the lord of green things and Viṣṇu is his presiding deity.
“Fish with a dot in its stomach” is yet another “fish” diacritic (see the second sign from the
left in the top row of Fig. 21.3a). What could the dot mean? Here one is reminded of a dot of
deep significance for Hindus, namely the red dot put on the forehead, especially by married
women whose husband is alive, known in Sanskrit and Hindi as a bindu, “dot, drop.” Men put
the bindu on the forehead after worshipping the Goddess; according to the Kālikā-Purāṇa, this
should be done with the blood from the sword used in decapitating the sacrificial victim, while
pronouncing the mantra: “Whomever I touch with my foot, whomever I see with my eye, he
must come into my power.” This “fish-with-a-dot” sign occurs prominently on the “fig deity”
seal which seems to depict an offering of a warrior’s head to the Goddess (see chapter 19)
(Fig. 21.9).
FIGURE 21.9 The “fig deity” seal M-1186 (DK 6847, NMP 50.295) from Mohenjo-daro. (This is the seal stamp, not its
impression.) After CISI 2:425. Photo JL, courtesy DAMGP.

In north India, the rite of applying red powder to the bride’s forehead and the parting of
the hair (sindūra-dāna) is sometimes the only marriage rite, and often the binding part of the
ritual. Archaeological evidence suggests that the custom has age-old Harappan roots, for
terracotta female figurines from Nausharo IB (2800–2600 BCE) have traces of red pigment in
their hair parting (cf. Kenoyer 1998:44–45,186). However, the Vedic manuals of domestic
ritual do not prescribe this custom as part of the marriage ritual, which strongly suggests that
it is of non-Aryan origin.
In Dravidian south India, the red dot put on the forehead is called poṭṭu (the word also
means “drop,” like Sanskrit bindu). The compound word poṭṭu-mīn has been recorded in the
Central Dravidian language Pengo with the meaning “a kind of fish,” identifiable as a carp
species, Cyprinus/Labeo rohita. In Sanskrit rohita denotes both “red” and “carp.” It has been
the custom in Karnataka that the bride and groom, at the time of their marriage, catch a carp
and use its red scales to mark each other’s forehead. Carp scales were also used to mark the
forehead in Kashmir. Perhaps the custom was connected with the fish-scale motif popular on
Early and Mature Harappan pottery.
Rohiṇī, the feminine form of rohita, “red,” is the Sanskrit name of the red star Aldebaran,
which rose together with the sun at the vernal equinox in 3054 BCE, in Early Harappan times.
As discussed in chapter 16, Rohiṇī was probably the original new-year star of the nakṣatra
calendar. The Gṛhyasūtras and Old Tamil texts mention Rohiṇī as the star most auspicious for
marriage, and Rohiṇī as the faithful wife of the moon is among the foremost models for a
Hindu wife. The Atharvaveda (AVŚ 13,1,22), however, makes Rohiṇī the mate of the rising sun,
Rohita, and a goddess of war. In Sanskrit, rohiṇī also denotes a marriageable young virgin who
has just attained menstruation. The red Rohiṇī star, therefore, may be imagined to be a drop of
menstrual blood, a most suitable symbol for the Goddess. Like the red dot on the forehead, it
also resembles the third eye of the virginal warrior Goddess, glaring red from anger. (In the
Mesopotamian tradition, Aldebaran represents the “eye” of Taurus, the Bull of Heaven.)
One recurring two-sign sequence, with the simple “fish” sign as its last member, begins
with a sign of which the pictorial meaning seems to be “fig tree” (Fig. 21.10). Can this
sequence, too, be a Dravidian compound word with an astral meaning?
FIGURE 21.10 The sequence of Indus signs “fig” + “fish” on two seals from Mohenjo-daro. (a) The seal M-414 (DK
3431, ASI 63.190.208), and (b) its impression. After CISI 3.1:409 (photo EL) and CISI 1:100 (photo Sind 18:587).
Courtesy ASI. (c) The seal M-172 (BJ4), and (d) its impression. After CISI 3.1:132 (photo Sind 6:I.67) and CISI 1:50
(photo Marshall 1931: pl. 106:71). Courtesy ASI.

The pictorial interpretation as “fig” is based on a comparison with a motif on Harappan

painted pottery, which depicts a “three-branched fig tree” from the Early through the Mature
to the Late phase. In one variant from Nausharo, phase I D, dating from c. 2600 BCE (about the
time the Indus script was created), four strokes are attached to either side of the middle stem
(Fig. 21.11). In the Indus sign, the fig tree is shown as three-branched, just as on the painted
pottery, except where the central “branch” is omitted to make a compound sign in which a
second sign (the “crab” sign) is inserted in place of the central “branch.” In some cases, the
branches end in fig leaves as they do on the painted pottery, but in others, the fig leaves are
simplified, even into straight lines (Fig. 21.12). The lines or strokes seem to represent the
rope-like air roots of the banyan fig, Ficus bengalensis or Ficus indica (Fig. 16.9).
FIGURE 21.11 The “three-branched fig” motif on a Harappan painted goblet from Nausharo, period I D, c. 2600–
2500 BCE. After Samzun 1992: 250, fig. 29.4 no. 2. Courtesy Mission Archéologique de l’Indus, CNRS/Musée
Guimet, Paris.

FIGURE 21.12 Variants of the “fig” and “fig” + “crab” signs. After Parpola 1994:235, fig. 13.15.

This mighty tree is native to South Asia and does not grow in the regions from which the
Indo-Aryan speakers came. A post-Vedic name for the banyan fig is vaṭa. Though a Sanskrit
word, vaṭa is a loanword from Dravidian, ultimately derived from Proto-Dravidian *vaṭam,
“rope, cord.” As a name of the banyan fig, vaṭam is short for the compound word vaṭa-maram,
“rope tree,” which is attested in Tamil. Vaṭam has a Proto-Dravidian homophone *vaṭa, “north,
northern.” This yields the expected astral meaning for the sign sequence “fig” + “fish.” Vaṭa-
mīn, “north star,” is attested in Old Tamil; according to the medieval commentaries, it is the
name of the star Alcor in Ursa Major. In Old Tamil texts, vaṭa-mīn is a symbol of marital fidelity
and during the wedding the star is pointed out to the bride as an object for emulation. As
discussed in chapter 16, vaṭa-mīn probably originally denoted the pole star, which in the third
millennium was the nearby star Thuban (Fig. 16.8). The pole star is of course the “immobile”
center of the rotating heavens, known in Sanskrit as dhruva, “fixed, firm, immovable,
constant.” It is a fitting symbol of fidelity; indeed in the Vedic marriage ritual the pole star is
pointed out to the bride as a model, in addition to Alcor.
This interpretation explains in a new way some peculiar cosmological conceptions of the
Purāṇa texts. In the first place, the mythical central mountain of the world, Meru, is
surrounded by four great mountains in the four cardinal directions, on the tops of which grows
a specific variety of gigantic tree. The tree on the northern mountain is the banyan fig (vaṭa).
Homophony connects vaṭa, “banyan,” with vaṭa, “north,” in Dravidian, but there is no such
association in Indo-Aryan languages, where the Sanskrit for “north” is uttarā or udīcī diś-,
literally, the “upper, upward” or “left” direction (north being on the left when one faces east,
the “forward” direction). Secondly, in reply to the question why the stars and planets do not
fall down from the sky, the Purāṇas say that the heavenly bodies are bound to the pole star
with invisible “ropes of wind.” In Dravidian vaṭa-mīn, besides being the name of the pole star,
can also be read as “rope star” and “banyan star.” Around 1000 BCE, a late hymn of the
Rigveda (1,24,7) speaks of the roots of a cosmic banyan tree being held up in the sky by
Varuṇa (see chapter 16). Both Vedic and Hindu texts repeatedly refer to a heavenly fig tree.
This conception seems to be reflected on an Indus tablet, which depicts an anthropomorphic
deity inside a fig tree. At the bottom the fig tree is flanked on either side by a star, suggesting
a heavenly connection for the tree (Fig. 16.10).
The above interpretation of the “fig” sign can be further checked by attempting to
understand the compound sign in which the “crab” sign has been substituted for the middle
branch of the “fig” sign. The “crab” sign has two variants, with legs and without legs, as
demonstrated by the similar context of the two variants in two different seals. On both seals
the compound sign is followed by the same two other signs (Fig. 21.13).

FIGURE 21.13 Two variants of the “fig” + “crab” sign (cf. Parpola 1994:232 fig. 13.12). (a) The seal stamp H-598
(13751, HM 200) from Harappa. After CISI 2:297. Photo AV, courtesy DAMGP. (b) Impression of the seal L-11 (4879,
LTH.SRG 1261) from Lothal. After CISI 1:241. Photo EL, courtesy ASI.

FIGURE 21.14 [See Fig. 21.12] Variants of the “crab” sign. After Parpola 1994:232.

Let us first consider the likely meaning of the “crab” sign. It occurs more than 150 times
as an uncompounded sign, mostly simplified to a round body with claws (Fig. 21.14). The
visual emphasis on the claws suggests that the sign expresses “grasping” or “seizing,” which
is consistent with the behavior of the crab in Indian folklore. Thus, in the Buddhist Baka-
Jātaka, a crab tells a heron who has promised to carry the crab away from a pond that is
drying up to some other pond (as a pretext for eating the crab): “You’d never be able to hold
me tight enough, friend heron; whereas we crabs have got an astonishingly tight grip.” Then
“the crab gripped hold of the heron’s neck with its claws, as with the pincers of a smith.”
Crab’s claws are compared with blacksmith’s pincers in Old Tamil texts, where the root
koḷ, “to grab, seize, take,” expresses the “seizing.” Indo-Aryan texts use the synonymous
Sanskrit root grah- / grabh-, which is related to English “grab.” The “crab” sign often occurs
close to “fish” signs and might therefore have an astral meaning. Proto-Dravidian *kōḷ,
“seizure,” means “planet” in Tamil. Surprisingly, this double meaning is also true of Sanskrit
graha—but not true of other Indo-European languages. Sanskrit graha is therefore more likely
to be a loan translation from Dravidian *kōḷ than kōḷ is to be a loan translation from Sanskrit
graha. In both the oldest Tamil and the oldest Sanskrit texts kōḷ and graha refer to the
invisible heavenly demon that causes eclipses, by seizing the sun and moon.
The planets are firmly believed in both Tamil and Sanskrit texts to “seize” people and
afflict them with ills. The Sanskrit author Daṇḍin, writing in about 700 CE, speaks of the
“terrifying stars and planets,” which the sorcerers control with magical diagrams. From the
eighteenth century comes a graphic description of such incantations:

The term graha, by which they are designated, signifies the act of seizing, that is, of
laying hold of those whom they are enjoined by the magical enchantments to torment.
… The magician … exclaims as though in a vehement rage, “Grasp it! Grasp it!” … No
sooner is this done than the grahas or planets take possession of the person against
whom such incantations are directed, and afflict him with a thousand ills. (Dubois
[1825] 1906:387f.)

We can check our interpretation of both the “crab” and “fig” signs by examining the
compound sign with the “crab” inside the “fig tree.” Luckily, this is among the few Indus signs
for which there are “pictorial bilinguals,” that is, tablets that mediate the sign’s intended
meaning visually with an accompanying iconographic motif. The 240 copper tablets from
Mohenjo-daro are a rare category of Indus objects, because they show a clear
interdependence between the inscription on the obverse and the iconographic animal- or
human-shaped motif on the reverse. Numerous duplicates form sets of identical tablets. In
some sets, a single sign appears on the reverse instead of an iconographic motif. By
comparing tablets having the same inscription on the obverse side but different reverse sides,
it is possible to link these single signs with their corresponding iconographic motifs. It
appears that the single sign represented the name of the divinity depicted in the motif. The
“crab inside fig” sign can thus be equated with a “horned archer,” a male figure armed with a
bow and arrows, anthropomorphic apart from its horns and a tail, and with protruding eyes
(Fig. 21.15).

FIGURE 21.15 A “pictorial translation” of an Indus sign: the identical inscriptions on the obverse sides correlate the
“horned archer” on the reverse of the type B-19 (there are fourteen identical tablets of this type) and the “fig” +
“crab” sign on the reverse of the type C-6 (there are seven identical tablets of this type) among the copper tablets
from Mohenjo-daro. After Parpola 1994:234, fig. 13.13.

In West Asian and Chinese scripts, an inserted sign often functions as a semantic or
phonetic determinative. The inserted “crab” sign could be a phonetic determinative indicating
that the “fig” sign is not to be read with its usual phonetic value as vaṭam, “banyan tree.”
While the meaning, “fig,” is retained, the phonetic shape of the word is similar to that
expressed by the “crab” sign, kōḷ. Proto-South-Dravidian possesses exactly such a word: *kōḷi.
This word denotes a fig tree, both as a grasping epiphytic plant that strangles its host tree and
as a plant that bears fruit without blossoming. In this latter sense kōḷi is related to Old Tamil
kōḷ, “the act of bearing fruit.” Both are derived from the root koḷ, “to take.” Sanskrit grabh-
has the same additional sense of “bearing fruit.” Even the structure of this compound sign
seems to express this ambivalent deity, who both “seizes or kills” and “bears fruit or
fructifies,” for the crab sign is placed inside the fig sign, just as anthropomorphic deities are
placed inside fig trees in Indus glyptics.
But how can this word *kōḷi be connected with the “horned archer” depicted on the copper
tablets? In early Vedic texts the grasping fig strangling its host tree and breaking buildings is
implored for help in crushing enemies. This suggests that the fig could be a symbol for Rudra,
the god who is described in Vedic texts as a cruel hunter and raider who, with his bow, shoots
arrows at animals and people. He is also called in Sanskrit Hara, “seizer, taker, robber,” which
could reflect the Dravidian word kōḷ, “seizure, plunder, robbery.” There is also a homonym kōḷ,
“hitting, killing,” from the root koḷ, “hit, shoot with bow, kill.” Rudra, whose original name was
probably *Rudhra, “Red,” seems to have been represented by the red planet Mars, called in
Sanskrit Rudhira, “red,” “blood,” and Aṅgāra, “live coal.” The latter name has a counterpart in
Proto-South-Dravidian *koḷḷi, “firebrand, glowing ember,” which may also be the name of the
Harappan archer-god intended to be expressed by the compound sign “fig” + “crab.”
Rudra is often equated in Vedic texts with the fire-god Agni, who is said to be either the
mate or the son of the Pleiades. Agni’s name has Indo-European ancestry (cf. Latin ignis,
“fire”), but the Vedic fire-god has absorbed attributes and myths likely to be of local Indian
origin since they are connected with local plants and animals. Thus the kindling stick was
made of the pipal fig or Ficus religiosa, which has flame-shaped leaves. The generation of fire
by churning the upright kindling stick in the hole of a horizontal wooden plank symbolized
sexual intercourse (chapter 11). Agni is called the “embryo of forest trees” and is appealed to
by prayer to place an embryo in the worshipper’s womb. In the Vedic ritual known as
“engendering male offspring” (puṁsavana), the wife wears a phallic amulet made of the shoot
of a banyan fig. The amulet should have two fruits symbolizing testicles. The shoot is to be cut
from the king of banyan figs that grows outside the village after propitiating the deity who
inhabits it. Jaimini-Gṛhyasūtra 1,5 prescribes: “Having shaped two beans and a barley corn
into shape of a male organ of procreation, he [the worshipper] should give it her [his wife] to
eat. … Then having with two threads … fastened a shoot of the banyan fig which has two
fruits, she should bear it on her throat. This, they say, is a sure means to get a son.”
On a tablet from Harappa we see a ram-headed but otherwise anthropomorphic deity
inside a fig tree (Fig. 21.16a). The god’s arms are covered in bangles and in their great length
the arms resemble the air roots of the banyan. This Harappan ram-headed fig deity has a
relatively little-known successor in later Indian tradition. Nejameṣa (meṣa means “ram”) is a
fertility god who should be appealed to by a woman who cannot conceive, while she touches
her genitals and utters this Vedic verse: “O Nejameṣa! Fly away, and fly hither again bringing
back a beautiful son; to me here who is longing for a son grant thou an embryo, and that a
male one” (RV Khila 30,1, translated Winternitz 1895:150–151). In later Hindu sources, a
related ram-faced deity, Naigameṣa, is said to be a constant companion of Skanda. Another
deity, Hari-Naigameṣin, is known from the Jain Kalpasūtra, where this deity transfers the
embryo of Mahāvīra Jina to the womb of his mother. According to the Kalpasūtra, Hari-
Naigameṣin is the leader of the divine army with a peacock as his mount; he is thus clearly a
double of the Hindu war-god Skanda who rides a peacock. On a second-century relief from
Mathurā, the goat- or ram-headed fertility god Naigameṣa is flanked on his left side by a baby
boy and three women (Fig. 21.16b). The relief is broken, but may originally have shown three
women on the god’s right side, too. Together, these six women would have represented the
Pleiades as mother goddesses.

FIGURE 21.16 Ram- or goat-headed fertility deity connected with the banyan fig. (a) A molded Indus tablet H-178 B
(7483, ASI 63.11.69) from Harappa. After CISI 1:209. Photo EL, courtesy ASI. (b) The god Naigameṣa flanked by a
child and three women. A stūpa railing from Mathurā, c. 100 CE. Now in Lucknow Museum (J 626). Photo courtesy
American Institute of Indian Studies.

The unexplained first part of the names of Neja-meṣa and Naiga-meṣa seems to derive from
the root nij-, “to wash, cleanse.” Sanskrit medical texts prescribe bathing the new-born baby
under the banyan tree, if its disease is diagnosed as caused by Naigameṣa, along with a bloody
offering (bali, perhaps originally of a ram or goat) to this deity at a banyan on the sixth day
after birth, accompanied by this prayer: “May God Naigameṣa, the child’s father, protect the
child, (this) greatly famed (Naigameṣa) who is goat-faced, who has moving eyes and eye-
brows, and who can take any shape at will.” (Suśruta-Saṁhitā, Uttaratantra 36,10–11)
The sections on the illnesses of new-born babies in medical texts speak of a group of nine
malignant demons keen to attack and seize infants, if proper respect is not shown to them and
the rules of cleanliness and nursing are not followed. The group’s name, nava-graha, is
identical with the classical group of “nine planets,” among whom the planet Mars represents
the war-god Skanda. The lord of the demons is Skanda-Graha, plus two forms of the goat-
headed Naigameṣa and six goddesses. The nine demons were created by the fire-god Agni and
Śiva along with the six Pleiades, in order to protect the new-born god Skanda.
The best-known version of the Pleiades myth describes the birth of the ever-youthful war-
god Skanda. Kāma, the god of love, shot his arrows of desire at Śiva, whose seed “leapt”
(caskanda) and fell into the River Ganges. The wives of the Seven Sages were bathing in this
heavenly river, and became either the mothers or the wet-nurses of the instantly born boy,
who rides a peacock.
Now, Skanda has a counterpart in the principal native deity of south India, the youthful
god of war, wisdom, and fertility, Murugan or Murukan, whom Tamils worship today as their
“national god.” Common people pray to Murukan for sons. His mount is a peacock and his
weapon a spear. In Old Tamil literature Murukan is both the hunter-god of the hill forests,
much like the Vedic Rudra, and the god of love and fertility. From 300 CE onwards, Murukan is
explicitly amalgamated with Skanda.
Given his importance to Dravidian speakers, Murukan’s name or names are likely to be
present in Indus texts. But how to locate them? Skanda’s association with the Pleiades offers a
clue. As already suggested, the Pleiades can be identified as the sign sequence “six” + “fish.”
Two Indus seals share a unique sequence of three signs (the first sign, which is very rare, is
represented by two allographs), suggesting that the two seals may be speaking about the
same thing (Fig. 21.17).

FIGURE 21.17 Two Indus texts sharing a unique three-sign sequence. (a) Impression of the seal M-112 (DK 11359,
ASI 63.10.81) from Mohenjo-daro. After CISI 1:40. Photo EL, courtesy ASI. (b) The seal M-241 (HR 5787, NMI 106)
from Mohenjo-daro in the collection of the National Museum of India, New Delhi; reversed to show the signs as they
would appear in an impression. After CISI 1:60. Photo EL, courtesy NMI.

In both of the seals the said rare three-sign sequence begins the inscription, and may
therefore contain a qualifying phrase—for in Dravidian languages the qualifier precedes the
qualified. The following, longer sequence in one of these two seals includes the signs “six” +
“fish,” and may therefore refer to “Skanda, son of the Pleiades.” Thus there is a chance,
because the initial phrase connects these two seals, that the corresponding latter part of the
other seal might include a name of Murukan. The three-sign sequence occurring here is found
very frequently in Indus inscriptions. Read in the normal direction of writing, the three signs
are the “two intersecting circles,” the “two long vertical strokes” and the “bull’s head.”
Contextual clues suggest that this sign sequence could indeed refer to a deity. For
example, on a tablet from Harappa, the only inscription (exactly repeated on the obverse and
the reverse) is the very same sign sequence, “two intersecting circles,” the “two long vertical
strokes,” and the “bull’s head,” along with various images (Fig. 21.18). On a tablet from
Mohenjo-daro, the same sign sequence concludes the inscription on the obverse, while the
reverse shows a god sitting on a throne, flanked on either side by a kneeling worshipper and a
cobra (Fig. 21.19). In South India, Murukan is associated with a phallic snake cult, and his
peacock feeds on snakes.

FIGURE 21.18 The sign sequence “two intersecting circles” + || occurs on both sides of the Indus tablet H-182 (201,
NMI 33) from Harappa in the collection of the National Museum of India, New Delhi: (a) with a drummer and a tiger
on the obverse, and (b) with swastikas on the reverse. After CISI 1:209. Photo EL, courtesy NMI.
FIGURE 21.19 A molded faience tablet M-453 (DK 7991, ASI 63.190.217) from Mohenjo-daro. (a) The obverse with
an inscription. (b) The reverse showing an erect cobra behind a kneeling worshipper (one extending a vessel of
offerings, the other with hands raised in prayer) on either side of an anthropomorphic deity seated on a throne.
After CISI 1:111 and 386. Photo EL, courtesy ASI.

If the “intersecting circles” sign really does express Murukan’s name, or at least the first
part of it, I suggest that it might be read in Proto-Dravidian as *muruku, “young man, baby
boy, Murukan.” This name is a synonym of Sanskrit kumāra, “youth, baby boy,” one of the
names of Skanda and Rudra. Muruku has an exact and ancient homophone, namely *muruku,
“ring, earring, bangle”—which nicely fits the pictorial meaning of “intersecting circles.”
This proposed reading, *muruku, for the “intersecting circles” sign is endorsed by its
frequent depiction on forty or more inscribed stoneware bangles. Several of these bangles are
inscribed with this sign—and no other sign. It is not unusual for ancient inscriptions to
mention the name of the inscribed object, especially if the object is a votive offering. These
stoneware bangles were manufactured with a difficult process, heated in closed and sealed
containers to 1200°C, and must have been very expensive. In this case, the “intersecting
circles” sign could denote not only the bangle itself but also—by homophony—the boy child
desired by the donor and/or the proper name of the child-granting divinity, himself the divine
child par excellence. Even today, many Tamil couples desiring a male child make a pilgrimage
to a shrine of Murukan and, following the birth, name their son after the god.
Furthermore, bangles have a strong association with pregnancy in many parts of India.
During pregnancy and childbirth mother and baby are both in danger of being attacked by
demons, and the bangle symbolizes an enclosed circle of protection. In Tamil Nadu, the
expectant mother is ritually adorned with bangles and blessed by older women in the seventh
month of the first pregnancy. In the Atharvaveda, bangles are charms to encourage
reproduction. Hymn 6,81 of the Atharvaveda (Śaunaka-śākhā) addresses pari-hasta, the
“bracelet” (literally, “what is around the arm”), as follows: “O bracelet, open up the womb,
that the embryo be put (into it)! Do thou … furnish a son, bring him here … The bracelet that
(goddess) Aditi wore, when she desired a son, (god) Tvaṣṭar shall fasten upon this woman,
intending that she shall beget a son” (translated Bloomfield 1897:96f.).
In Indian folk religion, Hindus, and even Muslims, offer pregnancy bangles to tree spirits.
People anxious to have children hang as many bangles as they can afford on the branches of a
sacred tree. If the tree spirit favors their wish, the tree “snatches up the bangles and wears
them on its arms” (Crooke 1926:417). This widespread folk custom is likely to go back to
Harappan traditions. The deity standing inside the fig tree in the “fig deity” seal from
Mohenjo-daro (Fig. 21.9) surely wears bangles on both arms. So do the seven
anthropomorphic figures at the bottom of this seal, wearing their hair in the traditional
fashion of Indian women, who likely represent the “Seven Mothers,” the wives of the Seven
Sages, famous as goddesses capable of child-granting and child-killing, like their son Skanda.
My tentative reading of “intersecting circles” as muruku might be corroborated by means
of the sign that frequently accompanies it, the “two long vertical strokes.” This lends itself to
various pictorial interpretations, and it is difficult to decide which of them, if any, is correct.
However, we may also consider the fact that the “two long vertical strokes” sign often
precedes the simple “fish” sign (Fig. 21.20).

FIGURE 21.20 Cross-checking Muruku and mīn. (a) The obverse of the molded tablet H-723 (H775, HM 290) from
Harappa. After CISI 2:319. Photo AV, courtesy DAMGP. (b) The seal H-669 (P I-44, Lahore Museum P-909) from
Harappa, reversed to show the signs as they would appear in an impression. After CISI 2:310. Photo AV, courtesy
Lahore Museum.

Let us collect together any attested Old Tamil compound words that start with muruku-
and all compound words that end with -mīn, that is, X in muruku-X and X-mīn. If these two
different sets of compound words reveal a word, X, that is common to the attested lists, we
may test X by asking whether its meaning adequately explains the pictorial shape, “two long
vertical strokes.”
Among the Old Tamil compound names of Murukan is Muruka-Vē ḷ. The word vēḷ means
“love, desire,” and is sometimes used alone as a name of Murukan, that is, Vēḷ. Old Tamil veḷ-
mīn denotes the planet Venus, the brightest star of the sky. In this case, veḷ means “white,
bright.” Its derivative veḷḷi denotes “Venus” in several Dravidian languages. The shared
component of these two compound words, muruka-vēḷ and veḷ-mīn, thus has the phonetic
shape veḷ or vēḷ. And these words have a further, so far unmentioned, homophone, Proto-
Dravidian *veḷ(i); its meaning, “space between,” such as “space between furrows,” matches
fairly well the pictorial shape, “two long vertical strokes.”
In the Indus compound word read as veḷ-mīn or veḷḷi-mīn, “bright star, Venus,” the “two
long vertical strokes” sign is used as an adjectival qualifier (“bright”) of the head word
(“star”). But in Tamil, veḷḷi is used not only as a qualifier of mīn but also as a synonym of mīn;
veḷḷi can mean both “bright” and “star (in general).” Thus, two Tamil dictionary renderings for
“star” are viṇ-mīn and vān-veḷḷi. The words viṇ and vān both mean “sky”; the prefix is there to
avoid confusion of the second part of the compound with other possible meanings such as mīn,
“fish,” and veḷḷi, “white metal, silver.” Synonymous use of veḷḷi and mīn is further attested in
two Tamil compounds, both meaning “the star of the dawn,” viṭi-mīn and viṭi-veḷḷi. That veḷḷi
was used as a synonym of mīn in the Indus language, too, can be seen by comparing two
inscriptions which otherwise share the same four-sign-long sequence, though three of the
signs have variant forms (Fig. 21.21). The “fish” sign is here preceded by the “fig” sign,
yielding the already-discussed compound word vaṭa-mīn, “north star.” We also have “fig” +
“two long vertical strokes,” yielding vaṭa-veḷḷi.

FIGURE 21.21 Impressions of two Indus seals. (a) M-172 (BJ 4) from Mohenjo-daro. After CISI 1:50, photo Marshall
1931: pl. 106.71. Courtesy ASI. (b) H-6 (P I.39, NMI 10) from Harappa in the collection of the National Museum of
India, New Delhi. After CISI 1:162. Photo EL, courtesy NMI.

One rare but delightful Indus sign has a very narrow pictorial meaning. From distinctly
carved occurrences, such as a seal from Nindowari, the sign can be recognized as depicting
the palm squirrel, head down, tail at right angles to its body, with its four paws clinging onto a
vertical tree trunk (Fig. 21.22a). In this typical pose, the palm squirrel can sleep for hours,
hence its Sanskrit name, “tree-sleeper” (vṛkṣa-śāyikā). It still inhabits the entire Indus Valley
and is represented in tiny faience figurines from Mohenjo-daro.
The “intersecting circles” sign, which we have interpreted as muruku, is followed by the
“palm squirrel” sign in three different inscriptions (Fig. 21.22b–e). May this sequence, too, be
read in Dravidian, so that the resulting compound is among the attested composite names of
Murukan? One of the inscriptions comprises just this sequence followed by the sign read as
Vēḷ, one of Murukan’s names.

FIGURE 21.22 The “squirrel” sign of the Indus script. (a) Part of the text engraved on the Indus seal Nd-1 (Exc.
Branch 1187) from Nindowari, reversed. After CISI 2:419 no. 5. Photo AV, courtesy DAMGP. (b) The seal M-1202 (DK
2797, Lahore Museum P-1749) from Mohenjo-daro, and (c) its modern impression. After CISI 2:143. Photo AV and
S.M. Ilyas, courtesy Lahore Museum. (d) Obverse of the molded tablet H-771 (657, HM 474) from Harappa. After
CISI 2:324. Photo AV, courtesy DAMGP. (e) First three signs carved on an Indus seal from Nausharo (excavation no.
NS 95.05.17.01), to be published in CISI 3.2. Reversed to show the signs as they would appear in an impression.
Photo Catherine Jarrige, courtesy Mission archéologique de l’Indus/Jean-François Jarrige.

In modern Tamil, the palm squirrel is called aṇil or aṇil-piḷḷai. The word piḷḷai means “child,
infant, son, boy,” as well as “young of animals and trees.” It is tacked on to aṇil to form an
affectionate diminutive, and even used on its own to refer to a squirrel. (It is used in the same
way to form words for “young mongoose” and “young parrot.”) Crucially for our argument, it
is also added to the various names of Murukan to form pet names that are popular as male
proper names in Jaffna Tamil, such as Muruka-piḷḷai. This Tamil usage goes back to Proto-
Dravidian, for the Central Dravidian languages preserve cognates of piḷḷai with the meaning
“squirrel.” As an honorific plural, Piḷḷaiyār is the Tamil name of the popular god known as
Gaṇeśa or Gaṇapati in Sanskrit (“Leader of the Host”). He is an ancient double of Skanda and
of Rudra (chapter 12).
The fact that even such uncommon Indus symbols as the palm squirrel, which have a
narrow pictorial meaning, find a natural and fitting explanation within a Dravidian linguistic
framework is a hopeful sign. As we have seen, a number of Dravidian-based rebus
interpretations interlock with each other and with external linguistic and cultural data,
making sense within ancient Indian cultural history and the Indus civilization. The
interpretations restrict themselves to ancient Indian astronomy and time-reckoning and its
associated mythology, the chief deities of Hindu and Old Tamil religion and the fertility cult
connected with fig trees. These contexts enable some progress to be made in spite of
difficulties, and suggest possible avenues for future progress. Although our knowledge of
Proto-Dravidian vocabulary, especially compound words, is deplorably defective, it
nevertheless allows some cross-checking.
Conclusion
22
Prehistory of Indo-Aryan Language and
Religion

In summarizing the book’s most important conclusions, I try to outline my research on the
early Aryan and Harappan religions and their survival in the Vedas and Hinduism. At the same
time I take the opportunity of making a few more suggestions, particularly in chapter 24,
hoping that these will stimulate further research.
The vocabulary common to the different Indo-European languages permits conclusions
about the natural and cultural environment of the PIE language. On this basis, the best-
informed scholars have long placed the PIE homeland in the Pontic-Caspian steppes, among
its pastoralist cultures—known as Khvalynsk, Srednij Stog, and Skelya—of the Neolithic and
Copper Age (c. 5000–3400 BCE). At present, leading researchers hold the Yamnaya cultures (c.
3300–2500 BCE), the immediately succeeding Early Bronze Age cultures of the Pontic-Caspian
steppes, as the archaeological correlate of the Late PIE phase. This hypothesis, however,
creates certain difficulties. The Corded Ware cultures, which must have played a central role
in the dispersal of the IE languages in northwestern Europe, cannot be credibly derived
directly from the Yamnaya cultures. Another problem is to explain where the PIE speakers got
their wheeled vehicles from. The PIE vocabulary relating to wheeled vehicles has emerged as
a particularly important clue to dating the PIE language. I make a suggestion that removes
these difficulties: the Early PIE speakers took over the Tripolye culture of Moldavia and
Ukraine, and the Late PIE language developed in the Late Tripolye culture (c. 4100–3400
BCE).
There is a fair degree of unanimity on the constituents necessary for this solution. It is
agreed that the pastoralist Skelya culture invaded the Balkans causing great destruction there
around 4300–4100 BCE, and a credible trail of archaeological cultures from the Pontic steppes
to Anatolia can be proposed as the archaeological correlate for the prevailing view that the
Anatolian languages were the first to separate from the PIE speech community and that they
are the only ones to preserve the laryngeal phoneme(s) reconstructed for Early PIE. It is also
agreed that the Skelya pastoralists, at the time of their invasion of the Balkans, also attacked
sites of the Tripolye culture, with which they had been longer in contact, the Tripolye culture
being the source of their metal and other prestige goods. The volume of the crude handmade
ceramics typical of the Skelya culture gradually increases within the Late Tripolye culture
until it is eventually the predominant type of Tripolye pottery, from which the Corded Ware
can be derived. It is further thought that the stone-headed maces, which are sometimes
shaped like a horse-head, have a steppe origin and were probably used as scepters, symbols of
power suggesting the importation of a chieftainship type of social structure that made the
Tripolye culture expand and flourish, but also led to internal strife.
The long-dominant view that the first wheeled vehicles were invented by the Sumerians in
the late fourth millennium BCE has been challenged by numerous new finds in Europe datable
to the second half of the fourth millennium. In the fourth millennium, the Tripolye culture
alone shows some evidence (in the form of model wheels and wagon-shaped drinking cups) for
wheeled vehicles before 3500 BCE. Being the most advanced culture anywhere in agriculture
and having the largest settlements of the Copper Age world, it also had the incentive to
improve the means of transport. There are many models of ox-pulled sledges, from which the
wheeled wagon is thought to have evolved from logs rolling under sledges with heavy loads.
Since inventors usually name their inventions, and the PIE vehicle terminology is derived from
PIE roots, by this time the language spoken in the Late Tripolye culture must have shifted
from non-Indo-European to Late PIE. The terms used for wheeled wagons in Kartvelian and
Sumerian make it appear that the invention spread to West Asia over the Caucasus, where the
steppe and Mesopotamian Uruk cultures met around 3500 BCE.
The Tripolye culture cultivated the most fertile fields in Europe with plows, and during the
fourth millennium its population increased phenomenally. When some kind of climatic crisis
around 3400 BCE brought this thriving culture to a sudden end, hundreds of Tripolye sites
were abandoned and tens of thousands of Late Tripolye people adopted the mobile pastoralist
lifestyle dispersing in wheeled vehicles in every direction. This dispersal of many people from
a central source created the two major cultural chains, both dated to c. 3300–2500 BCE, which
have long been held responsible for the spread of the PIE language to most places where its
branches first historically emerge: the Corded Ware or Battle Axe cultures, ranging from the
Netherlands to Russia across northwestern Europe, and the Yamnaya cultures, ranging from
the Danube to the Urals across southeastern Europe. Both the Corded Ware and the Yamnaya
cultures possessed wheeled vehicles, buried as prestige objects in elite graves, and pottery
and other artifacts that can be derived from the Late Tripolye culture.
The Yamnaya cultures came into being when the Late Tripolye people, probably speaking
Late PIE, spread to the Pontic-Caspian steppes. The archaic substratum is likely to have
differentiated the language of the Yamnaya cultures from the other early IE languages. The
Yamnaya languages spoken in the western half of the continuum (from the Danube to the
Dnieper, which has often functioned as a cultural border) are likely to have been ancestral to
Greek and Armenian, and those spoken in the eastern half (from the Dnieper to the Urals)
Indo-Iranian. These languages share the augment, the prefix *e functioning as a past-time
marker of verbs, and some other features not present in other IE languages. Unlike Greek,
Armenian is a satem language, so its ancestor is likely to have been closer to that of Indo-
Iranian and Balto-Slavic, which also share the innovative satemization. Because the ancestor
of Balto-Slavic with great probability has as its archaeological correlate the eastern variants of
the Corded Ware culture (Middle Dnieper, Fatyanovo, and Balanovo), the original homeland of
the Indo-Iranian languages cannot have been in Central Asia, as is often claimed, but has to be
in the Pontic-Caspian steppes. Only this location in the European steppes allows for the
development of the considerable differences that exist between the “Iranian” and “Indo-
Aryan” branches.
Another important, indeed crucial, component for the identification of the Proto-Aryan
homeland is the fact that Uralic (Finno-Ugric) languages contain about a hundred early Aryan
loanwords. Some of the words had been borrowed before some sound changes characteristic
of the Proto-Aryan language (as reconstructed on the bases of the later Indo-Iranian
languages) had taken place (Proto-West Uralic *kekrä vs. Proto-Aryan *cakra-), while some of
the loanwords had clearly come from Proto-Indo-Aryan (*mete-śišta-, “beeswax,” has an exact
counterpart in Sanskrit madhu-śiṣṭa, literally “what is left over of honey,” while even the verb
śiṣ-, “to leave,” lacks a counterpart in Iranian). The latest research suggests, moreover, that
Proto-Uralic had Aryan loanwords before its disintegration, which makes it possible to narrow
the possibilities for locating the homeland of Proto-Uralic, traditionally sought in the forested
eastern part of European Russia.
The above considerations have made it possible to suggest for the first time a more
definite network of archaeological correlations that comprises the whole web of Uralic and
Indo-Iranian languages (Parpola 2012 [2013]). Carpelan and Parpola (2001) proposed that the
Proto-Aryan community split into its two branches around 2300 BCE, when the Late Yamnaya
and Poltavka cultures of the forest steppes of the Upper Don and Lower Volga regions
expanded to the Mid-Volga and the Lower Kama rivers, forming the Abashevo culture (c.
2300–1850 BCE). The primary incentive for this expansion seems to have been the copper
resources of this area for which these presumably Proto-Indo-Aryan speakers fought with the
Balanovo extension of the Fatyanovo culture. The Abashevo people seem to have lived in
symbiosis with the local Garino-Bor people, who probably spoke Proto-Uralic.
Proto-Uralic has native terms both for “copper” (*wäśka) and for “tin” (*äsa), which occur
together in the compound *äsa-wäśka, “tin-bronze.” Tin-bronze was produced in the great
intercultural network of warrior traders that operated around 2100–1600 BCE and mediated
high-quality weapons and tools between the Altai Mountains (the source of the tin) and
Finland. This network came into being after the Abashevo culture had expanded further east,
to the other side of the Ural Mountains, and given rise to the Sintashta culture (c. 2100–1700
BCE) of the southern Urals and its eastern offshoot, the Petrovka culture (c. 2000–1700 BCE),
which reached the Altai Mountains. It seems likely that the European wing of this Sejma-
Turbino network was operated by Proto-Uralic speakers and was instrumental in the westward
expansion of the Uralic languages.
The Sintashta and Petrovka peoples were not only metallurgists but also pastoralists. They
started the Andronovo cultures that dominated the Asiatic steppes of Russia, Kazakhstan, and
Turkmenistan around 2000–1450 BCE. The elites of all these cultures possessed the horse-
drawn chariot, which was used for racing, hunting, and fighting, and buried together with its
owner. It appears that the horse was first yoked to pull the chariot in the Sintashta culture
around 2100–2000 BCE, after which it quickly spread in all directions.
From Sintashta both the domesticated horse and the chariot came to the semiurban
agriculturally based Bactria and Margiana Archaeological Complex (BMAC) in southern
Central Asia and to its extension around Tepe Hissar III in northern Iran. I have argued since
1988 that in the twentieth century BCE, rule in the BMAC was taken over by incoming Proto-
Indo-Aryan speakers, who apparently adopted wholesale this impressive culture that
preserved its original “Trans-Elamite” shape in its urban phase (c. 2300–1700 BCE). While the
distribution of the steppe cultures reaches the BMAC, it does not continue further south or
west, whereas the BMAC people came to the borders of the Indus Valley as early as around
1900 BCE and simultaneously spread to northern Iran and southwards to the Iranian plateau.
Since the twentieth century BCE, the BMAC people were in trade relations with Syria; Assyrian
merchants in Cappadocia were importing tin from southern Central Asia.
The easy adoption of a new culture by an incoming powerful minority was also assumed in
the case of the Late Tripolye culture, and is actually evidenced in the case of the Mitanni
kingdom. That the Mitanni rulers (c. 1500–1300 BCE) were Proto-Indo-Aryan speakers can be
seen only from texts. An archaeological correlate for the movement of Proto-Indo-Aryan
speakers to West Asia is the appearance of the Early West Iranian Grey Ware around 1500
BCE; its origins are seen in Gorgan Grey Ware that was connected with the BMAC extension at
Tepe Hissar around 1900 BCE. That the Mitanni Proto-Indo-Aryans initiated chariotry in West
Asia on a grand scale is seen above all from the adoption of the Indo-Aryan term marya-,
“chariot-warrior, nobleman,” with these meanings in the Hurrian, Assyrian, West Semitic, and
Egyptian languages. A newly found text suggests that the Hurrians had marya mercenaries as
early as about 1760 BCE (Eidem 2014). Before the westward move of the BMAC-related people
of Gorgan in northern Iran, they had probably fused with a new wave of Proto-Indo-Aryan-
speaking immigrants from the steppes of Kazakhstan, this time people of the Fëdorovo
Andronovo culture, who came to the BMAC in its late, posturban phase, around 1700 BCE.
Andronovo campsites surround practically all BMAC settlements, and several hybrid
Andronovo-BMAC cultures came into being in southern Central Asia.
In southern Central Asia, the BMAC was replaced by the Yaz I–related cultures (c. 1500–
1000 BCE). These belong to the “roller pottery” cultures that came from the European steppes.
After the departure of the Proto-Indo-Aryan branch (the formation of the Abashevo culture),
the Iranian branch remained in the Proto-Aryan homeland, the Pontic-Caspian steppes, where,
after the Yamnaya cultures, largely similar pastoralist cultures followed one another, of course
with varying distributions. But they did not move to the Asiatic steppes until around 1500 BCE,
when the newly formed “roller pottery” cultures suddenly expanded in every direction. This
seems to be connected with large-scale adoption of horse riding by Proto-Iranian speakers.
While Proto-Iranian has a special verb for “riding,” such a verb is not found in PIE or in Proto-
Aryan or in Proto-Indo-Aryan.
Having no known graves, the Yaz I people probably practiced “sky” burial (exposure to
carrion birds) and created the religion that would become, after its reform, Zoroastrianism,
around 1000 BCE. The cenotaphs that occur in increasing numbers in the BMAC suggest that
this kind of burial, which has no antecedents in the steppes, was adopted by Proto-Iranian
speakers in southern Central Asia. The terracotta figurines of horse riders from Pirak in
Baluchistan, datable to around 1500 BCE, are anthropomorphic except for their bird’s-beak
heads. I compare them with the mounted Saka horsemen from the frozen tombs of the Altai
Mountains (c. 500–200 BCE), who wear pointed felt caps, often topped by bird’s-beak heads.
The people of the Yaz I culture, who probably spoke Proto-East Iranian or Proto-Saka,
continued some traditions of the earlier BMAC culture. The Yaz I–related fortress of Tillya
Tepe in eastern Afghanistan is similar to the BMAC forts, and the tradition of building such
fortresses seems to have survived to the present day in the fortified manors of Pashto-
speakers in Waziristan. I suggest that the Yaz I forts are the Dāsa citadels conquered by the
Rigvedic Indo-Aryans on their way through the mountains from the Kandahar region (where
King Divodāsa was born on the banks of the Sarasvatī River) to Swat and the Punjab. Dāsa or
Dasa is an ethnic name of Saka speakers in the Indo-Iranian borderlands (meaning “man,
human being, hero” in Khotanese Saka and Wakhi), known as Dā(h)a and Daha in Greek and
Old Persian sources. Among the proper names of the Dāsa chiefs mentioned in the Rigveda is
the patronym Kānīta, which has been compared to the Scythian proper names Kanītēs and
Kanītos; some other names of Dāsa chiefs seem to come from the original non-Indo-European
language of the BMAC. This solution to the Dāsa problem provides a new basis for estimating
the date of the Rigvedic migration to South Asia; it may have taken place around the
fourteenth century BCE, which agrees with the usual estimates.
The Rigvedic Indo-Aryans would thus have arrived in South Asia some 500 years later than
the first Proto-Indo-Aryan-speaking immigrants coming from the “urban” phase of the BMAC. I
have been ascribing the linguistic and religious differences between the Rigveda and the
Atharvaveda to these two separate waves of Indo-Aryan speakers, whose gradual fusion is
reflected in the formation of the Vedic literature. This is a controversial issue, as traditionally
the evolution of the Vedic texts and religion has been seen as a unilinear development, the
differences between the Rigveda and the Atharvaveda being interpreted as due to a dichotomy
into a “hieratic” religion of the priests and a “popular” religion of the common people. This
view is mainstream opinion, so the question has been dealt with at length in this book. Among
the numerous linguistic differences, one is particularly important. In the original Rigvedic
dialect every original PIE *l had merged with *r; this rhotacism had happened already before
the Rigvedic Aryans entered South Asia. Although l was later reintroduced into the Rigvedic
language, where it occurs mainly in foreign names, the Rigvedic Sanskrit cannot be the source
of the Atharvavedic dialect, in which many words have preserved the original PIE *l. The
religious evidence will be discussed shortly.
The Vedic hymns were collected and the Vedic literature created during the rule of the
Kuru kings. Temporally, geographically, and culturally (in its use of the horse, iron, only a few
and small towns), the Kuru kingdom corresponds closely to the early phase (c. 1100–700 BCE)
of the culture distinguished by the Painted Grey Ware (PGW). The legendary war between the
Kauravas and the Pāṇḍavas, the main story of the Mahābhārata, seems to have taken place
during the latter phase (c. 700–350 BCE) of the PGW culture (which includes a number of
towns, such as Mathurā). None of the Pāṇḍava heroes is known to the Vedic literature before
the Gṛhyasūtras, while the grammarian Pāṇini (c. 350 BCE) mentions the Mahābhārata and
several key names associated with it.
I suggest connecting the Pāṇḍavas with the people who came to South Asia via Sindh,
Gujarat, and Rajasthan around 800 BCE and brought with them the so-called Megalithic
culture, adopting the earlier local Black-and-Red Ware as their pottery. The father of the five
Pāṇḍava brothers is Pāṇḍu, meaning “pale,” who was destined to be born pale-skinned; the
name (which is Sanskrit with a Dravidian etymology) seems to represent the local Indian
appellation for newly arrived foreigners with a slightly paler skin color. The Megalithic graves,
the horse equipment found in them, and the associated round, yurt-like houses have parallels
in Central Asia, the Caucasus, and western Iran, and suggest that the immigrants originally
spoke an Iranian language, a hypothesis supported by the un-Indian polyandric marriage of
the Pāṇḍavas, which can be compared with customs of Iranian tribes.
The Rāmāyaṇa seems to reflect the movement of the Megalithic culture to Sri Lanka and
southern India, and the takeover of power in these regions by Aryan-speaking adventurers
from the Indus Valley, Gujarat, and the Mathurā region. The Buddhist chronicles of Sri Lanka
link the Aryan conquest of the island with Prince Vijaya, who came from Gujarat and raided
towns on the west coast. The Pāṇḍya kings of Tamil Nadu and their capital Madurai, the
southern Madhurā, are connected with the northern Madhurā, that is, Mathurā.
There is general agreement that the “family books” 2–7 represent the oldest part of the
Rigveda and most truly reflect the original religion of the Rigvedic Aryans. In these old books,
the chief deity is Indra, who is worshipped by offering praise songs and the sacred drink soma.
The Aśvins or Nāsatyas, the twins who circle around the world in their flying chariot, are
divine medicine men, deities of lower rank who were originally excluded from the soma cult.
Mitra and Varuṇa, another divine dual, vie with Indra but have remarkably few hymns
addressed to them. Mitra and Varuṇa, Indra and the Nāsatyas (but no other Aryan god) are all
invoked by the Mitanni king Sātivāja to be the divine witnesses of his pact with the Hittite
king around 1350 BCE. They seem to have been the most important members of the Proto-
Indo-Aryan pantheon at that time, presumably around 1500 BCE, when Mitanni rule started.
Here Mitra and Varuṇa hold the first place, while in the Rigveda Indra is clearly the principal
deity.
My suggestion is that the religion of the two waves of Proto-Indo-Aryans that came to the
BMAC around 2000–1900 BCE and 1700–1600 BCE had already slightly drifted apart. The later
wave, that of the Fëdorovo Andronovo people, worshipped Indra. Their cultic drink *sauma
was neither of Proto-Indo-European nor—on the evidence of Herodotus and the Aryan
loanwords in Uralic languages—of Proto-Aryan origin, but seems to have been adopted
somewhere in Central Asia. (Mummies discovered in Xinjiang and connected with Andronovo
people have pouches filled with Ephedra twigs in their funeral dress.) Indra, however, may
have been the original Proto-Aryan head of the pantheon, the inherited PIE god of the “sky”
and the “day,” like Zeus (< *Dyeus) in Greece and Ju(piter) (< *Dius) in Rome. Indra is called
deva-, “god,” and has the epithet dyumat-, “shining, splendid, heavenly,” both from the root
dyav- / div- (PIE *dyev- / div-), “to shine.” The word for “(highest) god” in Proto-Uralic was
*juma (i.e., yuma) from Proto-Indo-Aryan *dyumat-. The name Indra itself has proved difficult
to explain etymologically. I propose deriving it from *Inmar, the Proto-Central-Uralic name for
the god of sky and thundery weather. The West-Uralic cognate Ilmari (from ilma, “air, weather,
atmosphere”) occurring in Finnic epic poems is the name of the divine smith who has made
the sky and its luminaries; he also created the sky-propping world pillar sampo / sampas <
Proto-Indo-Aryan *stambha-s. People belonging to the probably Proto-East-Uralic-speaking
Cherkaskul’ culture of the Mid-Urals collaborated with the Fëdorovo Andronovo people, and
some of them became pastoralists, moving long distances with Andronovo people, even to
southern Central Asia.
The first wave of Proto-Indo-Aryan immigrants to the BMAC came from Sintashta in the
south Urals, and are supposed to have invented the horse-drawn light chariot around 2100–
2000 BCE. The chariot must have enjoyed immense importance in their culture. The layout of
their fortified settlements is circular and resembles the chariot wheel. The horse-chariot was a
crucial marker of rank, being the most important prestige symbol of elite graves. With the
chariot there came into being new deities, who do not have a PIE ancestry as is usually
assumed. The new gods associated with the horse-chariot are called “the two sons of the Sky”
in the Indo-Aryan, Greek, and Baltic languages: ever since PIE times, the Sky had until this
time been the highest god, and his paternity gave prestige to the new gods, who were raised
to the top of the divine pantheon. It appears to me that the two-man chariot team was made
into a model for a real system of dual kingship, as in Dorian Greece; in India, this dual
kingship survives in the collaboration of the king with the priest of the royal house, the
purohita. The king was the chariot-warrior, his high priest the charioteer—as they were in
ancient India in bygone times, according to the Jaiminīya-Brāhmaṇa (3,94).
While the Aśvins clearly are the Indo-Aryan counterparts of the Dorian Greek Dioskouroi,
they are not dual kings. Instead, Mitra-and-Varu ṇa are dual kings. To explain this, I have
suggested that the Aśvins were dual kings when they came to the BMAC. But when the early
Proto-Indo-Aryans of the BMAC had trade relations with the Assyrians from the second half of
the twentieth century BCE, they probably became acquainted with the Assyrian religion, where
social virtues are represented as subsidiary deities, personifications of the qualities of the
high god Aššur, surrounding Aššur like a king is surrounded by his closest officials. I assume
that the Proto-Indo-Aryans of the BMAC imitated this Assyrian model and provided their own
highest gods (i.e., the Aśvins) with similar attributes, namely the Āditya gods. For some
reason, the two most important ones among these new attribute gods, Mitra and Varuṇa, came
to be associated with the Aśvins’ dual kingship, and the Aśvins were reduced to the role of
saviors, divine medicine men, and funeral gods. Yet they preserved their inherent connection
with the horse-drawn chariot and its two-man team.
In the steppes, chiefs were buried with their chariots. All three divine twins seem to have
been funeral gods, too, associated with a funeral chariot race both in Greece and India (if my
suggestion about Yama’s prize-race in Rigveda 1,116,2 is accepted), and a funeral horse-riding
race in the Baltics. The charioteer’s chief duty was to “save” his master, the chariot-warrior,
and “bring him safely home.” This latter sense is the meaning of the PIE root *nes-, from
which derives Greek Nestor, the Homeric hero famed as a charioteer; Greek nostos, “home-
coming,” corresponds with Sanskrit *nasati-, “homecoming,” which is implied by the Aśvins’
alternative name, Nāsatya. Like the Aśvins, the Greek Dioskouroi act as sōtêres, “saviors,”
who help people in dire need. In the Rigveda, the Aśvins save decrepit people—sometimes
described as lying underground, “like the dead”—by rejuvenating them, which suggests to me
that they look after the rebirth of the deceased.
The Aśvins’ function as funeral deities seems to be confirmed by the funerary “face urns”
of the Gandhāra Grave culture. The connection with the Aśvins is suggested by the urns’ lids,
which have horse-shaped handles, and above all by the large “nose” that stands out from the
“face” of many of these vessels. PIE *nes-, “to come safely home,” has become in Indo-Iranian
nas-, “to come safely home.” The existence of the homophonous word nas-, “nose,” has given
rise among Indo-Aryan-speakers to the myth of the Nāsatyas’ “nose-birth” (their conception is
said to have taken place when their mother, in the shape of a mare, sniffed the semen
inadvertently laid on the ground by the Aśvins’ stallion-shaped father). The myth in turn has
led to rituals of generation, in which material symbolizing seed is inserted into the wife’s nose.
The nose thus symbolizes (re)generation.
A “nose” similar to those of the Gandhāra Grave face-urns is prescribed also for the
mahāvīra or gharma pot sacred to the Aśvins; this gharma pot is said to be the head of a
decapitated hero. At the end of the gharma ritual this pot, along with other utensils of the
ritual, is to be laid out on ground in the shape of a human being. This ritual act has a parallel
in the Vedic funeral, where a deceased sacrificer’s ritual implements are laid on the various
parts of his corpse. The two officiating priests in charge of the gharma ritual are the adhvaryu
and the pratiprasthātar, who personify the two Aśvins among the Vedic sacrificial priests.
In the Rigveda, the worship of Indra with the soma cult clearly predominates, although the
inclusion of hymns to Mitra, Varuṇa, and the Aśvins in the “family” books suggests that their
cult had already been adopted in southern Central Asia, as does the Mitanni pact of around
1350 BCE. On their arrival in South Asia, however, the Rigvedic Aryans appear to have met
Indo-Aryan speakers originating from the urban phase of the BMAC whose principal Aryan
divinities were the Aśvins and Mitra-Varu ṇa, but who had not previously known Indra; on the
other hand, they had long been exposed to the substratum influence of the Indus people and
their descendants.
The Kāṇva and Āṅgirasa poets, whose traditions favored the tristich structure and the
gāyatrī and pragātha meters, belonged to the early wave of Proto-Indo-Aryan immigrants to
South Asia. It appears that the Atri clan of the fifth book of the Rigveda, which is part of the
“family” books and therewith belongs to the later, Rigvedic immigrants, adopted the worship
of the Aśvins in Gandhāra from the Kāṇvas. The Atri clan adapted the worship of Aśvins as
funeral gods to their newly introduced cremation burial by developing the gharma or
pravargya ritual, which seems to be connected with the “face urns” of the second phase of the
Gandhāra Grave culture (in 2014 dated to c. 1100–800 BCE).
Part of the Kāṇvas and Āṅgirasas appear to have joined the Indra-worshippers of the
Rigvedic Aryans more or less immediately after the latter’s entrance to South Asia, and
created the Sāmaveda. The earliest references to sāmans in the Rigveda are connected with
the gharma ritual, hence the worship of the Aśvins, whose original sacred drink appears to
have been honey-beer (madhu-surā). From the sautrāmaṇī ritual we know that the preparation
of surā included its purification through a filter of horsehair, while in the yasna ritual of the
Zoroastrians the haoma was filtered through bull’s hair. It seems to me that the sāmans
originally used in the Aśvin-related rites were now adapted to the worship of Indra and his
soma drink, resulting to the first great addition to the collection of Rigvedic hymns after the
“family” books. The principal priests of the Yajurveda are, again, the adhvaryu and the
pratiprasthātar, officiants of the gharma/pravargya ritual, said to be the human counterparts
of the two Aśvins. This suggests that priests who were once Aśvin-worshippers probably had a
major role in the creation of the Yajurveda.
Some Kāṇvas and Āṅgirasas remained faithful to their original traditions, however. Their
Atharvavedic hymns were accepted as part of the Vedic culture after they too had at least
nominally accepted the Indra cult. This happened in the last additions to the Rigveda,
especially in Book 10, where many hymns have an Atharvavedic character. But the core of this
earlier, Atharvavedic religion is in many respects rather different from the Rigvedic religion.
Particularly noteworthy is sorcery—harming enemies and rivals of the king by magic means—
which is among the primary duties of the royal priest, the purohita.
Consonant with this violent character of the Atharvaveda are the Vrātya rituals, first
described in the Atharvaveda (Book 15 in the Śaunaka recension) and the Brāhmaṇas of the
Sāmaveda. The vrātyastomas are performed at the beginning and end of raiding expeditions
by their participants. The details of the vrātyastomas connect them with archaic royal rituals
(the horse or human sacrifice, the royal consecration, the vājapeya, and the piling of the fire
altar, agnicayana), as well as the “popular” festival of mahāvrata performed at the end of the
year, usually coinciding with the completion of the fire altar. Besides the gods of ancient Aryan
background, the Aśvins, Varuṇa, and Bṛhaspati (who is a divine purohita and charioteer),
important divinities in these rites as well as in the Atharvaveda and the Brāhmaṇa literature
(which linguistically continues the Atharvavedic rather than Rigvedic tradition) there were the
creator god Prajāpati, the goddess Vāc and Rudra. It is these deities, in my opinion, who are
most likely to be survivals of the main gods of the Harappan pantheon, though we can also
expect that gods who are originally Aryan, such as Varuṇa, may have absorbed some of the
functions and attributes of their approximate counterparts in the religion of the Indus
civilization.
The Greek ambassador Megasthenes, who lived in Pāṭaliputra around 300 BCE, tells the
legend of Princess Pandaíē (i.e., Pāṇḍu’s daughter), who was married by his own father, the
Indian Heracles. The story resembles the Sāvitrī legend, which connects this myth with the
Vedic goddess Sītā-Sāvitrī, and thus with Sītā, the heroine of the Rāmāyaṇa. The Indian
Heracles can be identified as Bala-Rāma—according to Cicero, this Heracles was called Belus
—while in the Mahābhārata, Bala-Rāma is called simply Rāma, and seems to be originally
identical with the hero of the Rāmāyaṇa. In the Rāmāyaṇa, Sītā (whose name means “furrow”)
was born out of a furrow when King Janaka (“generator, father”) was plowing a field, and Bala-
Rāma is a deity whose characteristic weapon and attribute is the plow.
Mathurā is connected with the worship of Kṛṣṇa (“black”) and his elder brother Bala-Rāma
(whose skin-color is white). They belong to the Yādava clan, whose ancestor Yadu is connected
with one of the early tribes mentioned in the Rigveda, apparently belonging to the earlier
Proto-Indo-Aryan wave of Aśvin worshippers. The Pāṇḍavas united with Kṛṣna of Mathurā, and
Kṛṣṇa served as the charioteer of the main Pāṇḍava hero, Arjuna, whose name means “white.”
In my opinion this chariot team reflects the Aśvins, whose cult had survived among the
Yādavas of Mathurā. The same applies to the brothers Bala-Rāma and Kṛṣṇa, who are not on
good terms with Indra, for example in the Govardhana episode. The colors white and black
suit the Aśvins, who drive around the world in their airborne chariot in a day, apparently
personifying, like Mitra and Varuṇa, the day or sun, and the night or moon.
Naturally, this Aśvin heritage is only one side of these divinities. Kṛṣṇa as both a naughty
boy and a dark-skinned native cowherd strikingly resembles the low-caste Keralan deity
Kuṭṭiccāttan, whose myth and cult suggest how an originally low-caste or aboriginal deity
might have been adopted into the Brahmanical pantheon (Parpola 1999c). Bala-Rāma, in his
turn, as an addicted toddy-drinker seems to have replaced an earlier local deity worshipped
for the same (in)capacity, a demon in the shape of a wild ass (Dhenuka in Harivaṁsa 57). The
wild ass (Sanskrit gaura-, “wild ass,” not “bison” as in dictionaries) is an animal with a
phenomenal capacity to drink large amounts of water, mentioned as a model for the soma-
thirsty Indra in the Rigveda. It lives in the salt desert, and seems to have been worshipped in
the Indus civilization as a god of death and fertility (qualities also ascribed to salt) (Parpola &
Janhunen 2011). Bala-Rāma also seems to continue as an earlier agricultural divinity
connected with the plow, perhaps a predecessor of Hindu Śiva, whose names include
Lāṅgaleśvara, “Lord of the Plow.”
23
Harappan Religion in Relation to West and
South Asia

One of the most crucial problems in the investigation of the Indus civilization and its religion
is its linguistic identity. With the Indus population estimated at about one million, the
Harappan language could not possibly have disappeared without leaving traces in the Vedic
texts, which began to be composed in the Indus Valley toward the end of the second
millennium BCE. I think it can now be stated with fair confidence that there was only one
written Harappan language, and that it was an early member of the Dravidian language
family, by far the largest in South Asia after Indo-Iranian. This conclusion is supported by the
many Dravidian loanwords in the Vedic language, including the oldest available source, the
Rigveda. Among these early Dravidian loans (which include several new etymologies proposed
here) are religious terms that are important in both the Vedic and Hindu religions, such as the
sacred syllable oṁ. Furthermore, the present-day distribution of Dravidian languages in
central and southern India favors Dravidian as the Harappan language, since the Chalcolithic
cultures of Gujarat, Rajasthan, Maharashtra, and Karnataka are derived from the Early and
Mature Harappan cultures of the Indus Valley. In addition, there is the Dravidian kinship
system (radically different from the Aryan system of North India) and Dravidian toponyms,
which span from southern India up to Gujarat. None of the minority languages of South Asia,
distributed in marginal areas, is a serious competitor with Dravidian.
The Dravidian identification is borne out by the Indus inscriptions. They are written in a
language-based logosyllabic and pictographic script, based on the same principles as those of
other very early writing systems, such as Archaic Sumerian and the Egyptian hieroglyphic.
Using methods proven in successful decipherments, it is possible to interpret two or three
dozen Indus signs phonetically, even without bilinguals. The interpretations can be checked
because they form compound words that support each other like a crossword puzzle, and also
because they agree with external clues on Harappan objects and above all with compound
words attested in Dravidian languages.
Another problem concerns the cultural continuity from the Indus civilization to historical
times, which has often been questioned. One striking testimony to an unbroken continuity of
Harappan cults down to the present day is provided by the unique crocodile worship in fifty
tribal villages of southern Gujarat. Horizontal wooden crocodile images erected upon vertical
poles by these Gujarati tribals in their sanctuaries have a precise ancient antecedent painted
on a Mature Harappan potsherd excavated at Amri in Sindh, depicting a pair of gavials, fish-
eating river crocodiles, without back legs but with a projection sticking out of the body at a
ninety-degree angle, which anchors the crocodiles to the “ground” of the scene. The swelling
protuberance, which is at the tip of their long and narrow snout in the males, is hunted as a
powerful aphrodisiac, rendering the animal almost extinct. A tablet from Harappa depicts a
male gavial pushing its snout into the vagina of a human female with splayed legs. A tablet
from Dholavira seems to depict a sacrifice of little children to crocodiles. Votive offerings of
first-born babies to crocodiles were performed by couples wanting offspring in Bengal until
around 1810 CE; a similar promise to sacrifice his first-born son was forced upon King
Hariścandra when he approached god Varuṇa for an heir in the Vedic Śunaḥśepa legend
recited at the royal consecration. Varuṇa is a god of waters and the lord of aquatic animals,
who in Hindu iconography rides a crocodile. He is also connected with the night sky, and a
heavenly crocodile consisting of stars, the pole star being in its tail, is mentioned in Vedic and
Purāṇic texts.
Here we have entered the topic of “village Hinduism,” the essential components of which
appear to date back at least as early as the Early and Mature Harappan villages of the third
millennium BCE. A mother goddess as the guardian deity of the village, her husband or servant
symbolized by the bull or buffalo, and their worship in the shape of clay or stone images, are
basic elements of this age-old folk religion—likewise the cult of sacred trees and their divine
inhabitants. Fig trees, especially the pipal and banyan, the bull, the water buffalo, the fish,
and the peacock are some of the important motifs of Early and Mature Harappan painted
pottery, which are also among the principal religious symbols of “village Hinduism.” Its main
concerns are the promotion of human and animal fertility and averting disease, drought, and
other misfortunes. John Marshall (1931) suggested an unbroken continuity in Indian folk
religion from Harappan times to the present day, and this thesis has enjoyed wide scholarly
support.
In my efforts to penetrate deeper into Harappan ideology, I have tried to interpret the
manifestations of Early and Mature Harappan religion by relating them to the better- known
and better-understood religions of later South Asia, and also those of West Asia, which was in
direct contact with the Indus civilization and shared a comparable economic foundation.
During the first quarter of the third millennium BCE, the relatively short-lived Proto-
Elamite civilization spread from the region of Susa over most parts of the Iranian plateau,
giving rise to the “Trans-Elamite” cultures of Kerman, Seistan, and southern Central Asia,
reaching as far east as the Makran in Pakistan. The Proto-Elamites had a language-based
pictographic script. The idea of such a script was therefore within reach of the Early
Harappans, whose ceramics have been found at Shahr-i Sokhta in Iran. This site in Seistan has
yielded a tablet in the Proto-Elamite script. I suspect that the “yoga posture” of the Harappan
anthropomorphic deities in the seal iconography is derived from the “sitting bulls” of Proto-
Elamite seals. The horns of the Harappan gods may also have prototypes in Proto-Elamite
iconography (where human beings and gods are represented with animals), although gods and
kings with horns are also found in Mesopotamia and in many other places.
Indus merchants appear to have first arrived in northern Mesopotamia overland, following
caravan routes via which lapis lazuli was imported to West Asia from Bactria—called Aratta in
Sumerian sources, which seems to me the original location of the country of Āraṭṭa mentioned
in Sanskrit sources. In any case, Harappans would have been present in Mesopotamia towards
the end of the Early Dynastic period, since they mediated to the Indus Valley art motifs and
modes of dressing—particularly the royal hair-dress—that were current in West Asia in those
times.
From the Old Akkadian to Ur III periods (c. 2300–2000 BCE), cuneiform documents refer to
a sea trade with the Greater Indus Valley, called Meluhha. Sargon himself boasts that these
far-off foreigners moored their ships at the dock of his newly built capital city, Akkad. They
must have brought the great king living water buffaloes as jpgts, as buffaloes suddenly appear
in the iconography of royal cylinder seals, replacing the wild urus bull in the favorite Akkadian
motif. This “contest” motif had developed from the fight between the lion and the bull in
Proto-Elamite seals. In the earlier Late Uruk iconography of Mesopotamia and Egypt, the king
is depicted as holding back two lions with his bare hands, or as hunting these mighty beasts.
In the Indus Valley, too, a male figure holding back two tigers with his bare hands is likely to
have the same symbolism of power.
Some Meluhhans stayed in Mesopotamia and the Gulf for generations, becoming fully
integrated in the local society, speaking Sumerian and even giving their children Sumerian
names. At the same time, as trade agents, they also spoke their native Meluhhan language,
which was not generally understood in Mesopotamia. Translators were needed. Occasionally
Harappan agents wrote their foreign names in the Indus script on seals. Sumerians and
Akkadians, on the other hand, never seem to have come to the Indus Valley, though some Old
Akkadian kings in their war expeditions toward the east conquered the Kerman region in
southeastern Iran, the seat of the recently discovered Jiroft culture. One cuneiform text
reports that among the participants of a worldwide rebellion against the Old Akkadian ruler
Narām-Sīn was a king from Meluhha. Unfortunately the tablet is broken and we know only the
last part of this Meluhhan king’s name, “[. . .]-ib-ra.” This name is the one and only known
record of the political history of the Greater Indus Valley, in the form of a peripheral member
of the West Asian cultural sphere.
Cultural influences came from West Asia to the Indo-Iranian borderlands much earlier than
historical times. A settled way of life with agriculture and animal husbandry is supposed to
have started in Anatolia some ten thousand years ago or earlier (the newly discovered, and
impressive, temple at Göpekli Tepe in Anatolia is dated to 13,000 BCE!), from which it
gradually spread. Around 6000 BCE the village of Çatal Höyük in Anatolia had a temple of a
Mother Goddess, seated on a throne guarded by panthers, and worshipped with bull
sacrifices. During the following millennia a similar cult (with variations) is found all over West
Asia and the Mediterranean. The Goddess, usually associated with the lion and the planet
Venus, presides over birth and death, granting fertility and military success. She is the
personification of the Earth, which yields grain and other plants including fruit trees, and she
receives in her womb both the dead and the seed grains, “sacrificed” but reborn as the new
harvest. The bull personifies her husband, the Sky, who roars as thunder and fertilizes the
Earth with his rain-seed.
The marriage of this primeval couple was celebrated at the new year, the bull being
sacrificed and the people feasted with its meat. The king, who was responsible for the welfare
of the country, represented the son-husband of the Goddess, and as her protégé could count
on victory in war. He did not, however, personally perform this role in the sacrificial drama
enacted on the greatest festival of the year, the “sacred marriage” of the Goddess
(represented by her high priestess) at the new year. The drama ended with the death of the
bridegroom, which was ritually lamented. This and related rituals elsewhere in West Asia
developed into secret mysteries, in which participants killed human victims, ate the flesh, and
drank the blood.
In the Proto-Elamite “contest” seals, the lion and the bull are both alternately represented
as winners and losers. This has been interpreted to mean that they stand for opposite and
complementary cosmic powers, not only female (lion) and male (bull), but also sun/day (lion)
and moon/night (bull), fire (lion) and water (bull), and so on. The Sun and Moon are the
principal heavenly bodies, whose movements define the time, and are the prime objects of
study in astronomy/astrology, which in West Asia was the most important science, used for
making predictions, especially about the fate of kings and countries. Very significant, too,
were the five planets visible with the naked eye, distinguished by their independent movement
and brightness from the fixed stars, which were also important in marking the route through
the heavens of the sun, the moon, and the planets, which were all considered divinities in their
own right or astral aspects of gods known by other names.
Settled life in the Indo-Iranian borderlands starts around 7000 BCE. The principal religious
artifacts are terracotta images of human females and bulls, which presumably have the same
meaning as their counterparts in West Asia: Mother Earth and Father Sky. Similar images
have been made in South Asian villages ever since, and in most places the principal yearly
festival of the villagers has been the marriage of the guardian goddess of the village, involving
the sacrifice of the bull or buffalo who represented the goddess’s husband. Everywhere in
South India the water buffalo was until recently the preferred victim for the goddess, although
other male animals could be offered—nowadays they are chiefly goats or chickens. The
Daimabad hoard not far from Mumbai with its bronze statue of a water buffalo proves that it
was a sacred animal in central Maharashtra in Late Harappan times around 1800 BCE.
In the Old Akkadian seals, the water buffalo imported from the Indus civilization shares
the symbolism of the wild bull of the earlier Mesopotamian iconography. Harappans killed the
water buffalo by spearing it. The spearman’s hair is bound into a “double-bun” at the nape of
the neck, which was the hairdo of the royal warrior in West Asian art around 2400 BCE. The
Harappan buffalo-spearer also places his foot on the head of the beast, like the naked hero
vanquishing the buffalo on Akkadian seals. A seal from Chanhu-daro depicts the sexual
intercourse of a bison bull and an anthropomorphic female lying on the ground with open legs.
Thus in the Harappan “sacred marriage” the sacrificed bridegroom of the goddess was a
water buffalo or bison bull. The sacrificial buffalo may be assumed to personify the important
Harappan god called Proto-Śiva, who wears the horns of the water buffalo, and his human
counterpart, the king.
The Harappan “sacred marriage” as a royal institution was taken over by the first
(“Atharvavedic”) wave of Indo-Aryan speaking immigrants, who replaced the water buffalo as
the sacrificed bridegroom with the horse, their own most prestigious sacrificial animal. A
reminiscence of the Harappan buffalo is the title of the king’s chief queen who lies with the
horse, mahiṣī, “water buffalo cow.” The chief queen is said to personify the Earth and the most
important goddess of the Vedic Brāhmaṇa texts, Vāc, “Sound.” As the goddess of victory, Vāc is
called siṁhī, “lioness,” and is worshipped at the mahāvrata festival with the sound of war-
drums and other musical instruments. Vāc is the daughter of the creator god Prajāpati,
approached by her father in an incestuous “sacred marriage” that created all beings, but led
to Prajāpati’s slaughter as punishment for his sin. This marriage is imitated by couples
participating in the mahāvrata, who do not need to care for social differences.
The mahāvrata was originally celebrated at the autumnal equinox, the traditional time of
starting war expeditions after the rains. As a vrātya ritual, it had the same function as the
vrātyastoma, uniting the warriors leaving for war. It survives to the present day in the
navarātri/daśahrā festival of Goddess Durgā, who, on the “tenth day of victory,” vanquishes
her ferocious opponent, the Buffalo Demon, who according to some variants of the myth, made
amorous passes or even married her on false pretexts, thereby leading to his slaughter. Durgā
has been the guardian goddess of many Indian royal houses. On leaving for a war expedition,
kings have lustrated their armies at the navarātri/daśahrā festival, and sometimes offered
hundreds of water buffaloes to the goddess to capture her favor and obtain victory.
The myth of the lion-riding warrior goddess Durgā’s fight with the buffalo demon Mahiṣa
Asura emerges in Indian literature very late, around 700 CE, but by means of iconographical
representations it can be traced back to the first centuries CE, and is supposed to have been
introduced to India by the Kuṣāṇas, who a little earlier had in Afghanistan adopted the local
worship of the lion-escorted goddess Nanaya. Nanaya’s cult was brought to Afghanistan by
about 2000 BCE from West Asia, where she was a variant of Inanna/ Ištar, the goddess of love
and war. But a tiger-riding goddess of war is attested also in a Harappan cylinder seal from
Kalibangan. A few Rigvedic hymns, in which hundreds of water buffaloes are killed for Indra
to consume before a battle, prove the existence of such rituals for the sake of victory in the
Indus Valley at the end of the second millennium BCE, when the Rigvedic Aryans were
reaching South Asia. The water buffalo is a South Asian animal, and the almost complete later
silence about the sacrifice of the water buffalo in subsequent Vedic texts shows that it was an
alien custom. The water buffalo sacrifice is mentioned in later Vedic literature in one context
only: the animal is one of the many subsidiary victims in the horse sacrifice, sacred to Varuṇa,
the divine king.
In the BMAC-related religion of the Indo-Aryan speakers, Varuṇa seems to have originally
represented the “priest-king,” the charioteer of the two-man royal chariot team of the Aśvins,
associated with the moon, night, and water (while Mitra was the chariot-warrior associated
with the sun, day, and fire). The first Indo-Aryan speakers coming to South Asia substituted
Varuṇa for the divine king of the Harappan religion, who was the god of waters, and the god of
creation and death, surviving in the Brāhmaṇa texts as the creator god Prajāpati, who was
killed in punishment of his incest (thus becoming the first dead person and the king of the
deceased, like the god Yama of the originally Iranian tradition also preserved in the Veda).
That Varuṇa inherited the position and attributes of the Harappan “priest-king” is also
suggested by Varuṇa’s tārpya garment, with which the Vedic king is invested in the ritual of
royal consecration. The tārpya garment is said to be decorated with images of dhiṣṇyas, a term
that denotes both a “sacrificial hearth” and a “star,” and agrees with the Vedic conception that
the stars are ancient sages transferred to the sky who stand by their sacrificial hearths. The
Priest-King statuette from Mohenjo-daro is draped in a garb decorated with trefoil motifs
originally filled with a red paste. The trefoil has an astral significance in West Asia, where it
covers statuettes depicting the “Bull of Heaven”; a fragment of a trefoil-decorated bull-
figurine has been excavated at Mohenjo-daro, too. Mesopotamian kings and gods were
dressed in a garment decorated with the figures of stars, rosettes, and crosses, called a “sky-
garment,” in imitation of the starry night sky.
In the elite religion of the Indus civilization, as in Mesopotamia and other early urban
civilizations, astronomy and time-reckoning played an important role from the Early Harappan
period. This is clear from the orientation of the towns according to the cardinal directions,
from sun-like art motifs and from what seem to be gnomon stands. Vedic and epic traditions
connected with the nakṣatra calendar suggest that its origins date to about 3000 BCE when the
new-year star was Aldebaran (Sanskrit rohiṇī), observed when it rose together with the sun at
the vernal equinox. Around 2400 BCE, the calendar was apparently revised by the Indus
people, who made it luni-solar and shifted the new year to the sun’s conjunction with the
Pleiades (Sanskrit kṛttikāḥ). This shift seems to be reflected in the principal astral myths of the
Veda and epic-purāṇic Hinduism, the abduction of Tārā (“star”), the wife of Bṛhaspati, by Soma
the Moon; and the birth of Skanda-Kumāra in the place where the Kṛttikās were bathing in the
(heavenly) Ganges. Names of the planets known from Old Tamil texts can be plausibly read in
Indus texts, while the Vedic tradition is curiously silent about these luminaries of the sky.
Particularly important is the identification of the Old Tamil name of the pole star in the
Indus inscriptions: vaṭa-mīn, which literally means “north star,” through Dravidian homophony
(utilized in the Indus script) also means “rope star” and “banyan star.” These meanings make
it possible to understand the concepts of a heavenly banyan tree, which the Rigveda (1,24,7)
connects with the divine king Varuṇa, and the idea fundamental to the Purāṇic cosmology,
according to which all the stars and planets are maintained in the heavens by invisible ropes
binding them with the pole star. Mantras addressed to the pole star by the Vedic bridegroom
preserve a tradition that probably once was part of the royal consecration, making the pole
star a symbol of kingship as it was in China.
The banyan, the mightiest native tree of South Asia, got its name vaṭa- (not attested in
Sanskrit before the epics) from the Proto-Dravidian word *vaṭam, “rope,” which refers to the
hanging air roots characteristic of the banyan. From RV 1,24,7 we know that Vedic people
thought the invisible air roots of the heavenly banyan were attached to living beings,
supplying them with life-energy. The Sanskrit word raśmi-, which is used of them, means not
only “rope” but also “ray,” a meaning confirmed by the word ketavaḥ, “rays,” applied to them
in this context in RV 1,24,7. Thus they can fully be compared to the rays of the sun, which
according to the Vedic and Purāṇic conception feed the heavenly bodies and living beings. This
parallelism between the rays of the sun and the air roots of the heavenly banyan tree strongly
suggests a Harappan-Dravidian ancestry to some related cosmological conceptions. They also
provide new evidence for the hypothesis that beginnings of yoga lie in the Indus civilization.
Thus the Purāṇa texts have specific names, suṣumṇa- and others, for the principal rays of
the sun, which feed the moon and the five planets. Suṣumṇa- is known as the name of a solar
ray in the earliest Yajurvedic texts, and the old Upaniṣads know suṣumṇā- as the name of the
principal vein in the human body, which is said to have 72,000 veins. Moreover, the veins of
the human body are linked with the rays of the sun, and they function as paths through which
the soul can move from human heart to the sun in deep sleep and at the moment of death, the
sacred syllable oṁ being connected with this passage. These ideas are clear precedents of
concepts central to the later haṭha- or kuṇḍalinī-yoga, and their Harappan origin is further
supported by the fact that the old Upaniṣads ascribe to both veins and solar rays the same five
colors that the five planets have in Indian iconography, and from which derive their Dravidian
names, attested both in Old Tamil texts and in Indus inscriptions.
The parallelism of the air roots of the heavenly banyan and the rays of the sun also make it
likely that the pole star is the nightly counterpart of the midday sun of the zenith. Vedic texts
have preserved the conception that the sun does not go beneath the horizon in the evening
but simply turns around so that its dark side is toward the earth and then returns invisible
through the sky to the east, where it turns its bright side toward the earth. The Brāhmaṇa
texts mention as a symbol of the sun the tortoise. The starred tortoise of South Asia has a
black shell studded with several bright star-like ornaments.
Several Dravidian names of the tortoise are derived from verbs denoting “contraction,”
which refers to the tortoise’s habit of drawing its limbs beneath its shell. These verbs also
mean “to go into hiding” and, when speaking of the sun, “to set,” as the setting sun draws in
its rays like the tortoise its limbs. Significantly, the Dravidian verbs just mentioned are also
used of an ascetic or yogi who is curbing his mind by drawing the senses back from their
objects, and also the Bhagavadgītā and other texts compare this act of pratyāhāra to the
“contraction” of the tortoise.
The Brāhmaṇa texts specifically connect the tortoise with the head and its seven openings,
which represent the sense organs and are associated with the Seven Sages. The Seven Sages
are identified with the seven stars of Ursa Major, next to the pole star, whose place in the
zenith corresponds to the suture at the top of the human skull, the brahma-randhra through
which the soul is supposed to exit from the body, corresponding to the sun as the door to the
world of Brahman. The pole star is also the pivot on which the night sky turns, and all post-
Vedic texts, Hindu, Buddhist, and Jaina, assume that the cosmic mountain Meru connects the
center of the earth with the pole star. Together these concepts may be behind the myth of the
churning of the cosmic ocean (i.e., the ocean of the sky), in which Mount Meru was the
churning stick and the tortoise the steady pivot (i.e., the pole star as dhruva, “fixed, immobile,
constant”). I have proposed tentative Dravidian etymologies for the Sanskrit names of Meru
and the (cosmic) tortoise, kūrma.
The proverbially slow tortoise is also the vehicle of the slow planet Saturn in Buddhist
iconography. In chapter 21 it was suggested that while the Harappans wrote Saturn’s Proto-
Dravidian name, *may-mīn, “black star,” they wanted to refer to the planet not only
phonetically but also pictorially, choosing rebuses that mean “roof-fish”: the tortoise is an
aquatic animal and hence a kind of “fish,” which carries a roof over itself. But Saturn is also
usually very dim and is assigned the dark blue or black color. Saturn is supposed to be son of
the Sun-god, undoubtedly representing the Sun’s nightly aspect. This may be the same as
Saturn’s overlord Yama, the god of death, who is called “the Righteous King” (dharma-rāja) as
the judge of the deceased. In this respect he is like Varuṇa, the divine king who guards the law
(ṛta-) and punishes the sinners. Varuṇa is also the god and husband of the waters and the lord
of all aquatic animals.
It should by now have become clear that I argue, against the prevailing wisdom, that the
Indus civilization had kings or priest-kings, even though their palaces are not as easily
recognizable as in other ancient civilizations. (The complex of buildings in Mohenjo-daro, in
which both the “Proto-Śiva” seal and the “fig deity” seal were excavated, is a likely candidate
for a palace. See Possehl 2002:208−210; Vidale 2010.) From the Early Harappan period, there
is evidence of an increasing concentration of power, which enabled wide standardization of
culture and preplanned towns. The Kot Diji seals, with their four or five sets of concentric
circles, and the Mehrgarh pot painted with the motif “sun in four quadrants,” point to the
ideology of a solar ruler, governing the four regions and the zenith, an ideology reflected in
Vedic royal consecration and in the epics.
I assume that the Harappans had two principal male gods connected with kingship, the
“old” divine king (comparable to Vedic Varuṇa and Prajāpati and to Hindu Śiva in his ascetic
aspect), ruling in his central palace or capital, and his “young” alter ego, the crown prince
(comparable to Vedic Rudra < *Rudh(i)ra, “Red,” and also with Prince Rohita, “Red,” in the
Śunaḥśepa legend), who was in charge of military expeditions, as in the Vedic royal
consecration. This is the “young sun,” the red, newly born sun of the morning, and the “new
year,” the sun (re)born at the asterism of the Pleaides, the war-god Skanda nursed by the
Kṛttikās, also called Kumāra, “baby, youth,” also the name of Skanda’s Vedic counterpart. The
Old Tamil name of this divinity is Muruku or Murukan, “baby, youth,” which I have suggested
reading in the Indus inscriptions, called in Tamil also Cēy, “the Red one,” yet to be identified in
the Indus texts.
In my view these two “royal” deities were connected with the two halves of the year, which
ended, respectively, in the mahāvrata and viṣuvat festivals at the autumnal and vernal
equinoxes, and in both of which a “sacred marriage” took place. The old king slept with his
young virgin daughter (like Vedic Prajāpati with his own daughter Vāc), while the young
prince (Rudra, a harp player and singer, a kinnara) slept with his old mother (mahiṣī,
symbolizing the earth, the mother of the planet Mars, Rudhira). In both cases, the male
partner was killed. As in ancient Mesopotamia and in the Vedic horse/human sacrifice, these
rituals were enacted not by the king himself but by substitute victims.
The goddess with whom these two male gods fatally united, had two aspects like the
Mesopotamian goddess Inanna/Ištar, symbolized by the planet Venus. As the morning star,
Ištar was a bloodthirsty warrior goddess; as the evening star, she was a prostitute dedicated
to sexual love. The Vedic goddess Vāc as the virginal daughter of Prajāpati is said to represent
the Dawn. Vāc, also said to be a lioness and goddess of victory, was thus a predecessor of
Durgā, the virginal, lion- or tiger-riding Hindu goddess of victory. The most ancient new-year
star, Rohiṇī (“the red female”; the Sanskrit word rohiṇī also denotes a young girl just coming of
age), was according to the Atharvaveda (13,1,22) a goddess of war. But Vāc is also the Indian
counterpart of the Akkadian Ištar as a prostitute. In the Vedic mahāvrata, Vāc is worshipped
with sexual intercourse, sumptuous feasting, and the sounding of many musical instruments.
The Śābara-utsava of the Hindu navarātri/daśahrā festival of Durgā is a direct continuation of
the Vedic mahāvrata. Another Hindu counterpart of Vedic Vāc is goddess Sarasvatī, Vāg-īśvarī,
“Mistress of Sound,” who plays harp and is the guardian deity of the courtesans and the
patroness of fine arts.
I would like to conclude this chapter with some speculations on the Vedic name of the
Goddess Vāc, which means “sound, voice,” being a cognate of Latin vox. It is connected with
music making, which clearly played an important part in her “Atharvavedic” cult, while
instrumental music seems to have had no place in the Rigvedic tradition. This is particularly
true of the Goddess’s warrior aspect evidenced in the sounding of the “earth drum” in the
mahāvrata (see chapter 12), while the war drum is a key object of worship in the Old Tamil
religion. But I believe the full meaning of the name Vāc can be understood only if we can find
the original Proto-Dravidian word from which Vāc was translated. For this I propose two
homophonous Proto-Dravidian roots having the shape *viḷ / *veḷ. One root spans the meanings
“to say, speak, cry (a war-cry), summon, proclaim, reveal, make known; to sound, to sing,”
while the other root means “to break (as the day), become bright, white, clear,” with derived
nouns denoting “light, lamp, brightness, dawn, the planet Venus, purity.” The interference
between these two roots can be seen from terms intimately connected with the worship of the
Goddess Bhadra-Kāḷī or Durgā in Kerala. The Goddess is represented by a sword-bearing
oracle priest called in Malayalam veḷiccappāṭu, through whom the goddess “speaks out” and
“reveals her heart.” The priest’s name is derived from the word veḷiccam, “light of lamp or
dawn, showing oneself, revelation.”
The lamp, viḷakku, is one of the most important symbols of the Goddess in South India. A
lamp is lighted on the severed head of the buffalo sacrificed to the Goddess (see chapter 15).
It represents her as the vanquisher of the dark forces of night and death that the black buffalo
symbolizes. In Mesopotamia, Goddess Inanna-Ištar is symbolized by the planet Venus, which
as the morning star brings the first light to the darkness of night, being in this sense
comparable to the dawn, imagined as the solar lion of the morning who kills the black buffalo
of night. It therefore seems possible that the planet Venus was the star of the Goddess also in
the ancient Indus Valley (in addition to Rohiṇī). In modern Hinduism, Friday, the day of the
planet Venus, is sacred to the Goddess, but the seven-day week came to India only with
Hellenistic astronomy and astrology.
24
Retrospect and Prospect

The summaries in chapters 22 and Chapter 23 have not reported all of the new research into
the religions of the Indus civilization, the Vedas, and Hinduism published in this book. I have
concentrated on major topics, with the aim of showing that some key issues can be
satisfactorily solved, although further work will undoubtedly refine the proposals offered here.
The trail of the Indo-Iranian languages from their East European homeland to their historical
speaking areas can be followed in the archaeological record. I further claim that the Harappan
language was Proto-Dravidian or close to it, and that Old Tamil literature has preserved
important elements of Harappan heritage, such as the names of the planets and some stars,
notably the pole star, and such original Harappan names of deities as Muruku and Vēḷ. I also
maintain that a decisive beginning has been made in the decipherment of the Indus script,
with a number of cross-checked interpretations that are in full agreement with the Harappan
context and the wider cultural traditions of West and South Asia. Applying the same methods
and working hypotheses, it should be possible to penetrate the script further, although a
complete decipherment is clearly impossible with presently available materials. Among the
“sects” of later Hinduism, Śaiva-Śākta Tantrism seems to me to be closest to the Harappan
and “Atharvavedic” religions, although it, too, has many elements from external sources,
including those introduced by the originally Proto-Saka-speaking Dāsas.
There is, however, much more to be discovered. I should like to conclude with some
speculations on the social structure of the Indus civilization and the mode of thinking among
its priestly administrators. My starting point is the five visible planets. It appears that the
planets were connected with five different colors in their original Proto-Dravidian names, for
such names are preserved in Tamil texts and seem to occur in the Indus script:

*pon-mīn, “golden star” = Jupiter

*veḷ(ḷi)-mīn, “white star” = Venus
*kem-mīn, “red star” = Mars
*pacu-mīn, “green star” = Mercury
*may-mīn, “black star” = Saturn

The Indus signs for Mercury, Saturn, and Venus are identified and briefly explained in chapter
21; some comments on the names of Jupiter and Mars follow. The compound pon-mīn is
attested in Tamil, but only as the name of a “golden fish,” namely mahsir, a large carp fish
(Barbus/Cyprinus tor). However, pon, “gold,” por-kōḷ, “golden planet,” and ponnan and
ponnavan, “he who is golden,” are Tamil names of the planet Jupiter, which is yet to be
identified among the “fish” signs of the Indus script.
The expected Proto-Dravidian name of Mars, *kem-mīn, “red star,” survives (with the
Tamil-Malayalam affrication *k > c) in Old Tamil cem-mīn, denoting Mars in Puranānūru 60,2.
Cem-mīn (“red fish”) is used in Tamil of the sperm whale (Euphysetes macrocephalus), and in
Malayalam of “prawn, shrimp.” But neither of these meanings explains the “fish” pictogram,
which I suspect denotes Mars (see the middlemost “fish” sign in Fig. 21.3a), because this
particular “fish” sign is found immediately before the sign sequence read as Muruku-Vē ḷ (as in
Fig. 21.19a—here the reading of this “fish” sign is uncertain, but there are other clear
parallels). Muruku is the Tamil counterpart of the Hindu war-god Skanda, whom the planet
Mars represents according to Jaiminīya-Gṛhyasūtra 2,9 (vidyād . . . skandam aṅgārakam);
moreover, both Muruku and the planet Mars are called in Old Tamil cēy, “redness.” Only while
writing this chapter did I discover that in Kannada kem-mīn is the name of a locally restricted
red-colored carp fish (Maclean 1893:410b;763a). Its Sanskrit synonym, rohita-matsya, “red
fish,” denotes Labeo rohita, a delicious carp fish common everywhere in the subcontinent
except South India; its absence in the south explains why the inherited Dravidian name is
applied to several other varieties of fish in South Dravidian languages. The suspected Indus
sign can easily be understood as a picture of a carp fish, with characteristic barbels hanging
from the sides of the mouth, as suggested by Nikita Gurov (1968:40–42) long ago. Gurov
proposed to read it in Proto-Dravidian as *kay/*key, “carp,” (on the basis of Tamil kayal, cēl,
“carp,” and cognates in DEDR no. 1252) to stand for Proto-Dravidian *key/*kēy/*ke-, “red.”
About half a dozen times the “carp” sign is followed by the plain “fish” sign, and these sign
sequences Gurov interpreted to form the compound *kem-mīn, “planet Mars.” (In my opinion,
the word *mīn is included in the “carp” sign, as seems to be the case with other “fish” signs
having diacritics, see chapter 21: the “carp” sign stands for *kem-mīn, “red fish” as well as
“red star.”)
The above Dravidian planetary colors agree with those connected with the planets in
Sanskrit texts dealing with planetary worship, starting with the late Vedic Gṛhyasūtras.
In the Purāṇic cosmology, as we have seen in chapter 16, the planets are fed by specific
rays of the sun having five different colors. These solar rays correspond to veins of five
different colors in the human body, according to the Upaniṣads and much later texts of the
Yoga tradition.
In ancient Mesopotamia and China, too, the planets are connected with different colors.
The Babylonians correlated the planets with different metals, with the quarters of space, and
with the times of the sun’s daily and yearly circulation (Jeremias 1929). In the third century
BCE, the Chinese naturalists had a whole system of correlative thinking involving—in addition
to the planets, the colors, the cardinal points, and the seasons—the five elements, tastes,
smells, numbers, musical notes, and many other things.
In the Vedic Brāhmaṇa texts, such correlative thinking plays a major role. Different
conceptual categories are classified hierarchically, and the correspondences between them
are interpreted in terms of equivalences. One of the main purposes of this “classification of
the universe,” as suggested by Brian K. Smith (1994), is to justify with cosmic parallels the
highest social position occupied by the Brahmins. In this classification of the Brāhmaṇa texts,
too, the colors play a major role; indeed, each of the traditional four social classes of ancient
India is symbolized by a specific color, and the Sanskrit word varṇa, “color,” also denotes
“social class.” Much of the Brāhmaṇa classification is ad hoc and unsystematic, with several
different systems having different numbers of members. However, the systems with four and
five members stand out as most important, and can at least partially be traced back to the
latest parts of the Rigveda. Thus in the Puruṣa hymn (RV 10,90), the four social classes are
correlated with the different body parts of the “cosmic man”: the brāhmaṇas with the head,
the kṣatriyas with the arms, the vaiśyas with the loins, and the śūdras with the feet. This idea
of the cosmos originally consisting of a primeval man who was sacrificed by the gods is
unknown in older Rigvedic tradition, but closely connected with religious concepts that seem
to have a Harappan origin, such as the creator-god Prajāpati and the human sacrifice, in
which the victim was the bridegroom of a “sacred marriage” (chapters 16 and 17).
Table 24.1 shows some basic equivalences. Each column consists of a category: numbers,
directions, times of day, seasons, colors, elements, and social classes. Each row comprises
equivalent concepts.
TABLE 24.1

Could these most basic correlative classifications of the Brāhmaṇas derive from the
Harappans? In that case the planets—connected with the colors—are likely to have been
among the ancient categories, too. Colors are associated with the directions of space in that
type of Tantric maṇḍala (Tucci 1961), which has its prototype in the BMAC “palace” of Dashly-
3 (chapter 20). In Tantric iconography, the square maṇḍala is the palace or capital of the
divinity depicted as the ruler in its center, with palace or city gates in the four cardinal
directions. As a cosmogram, the maṇḍala also represents the square earth (RV 10,58,2) of the
Vedic cosmos, divided according to the cardinal directions (Kirfel 1920:10).
The four cardinal directions form the pattern of cross, providing the basic framework for
the fourfold classification. The center of the cross, symbolizing the fifth direction to the zenith,
unites and exceeds the four; it is therefore an appropriate symbol for the king who unites the
four social classes and is above them. This is reflected in the Vedic royal consecration, where
the king takes a step in each of the five directions, thereby ascending to the zenith, the
exalted place of the midday sun and the pole star (chapter 16). One is reminded of the Early
Harappan seals that have four or five sets of concentric circles (Fig. 16.3), which can be
assumed to symbolize the sun and its daily and yearly rotation through the four directions and
the four seasons defined by the equinoctial and solstitial points.
If the east is in the front, as it is in the traditional orientation of the Vedic texts, the
horizontal axle of the cross contains the colors black (in the north) and red (in the south),
corresponding to the planets Saturn and Mars, counterparts of the evening/night sun and the
morning/day sun. They represent the solar/Śaiva gods Kāla, “Black,” and Rud[h]​ra, “Red.”
Śiva is worshipped in the ancient temple of Ujjain (praised by Kālidāsa in Meghadūta) as
Mahā-Kāla; Rudra is the youthful Kumāra. These male deities have female counterparts in the
black Kālī (often represented as an ugly old hag) and the red Rohiṇī/Durgā (a youthful and
beautiful virgin, Kumārī). In this connection it is very significant to note that Sanskrit kāla (not
attested before the fourth century BCE) and Pali kāḷa, “black,” almost certainly are derived
from Proto-Dravidian *kāẓ, “black” (DEDR no. 1494).
The vertical axle of the cross contains the colors white (in the east) and greenish-yellow (in
the west), corresponding to the planets Venus and Mercury, which probably primarily
represent the new moon and the full moon. I suspect that besides the harsh solar gods the
Indus people also worshipped milder “lunar” gods. Chapter 13 suggested that the Vaiṣṇava
gods Bala-Rāma (white) and Kṛṣṇa (black) are continuations of the Aśvins, divine twins of
Aryan origin, connected with the chariot team as well as the sun/day and moon/night. But
behind these Vaiṣṇava gods are undoubtedly also earlier local divinities, partly surviving in the
texts as “demons” vanquished and replaced by Bala-Rāma and Kṛṣṇa.
The color white characterizes the demons who had the shape of a conch (Pañcajana) and
the wild ass (Dhenuka), while the white cosmic snake Śeṣa (used as the rope in the churning
of the cosmic milk ocean) is expressly identified with Bala-Rāma. Bala-Rāma’s weapon, the
plowshare, resembles the sickle of the new moon, while Kṛṣṇa is said to be the “full”
incarnation of Viṣṇu, a hint of the full moon. Kṛṣṇa’s circular dance with the milkmaids seems
to represent the full moon’s conjunctions with the various asterisms along its circular path:
according to the Śatapatha-Brāhmaṇ̣a (9,4,1,9), the moon is a heavenly playboy, a gandharva,
and the asterisms are his heavenly mistresses, apsarases (chapter 16). Kṛṣṇa’s name meaning
“black” connects him with the dark-hued aboriginal cowboys of the Vraja tract. But he is
actually often depicted as green-hued and called Hari, “greenish-yellow,” which is the color of
the moon (cf. Hariścandra in chapter 16) and honey: Madhu, “honey,” is a demon vanquished
by Kṛṣṇa, and symbol for the sweet pleasures of love associated with Kṛṣṇa.
With these hopefully thought-provoking speculations, the book must close.
Bibliographical Notes

Detailed documentary annotation and copious references to further literature can be found in
my previous publications, on which this book is mostly based. Unfortunately space forbids
giving this apparatus in the present book and thus fully acknowledging here as well my
indebtedness to all the scholars involved. Therefore, listed below are references to the
principal sources of chapters 6 to Chapter 21: first my own publications on the topic
(including earlier versions, which may be partially dated, especially with regard to the Dāsas
or the first wave(s) of Aryan immigrants to South Asia; the principal ones are in bold),
followed by some other works of fundamental importance, in alphabetical order.

CHAPTER 6 PROTO-INDO-EUROPEAN HOMELANDS

Parpola 2008a; 2012 (2013); Carpelan & Parpola 2001.
Anthony 1995; 2007. Burmeister 2004. Chernykh 1992. Dergachev 1998; 2007. Gusev 1998. Hänsel & Zimmer
1994. Kohl 2007. Levine, Renfrew, & Boyle 2003. Littauer & Crouwel 1979; 2002. Mallory 1989 through 2002.
Mallory & Adams 1997. Maran 2004a; 2004b. Piggott 1983; 1992. Rassamakin 1999; 2004. Videiko 1994; 1995
(1996).

CHAPTER 7 EARLY INDO-IRANIANS ON EURASIAN STEPPES

Parpola 1974; 1988a; 1995; 1998a; 1999a; 2002a; 2012 (2013); Carpelan & Parpola 2001.
Anthony & Vinogradov 1995. Bader, Krajnov, & Kosarev 1987. Chernykh 1992; 2007. Chernykh & Kuz’minykh 1987;
1989. Chlenova 1984. Gening, Zdanovich & Gening 1992. Häkkinen 2009. Joki 1973. Koivulehto 2001. Koryakova
& Epimakhov 2007. Kovtun 2013. Kuz’mina 2007. Kuz’minykh 2011. Parzinger 2006. Pryakhin & Khalikov 1987.
Rédei 1988; 1988–1991. Sinor 1988. Young 1985. Zdanovich & Zdanovich 2002.

CHAPTER 8 THE BMAC OF CENTRAL ASIA AND THE MITANNI OF SYRIA

Parpola 1988a; 1995; 1999a; 2002a, 2002b; 2004; 2005d.
Amiet 1979; 1986. Baghestani 1997. Bobomulloev 1997. Brisch & Bartl 1995. Burrow 1973. Collon 1987. Francfort
2005. Ghirshman 1977. Gubaev, Koshelenko, & Tosi 1998. Hiebert 1994. Jarrige 1991. Lamberg-Karlovsky 1986.
Lawergren 2003. Ligabue & Salvatori 1989. Lubotsky 2001. Mayrhofer 1966; 1974. Moran 1992. Raulwing 2005
(2006). Sarianidi 1977 through 2007; Sarianidi et al. 2011 (2012). Schmidt 1933; 1937. Shinde, Possehl, & Ameri
2005. Stein 1994. Thieme 1960. Teufer 2012. von Dassow 2008. Wilhelm 1994. T. C. Young 1985.

CHAPTER 9 THE RIGVEDIC INDO-ARYANS AND THE DĀSAS

Parpola 1988a; 2002a; 2002b; 2004; 2012a.
Bailey 1959 (1960). Burrow 1976 (1977). Chernykh 1992. Falk 1994. Francfort 2001; 2005. Hillebrandt 1927–1929.
Hintze 1998. Jamison & Brereton 2014. Jarrige 1985. Rau 1976. Schmitt 1989; 2000. Szabo & Barfield 1991.
Teufer 2013. Wheeler 1947. Witzel 1995b.

CHAPTER 10 THE AŚVINS AND MITRA-VARU ṆA

Parpola 2005a.
Brereton 1981. Falk 1989. Frame 1978; 2010. Gonda 1974; 1975; 1977; 1978. Hillebrandt 1927–1929. Hopkins
1889. Jamison & Brereton 2014. Jamison & Witzel 1992. Lüders 1951–1959. Macdonell 1897. Nyberg 1995.
Oberlies 1993; 2012. West 2007. Zeller 1990.

CHAPTER 11 THE AŚVINS AS FUNERARY GODS

Parpola 2005a; 2005c; 2009.
Caland 1896; 1914. Hopkins 1896. Houben 1991; 2000. Oldenberg 1884. Rönnow 1929. Sparreboom 1985. Stacul
1970; 1971. van Buitenen 1968.

CHAPTER 12 THE ATHARVAVEDA AND THE VRĀTYAS

Parpola 1983; 1988a; in press b,e.
Bloomfield 1899. Caillat 1989. Caland 1896; 1900; 1908; 1931. Charpentier 1911. Emeneau 1966. Falk 1986. Gonda
1971. Hauer 1927. Heesterman 1957; 1962; 1967. Hillebrandt 1897. Hopkins 1896. Horsch 1966. Insler 1998.
Kuiper 1979. Oldenberg 1884. Rolland 1973. von Hinüber 2001. Wackernagel 1896. Wackernagel & Debrunner
1954. Weber 1849. Witzel 1989.

CHAPTER 13 THE MEGALITHIC CULTURE AND THE GREAT EPICS

Parpola 1984b; 1998b; 1999c; 2002c; 2004 (2006).
Allchin 1995. Brockington 1998. Coningham & Allchin 1995. Ghosh 1989. Goldman 1984. Hopkins 1915. Horsch
1966. Jaiswal 1981. Leshnik 1974. Maloney 1970. van Buitenen 1973–1978. Weber 1850; 1870 (1871). Witzel
1995b; 1997.

CHAPTER 14 THE LANGUAGE OF THE INDUS CIVILIZATION

Parpola 1981; 1994a; 1997a; 1997b; 2003; 2010; 2011a; in press c, d; Parpola & Janhunen 2011.
Anderson 2008. Burrow & Emeneau 1984. Caldwell 1856; 1875. Hart 1975. Krishnamurti 2003. Osada 2006. Pinnow
1959. Trautmann 1981. Zvelebil 1970 through 1990.

CHAPTER 15 FERTILITY CULTS IN FOLK RELIGION

Parpola 1994a; 2003; 2011a; 2011b.
Biardeau 2004. Charpentier 1926. Coomaraswamy 1928–1931. Crooke 1906; 1926. Fischer & Shah 1971. Hiltebeitel
1978. Jarrige et al. 1995. Marshall 1931a; 1931b. Ward 1811. Whitehead 1921.

CHAPTER 16 ASTRONOMY, TIME-RECKONING, AND COSMOLOGY

Parpola 1985; 1990; 1994a; 2013; 2014.
Filliozat 1962. Geldner 1889. Kirfel 1920; 1952; 1954. Liebert 1968 (1969). Mackay 1938. Needham 1959.
Pankenier 2004; 2013. Scherer 1953. Uesugi 2011. Vahia & Menon 2011. Wanzke 1987. Weber 1853; 1861 (1862).
Winternitz 1892.

CHAPTER 17 DILMUN, MAGAN, AND MELUHHA

Parpola 1994a; 1994b; 2002; Parpola, S., A. Parpola, & Brunswig 1977.
Bibby 1972. Franke-Vogt 1995. Laursen 2010. Maekawa & Mori 2011. Olijdam & Spoor 2008. Potts 2008. Reade
1995; 2008. Steinkeller 1982. Vermaak 2008.

CHAPTER 18 ROYAL SYMBOLS FROM WEST ASIA

Parpola 1984a; 1985; 1994a; 2011b; 2011c; in press c.
Amiet 1956; 1980. Black & Green 1992. Boehmer 1974; 1975. Collon 1987. Franke-Vogt 1991 (1992). Heesterman
1957; 1989. Leick 1991. Mode 1944. Petrie 1939. Porada 1950. Rothman 2001. Sievertsen 1992; 1999.

CHAPTER 19 THE GODDESS AND THE BUFFALO

Parpola 1983; 1984a; 1985; 1988a; 1989; 1992a; 1994a; 1998b; 1999b; 2002b&c; 2011c; 2012b; in press e.
Beck 1981. Biardeau 1981; 2004. C. J. Fuller 1992. Hiltebeitel 1978. Kane 1941–1958. Kinsley 1986. Shulman 1980.
Silva 1955. van Kooij 1972. Vermaseren 1977. von Stietencron 1983. Whitehead 1921.

CHAPTER 20 EARLY IRANIANS AND “LEFT-HAND” TANTRISM

Parpola 1988a; 2002b.
Azarpay 1976. Brentjes 1983. Degener 2002. Jettmar 1986. Potts 2001.

CHAPTER 21 RELIGION IN THE INDUS SCRIPT

Parpola 1975a; 1975b; 1986a; 1986b; 1988b; 1990a; 1990b; 1992b; 1994a; 1998b; 2003; 2005b; 2008b; 2011a;
2011c; Parpola & Janhunen 2011; CISI.
Baines 1999. Chadwick 1958. Daniels 1996. Geldner 1889. Hart 1975. Robinson 2002. Van Buren 1948. Winternitz
1895.

CHAPTER 24 RETROSPECT AND PROSPECT

Parpola 1999d. Parpola & Janhunen 2011.
Maclean 1893. Smith 1994.
References

Allchin, Bridget, & Raymond Allchin. 1968. The birth of Indian civilization: India and Pakistan before 500 B.C.
Harmondsworth: Penguin.
_____. 1982. The rise of civilization in India and Pakistan. Cambridge: Cambridge University Press.

Allchin, F. R. 1995. The archaeology of early historic South Asia: The emergence of cities and states. With contributions
from George Erdosy, R. A. E. Coningham, D. K. Chakrabarti, and Bridget Allchin. Cambridge: Cambridge University
Press.
Amiet, Pierre. 1956. Le symbolisme cosmique du répertoire animalier en Mésopotamie. Revue d’Assyriologie et
d’Archéologie Orientale 50:113–126.
_____. 1979. Archaeological discontinuity and ethnic duality in Elam. Antiquity 53:195–204.
_____. 1980. La glyptique mésopotamienne archaïque. Deuxième édition revue et corrigée avec un supplément. Paris:
Éditions du Centre National de la Recherche Scientifique.
_____. 1986. L’âge des échanges inter-iraniens 3500–1700 avant J.-C. Paris: Éditions de la Réunion des musées nationaux.

Anderson, Gregory D. S. 2008. Introduction to the Munda languages. Pp. 1–10 in: Gregory D. S. Anderson (ed.), The
Munda languages. (Routledge Language family series.) London and New York: Routledge.
Anthony, David W. 1995. Horse, wagon and chariot: Indo-European languages and archaeology. Antiquity 69:554–565.
_____. 2007. The horse, the wheel, and language: How Bronze-Age riders from the Eurasian steppes shaped the modern
world. Princeton, NJ: Princeton University Press.
Anthony, David W., & N. B. Vinogradov. 1995. Birth of the chariot. Archaeology 48(2):36–41.
Ardeleanu-Jansen, Alexandra. 1989. A short note on a steatite sculpture from Mohenjo-Daro. Pp. 196–210 in: Karen
Frifelt & Per Sørensen (eds.), South Asian Archaeology 1985. (Scandinavian Institute of Asian Studies, Occasional
Papers 4.) London: Curzon.
Aruz, Joan (ed.). 2003. Art of the first cities: The third millennium B.C. from the Mediterranean to the Indus. New York:
Metropolitan Museum of Art; New Haven, CT: Yale University Press.
Avanesova, N. A. 2010. Zarafshanskaya kul’turnaya provintsiya Baktrijsko-Margianskoj tsivilizatsii. Pp. 334–364 in: P.
M. Kozhin, M. F. Kosarev, & N. A. Dubova (eds.), Na puti otkrytiya tsivilizatsii: Sbornik statej k 80-letiyu V. I.
Sarianidi. (Trudy Margianskoj arkheologicheskoj èkspeditsii [3]​.) St. Petersburg: Aletejya.
Azarpay, G. 1976. Nanā, the Sumero-Akkadian goddess of Transoxiana. Journal of the American Oriental Society
96(4):536–542.
Bader, O. N., D. A. Krajnov, & M. F. Kosarev (eds.). 1987. Arkeologiya SSSR: Èpokha bronzy lesnoj polosy SSSR.
Moscow: Nauka.
Baghestani, Susanne. 1997. Metallene Compartimentsiegel aus Ost-Iran, Zentralasien und Nord-China. (Archäologie in
Iran und Turan 1.) Rahden, Westfahlen: Marie Leidorf.
Bailey, Harold Walter. 1959 (1960). Iranian Arya- and Daha-. Transactions of the Philological Society 1959:71–115.
Baines, John. 1999. Writing, invention and early development. Pp. 882–885 in: Kathryn A. Bard (ed.), Encyclopedia of
the archaeology of ancient Egypt. London and New York: Routledge.
Barnett, Lionel D. (trans.). 1907. The Antagaḍa-dasāo and Aṇuttarovavāiya-dasāo. Translated from the Prakrit.
(Oriental Translation Fund NS 17.) London: Royal Asiatic Society.
Bartl, Karin, Reinhard Bernbeck, & Marlies Heinz (eds.). 1995. Zwischen Euphrat und Indus: Aktuelle
Forschungsprobleme in der Vorderasiatischen Archäologie. Hildesheim: Georg Olms.
Beck, Brenda E. F. 1981. The goddess and the demon: A local South Indian festival and its wider context. Pp. 83–136 in:
Biardeau 1981.
Benfey, Theodor. 1869. Geschichte der Sprachwissenschaft und orientalischen Philologie in Deutschland seit dem
Anfange des 19. Jahrhunderts, mit einem Rückblick auf die früheren Zeiten. (Geschichte der Wissenschaften in
Deutschland 8.) Munich: J. G. Cotta.
Besenval, Roland. 2011. Between East and West: Kech-Makran (Pakistan) during Protohistory. Pp. 41–164 in: Osada &
Witzel 2011.
Biardeau, Madeleine (ed.). 1981. Autour de la déesse hindoue. (Puruṣārtha 5.) Paris: Éditions de l’École des Hautes
Études en Sciences Sociales.
_____. 2004. Stories about posts: Vedic variations around the Hindu Goddess. Translated by Alf Hiltebeitel, Marie-Louise
Reiniche, & James Walker, edited by Alf Hiltebeitel & Marie-Louise Reiniche. Chicago: University of Chicago Press.
Bibby, Geoffrey. 1972. Looking for Dilmun. Harmondsworth: Penguin.
Black, Jeremy, & Anthony Green. 1992. Gods, demons and symbols of ancient Mesopotamia: An illustrated dictionary.
Illustrations by Tessa Rickards. London: British Museum Press.
Bloomfield, Leonard. 1933. Language. New York: Henry Holt.
Bloomfield, Maurice (trans.). 1897. Hymns of the Atharva-Veda. (Sacred Books of the East 42.) Oxford: Clarendon
Press.
_____. 1899. The Atharva-Veda and the Gopatha-Brāhmaṇa. (Grundriss der indo-arischen Philologie und Altertumskunde
2:1:B.) Strasbourg: Karl J. Trübner.
Bobomulloev, Saïdmurâd. 1997. Ein bronzezeitliches Grab aus Zardča Chalifa bei Pendžikent (Zeravšan-Tal).
Archäologische Mitteilungen aus Iran und Turan 29:121–134.
Boehmer, Rainer Michael. 1974. Das Rollsiegel im prädynastischen Ägypten. Archäologischer Anzeiger 1974:495–514.
_____. 1975. Das Auftreten des Wasserbüffels in Mesopotamien in historischer Zeit und seine sumerische Bezeichnung.
Zeitschrift für Assyriologie 64:1–19.
Bopp, Franz. 1816. Über das Conjugationssystem der Sanskritsprache in Vergleichung mit jenem der griechischen,
lateinischen, persischen und germanischen Sprache. Herausgegeben und mit Vorerinnerungen begleitet von K. J.
Windischmann. Frankfurt am Main: Andreäische Buchhandlung.
Boyle, Katie, Colin Renfrew, & Marsha Levine (eds.). 2002. Ancient interactions: East and west in Eurasia. Cambridge,
UK: McDonald Institute for Archaeological Research.
Brentjes, Burchard. 1983. Das “Ur-Mandala” (?) von Daschly-3. Iranica Antiqua 18:25–49 and Table I.
Brereton, Joel Peter. 1981. The Ṛgvedic Ādityas. (American Oriental Series 63.) New Haven, CT: American Oriental
Society.
Brisch, Nicole, & Karin Bartl. 1995. Die altassyrischen Handelskolonien in Anatolien. Pp. 134–147 in: Bartl, Bernbeck,
& Heinz 1995.
Brockington, John. 1998. The Sanskrit epics. (Handbuch der Orientalistik, Abt. 2: Indien 12.) Leiden: Brill.
Brunt, P. A. 1976–1983. Arrian with an English translation I–II. (Loeb classical library 236, 269.) Cambridge, MA:
Harvard University Press; London: William Heinemann.
Bryant, Edwin. 2001. The quest for the origins of Vedic culture: The Indo-Aryan migration debate. New York: Oxford
University Press.
Burlingame, Eugene Watson (trans.). 1921. Buddhist legends I–III. Translated from the original Pali text of the
Dhammapada commentary. (Harvard Oriental Series 28–30.) Cambridge, MA: Harvard University Press.
Burmeister, Stefan. 2004. Der Wagen im Neolithikum und in der Bronzezeit: Erfindung, Ausbreitung und Funktion der
ersten Fahrzeuge. Pp. 13–40 in: Fansa & Burmeister 2004.
Burrow, Thomas. 1973. The Proto-Indoaryans. Journal of the Royal Asiatic Society of Great Britain and Ireland
1973(2):123–140.
_____. 1976 (1977). Review of Rau 1976. Kratylos 21:72–76.

Burrow, Thomas, & M. B. Emeneau. 1984. A Dravidian etymological dictionary. 2nd ed. Oxford: Clarendon Press. (=
DEDR)
Caillat, Colette (ed.). 1989. Dialectes dans les littératures indo-aryennes. (Publications de l’Institut de civilisation
indienne, Série in-8° 55.) Paris: Collège de France, Institut de civilisation indienne.
Caland, Willem. 1896. Die altindischen Todten- und Bestattungsgebräuche. (Verhandelingen der Koninklijke Akademie
van Wetenschappen te Amsterdam, Afdeeling Letterkunde, I:6.) Amsterdam: Johannes Müller.
______. 1900. Altindisches Zauberritual: Probe einer Uebersetzung der wichtigsten Theile des Kauśika Sūtra.
(Verhandelingen der Koninklijke Akademie van Wetenschappen te Amsterdam, Afdeeling Letterkunde, Nieuwe Reeks,
III:2.) Amsterdam: Johannes Müller.
_____. 1908. Altindische Zauberei: Darstellung der altindischen “Wunschopfer.” (Verhandelingen der Koninklijke
Akademie van Wetenschappen te Amsterdam, Afdeeling Letterkunde, N.R. 10:1.) Amsterdam: Johannes Müller.
_____. 1914. Die vorchristlichen baltischen Totengebräuche. Archiv für Religionswissenschaft 17:476–512.
_____ (trans.). 1931. Pañcaviṃśa-Brāhmaṇa: The Brāhmaṇa of twenty five chapters. (Bibliotheca Indica, Work 255.)
Calcutta: Asiatic Society of Bengal.
Caldwell, Robert. 1856. A comparative grammar of the Dravidian or South-Indian family of languages. London:
Harrison.
_____. 1875. A comparative grammar of the Dravidian or South-Indian family of languages. 2nd ed. Madras: University of
Madras.
Campbell, Lyle. 2004. Historical linguistics: An introduction. 2nd ed. Edinburgh: Edinburgh University Press;
Cambridge, MA: MIT Press.
Cardona, George. 1976. Pāṇini: A survey of research. (Trends in Linguistics, State-of-the-art Reports 6.) The Hague:
Mouton.
Carpelan, Christian, & Asko Parpola. 2001. Emergence, contacts and dispersal of Proto-Indo-European, Proto-Uralic
and Proto-Aryan in archaeological perspective. Pp. 55–150 in: Carpelan, Parpola, & Koskikallio 2001.
Carpelan, Christian, Asko Parpola, & Petteri Koskikallio (eds.). 2001. Early contacts between Uralic and Indo-European:
Linguistic and archaeological considerations. (Mémoires de la Société Finno-Ougrienne 242.) Helsinki: Suomalais-
Ugrilainen Seura.
Casal, Jean-Marie. 1964. Fouilles d’Amri I–II. (Publications de la Commission des Fouilles Archéologiques: Fouilles du
Pakistan.) Paris: C. Klincksieck.
Chadwick, John. 1958. The decipherment of Linear B. Cambridge: Cambridge University Press.
Charpentier, Jarl. 1911. Bemerkungen über die vrātya’s. Wiener Zeitschrift für die Kunde des Morgenlandes 25:355–
388.
_____
. 1926. Über den Begriff und die Etymologie von pūjā. Pp. 276–297 in: Willibald Kirfel (ed.), Beiträge zur
Literaturwissenschaft und Geistesgeschichte Indiens: Festgabe Hermann Jacobi zum 75. Geburtstag dargebracht von
Freunden, Kollegen und Schülern. Bonn: Fritz Klopp.
Chernykh, E. N. 1970. Drevnejshaya metallurgiya Urala i Povolzh’ya. Moscow: Nauka.
_____. 1992. Ancient metallurgy in the USSR: The early metal age. Translated by Sarah Wright. (New Studies in
Archaeology.) Cambridge: Cambridge University Press.
_____
. 2007. Kargaly: Fenomen i paradoksy razvitiya. Kargaly v sisteme metallurgicheskikh provintsij. Potaennaya
(sakral’naya) zhizn’ arkhaichnykh gornyakov i metallurgov. (Kargaly 5.) Moscow: Yazyki slavyanskoj kul’tury.
Chernykh, E. N., & S. V. Kuz’minykh. 1987. Pamyatniki sejminsko-turbinskogo tipa v Evrazii. Pp. 84–105 in: Bader,
Krajnov, & Kosarev 1987.
_____. 1989. Drevnyaya metallurgiya severnoj Evrasii (Sejminsko-Turbinskij fenomen). Moscow: Nauka.

Childe, V. Gordon. 1926. The Aryans: A study of Indo-European origins. London: Kegan Paul.
Chlenova, N. L. 1984. Arkheologicheskie materialy k voprosu ob irantsakh doskifskoj èpokhi i indo-irantsakh.
Sovetskaya Arkheologiya 1984(1):88–103.
Cierny, Jan, Thomas Stöllner, & Gerd Weisgerber. 2005. Zinn in und aus Mittelasien. Pp. 431–448 in: Bergbau,
Metallurgie, Industriezentren und Werkstätten von Handelsgütern. Bochum: Deutsches Bergbau-Museum.
CISI = Corpus of Indus Seals and Inscriptions. 1 = Joshi & Parpola 1987. 2 = Shah & Parpola 1991. 3.1 = Parpola,
Pande, & Koskikallio 2010.
Clauson, Gerard, & John Chadwick. 1969. The Indus script deciphered? Antiquity 43(71):200–207.
Coningham, R. A. E., & F. R. Allchin. 1995. The rise of cities in South Asia. Pp. 152–184 in: Allchin 1995.
Coeurdoux, Gaston-Laurent. 1777. Moeurs et coutumes des Indiens. See Murr 1987.
Collon, Dominique. 1987. First impressions: Cylinder seals in the ancient Near East. London: British Museum
Publications.
Coomaraswamy, Anand K. 1928–1931. Yakṣas I–II. (Smithsonian Institution Publications 2926, 3059.) Washington, DC:
Smithsonian Institution.
Crooke, William. 1906. Things Indian, being discursive notes on various subjects connected with India. London: John
Murray.
_____
. 1926. Religion and folklore of Northern India. Prepared for the press by R. E. Enthoven. London: Oxford
University Press.
Cunningham, Alexander. 1875. Harapâ. Pp. 105–108 and pl. XXXII–XXXIII in: Archaeological Survey of India, Report for
the year 1872–73. ([Reports, Old series] vol. V.) Calcutta: Office of the Superintendent of Government Printing, India.
Daniel, J. C. 1983. The book of Indian reptiles. Bombay: Bombay Natural History Society and Oxford University Press.
Daniels, Peter T. 1996. Methods of decipherment. Pp. 141–159 in: Daniels & Bright 1996.
Daniels, Peter T., & William Bright (eds.). 1996. The world’s writing systems. New York: Oxford University Press.
DEDR = Burrow & Emeneau 1984.
Degener, Almuth. 2002. The Nuristani languages. Pp. 103–117 in: Sims-Williams 2002.
Dergachev, Valentin A. 1998. Kulturelle und historische Entwicklungen im Raum zwischen Karpaten und Dnepr: Zu den
Beziehungen zwischen frühen Gesellschaften im nördlichen Südost- und Osteuropa. Pp. 27–64 in: Bernhard Hänsel &
Jan Machnik (eds.), Das Karpatenbecken und die osteuropäische Steppe. (Südosteuropa-Schriften 20; Prähistorische
Archäologie in Südosteuropa 12.) Munich: Südosteuropa-Gesellschaft; Rahden, Westfalen: Marie Leidorf.
_____. 2007. O skipetrakh, o loshadyakh, o vojne: ètyudy v zashchitu migratsionnoj kontseptsii M. Gimbutas. St.
Petersburg: Nestor.
Didier, Aurore. 2013. La production céramique du Makran (Pakistan) à l’âge du bronze ancien: Contribution à l’étude
des peuplements des régions indo-iraniennes. (Mémoires des Missions archéologiques françaises en Asie centrale et
en Asie moyenne 14. Série Indus-Balochistan.) Paris: De Boccard.
Dixit, Yama, David A. Hodell, & Cameron A. Petrie. 2014. Abrupt weakening of the summer monsoon in northwest India
~4100 yr ago. Geology 42(4), 24 Feb 2014. http://www.cam.ac.uk/research/news/decline-of-bronze-age-megacities-
linked-to-climate-change.
Doniger, Wendy. 2009. The Hindus: An alternative history. New York: Penguin.
Dubois, Jean Antoine. [1825] 1906. Hindu manners, customs and ceremonies. Translated from the author’s later French
MS. and edited with notes, corrections, and biography by Henry K. Beauchamp. 3rd ed. Oxford: Clarendon Press.
[Dubois pirated the unpublished work of Coeurdoux 1777, see Murr 1987.]
During Caspers, E. C. L. 1979. Sumer, coastal Arabia and the Indus Valley in Protoliterate and Early Dynastic eras:
Supporting evidence for a cultural linkage. Journal of the Economic and Social History of the Orient 22(2):121–135.
Durrani, F. A. 1988. Excavations in the Gomal Valley: Rahman Dheri excavation report no. 1. (Ancient Pakistan 6.)
Peshawar: Department of Archaeology, University of Peshawar.
Durrani, F. A., I. Ali, & G. Erdosy. 1994–1995. Seals and inscribed sherds. Pp. 198–223 in: Taj Ali (ed.), Excavations in
the Gomal Valley: Rehman Dheri report no. 2. (Ancient Pakistan 10.) Peshawar: Department of Archaeology,
University of Peshawar.
Eder, Christian. 1996. Die ägyptischen Motive in der Glyptik des östlichen Mittelmeerraumes zu Anfang des 2. Jts.
v.Chr. (Orientalia Lovanensia analecta 71.) Leuven: Peeters.
Edzard, Dietz Otto. 1968. Die Inschriften der altakkadischen Rollsiegel. Archiv für Orientforschung 22:12–20.
Eggeling, Julius (trans.). 1882–1900. The Śatapatha-Brāhmaṇa according to the text of the Mādhyandina school I–V.
(Sacred Books of the East 12, 26, 41, 43, 44.) Oxford: Clarendon Press.
Ehlers, Gerhard. 2007. Sieben Seher, sieben Wasser. Pp. 145–150 in: Konrad Klaus & Jens-Uwe Hartmann (eds.), Indica
et Tibetica: Festschrift für Michael Hahn zum 65. Geburtstag von Freunden und Schülern überreicht. (Wiener
Studien zur Tibetologie und Buddhismuskunde 66.) Vienna: Arbeitskreis für tibetische und buddhistische Studien,
Universität Wien.
Eichmann, Ricardo, & Hermann Parzinger (eds.). 2001. Migration und Kulturtransfer: Der Wandel vorder- und
zentralasiatischer Kulturen im Umbruch vom 2. zum 1. vorchristlichen Jahrtausend. Bonn: Dr Rudolf Habelt.
Eidem, Jesper. 2014. The kingdom of Šamšī-Adad and its legacies. Pp. 137–146 in: Eva Cancik-Kirschbaum, Nicole
Brisch & Jesper Eidem (eds.), Constituent, confederate and conquered space: The emergence of the Mittani state.
(Topoi 17.) Berlin: Walter de Gruyter.
Ellis, Francis Whyte. 1816. Dissertation on the Telugu language [= “Note to the introduction.”]. Pp. 1–31 in: A. D.
Campbell, A grammar of the Teloogoo language. Madras: College Press.
Eltsov, Piotr Andreevich. 2008. From Harappa to Hastinapura: A study of the earliest South Asian city and civilization.
(American School of Prehistoric Research Monograph series.) Leiden: Brill.
Emeneau, M. B. 1966. The dialects of Old Indo-Aryan. Pp. 123–138 in: Henrik Birnbaum & Jaan Puhvel (eds.), Ancient
Indo-European dialects. Berkeley: University of California Press.
Erdosy, George (ed.). 1995. The Indo-Aryans of ancient South Asia: Language, material culture and ethnicity. (Indian
Philology and South Asian Studies 1.) Berlin: Walter de Gruyter.
Falk, Harry. 1986. Bruderschaft und Würfelspiel. Untersuchungen zur Entwicklungsgeschichte des vedischen Opfers.
Freiburg: Hedvig Falk.
_____
. 1989. Soma I and II. Bulletin of the School of Oriental and African Studies, University of London 52(1):77–90.
_____. 1993. Schrift im alten Indien: Ein Forschungsbericht mit Anmerkungen. (ScriptOralia 56.) Tübingen: Gunter Narr.
_____
. 1994. Das Reitpferd im vedischen Indien. Pp. 91–101 in: Hänsel & Zimmer 1994.
Fansa, Mamoun, & Stefan Burmeister (eds.). 2004 Rad und Wagen: Der Ursprung einer Innovation. Wagen im Vorderen
Orient und Europa. (Beiheft der Archäologischen Mitteilungen aus Nordwestdeutschland 40.) Mainz: Philipp von
Zabern.
Farmer, Steve, Richard Sproat, & Michael Witzel. 2004. The collapse of the Indus script thesis: The myth of a literate
Harappan Civilization. Electronic Journal of Vedic Studies 11(2):19–57.
Farnell, Lewis Richard. 1921. Greek hero cults and ideas of immortality. Oxford: Clarendon Press.
Filliozat, Jean. 1962. Notes d’astronomie ancienne de l’Iran et de l’Inde (I, II et III). Journal Asiatique 250:325–350.
Fiordalis, David V. 2013. The smile of the Buddha: Enigma, illumination, and majesty. Paper read at the 223rd annual
meeting of the American Oriental Society in Portland, Oregon, March 15, 2013.
Fischer, Eberhard, & Haku Shah. 1971. Mogra Dev, tribal crocodile gods: Wooden crocodile images of Chodhri, Gamit
and Vasava tribes, South Gujarat (India). (Art for tribal rituals in South Gujarat 1.) Ahmedabad: Gujarat Vidyapith.
Flattery, David Stophlet, & Martin Schwartz. 1989. Haoma and harmaline: The botanical identity of the Indo-Iranian
sacred hallucinogen “Soma” and its legacy in religion, language, and Middle Eastern folklore. (University of
California publications, Near Eastern studies 21.) Berkeley: University of California Press.
Fortson, Benjamin W., IV. 2010. Indo-European language and culture: An introduction. 2nd ed. Oxford: Blackwell.
Frame, Douglas. 1978. The myth of return in early Greek epic. New Haven, CT: Yale University Press.
_____. 2010. Hippota Nestor. (Hellenic Studies 37.) Washington, DC: Center for Hellenic Studies.

Francfort, Henri-Paul. 1989. Fouilles de Shortughaï: Recherches sur l’Asie Centrale Protohistorique I–II. (Mémoires de
la Mission Archéologique Française en Asie Centrale 2.) Paris: Diffusion De Boccard.
Francfort, Henri-Paul. 2001. The cultures with painted ceramics of south Central Asia and their relations with the
northeastern steppe zone (late 2nd/early 1st millennium BC). Pp. 221–235 in: Eichmann & Parzinger 2001.
_____. 2005. La civilisation de l’Oxus et les Indo-Iraniens et Indo-Aryens. Pp. 253–328 in: Fussman et al. 2005.

Franke, Ute. 2010. From the Oxus to the Indus: Two compartmented seals from Mohenjo-daro (Pakistan). Pp. xvii–xliii
in: CISI 3.1.
Franke-Vogt, Ute. 1991 (1992). Die Glyptik aus Mohenjo-Daro. Uniformität und Variabilität in der Induskultur:
Untersuchungen zur Typologie, Ikonographie und räumlichen Verteilung I–II. (Baghdader Forschungen 13.) Mainz:
Philipp von Zabern.
_____. Der Golfhandel im späten 3. und frühen 2. Jt. v. Chr. Pp. 114–133 in: Bartl, Bernbeck, & Heinz 1995.

Frankfort, Henri. 1955. Stratified cylinder seals from the Diyala region. (University of Chicago Oriental Institute
Publications 72.) Chicago: University of Chicago Press.
Fujii, Masato. 2012. The Jaiminīya Sāmaveda traditions and manuscripts in South India. Pp. 99–118 in: Saraju Rath
(ed.), Aspects of manuscript culture in South India. (Brill’s Indological library 40.) Leiden: Brill.
Fuller, C. J. 1992. The camphor flame: Popular Hinduism and society in India. Princeton, NJ: Princeton University Press.
Fuller, Dorian Q. 2006. Agricultural origins and frontiers in South Asia: A working synthesis. Journal of World
Prehistory 20(1):1–86.
Fussman, Gérard, Jean Kellens, Henri-Paul Francfort, & Xavier Tremblay. 2005. Âryas, Aryens et iraniens en Asie
centrale. (Collège de France, Publications de l’Institut de Civilisation Indienne 72.) Paris: Diffusion De Boccard.
Gadd, C. J., & Sidney Smith. 1924. The new links between Indian and Babylonian civilizations. Illustrated London News,
October 4, 1924:614–616.
Gamkrelidze, T. V., & V. V. Ivanov. 1995. Indo-European and the Indo-Europeans I–II. Translated by Johanna Nichols.
(Trends in Linguistics, Studies and Monographs 80.) Berlin: Mouton de Gruyter.
Geldner, Karl F. 1889. Ficus Indica in RV. 1,24,7. Pp. 113–115 in: Richard Pischel & Karl F. Geldner, Vedische Studien I.
Stuttgart: W. Kohlhammer.
Gening, V. F., G. B. Zdanovich, & V. V. Gening. 1992. Sintashta: Arkheologicheskie pamyatniki arijskikh plemen Uralo-
Kazakhstanskikh stepej I. Chelyabinsk: Yuzhno-Ural’skoe knizhnoe izdatel’stvo.
Ghirshman, Roman. 1977. L’Iran et la migration des Indo-Aryens et des Iraniens. Leiden: Brill.
Ghosh, A. (ed.). 1989. An encyclopaedia of Indian archaeology I–II. New Delhi: Munshiram Manoharlal.
Gimbutas, Marija. 1997. The Kurgan Culture and the Indo-Europeanization of Europe: Selected articles from 1952 to
1993. Edited by Miriam Robbins Dexter and Karlene Jones-Bley. (Journal of Indo-European Studies Monograph 18.)
Washington, DC: Institute for the Study of Man.
Goldman, Robert P. 1984. The Rāmāyaṇa of Vālmīki: An epic of ancient India I: Bālakāṇḍa. Introduction and translation
by Robert P. Goldman. Annotation by Robert P. Goldman and Sally J. Sutherland. Princeton, NJ: Princeton University
Press.
Gonda, Jan. 1971. Old Indian. (Handbuch der Orientalistik II.1.1.) Leiden: Brill.
_____. 1974. The dual deities in the religion of the Veda. (Verlandelingen der Koninklijke Nederlandse Akademie van
Wetenschappen, Afdeeling Letterkunde, Nieuwe Reeks 81.) Amsterdam: North-Holland.
_____. 1975. Vedic literature (Saṃhitās and Brāhmaṇas). (A history of Indian literature I:1.) Wiesbaden: Otto
Harrassowitz.
_____. 1977. The ritual Sūtras. (A history of Indian literature I:2.) Wiesbaden: Otto Harrassowitz.
_____. 1978. Die Religionen Indiens I: Veda und älterer Hinduismus. 2nd ed. (Die Religionen der Menschheit 11.)
Stuttgart: W. Kohlhammer.
Gordon, D. H. 1960. The Prehistoric background of Indian culture. 2nd ed. Bombay: Bhulabhai Memorial Institute.
Grimm, Jakob. 1822. Deutsche Grammatik I. Zweite Ausgabe. Göttingen: Dieterichische Buchhandlung.
Gubaev, A., G. Koshelenko, & M. Tosi (eds.). 1998. The archaeological map of the Murghab Delta: Preliminary reports
1990–95. (Istituto Italiano per l’Africa e l’Oriente, Centro Scavi e ricerche archaeologiche, Reports and Memoirs,
Series Minor 3.) Rome: IsIAO.
Gundert, Hermann. 1869. Die dravidischen Elemente in Sanskrit. Zeitschrift der Deutschen Morgenländischen
Gesellschaft 23:517–530.
Gurov, Nikita V. 1968. Prospects for the linguistic interpretation of the Proto-Indian texts (on the basis of Dravidian
languages). Pp. 28–54 in: Proto-Indica: 1968. Brief report on the investigation of the Proto-Indian texts. (VIII
International Congress of Anthropological and Ethnographical Sciences, Tokio, September, 1968.) Moscow: Institute
of Ethnography, Academy of Sciences of USSR.
_____. 2000. On the substratum heritage in the vocabulary of the Ṛgveda (in connection with recent works by F. B. J.
Kuiper and M. Witzel). Paper read at the 21st Zograph conference “Traditional Indian texts: Problems of
interpretation”, St. Petersburg, April 24–26, 2000.
Gusev, S. A. 1998. K voprosu o transportnykh sredstvakh tripol’skoj kul’tury. Rossijskaya Arkheologiya 1998(1):15–28.
Häkkinen, Jaakko. 2009. Kantauralin ajoitus ja paikannus: perustelut puntarissa. Journal de la Société Finno-Ougrienne
92:9–56.
Hänsel, Bernhard, & Stefan Zimmer (eds.). 1994. Die Indogermanen und das Pferd. (Archaeolingua 4.) Budapest:
Archaeolingua.
Hansen, Donald P. 2003a. Art of the early city-states. Pp. 21–42 in: Aruz 2003.
_____. 2003b. Art of the Akkadian Dynasty. Pp. 189–233 in: Aruz 2003.

Hart, George L. 1975. The poems of ancient Tamil: Their milieu and their Sanskrit counterparts. Berkeley: University of
California Press.
Hauer, J. W. 1927. Der Vrātya: Untersuchungen über die nichtbrahmanische Religion Altindiens I. Stuttgart: W.
Kohlhammer.
Heesterman, J. C. 1957. The ancient Indian royal consecration: The rājasūya described according to the Yajus texts and
annotated. (Disputationes Rheno-Trajectinae 2.) ’s-Gravenhage: Mouton.
_____
. 1962. Vrātya and sacrifice. Indo-Iranian Journal 6(1):1–37.
_____. 1967. The case of the severed head. Wiener Zeitschrit für die Kunde Süd- und Ostasiens 11:22–43.
_____. 1989. King and warrior. Pp. 97–122 in: J. C. Galey (ed.), Kingship and the kings. (History and anthropology 4.)
London: Harwood Academic.
Heras, Henry. 1953. Studies in Proto-Indo-Mediterranean Culture I. (Studies in Indian History 19.) Bombay: Indian
Historical Research Institute.
Hévesy, Guillaume de [Vilmos Hevesy]. 1933. Océanie et Inde préaryenne: Mohenjo-daro et l’Ile de Pâques. Bulletin de
l’Association française des amis de l’orient 14–15:29–50.
Hickmann, Hans 1946. La trompette dans l’Egypte ancienne. (Supplément aux Annales du Service des Antiquités de
l’Egypte, Cahier 1.) Cairo: [n.p.].
Hiebert, Fredrik Talmage. 1994. Origins of the Bronze Age oasis civilizations of Central Asia. (Bulletin of the American
School of Prehistoric Research 42.) Cambridge, MA: Peabody Museum.
_____. 1995. South Asia from a Central Asian perspective. Pp. 192–205 in: Erdosy (ed.) 1995.

Hillebrandt, Alfred. 1897. Ritual-Litteratur, vedische Opfer und Zauber. (Grundriss der Indo-Arischen Philologie und
Altertumskunde III:2.) Strasbourg: Karl J. Trübner.
_____. 1927–1929. Vedische Mythologie I–II. Zweite veränderte Auflage. Breslau: M. & H. Marcus.

Hiltebeitel, Alf. 1978. The Indus Valley “Proto-Śiva”, reexamined through reflections on the goddess, the buffalo, and
the symbolism of vāhanas. Anthropos 73:767–797.
Hintze, Almut. 1998. The migrations of the Indo-Iranians and the Iranian sound change s > h. Pp. 139–153 in: Wolfgang
Meid (ed.), Sprache und Kultur der Indogermanen. (Innsbrucker Beiträge zur Sprachwissenschaft 93.) Innsbruck:
Institut für Sprachwissenschaft der Universität Innsbruck.
Hoffmann, Karl. 1940. Vedische Namen. Wörter und Sachen 21:139–161.
Honoré, Emmanuelle. 2007. Earliest cylinder-seal glyptic in Egypt: From Greater Mesopotamia to Naqada. Pp. 31–45
in: Hany Hanna (ed.), The International conference on heritage of Naqada and Qus region, January 22–28, 2007,
Naqada, Egypt I. Alexandria: International Council of Museums (ICOM).
Hopkins, E. Washburn. 1889. The social and military position of the ruling caste in ancient India, as represented by the
Sanskrit epic; with an appendix on the status of woman. Journal of the American Oriental Society 13:57–376.
_____
. 1896. Prāgāthikāni I. Journal of the American Oriental Society 17:23–92.
_____
. 1915. Epic mythology. (Grundriss der indo-arischen Philologie und Altertumskunde III:1:B.) Strasbourg: Karl J.
Trübner.
Horsch, Paul. 1966. Die vedische Gāthā- und Śloka-Literatur. Bern: Francke.
Houben, Jan E. M. (ed. and trans.). 1991. The Pravargya Brāhmaṇa of the Taittirīya Āraṇyaka: An ancient commentary
on the Pravargya ritual. Introduction, translation, and notes. Delhi: Motilal Banarsidass.
_____
. 2000. On the earliest attestable forms of the Pravargya ritual: Ṛg-Vedic references to the Gharma-Pravargya,
especially in the Atri-family book (book 5). Indo-Iranian Journal 43:1–25.
Hume, Robert Ernest (trans.). 1931. The thirteen principal Upanishads. London: Oxford University Press.
Hunter, G. R. 1934. The script of Harappa and Mohenjo-daro and its connection with other scripts. (Studies in the
History of Culture 1.) London: Kegan Paul, Trench, Trubner.
Indus Civilization Exhibition, Tokyo Metropolitan Art Museum, 5th August – 3rd December, 2000. Tokyo: NHK, NHK
Promotions, 2000.
Insler, Stanley. 1996. Avestan vāz and Vedic vāh. Studien zur Indologie und Iranistik 20:169–186.
_____. 1998. On the recensions of the Atharva Veda and Atharvan hymn composition. Wiener Zeitschrift für die Kunde
Südasiens 42:5–21.
Jacobsen, Knut A. (ed.). 2009–2013. Brill’s Encyclopedia of Hinduism I–V. (Handbook of Oriental Studies, Section 2:
South Asia 22.) Leiden: Brill.
Jaiswal, Suvira. 1981. Origin and development of Vaiṣṇavism: Vaiṣṇavism from 200 BC to AD 500. Second revised and
enlarged edition. New Delhi: Munshiram Manoharlal.
Jamison, Stephanie W., & Joel P. Brereton (ed. and trans.). 2014. The Rigveda I–III. New York: Oxford University Press.
Jamison, S. W., & M. Witzel 1992. Vedic Hinduism. http://www.people.fas.harvard.edu/~witzel/vedica.pdf.
Jansen, Michael. 1993. Mohenjo-Daro: Stadt der Brunnen und Kanäle: Wasserluxus vor 4500 Jahren. (Schriftenreihe
der Frontinus-Gesellschaft, Supplementband 2.) Bergisch Gladbach: Frontinus-Gesellschaft.
Jansen, Michael, Máire Mulloy, & Günter Urban (eds.). 1991. Forgotten cities on the Indus: Early civilization in Pakistan
from the 8th to the 2nd millennium B.C. Mainz: Philipp von Zabern.
Jarrige, Catherine, Jean-François Jarrige, Richard H. Meadow, & Gonzague Quivron. 1995. Mehrgarh field reports,
1974–1985: From Neolithic times to the Indus Civilization. Karachi: Sind Culture Department.
Jarrige, Jean-François. 1985. The Indian world: The archaeological background. Pp. 238–247 in: The world atlas of
archaeology. London: Mitchell Beazley.
_____. 1991. The cultural complex of Mehrgarh (Period VIII) and Sibri. The “Quetta hoard.” Pp. 94–103 in: Jansen,
Mulloy, & Urban 1991.
Jeremias, Alfred. 1929. Handbuch der altorientalischen Geisteskultur. Zweite Aufl. Berlin: Walter de Gruyter.
Jettmar, Karl. 1986. The religions of the Hindukush I: The religion of the Kafirs. The pre-Islamic heritage of Afghan
Nuristan. Translated by Adam Nayyar. With contributions from Schuyler Jones and Max Klimburg and a glossary by
Peter S. C. Parkes. Warminster: Aris & Phillips.
Joki, Aulis J. 1973. Uralier und Indogermanen: Die älteren Berührungen zwischen den uralischen und
indogermanischen Sprachen. (Mémoires de la Société Finno-Ougrienne 151.) Helsinki: Suomalais-Ugrilainen Seura.
Jones, William. 1788. On the Hindús. The third anniversary discourse [delivered at the Asiatic Society, Calcutta, on
February 2, 1786]. Asiatick Researches 1:415–431.
Joshi, Jagat Pati, & Asko Parpola. 1987. Corpus of Indus seals and inscriptions I: Collections in India. (Annales
Academiae Scientiarum Fennicae B239.) Helsinki: Suomalainen Tiedeakatemia. (= CISI 1)
Kane, P. V. 1941–1958. History of Dharmaśāstra I–V. Partly 2nd ed. (Government Oriental Series B6.) Poona:
Bhandarkar Oriental Research Institute.
Karttunen, Klaus, & Petteri Koskikallio (eds.). 2001. Vidyārṇavavandanam: Essays in honour of Asko Parpola. (Studia
Orientalia 94.) Helsinki: Finnish Oriental Society.
Kaw, R. N. 1989. Burzahom. Pp. 86–89 in: Ghosh 1989 II.
Keith, Arthur Berriedale (trans.). 1914. The Veda of the Black Yajus school entitled Taittirīya Sanhitā I–II. (Harvard
Oriental Series 18–19.) Cambridge, MA: Harvard University Press.
_____
(trans.). 1920. Rigveda Brahmanas: The Aitareya and Kauṣītaki Brāhmaṇas of the Rigveda. (Harvard Oriental
Series 25.) Cambridge, MA: Harvard University Press.
Kenoyer, Jonathan Mark. 1998. Ancient cities of the Indus Valley civilization. Karachi: Oxford University Press.
Khan, F. A. 1965. Excavations at Kot Diji. Pakistan Archaeology 2:11–85.
Kharakwal, J. S., Y. S. Rawat, & Toshiki Osada (eds.). 2012. Excavation at Kanmer 2005–06 – 2008–09. Kanmer
Archaeological Research Project: An Indo-Japanese collaboration. Kyoto: Indus Project, Research Institute for
Humanity and Nature.
Kinsley, David. 1986. Hindu goddesses: Visions of the divine feminine in the Hindu religious tradition. Berkeley:
University of California Press.
Kirfel, Willibald. 1920. Die Kosmographie der Inder, nach den Quellen dargestellt. Bonn and Leipzig: Kurt Schroeder.
_____. 1952. Der Mythos von der Tārā und der Geburt des Budha. Zeitschrift der Deutschen Morgenländischen
Gesellschaft 102:66–90.
_____. 1954. Das Purāṇa vom Weltgebäude (Bhuvanavinyāsa): Die kosmographischen Traktate der Purāṇa’s, Versuch
einer Textgeschichte. (Bonner Orientalistische Studien N.S.1.) Bonn: Selbstverlag des Orientalischen Seminars der
Universität.
Knorozov, Yurij V. 1952. Drevnyaya pis’mennost’ tsentral’noj Ameriki. Sovetskaya ètnografiya 1952(3):100–118.
Knorozov, Yuri V., M. F. Albedil, & B. Ya. Volchok. 1981. Proto-Indica: 1979. Report on the investigation of the Proto-
Indian texts. Moscow: Nauka.
[Knorozov, Yuri V., et al.] Proto-Indica: 1968. Brief report on the investigation of the Proto-Indian texts. Moscow 1968.
[Presented at the VIII International Congress of Anthropological and Ethnographical Sciences, Tokyo, September
1968.]
Kohl, Philip L. 2007. The making of Bronze Age Eurasia. Cambridge: Cambridge University Press.
Koivulehto, Jorma. 2001. The earliest contacts between Indo-European and Uralic speakers in the light of lexical loans.
Pp. 235–263 in: Carpelan, Parpola, & Koskikallio 2001.
Koryakova, Ludmila, & Andrej Epimakhov. 2007. The Urals and Western Siberia in the Bronze and Iron Ages.
(Cambridge World archaeology.) Cambridge: Cambridge University Press.
Koskenniemi, Kimmo. 1981. Syntactic methods in the study of the Indus script. Studia Orientalia 50:125–136.
Koskenniemi, Kimmo, & Asko Parpola. 1979. Corpus of texts in the Indus script. (Department of Asian and African
Studies, University of Helsinki, Research Reports 1.) Helsinki.
_____. 1980. Documentation and duplicates of the texts in the Indus script. (Department of Asian and African Studies,
University of Helsinki, Research Reports 2.) Helsinki.
_____. 1982. A concordance to the texts in the Indus script. (Department of Asian and African Studies, University of
Helsinki, Research Reports 3.) Helsinki.
Koskenniemi, Seppo, Asko Parpola, & Simo Parpola. 1970. A method to classify characters of unknown ancient scripts.
Linguistics 61:65–91.
_____
. 1973. Materials for the study of the Indus script I: A concordance to the Indus inscriptions. (Annales Academiae
Scientiarum Fennicae B185.) Helsinki: Suomalainen Tiedeakatemia.
Kovtun, I. V. 2013. Predystoriya indoarijskoj mifologii. Kemerovo: Rossijskaya Akademiya Nauk, Sibirskoe otdelenie,
Institut èkologii cheloveka.
Kramrisch, Stella. 1954. The art of India: Traditions of Indian sculpture, painting and architecture. London: Phaidon.
Krishnamurti, Bhadriraju. 2003. The Dravidian languages. (Cambridge Language Surveys.) Cambridge: Cambridge
University Press.
Kriyā. 2008. Kriyāvi nta rkālat tami lakarāti: Tami l-tami l-āṅkilam. 2nd ed. Chennai: Kriyā (Cre-A).
Kuiper, F. B. J. 1979. Varuṇa and Vidūṣaka: On the origin of the Sanskrit drama. (Verhandelingen der Koninklijke
Nederlandse Akademie van Wetenschappen, Afd. Letterkunde, Nieuwe Reeks 100.) Amsterdam: North Holland.
_____
. 1991. Aryans in the Rigveda. (Leiden Studies in Indo-European 1.) Amsterdam: Rodopi.
Kuz’mina, Elena E. 2007. The origin of the Indo-Iranians. Translated by S. Pitina and P. Prudovsky. Edited by James P.
Mallory. (Leiden Indo-European Etymological Dictionary Series 3.) Leiden: Brill.
Kuz’minykh, S. V. 2011. Sejminsko-turbinskaya problema: Novye materialy. Kratkie soobshcheniya Instituta Arkheologii
RAN 225:240–263.
Lamberg-Karlovsky, C. C. 1986. Third millennium structure and process: From the Euphrates to the Indus and the Oxus
to the Indian Ocean. Oriens Antiquus 25(3–4):189–219.
_____. 2002. Archaeology and language: The Indo-Iranians. Current Anthropology 43(1):63–88.

Lambrick, H. T. 1964. Sind: A general introduction. Hyderabad (Sind): Sindhi Adabi Board.
Lassen, Christian. 1847–1852. Indische Alterthumskunde I–II. Bonn: H. B. Koenig.
Laursen, Steffen Terp. 2010. The westward transmission of Indus Valley sealing technology: Origin and development of
the “Gulf Type” seal and other administrative technologies in Early Dilmun, c. 2100–2000 BC. Arabian Archaeology
and Epigraphy 21:96–134.
Law, Randall William. 2011. Inter-regional interaction and urbanism in the ancient Indus Valley: A geologic provenience
study of Harappa’s rock and mineral assemblage. (Linguistics, archaeology and the human past, Occasional Paper
11.) Kyoto: Indus Project, Research Institute for Humanity and Nature.
Lawergren, Bo. 2003. Oxus trumpets, ca. 2200–1800 BCE: Material overview, usage, societal role and catalogue. Iranica
Antiqua 38:41–118.
Lawler, Andrew. 2004. The Indus script: Write or wrong? Science 306 (December 17, 2004):2026–2029.
Legrain, L. 1921. Empreintes de cachets élamites. (Mémoires de la Mission Archéologique de Perse 16.) Paris: Ernest
Leroux.
Leick, Gwendolyn. 1991. A dictionary of ancient Near Eastern mythology. London and New York: Routledge.
Leshnik, Lawrence S. 1974. South Indian “Megalithic” burials: The Pandukal complex. Wiesbaden: Franz Steiner.
Liebert, Gösta. 1968 (1969). Beitrag zur Frage des Polarsterns in der altindischen Literatur. Orientalia Suecana
17:155–170.
Levine, Marsha, Colin Renfrew, & Katie Boyle (eds.). 2003. Prehistoric steppe adaptation and the horse. (McDonald
Institute Monographs.) Cambridge: McDonald Institute for Archaeological Research.
Ligabue, Giancarlo, & Sandro Salvatori (eds.). [1989]. Bactria: An ancient oasis civilization from the sands of
Afghanistan. (Studies and documents 3.) Venice: Erizzo.
Littauer, M. A., & J. H. Crouwel. 1979. Wheeled vehicles and ridden animals in the ancient Near East. Drawings by J.
Morel. (Handbuch der Orientalistik 7:1:2:B:1.) Leiden: Brill.
_____. 2002. Selected writings on chariots and other early vehicles, riding and harness. Edited by Peter Raulwing.
(Culture and history of the ancient Near East 6.) Leiden: Brill.
Lubotsky, Alexander. 2001. The Indo-Iranian substratum. Pp. 301–317 in: Carpelan, Parpola, & Koskikallio 2001.
Lüders, Heinrich. 1951–1959. Varuṇa I–II. Aus dem Nachlass herausgegeben von Ludwig Alsdorf. Göttingen:
Vandenhoeck & Ruprecht.
Macdonell, A. A. 1897. Vedic mythology. (Grundriss der indo-arischen Philologie und Altertumskunde III:1:A.)
Strasbourg: Karl J. Trübner.
Mackay, Ernest J. H. 1938. Further excavations at Mohenjo-daro I–II. Delhi: Manager of Publications, Government of
India.
_____
. 1943. Chanhu-daro excavations, 1935–36. (American Oriental Series 20.) New Haven, CT: American Oriental
Society.
Maclean, C. D. 1893. Glossary of the Madras Presidency. (Manual of the administration of the Madras Presidency 3.)
Madras: Government Press.
Maekawa, Kazuya, & Wakaha Mori. 2011. Dilmun, Magan, and Meluhha in early Mesopotamian history: 2500–1600 BC.
Pp. 245–269 in: Osada & Witzel 2011.
Mahadevan, Iravatham. 1977. The Indus script: Texts, concordance and tables. (Memoirs of the Archaeological Survey
of India 77.) New Delhi: Archaeological Survey of India.
_____
. 2003. Early Tamil epigraphy: From the earliest times to the sixth century A.D. (Harvard Oriental Series 62.)
Chennai: Cre-A; Cambridge, MA: Department of Sanskrit and Indian Studies, Harvard University.
Mahapatra, Piyushkanti. 1972. The folk cults of Bengal. Calcutta: Indian Publications.
Mair, Victor H. (ed.). 1998. The Bronze and Iron Age peoples of eastern Central Asia I–II. (Journal of Indo-European
Studies Monograph 26.) Washington, DC: Institute for the Study of Man.
Mallory, J. P. 1989. In search of the Indo-Europeans: Language, archaeology and myth. London: Thames and Hudson.
_____
. 1994–1995. The Indo-European homeland: An Asian perspective. Bulletin of the Deccan College Post-Graduate and
Research Institute 54–55:237–254.
_____. 1996. The Indo-European homeland: A matter of time. Pp. 1–22 in: Karlene Jones-Bley & Martin E. Huld (eds.),
The Indo-Europeanization of northern Europe. (Journal of Indo-European Studies Monograph 17.) Washington, DC:
Institute for the Study of Man.
_____
. 1997. The homelands of the Indo-Europeans. Pp. 93–121 in: Roger Blench & Matthew Spriggs (eds.), Archaeology
and language I. London: Routledge.
_____. 1998. A European perspective on Indo-Europeans in Asia. Pp. 175–201 in: Mair 1998:I.
_____
. 2001. Uralics and Indo-Europeans: Problems of time and space. Pp. 345–366 in: Carpelan, Parpola, & Koskikallio
2001.
_____. 2002a. Archaeological models and Asian Indo-Europeans. Pp. 19–42 in: Sims-Williams 2002.
_____
. 2002b. Comment. Pp. 79–80 in: Lamberg-Karlovsky 2002.
Mallory, J. P., & D. Q. Adams (eds.). 1997. Encyclopedia of Indo-European culture. London and Chicago: Fitzroy
Dearborn.
Mallory, J. P., & Victor H. Mair. 2000. The Tarim mummies: Ancient China and the mystery of the earliest peoples from
the west. London: Thames and Hudson.
Mallowan, M. E. L. 1965. Early Mesopotamia and Iran. (Library of the early civilizations.) London: Thames and Hudson.
Maloney, Clarence. 1970. The beginnings of civilization in South India. Journal of Asian Studies 29(3):603–616.
Maran, Joseph. 2004a. Die Badener Kultur und ihre Räderfahrzeuge. Pp. 265–282 in: Fansa & Burmeister 2004.
_____. 2004b. Kulturkontakte und Wege der Ausbreitung der Wagentechnologie im 4. Jahrtausend v.Chr. Pp. 429–442 in:
Fansa & Burmeister 2004.
Marshall, John. 1924. First light on a long-forgotten civilization. Illustrated London News September 20, 1924:528–532.
_____
. 1931a. Religion. Pp. 48–78 in: Marshall 1931b:I.
_____, (ed.). 1931b. Mohenjo-Daro and the Indus Civilization I–III. London: Arthur Probsthain.

Mayrhofer, Manfred. 1966. Die Indo-Arier im Alten Vorderasien. Mit einer analytischen Bibliographie. Wiesbaden: Otto
Harrassowitz.
_____. 1974. Die Arier im Vorderen Orient—ein Mythos? Mit einem bibliographischen Supplement. (Österreichische
Akademie der Wissenschaften, Philos.-hist. Klasse, Sitzungsberichte 294:3.) Vienna: Verlag der Österreichischen
Akademie der Wissenschaften.
_____
. 1992–2001. Etymologisches Wörterbuch des Altindoarischen I–III. Heidelberg: Universitätsverlag Carl Winter.
_____. 2002 (2004). Zur Vertretung der indogermanischen Liquiden in den indo-iranischen Sprachen. Indologica
Taurinensia 28:149–161.
McIntosh, Jane R. 2008. The ancient Indus Valley: New perspectives. (Understanding ancient civilizations.) Santa
Barbara, CA: ABC-CLIO.
Meadow, Richard H. 1991. The domestication and exploitation of plants and animals in the Greater Indus Valley, 7th–
2nd millennium B.C. Pp. 51–58 in: Jansen, Mulloy, & Urban 1991.
Meadow, Richard H., & Ajita K. Patel. 1997. A comment on “Horse remains from Surkotada” by Sandor Bökönyi. South
Asian Studies 13:308–315.
Meriggi, Piero. 1934. Zur Indus-Schrift. Zeitschrift der Deutschen Morgenländischen Gesellschaft 87(N.F.12):198–241.
Mode, Heinz. 1944. Indische Frühkulturen und ihre Beziehungen zum Westen. Basel: Benno Schwabe.
Monier-Williams, Monier. 1885. Religious thought and life in India: An account of the religions of the Indian peoples,
based on a life’s study of their literature and on personal investigations in their own country I: Vedism, Brāhmanism
and Hindūism. 2nd ed. London: John Murray.
Moran, William L. (ed. and trans.). 1992. The Amarna letters. Baltimore, MD: Johns Hopkins University Press.
Morgenstierne, Georg. 1953. Some Kati myths and hymns. Acta Orientalia 21: 161–189.
Mughal, Mohammad Rafique. 1970. The Early Harappan period in the Greater Indus Valley and Northern Baluchistan
(c. 3000–2400 B.C.). (PhD dissertation in anthropology, University of Pennsylvania, Philadelphia.) Ann Arbor:
University Microfilms.
Müller, Max. 1859. A history of ancient Sanskrit literature, so far as it illustrates the primitive religion of the Brahmans.
London: Williams and Norgate.
Murr, Sylvia. 1987. L’Inde philosophique entre Bossuet et Voltaire. I: Moeurs et Coutumes des Indiens (1777). Un inédit
du Père G.-L. Coeurdoux s.j. dans la version de N.-J. Desvaulx, texte établi et annoté par Sylvia Murr. II: L’Indologie du
Père Coeurdoux: Stratégies, apologétique et scientificité. (Publications de l’École Française d’Extrême-Orient 146.)
Paris: École Française d’Extrême-Orient. [I = Coeurdoux 1777]
Needham, Joseph. 1959. Science and civilisation in China III: Mathematics and the sciences of the heavens and the
earth. Cambridge: Cambridge University Press.
Nelson, Harold Hayden, et al. 1932. Medinet Habu II: Later historical records of Ramses III. (Oriental Institute
Publications 9.) Chicago: University of Chicago Press.
NHK 2000. See Indus Civilization Exhibition.
Nichols, Johanna. 1997. Modeling ancient population structures and movement in linguistics. Annual Review of
Anthropology 26:359–384.
Nyberg, Harri. 1995. The problem of the Aryans and the Soma: The botanical evidence. Pp. 382–406 in: Erdosy 1995.
Oates, Joan. 2003. A note on the early evidence for horse and the riding of equids in Western Asia. Pp. 115–125 in:
Levine, Renfrew, & Boyle 2003.
Oberlies, Thomas. 1993. Die Aśvin: Götter der Zwischenbereiche. Studien zur Indologie und Iranistik 18:169–189.
_____. 2012. Der Rigveda und seine Religion. Berlin: Verlag der Weltreligionen im Insel Verlag.
Oldenberg, Hermann. 1884. Ṛigveda-Saṃhitā and Sāmavedārcika. Nebst Bemerkungen über die Zerlegung der
Ṛigveda-Hymnen in Theilhymnen und Strophen sowie über einige verwandte Fragen. Zeitschrift der Deutschen
Morgenländischen Gesellschaft 38:439–480.
Olijdam, Eric, & Richard H. Spoor (eds.). 2008. Intercultural relations between South and Southwest Asia: Studies in
commemoration of E.C.L. During Caspers (1934–1996). (BAR International Series 1826). Oxford: Archaeopress.
Olivelle, Patrick. 1998. The early Upaniṣads: Annotated text and translation. New York: Oxford University Press.
Osada, Toshiki. 2006. How many Proto-Munda words in Sanskrit? With special reference to agricultural vocabulary. Pp.
151–174 in: Toshiki Osada (ed.), Proceedings of the Pre-symposium of RIHN and the 7th ESCA Harvard-Kyoto
Roundtable. Kyoto: Research Institute for Humanity and Nature.
Osada, Toshiki & Hitoshi Endo (eds.) 2011. Linguistics, Archaeology and the Human Past: Occasional Paper 12. Kyoto:
Indus Project, Research Institute for Humanity and Nature. Reprinted 2012 with the title: Current Studies on the
Indus Civilization, IX. New Delhi: Manohar.
Osada, Toshiki, & Michael Witzel (eds.). 2011. Cultural relations between the Indus and the Iranian Plateau during the
third millennium BCE. Indus Project, Research Institute for Humanities and Nature, June 7–8, 2008. (Harvard
Oriental Series, Opera Minora 7.) Cambridge, MA: Department of Sanskrit and Indian Studies, Harvard University.
Pankenier, David W. 2004. A brief history of Beiji (Northern Culmen). Culture and Cosmos 8(1–2):287–308.
_____. 2013. Astrology and cosmology in early China: Conforming Earth to Heaven. Cambridge: Cambridge University
Press.
Parpola, Asko. 1967 (1968). On the Jaiminīyaśrautasūtra and its annexes. Orientalia Suecana 16:181–214.
_____. 1968–1969. The Śrautasūtras of Lāṭyāyana and Drāhyāyaṇa and their commentaries: An English translation and
study I. 1: General introduction and appendices to vol. 1. 2: The agniṣṭoma (LŚS I-II, DŚS I-VI). (Commentationes
Humanarum Litterarum 42:2 and 43:2.) Helsinki: Societas Scientiarum Fennica.
_____
. 1973. The literature and study of the Jaiminīya Sāmaveda in retrospect and prospect. (Studia Orientalia 43:6.)
Helsinki: Finnish Oriental Society.
_____. 1974. On the protohistory of the Indian languages in the light of archaeological, linguistic and religious evidence:
An attempt at integration. Pp. 90–100 in: J. E. van Lohuizen-de Leeuw & J. M. M. Ubaghs (eds.), South Asian
Archaeology 1973. Leiden: Brill.
_____
. 1975a. Tasks, methods and results in the study of the Indus script. Journal of the Royal Asiatic Society of Great
Britain and Ireland 1975(2):178–209.
_____. 1975b. Isolation and tentative interpretation of a toponym in the Harappan inscriptions. Pp. 121–143 in: Jean
Leclant (ed.), Le déchiffrement des écritures et des langues: Colloque du XXIXe Congrès International des
Orientalistes. Paris: L’Asiathèque.
_____
. 1981. On the primary meaning and etymology of the sacred syllable Ōm. Studia Orientalia 50:195–213.
_____. 1983. The pre-Vedic Indian background of the Śrauta rituals. Pp. 41–75 in: F. Staal 1983:II.
_____
. 1984a. New correspondences between Harappan and Near Eastern glyptic art. Pp. 176–195 in: Bridget Allchin
(ed.), South Asian Archaeology 1981. (University of Cambridge Oriental Publlications 35.) Cambridge: Cambridge
University Press.
_____. 1984b. On the Jaiminīya and Vādhūla traditions of South India and the Paṇḍu / Pāṇḍava problem. Studia Orientalia
55:429–468.
_____
. 1985. The Sky-Garment: A study of the Harappan religion and its relation to the Mesopotamian and later Indian
religions. (Studia Orientalia 57.) Helsinki: Finnish Oriental Society.
_____. 1986a. The size and quality of the Indus seals and other clues to the royal titles of the Harappans. Tamil
Civilization 4(3–4):144–156.
_____
. 1986b. The Indus script: A challenging puzzle. World Archaeology 17(3):399–419.
_____. 1988a. The coming of the Aryans to Iran and India and the cultural and ethnic identity of the Dāsas. Studia
Orientalia 64:195–302.
_____
. 1988b. Religion reflected in the iconic signs of the Indus script: Penetrating into long-forgotten picto+graphic
messages. Pp. 114–135 in: Visible Religion (Annual for Religious Iconography) VI: The Image in Writing. Leiden: Brill.
_____. 1989. Witnessing a hog sacrifice to the Goddess in South Kanara. NIF Newsletter 1989(4):20–24. Turku: Nordic
Institute of Folklore.
_____
. 1990a. Astral proper names in India: An analysis of the oldest sources, with argumentation of an ultimately
Harappan origin. Adyar Library Bulletin 53:1–53.
_____
. 1990b. Bangles, sacred trees and fertility: Interpretations of the Indus script related to the cult of Skanda-Kumāra.
Pp. 263–284 in: Maurizio Taddei (ed.) South Asian Archaeology 1987 I. (Serie Orientale Roma 66:1.) Rome: Istituto
Italiano per il Medio ed Estremo Oriente.
_____
. 1992a. The metamorphoses of Mahiṣa Asura and Prajāpati. Pp. 275–308 in: A.W. van den Hoek, D.H.A. Kolff, &
M.S. Oort (eds.), Ritual, state and history in South Asia: Essays in honour of J. C. Heesterman. (Memoirs of the Kern
Institute 5.) Leiden: Brill.
_____. 1992b. The “fig deity seal” from Mohenjo-daro: Its iconography and inscription. Pp. 277–236 in: Catherine Jarrige
(ed.), South Asian Archaeology 1989 I. (Monographs in World Archaeology 14.) Madison, WI: Prehistory.
_____
. 1994a. Deciphering the Indus script. Cambridge: Cambridge University Press.
_____. 1994b. Harappan inscriptions: An analytical catalogue of the Indus inscriptions from the ancient Near East. Pp.
304–315 and (bibliography) 483–492 in: Flemming Højlund & H. Hellmuth Andersen, Qala’at al-Bahrain I: The
northern city wall and the Islamic fortress. (Jutland Archaeological Society Publications 30:1.) Aarhus: Jutland
Archaeological Society.
_____
. 1995. The problem of the Aryans and the Soma: Textual-linguistic and archaeological evidence. Pp. 353–381 in:
Erdosy 1995.
_____. 1997a. Dravidian and the Indus script: On the interpretation of some pivotal signs. Studia Orientalia 82:167–191.
_____
. 1997b. Deciphering the Indus script: Methods and select interpretations. Keynote address delivered at the 25th
Annual South Asia Conference, University of Wisconsin, Madison, Wisconsin, October 18–20, 1996. (Occasional
Papers Series 2.) Madison: Center for South Asia, University of Wisconsin-Madison.
_____. 1997c. The Aryan languages and archaeology, with an excursus on Botaj. Pp. 291–308 in: Raymond Allchin &
Bridget Allchin (eds.), South Asian Archaeology 1995 I. New Delhi: Oxford & IBH.
_____
. 1998a. Aryan languages, archaeological cultures, and Sinkiang: Where did Proto-Iranian come into being, and how
did it spread? Pp. 114–147 in: Mair 1998 I.
_____. 1998b. Sāvitrī and resurrection. Pp. 167–312 in: Asko Parpola & Sirpa Tenhunen (eds.), Changing patterns of
family and kinship in South Asia. (Studia Orientalia 84.) Helsinki: Finnish Oriental Society.
Parpola, Asko. 1999a. The formation of the Aryan branch of Indo-European. Pp. 180–207 in: Roger Blench and Matthew
Spriggs (eds.), Combining archaeological and linguistic aspects of the past. (World Archaeology: Archaeology and
Language series 3.) London: Routledge.
_____
. 1999b. Vāc as a goddess of victory in the Veda and her relation to Durgā. Zinbun 34(2):101–143.
_____
. 1999c. The iconography and cult of Kuṭṭiccāttan: Field research on the sanskritization of local folk deities in
Kerala. Pp. 175–205 in: Johannes Bronkhorst & Madhav M. Deshpande (eds.), Aryan and Non-Aryan in South Asia:
Evidence, interpretation and ideology. (Harvard Oriental Series, Opera Minora 3.) Cambridge, MA: Department of
Sanskrit and Indian Studies, Harvard University.
_____
. 1999d. Conceptual categories and their classification in Middle Vedic texts: A review of Brian K. Smith’s two
recent books. Pp. 5–22 in: Folke Josephson (ed.), Categorisation and interpretation: Indological and comparative
studies. A volume dedicated to the memory of Gösta Liebert. (Meijerbergs arkiv för svensk ordforskning 24.)
Göteborg: Styrelsen för Meijerbergs institut vid Göteborgs universitet.
_____. 2002a. From the dialects of Old Indo-Aryan to Proto-Indo-Aryan and Proto-Iranian. Pp. 43–102 in: Sims-Williams
2002.
_____
. 2002b. Pre-Proto-Iranians of Afghanistan as initiators of Śākta Tantrism: On the Scythian/Saka affiliation of the
Dāsas, Nuristanis and Magadhans. Iranica Antiqua 37:233–324.
_____. 2002c. Pandaíē and Sītā: On the historical background of the Sanskrit epics. Journal of the American Oriental
Society 122(2):361–373.
_____
. 2003. Sacred bathing place and transcendence: Dravidian kaṭa(vuḷ) as the source of Indo-Aryan ghāṭ, tīrtha,
tīrthankara and (tri)vikrama. Pp. 523–574 in: Olle Qvarnström (ed.), Jainism and early Buddhism: Essays in honor of
Padmanabh S. Jaini. Fremont, CA: Asian Humanities Press.
_____
. 2004. From archaeology to a stratigraphy of Vedic syncretism: The banyan tree and the water buffalo as
Harappan-Dravidian symbols of royalty, inherited in succession by Yama, Varuṇa and Indra, divine kings of the first
three layers of Aryan speakers in South Asia. Pp. 479–515 in: Arlo Griffiths & Jan E. M. Houben (eds.), The Vedas:
Texts, language & ritual. (Groningen Oriental Studies 20.) Groningen: Egbert Forsten.
_____. 2004 (2006). Bala-Rāma and Sītā: On the origins of the Rāmāyaṇa. Proceedings of the XIth World Sanskrit
Conference (Turin, April 3rd–8th, 2000) II. Indologica Taurinensia 30:185–200.
_____
. 2005a. The Nāsatyas, the chariot and Proto-Aryan religion. Journal of Indological Studies 16–17(2004–2005):1–63.
_____. 2005b. Study of the Indus script. Transactions of the International Conference of Eastern Studies 50:28–66.
_____. 2005c. Gandhara Graves and the Gharma pot, the Nasatyas and the Nose: In pursuit of the Chariot Twins. Ancient
Pakistan 16:73–107.
_____
. 2005d. Administrative contact and acculturation between Harappans and Bactrians: Evidence of sealings and
seals. Pp. 267–274 in: Catherine Jarrige & Vincent Lefèvre (eds.), South Asian Archaeology 2001 I: Prehistory. Paris:
Éditions Recherche sur les Civilisations.
_____
. 2008a. Proto-Indo-European speakers of the Late Tripolye culture as the inventors of wheeled vehicles: Linguistic
and archaeological considerations of the PIE homeland problem. Pp. 1–59 in: Karlene Jones-Bley, Martin E. Huld,
Angela Della Volpe, & Miriam Robbins Dexter (eds.), Proceedings of the Nineteenth Annual UCLA Indo-European
Conference, November 2–3, 2007. (Journal of Indo-European Studies Monograph 54.) Washington, DC: Institute for
the Study of Man.
_____. 2008b. Copper tablets from Mohenjo-daro and the study of the Indus script. Pp. 132–139 in: Olijdam & Spoor
2008.
_____
. 2008c. Is the Indus script indeed not a writing system? Pp. 111–131 in: Airavati: Felicitation volume in honour of
Iravatham Mahadevan. Chennai: Varalaaru.com. http://www.harappa.com/script/indus-writing.pdf.
_____. 2009. The face urns of Gandhāra and the Nāsatya cult. Pp. 149–162 in: Michael Willis (gen. ed.), Migration, trade,
and peoples (European Association of South Asian Archaeologists, Proceedings of the Eighteenth Congress, London,
2005), Part 3: Aryans and nomads, edited by Asko Parpola. London: British Association for South Asian Studies,
British Academy.
_____. 2010. A Dravidian solution to the Indus script problem. (Kalaignar M. Karunanidhi Classical Tamil Research
Endowment Lecture, World Classical Tamil Conference 25–6–2010 Coimbatore.) Chennai: Central Institute of
Classical Tamil.
_____
. 2011a. Crocodile in the Indus civilization and later South Asian traditions. Pp. 1–57 in: Osada & Endo 2011.
_____. 2011b. The Harappan unicorn in Eurasian and South Asian perspectives. Pp. 125–188 in: Osada & Endo 2011.
_____. 2011c. Motifs of early Iranian, Mesopotamian and Harappan art (and script), reflecting contacts and ideology. Pp.
271–357 in: Osada & Witzel 2011.
_____
. 2011d. Three ways of chanting in a sacrificial laud: Chapter two of Jaimini-Paryadhyāya (Jaiminīya-Śrautasūtra III)
with Bhavatrāta’s commentary: Sanskrit text with an annotated English translation. Pp. 141–163 in: Bertil Tikkanen
& Albion M. Butters (eds.), Pūrvāparaprajñābhinandanam: East and West, past and present. Indological and other
essays in honour of Klaus Karttunen. (Studia Orientalia 110.) Helsinki: Finnish Oriental Society.
_____
. 2011e. Codification of Vedic domestic ritual in Kerala: Pārvaṇa-sthālīpāka— the model of rites with fire offerings—
in Jaiminīya-Gṛhyasūtra 1,1-4 and in the Malayāḷam manual of the Sāmaveda Nampūtiri Brahmins of Kerala, the
Sāma-Smārtta-Caṭaṅṅu. Pp. 261–354 in: Jan E. M. Houben & Julieta Rotaru (eds.), Veda-Vedāṅga and Avesta between
orality and writing. (Travaux de symposium international Le Livre, La Roumanie, L’Europe, Troisième édition—20 à
24 Septembre 2011, Tome III: La troisième section—Études euro- et afro-asiatiques, Section IIIA.) Bucarest: Éditeur
Bibliothèque de Bacarest.
http://www.bibliotecametropolitana.ro/Uploads/Simpozionul%20International_Cartea_Romania_Europa_III_V3_mic.pdf.
_____
. 2012a. The Dāsas of the Ṛgveda as Proto-Sakas of the Yaz I -related cultures. With a revised model for the
protohistory of Indo-Iranian speakers. Pp. 221–264 in: Martin E. Huld, Karlene Jones-Bley, & Dean Miller (eds.),
Archaeology and Language: Indo-European Studies presented to James P. Mallory. (Journal of Indo-European Studies
Monograph 60.) Washington, DC: Institute for the Study of Man.
_____
. 2012b. Indus civilization (–1750 BCE). Pp. 3–18 in: Jacobsen 2012 IV.
_____
. 2012c. Chapter headings of the Jaimini-Paryadhyāya (Jaiminīya-Śrautasūtra III). Pp. 423–430 in: L. I. Kulikov & M.
A. Rusanov (eds.), T. Ya. Elizarenkova memorial volume, Book 2. (Orientalia et Classica 40.) Moscow: Russian State
University for the Humanities.
_____
. 2012 (2013). Formation of the Indo-European and Uralic (Finno-Ugric) language families in the light of
archaeology: Revised and integrated “total” correlations. Pp. 119–184 in: Riho Grünthal & Petri Kallio (eds.),
Linguistic Map of Prehistoric Northern Europe. (Mémoires de la Société Finno-Ougrienne 266.) Helsinki: Suomalais-
Ugrilainen Seura.
_____. 2013. Beginnings of Indian astronomy, with reference to a parallel development in China. History of Science in
South Asia 1:21–78. http://hssa.sayahna.org/.
_____
. 2014. Beginnings of the Indian and Chinese calendrical astronomy. Journal of the American Oriental Society
134(1):107–112.
_____. in press a. Indus seals and glyptic: An overview. In: Marta Ameri, Sarah Kielt Costello, Gregg M. Jamison, & Sarah
Jarmer Scott (eds.), Small Windows: New Approaches to the Study of Seals and Sealing as Tools of Identity, Political
Organization and Administration in the Ancient World. Cambridge: Cambridge University Press.
_____
. in press b. Sāmaveda and sorcery. In: V. Sadovsky & J. Rotaru (eds.), Language and ritual in Indo-Iranian
traditions. Vienna: Österreichische Akademie der Wissenschaften.
_____. in press c. Sanskrit kaparda- “braided hair”: Yet another Harappan symbol of royalty surviving in Vedic “Vrātya
rituals”. In: Tiziana Pontillo et al. (eds.), The volatile world of sovereignty: The Vrātya problem and kingship in South
Asia. New Delhi: D. K. Printworld.
_____
. in press d. Sanskrit śakaṭī- / śakaṭa- “ox-cart”. In: Gerd J. R. Mevissen & Vijay Kumar Gupta (eds.), Indology’s
pulse: Arts in context. Essays presented to Doris Meth Srinivasan in admiration of her scholarly research. New Delhi:
Aryan Publications International.
_____
. in press e. Rudra: “Red” and “cry” in the name of the destroyer god. In a forthcoming Festschrift.
Parpola, Asko. in press f. References to ritual authorities and Vedic schools in the Jaiminīya-Śrautasūtra, collected and
evaluated. Pp. 32 in: Jan. E. M. Houben, Julieta Rotaru, & Michael Witzel (eds.), Vedic śākhās, past, present, future.
Proceedings of the Fifth international Vedic workshop, Bucharest 2011.
Parpola, Asko, & Juha Janhunen. 2011. On the Asiatic wild asses and their vernacular names. Pp. 59–124 in: Osada &
Endo 2011.
Parpola, Asko, Seppo Koskenniemi, Simo Parpola, & Pentti Aalto. 1969a. Decipherment of the Proto-Dravidian
inscriptions of the Indus civilization: A first announcement. (Scandinavian Institute of Asian Studies, Special
Publications 1.) Copenhagen.
_____. 1969b. Progress in the decipherment of the Proto-Dravidian Indus script. (Scandinavian Institute of Asian Studies,
Special Publications 2.) Copenhagen.
_____
. 1970. Further progress in the Indus script decipherment. (Scandinavian Institute of Asian Studies, Special
Publications 3.) Copenhagen.
Parpola, Asko, & Petteri Koskikallio (eds.). 1994. South Asian Archaeology 1993: Proceedings of the Twelfth
international conference of the European association of South Asian archaeologists held in Helsinki University, 5–9
July 1993. I–II. (Annales Academiae Scientiarum Fennicae, Series B 271.) Helsinki: Suomalainen Tiedeakatemia.
Parpola, Asko, B. M. Pande, & Petteri Koskikallio. 2010. Corpus of Indus seals and inscriptions 3: New material,
untraced objects, and collections outside India and Pakistan 1: Mohenjo-daro and Harappa, in collaboration with
Richard H. Meadow & J. Mark Kenoyer. (Annales Academiae Scientiarum Fennicae, Humaniora 359.) Helsinki:
Suomalainen Tiedeakatemia. (= CISI 3.1)
Parpola, Asko, Simo Parpola, & Seppo Koskenniemi. 1966. Computing approach to Proto-Indian, 1965: An interim
report. [Helsinki: The authors.]
Parpola, Marjatta. 2000. Kerala Brahmins in Transition: A Study of a Nampūtiri Family. (Studia Orientalia 91.) Helsinki:
The Finnish Oriental Society.
Parpola, Päivikki. 1987. Väline merkkijoukkojonoa kuvaavan kieliopin löytämiseksi askelittain. Unpublished M.Sc.
thesis, Department of computer science, University of Helsinki.
_____. 1988. On the synthesis of context-free grammars. Pp. 133–141 in: Matti Mäkelä, Seppo Linnainmaa & Esko
Ukkonen (eds.), STeP-88 (Finnish artificial intelligence symposium, University of Helsinki, August 15–18, 1988) I.
Helsinki: Suomen tekoälyseura.
Parpola, Simo, Asko Parpola, & Robert H. Brunswig, Jr. 1977. The Meluḥḥa village: Evidence of acculturation of
Harappan traders in late third millennium Mesopotamia? Journal of the Economic and Social History of the Orient
20(2):129–65.
Parrot, André. 1960. Sumer. Translated by Stuart Gilbert and James Emmons. (The arts of mankind 1.) London: Thames
and Hudson.
Parzinger, Hermann. 2006. Die frühen Völker Eurasiens: Vom Neolithikum bis zum Mittelalter. Munich: C. H. Beck.
Pennington, Brian K. 2005. Was Hinduism invented? Britons, Indians, and the colonial construction of religion. New
York: Oxford University Press.
Petrie, Flinders. 1939. The making of Egypt. London: Sheldon.
Piggott, Stuart. 1950. Prehistoric India to 1000 B.C. Harmondsworth: Penguin.
_____
. 1983. The earliest wheeled transport: From the Atlantic Coast to the Caspian Sea. London: Thames and Hudson.
_____
. 1992. Wagon, chariot and carriage: Symbol and status in the history of transport. London: Thames and Hudson.
Pinault, Georges-Jean. 1989. Reflets dialectaux en védique ancien. Pp. 35–96 in: Caillat 1989.
Pinnow, Heinz-Jürgen. 1959. Versuch einer historischen Lautlehre der Kharia-Sprache. Wiesbaden: Otto Harrassowitz.
Pirart, Éric. 1995. Les Nāsatya I: Les noms des Aśvin. Traduction commentée des strophes consacrées aux Aśvin dans
le premier maṇḍala de la Ṛgvedasamhitā. (Bibliothèque de la Faculté de Philosophie et Lettres de l’Université de
Liège 261.) Geneva: Diffusion Librairie Droz.
Pittman, Holly. 1984. Art of the Bronze Age southeastern Iran, western Central Asia, and the Indus Valley. With
comments on style and iconography by Edith Porada. New York: Metropolitan Museum of Art.
Polos’mak, Nataliya V. 2001. Zur Kleidung der Pazyryk-Bevölkerung aus Ukok, Südaltaj. Pp. 101–126 in: Eichmann &
Parzinger 2001.
Porada, Edith. 1950. A leonine figure of the Protoliterate period of Mesopotamia. Journal of the American Oriental
Society 70:223–226.
Possehl, Gregory L. 1999. Indus Age: The beginnings. New Delhi: Oxford & IBH Publishing.
_____
. 2002. The Indus Civilization: A contemporary perspective. Walnut Creek, CA: Alta Mira.
Potts, Daniel T. 2001. Nana in Bactria. Silk Road Art and Archaeology 7:23–35.
_____. 2008 (2009). Puzur-Inšušinak and the Oxus Civilization (BMAC): Reflections on Šimaški and the geo-political
landscape of Iran and Central Asia in the Ur III period. Zeitschrift für Assyriologie und Vorderasiatische Archäologie
95:165–194.
Preston, James J. 1980. Cult of the goddess: Social and religious change in a Hindu temple. New Delhi: Vikas.
Pritchard, James B. 1969. The ancient Near East in pictures relating to the Old Testament. 2nd ed. with supplement.
Princeton, NJ: Princeton University Press.
Proferes, Theodore N. 2007. Vedic ideals of sovereignty and the poetics of power. (American Oriental Series 90.) New
Haven, CT: American Oriental Society.
Pryakhin, A. D., & A. Kh. Khalikov 1987. Abashevskaya kul’tura. Pp. 124–131 in: Bader, Krajnov, & Kosarev 1987.
Quivron, Gonzague. 1980. Les marques incisées sur les poteries de Mehrgarh au Baluchistan, du milieu du IVe
millénaire à la première moitie du IIIe millénaire. Paléorient 6:269–280.
Raghu Vira & Lokesh Chandra (eds.) 1995. Tibetan maṇḍalas (Vajrāvalī and Tantra-samuccaya). (Śata-Piṭaka Series
383.) New Delhi: International Academy of Indian Culture.
Rao, S R. 1982. The decipherment of the Indus script. Bombay: Asia Publishing House.
_____
. 1991. Dawn and devolution of the Indus Civilization. New Delhi: Aditya Prakashan.
Rask, Rasmus. 1818. Undersøgelse om den gamle Nordiske eller Islandske Sprogs Oprindelse. Copenhagen: Gyldendal.
Rassamakin, Yuri. 1999. The Eneolithic of the Black Sea Steppe: Dynamics of cultural and economic development,
4500–2300 BC. Pp. 59–182 in: Marsha Levine, Yuri Rassamakin, Aleksandr Kislenko & Nataliya Tatarintseva (eds.),
Late Prehistoric exploitation of the Eurasian steppe. Cambridge: McDonald Institute for Archaeological Research.
_____
. 2004. Die nordpontische Steppe in der Kupferzeit: Gräber aus der Mitte des 5. Jts. bis Ende des 4. Jts. v.Chr. I–II.
(Archäologie in Eurasien 17.) Mainz: Philipp von Zabern.
Ratnagar, Shereen. 1991. Enquiries into the political organization of Harappan society. Pune: Ravish.
Rau, Wilhelm. 1976. The meaning of pur in Vedic literature. (Abhandlungen der Marburger Gelehrten Gesellschaft,
Jahrgang 1973:1.) Munich: Wilhelm Fink.
Raulwing, Peter. 2005 (2006). The Kikkuli Text (CTH 284): Some interdisciplinary remarks on Hittite training texts for
chariot horses in the second half of the 2nd millennium B.C. Pp. 61–75 in: Armelle Gardeisen (ed.), Les équides dans
le monde méditerranéen antique. (Monographies d’archéologie méditerranéen 1.) Lattes: èditions de l’Association
pour le développement de l’archéologie en Languedoc-Rousillon.
Raverty, H. G. 1859. Notes on Kafiristan. Journal of the Asiatic Society of Bengal 28:317–368.
Rawlinson, George (trans.). [1860] 1942. Herodotus, The Persian Wars. (Modern Library 255.) New York: Random
House.
Reade, Julian. 1995. Magan and Meluhha merchants at Ur? Pp. 597–599 in: U. Finkbeiner, R. Dittmann & H.
Hauptmann (eds.), Beiträge zur Kulturgeschichte Vorderasiens: Festschrift für Rainer Michael Boehmer. Mainz:
Philipp von Zabern.
_____
. 2008. The Indus-Mesopotamia relationship reconsidered. Pp. 12–18 in: Olijdam & Spoor 2008.
Rédei, Károly. 1988. Die ältesten indogermanischen Lehnwörter der uralischen Sprachen. Pp. 638–664 in: Sinor 1988.
_____. 1988–1991. Uralisches etymologisches Wörterbuch I–III. Wiesbaden: Otto Harrassowitz.

Renger, Johannes. 2004. Die naturräumlichen Bedingungen im Alten Orient im 4. und frühen 3. Jt. v. Chr. Pp. 41–48 in:
Fansa & Burmeister 2004.
Renou, Louis. 1953. Religions of ancient India. London: Athlone.
Roberts, T. J. 1977. The mammals of Pakistan. London: Ernest Benn.
Robinson, Andrew. 2002. Lost languages: The enigma of the world’s undeciphered scripts. New York: McGraw Hill.
Rolland, Pierre. 1973. Le Mahāvrata. Nachrichten der Akademie der Wissenschaften in Göttingen, Philologisch-
historische Klasse, Jahrgang 1973:51–79.
Rönnow, Kasten. 1929. Zur Erklärung des Pravargya, des Agnicayana und der Sautrāmaṇī. Le Monde Oriental 23:69–
173.
Rosenfield, John M. 1967. The dynastic arts of the Kushans. Berkeley: University of California Press.
Rothman, Mitchell S. (ed.). 2001. Uruk Mesopotamia & its neighbors: Cross-cultural interactions in the era of state
formation. Santa Fe: School of American Research Press.
Ryzhov, S. M. 2003. Glynyani modeli sanej Kukuten’-Trypil’s’koï spil’nosti. Arkheologichnyj litopys livoberezhnoï
Ukraïny 2002:3 & 2003:1:56–57.
Sajnovics, Joannes [János]. 1770. Demonstratio idioma Ungarorum et Lapponum idem esse. Turnaviae: Typis Collegii
Academici Societatis Jesu.
Salomon, Richard G. 1996. Brahmi and Kharoshthi. Pp. 373–383 in: Daniels & Bright 1996.
Samzun, Anaick 1992. Observations on the characteristics of the Pre-Harappan remains, pottery, and artifacts at
Nausharo, Pakistan (2700–2500 BC). Pp. 245–252 in: Catherine Jarrige (ed.), South Asian Archaeology 1989.
(Monographs in World Archaeology.) Madison, WI: Prehistory.
Sarianidi, Viktor I. 1977. Drevnie zemledel’tsy Afganistana: Materialy Sovetsko-Afganskoj èkspeditsii 1969–1974 gg.
Moscow: Nauka
_____
. 1986. Die Kunst des alten Afghanistan: Architektur, Keramik, Siegel, Kunstwerke aus Stein und Metall. Übersetzt
aus dem Russischen von Sabine Grebe. Weinheim: VCH Verlagsgesellschaft.
_____. 1987. South-west Asia: Migrations, the Aryans and Zoroastrians. Information Bulletin of the International
Association for the Study of the Cultures of Central Asia 13:44–56.
_____
. 1989. Khram i nekropol’ Tillyatepe. Moscow: Nauka.
_____
. 1998a. Margiana and Protozoroastrism. Translated from Russian by Inna Sarianidi. Athens: Kapon.
_____. 1998b. Myths of ancient Bactria and Margiana on its seals and amulets. Moscow: Pentagraphic.
_____
. 2001. The Indo-Iranian problem in the light of the latest excavations in Margiana. Pp. 417–441 in: Karttunen &
Koskikallio 2001.
_____. 2002. Marguş: Murgap derýasynyn köne hanasynyn aýagyndaky gadymy gündogar şalygy / Drevnevostochnoe
tsarstvo s staroj del’te reki Murgab / Ancient Oriental kingdom in the old delta of Murghab river. Aşgabat:
Türkmendöwlethabarlary.
_____
. 2007. Necropolis of Gonur. English translation by Inna Sarianidi. Athens: Kapon.
Sarianidi, V. I., P. M. Kozhin, M. F. Kosarev, & N. A. Dubova (eds.). 2011 (2012). Trudy Margianskoj arkheologicheskoj
èkspeditsii 4. Moscow: Institut ètnologii i antropologii im. N. N. Miklukho-Maklaya, RAN.
Savarkar, V. D. 1938. Hindutva: Who is a Hindu? New Delhi: Central Hindu Yuvak Sabha.
Sayce, A. H. 1924. Remarkable discoveries in India. Illustrated London News, September 27, 1924:526.
Scherer, Anton. 1953. Gestirnnamen bei den indogermanischen Völkern. (Indogermanische Bibliothek 3.1.) Heidelberg:
Carl Winter–Universitätsverlag.
Schlerath, Bernfried. 1960. Das Königtum im Rig- und Atharvaveda. Ein Beitrag zur indogermanischen
Kulturgeschichte. (Abhandlungen für die Kunde des Morgenlandes 33:3.) Wiesbaden: Kommissionsverlag Franz
Steiner.
Schmidt, Erich F. 1933. Tepe Hissar excavations, 1931. Museum Journal 23(4):323–487.
_____
. 1937. Excavations at Tepe Hissar, Damghan. Philadelphia, PA: University Museum.
Schmitt, Rüdiger. 1967. Dichtung und Dichtersprache in indogermanischer Zeit. Wiesbaden: Otto Harrassowitz.
_____
, (ed.). 1989. Compendium Linguarum Iranicarum. Wiesbaden: Dr Ludwig Reichert.
_____
. 2000. Die iranischen Sprachen in Geschichte und Gegenwart. Wiesbaden: Dr Ludwig Reichert.
Schrader, Otto. 1907–1906. Sprachvergleichung und Urgeschichte: Linguistisch-historische Beiträge zur Erforschung
des indogermanischen Altertums I–II. Dritte neubearbeitete Auflage. Jena: Hermann Costenoble.
Schrader, O. [& A. Nehring]. 1917–1929. Reallexikon der indogermanischen Altertumskunde. Zweite, vermehrte und
umgearbeitete Auflage herausgegeben von A. Nehring. I–II. Berlin: Walter de Gruyter.
Shah, Sayid Ghulam Mustafa, & Asko Parpola. 1991. Corpus of Indus seals and inscriptions 2: Collections in Pakistan.
(Annales Academiae Scientiarum Fennicae B240.) Helsinki: Suomalainen Tiedeakatemia. (= CISI 2)
Shinde, V., G. L. Possehl, & M. Ameri. 2005. Excavations at Gilund 2001–2003: The seal impressions and other finds.
Pp. 159–169 in: Ute Franke-Vogt & Hans-Joachim Weisshaar (eds.), South Asian Archaeology 2003. (Forschungen zur
Archäologie aussereuropäischer Kulturen 1.) Aachen: Linden Soft.
Shulman, David Dean. 1980. Tamil temple myths: Sacrifice and divine marriage in the South Indian Śaiva tradition.
Princeton, NJ: Princeton University Press.
Sievertsen, Uwe. 1992. Das Messer vom Gebel el-Arak. Baghdader Mitteilungen 23:1–75.
_____. 1999. Early buttress-recess architecture in Mesopotamia and Syria. Baghdader Mitteilungen 30:7–20.

Silva, Severine. 1955. Traces of human sacrifice in Kanara. Anthropos 50:577–592.

Sims-Williams, Nicholas (ed.). 2002. Indo-Iranian languages and peoples. (Proceedings of the British Academy 116.)
Oxford: Oxford University Press.
Sinor, Denis (ed.). 1988. The Uralic languages: Description, history, and foreign influences. (Handbuch der Orientalistik
8:1.) Leiden: Brill.
Smith, Brian K. 1994. Classifying the universe: The ancient Indian varṇa system and the origins of caste. New York:
Oxford University Press.
Sontheimer, Günther-Dietz. 1989. Pastoral deities in western India. Translated by Anne Feldhaus. New York: Oxford
University Press.
Southworth, Franklin C. 2005. Linguistic archaeology of South Asia. London: RoutledgeCurzon.
Sparreboom, Marcus. 1985. Chariots in the Veda. (Memoirs of the Kern Institute 3.) Leiden: Brill.
Speyer, J. S. (ed.). 1906–1909. Avadānaçataka: A century of edifying tales belonging to the Hīnayāna I–II. (Bibliotheca
Buddhica 3.) St. Pétersbourg: Académie Impériale des Sciences.
Staal, J. F. 1961. Nambudiri Veda recitation. (Disputationes Rheno-Traiectinae 5.) ’s-Gravenhage: Mouton.
_____ (ed.) 1983. Agni: The Vedic ritual of the Fire Altar I–II. Berkeley: Asian Humanities Press.

Stacul, Giorgio. 1970. The Grey Pottery in the Swat Valley and the Indo-Iranian connections (c. 1500–300 B.C.). East
and West NS 20(1–2):92–102.
_____
. 1971. Cremation graves in northwest Pakistan and their Eurasian connections: Remarks and hypotheses. East and
West NS 21(1–2):9–19.
Stein, Diana L. 1994. Mittan(n)i. B. Bildkunst und Architektur. Pp. 296–299 in: Reallexikon der Assyriologie und
Vorderasiatischen Archäologie 8(3–4). Berlin: Walter de Gruyter.
Steinkeller, Piotr. 1982. The question of Marḥaši: A contribution to the historical geography of Iran in the third
millennium B.C. Zeitschrift für Assyriologie und vorderasiatische Archäologie 72:237–265.
Strommenger, Eva. 1962. Fünf Jahrtausende Mesopotamien: Die Kunst von den Anfangen um 5000 v.Chr. bis zu
Alexander dem Grossen. Aufnahmen von Max Hirmer. Munich: Hirmer.
Szabo, Albert, & Thomas J. Barfield. 1991. Afghanistan: An atlas of indigenous domestic architecture. Austin: University
of Texas Press.
Tallgren, A. M. 1928. [Review of articles from 1927 by S. Teploukhov, M. P. Gryaznov, P. Rykov, and S. Dubrilin.] Eurasia
Septentrionalis Antiqua 3:186–188.
Tamil lexicon. 1924–1939. I–VI and Supplement. Madras: University of Madras.
Teufer, Mike. 2012. Der Streitwagen: Eine “indo-iranische” Erfindung? Zum Problem der Verbindung von
Sprachwissenschaft und Archäologie. Archäologische Mitteilungen aus Iran und Turan 44:270–313.
_____
. 2013. Vom Grab als Totenhaus zum “Haus des Lobliedes” im Jenseits. Überlegungen zum Aufkommen des
Zoroastrismus aus archäologischer Sicht. Pp. 1–44 in: Gunvor Lindström, Svend Hansen, Alfried Wieczorek &
Michael Tellenbach (eds.), Zwischen Ost und West: Neue Forschungen zum antiken Zentralasien. Darmstadt: Philipp
von Zabern.
Thapar, B. K. 1985. Recent archaeological discoveries in India. Paris: UNESCO.
Thieme, Paul. 1957. Mitra and Aryaman. Transactions of the Connecticut Academy of Arts and Sciences 41:1–96.
_____
. 1960. The “Aryan” gods of the Mitanni treaties. Journal of the American Oriental Society 80:301–317.
Tod, James. 1835. Comparison of the Hindu and Theban Hercules, illustrated by an ancient Hindu Intaglio.
Transactions of the Royal Asiatic Society of Great Britain and Ireland 3:139–159.
Tokunaga, Muneo (ed.) 1997. The Bṛhaddevatā: Text reconstructed from the manuscripts of the shorter recension, with
introduction, explanatory notes, and indices. Kyoto: Rinsen.
Trautmann, Thomas R. 1981. Dravidian kinship. (Cambridge Studies in Social Anthropology 36.) Cambridge:
Cambridge University Press.
_____
(ed.) 2005. The Aryan debate. New Delhi: Oxford University Press.
Tucci, Giuseppe. 1961. The theory and practice of the maṇḍala. Translated by Alan Houghton Brodrick. London: Rider.
Tull, Herman. 1989. The Vedic origins of karma: Cosmos as man in ancient Indian myth and ritual. (SUNY series in
Hindu studies.) Albany: State University of New York Press.
Turner, Ralph L. 1966. A comparative dictionary of the Indo-Aryan languages. London: Oxford University Press.
Uesugi, Akinori. 2011. Development of the inter-regional interaction system in the Indus Valley and beyond: A
hypothetical view towards the formation of the urban society. Pp. 359–380 in: Osada & Witzel 2011.
Usachuk, A. N. (ed.) 2004. Psalii: Èlementy upryazhi i konskogo snaryazheniya v drevnosti. (Arkheologicheskij
al’manakh 15.) Donetsk: Institut arkheologii NAN Ukrainy.
_____
. 2013. Drevnejshie psalii (izgotovlenie i ispol’zovanie). Kiev and Donetsk: Institut arkheologii, Natsional’naya
Akademiya Nauk Ukrainy.
Vahia, Mayank N., & Srikumar M. Menon. 2011. A possible astronomical observatory at Dholavira.
http://www.tifr.res.in/~archaeo/papers/Harappan%20Civilisation/Observatory%20at%20Dholavira.pdf.
van Buitenen, J. A. B. 1968. The Pravargya: An ancient Indian iconic ritual described and annotated. (Deccan College
Building Centenary and Silver Jubilee Series 58.) Poona: Deccan College.
_____ (trans. and ed.). 1973–1978. The Mahābhārata I–III. Chicago: University of Chicago Press.
Van Buren, Elizabeth Douglas. 1948. Fish-offerings in ancient Mesopotamia. Iraq 10: 101–121.
van Kooij, K. R. 1972. Worship of the Goddess according to the Kālikāpurāṇa I. (Orientalia Rheno-Traiectina 14.)
Leiden: Brill.
Vasil’ev, I. B., P. F. Kuznetsov, & A. P. Semenova. 1994. Potapovskij kurgannyj mogil’nik indoranskikh plemen na Volge.
Samara: Samarskij universitet.
Vermaak, P. S. 2008. Guabba, the Meluhhan village in Mesopotamia. Journal for Semitics 17(2):454–471.
Vermaseren, Maarten J. 1977. Cybele and Attis: The myth and the cult. London: Thames and Hudson.
Vidale, Massimo. 2007. The collapse melts down: A reply to Farmer, Sproat & Witzel [2004]. East and West 57(1–
4):333–366.
_____
. 2010. Aspects of palace life at Mohenjo-Daro. South Asian Studies 26(1):59–76.
Vidale, Massimo, Roberto Micheli, Luca Maria Olivieri, & Muhammad Zahir. 2014. New radiocarbon dates and the
chronology of protohistorical graves of Swat Valley: The latest evidence from Godgara IV and Udegram (Khyber
Pakhtunkhwa, Pakistan). Paper read at the 22nd conference of the European Association for South Asian Archaeology
and Art in Stockholm, Sweden, June 30 to July 4, 2014.
Videiko, Mihailo Y. 1994. Tripolye—“pastoral” contacts. Facts and character of the interactions: 4800–3200 BC. Baltic-
Pontic Studies 2:5–28.
_____. 1995 (1996). Grossiedlungen der Tripol’e-Kultur in der Ukraine. Eurasia Antiqua 1:45–80.

Vinogradov, N. B. 2003. Mogil’nik bronzovogo veka Krivoe Ozera v yuzhnom Zaural’e. Chelyabinsk: Yuzhno-Ural’skoe
knizhnoe izdatel’stvo.
von Dassow, Eva. 2008. State and society in the Late Bronze Age Alalah under the Mittani empire. (SCCNH = Studies
on the civilization and culture of Nuzi and the Hurrians 17.) Bethesda, MD: CDL Press.
von Hinüber, Oskar. 2001. Das ältere Mittelindisch im Überblick. Zweite Auflage. (Österreichische Akademie der
Wissenschaften, Philosophisch-historische Klasse, Sitzungsberiche 467; Veröffentlichungen der Kommission für
Sprachen und Kulturen Südasiens, 20.) Vienna: Österreichische Akademie der Wissenschaften.
von Stietencron, Heinrich. 1983. Die Göttin Durgā Mahiṣāsuramardinī: Mythos, Darstellung und geschichtliche Rolle
bei der Hinduisierung Indiens. Visible Religion 2:118–66.
_____
. 1989. Hinduism: On the proper use of a deceptive term. Pp. 11–27 in: Günther D. Sontheimer & Hermann Kulke
(eds.), Hinduism reconsidered. (South Asian Studies 24.) New Delhi: Manohar.
Wace, Alan J. B., & Frank H. Stubbings (eds.). 1962. A companion to Homer. London: Macmillan.
Wackernagel, Jakob. 1896. Altindische Grammatik I. Göttingen: Vandenhoeck & Ruprecht.
Wackernagel, Jacob [& Albert Debrunner]. 1954. Altindische Grammatik, II,2: Die Nominalsuffixe, von Albert
Debrunner. Göttingen: Vandenhoeck & Ruprecht.
Waddell, L. A. 1925. The Indo-Sumerian seals deciphered: Discovering Sumerians of the Indus Valley as Phoenicians,
Barats, Goths, and famous Vedic Aryans, 3100 to 2300 B.C. London: Luzac.
Wanzke, Holger. 1987. Axis systems and orientation at Mohenjo-Daro. Pp. 33–44 in: Michael Jansen & Günter Urban
(eds.), Interim reports 2: Reports on field work carried out at Mohenjo-Daro, Pakistan, 1983–84 by the IsMEO—
Aachen-University Mission. Aachen: RWTH, and Rome: IsMEO.
Ward, William (trans.). 1811. Account of the writings, religion, and manners, of the Hindoos, including translations
from their principal works I–IV. Serampore: Mission.
Watkins, Calvert. 1995. How to kill a dragon: Aspects of Indo-European poetics. New York: Oxford University Press.
Weber, Albrecht. 1849. Ueber die Literatur des Sāmaveda. Indische Studien 1(1):25–67.
_____. 1850. Zwei Sagen aus dem Çatapatha-Brāhmaṇa über Einwanderung und Verbreitung der Ārier in Indien, nebst
einer geographisch-geschichtlichen Skizze aus dem weissen Yajus. Indische Studien 1(2):161–232.
_____. 1853. Zur Geschichte der indischen Astrologie. Indische Studien 2:236–287.
_____
. 1861 (1862). Die vedischen Nachrichten von den naxatra (Mondstationen) II. (Abhandlungen der Königlichen
Akademie der Wissenschaften zu Berlin, aus dem Jahre 1861:267–400.) Berlin: Druckerei der Königlichen Akademie
der Wissenschaften.
_____. 1870 (1871). Über das Rāmāyaṇa. Abhandlungen der Königlichen Akademie der Wissenschaften zu Berlin, Aus
dem Jahre 1870, Philosophisch-historische Klasse: 1–88.
Wessels-Mevissen, Corinna. 2001. The gods of the directions in ancient India: Origin and early development in art and
literature (until c. 1000 A.D.). (Monographien zur indischen Archäologie, Kunst und Philologie, Band 14.) Berlin:
Dietrich Reimer.
West, M. L. 2007. Indo-European poetry and myth. Oxford: Oxford University Press.
Wheeler, Mortimer. 1947. Harappa 1946: The defences and cemetery R37. Ancient India 3:58–130, and pl. XV–LX.
_____. 1968. The Indus civilization. 3rd ed. Cambridge: Cambridge University Press.

Whitehead, Henry. 1921. The village gods of South India. 2nd ed., revised and enlarged. (The religious life of India.)
Calcutta: Association Press (YMCA); London: Oxford University Press.
Wilhelm, Gernot. 1994. Mittan(n)i, Mitanni, Maitani. A. Historisch. Pp. 286–296 in: Reallexikon der Assyriologie und
Vorderasiatischen Archäologie 8(3–4). Berlin: Walter de Gruyter.
Winkelmann, Sylvia. 2004. Seals of the oasis from the Ligabue collection. With an introduction by Pierre Amiet and a
postscriptum by Gabriele Rossi-Osmida. Padova: Il Punto Edizioni.
Winternitz, Moriz. 1892. Das altindische Hochzeitsrituell nach dem Āpastamba-Gṛhyasūtra und einigen anderen
verwandten Werken. (Denkschriften der Kaiserlichen Akademie der Wissenschaften in Wien, Philos.-hist. Classe
40:1.) Vienna: In Commission bei F. Tempsky.
_____. 1895. Nejamesha, Naigamesha, Nemeso. Journal of the Royal Asiatic Society of Great Britain and Ireland
1895:149–155.
_____. 1927. A history of Indian literature I. Translated by S. Ketkar and revised by the author. Calcutta: University of
Calcutta.
Witzel, Michael. 1989. Tracing the Vedic dialects. Pp. 97–265 in: Caillat 1989.
_____
. 1995a. Early Indian history: Linguistic and textual parameters. Pp. 85–125 in: Erdosy 1995.
_____
. 1995b. Rgvedic history: Poets, chieftains and polities. Pp. 307–352 in: Erdosy 1995.
_____. 1997. The development of the Vedic canon and its schools: The social and political milieu. (Materials on Vedic
Śākhās 8.) Pp. 257–345 in: Michael Witzel (ed.), Inside the texts, beyond the texts: New approaches to the study of
the Vedas. (Harvard Oriental Series, Opera Minora 2.) Cambridge, MA: Department of Sanskrit and Indian Studies,
Harvard University.
_____. 2000. The home of the Aryans. Pp. 283–338 in: Almut Hintze & Eva Tichy (eds.), Anusantatyai: Festschrift für
Johanna Narten zum 70. Geburtstag. (Münchener Studien zur Sprachwissenschaft, Beihefte NF 19.) Dettelbach: J. H.
Röll.
Witzel, Michael. 2005. Central Asian roots and acculturation in South Asia: Linguistic and archaeological evidence from
western Central Asia, the Hindukush and northwestern South Asia for early Indo-Aryan language and religion. Pp.
87–211 in: Toshiki Osada (ed.), Linguistics, Archaeology and the Human Past: Occasional Paper 1. Kyoto: Indus
Project, Research Institute for Humanity and Nature.
Young, T. C., Jr. 1985. Early Iron Age Iran revisited: Preliminary suggestions for the reanalysis of old constructs. Pp.
361–377 in: J.-L. Huot, M. Yon & Y. Calvet (eds.), De l’Indus aux Balkans: Recueil à la mémoire de Jean Deshayes.
Paris: Éditions Recherche sur les Civilisations.
Young, Thomas. 1813. Adelung’s General History of Languages. Quarterly Review (London) 10(19):250–292.
Zdanovich, Gennady B., & Dmitry G. Zdanovich. 2002. The “Country of Towns” of southern Trans-Urals and some
aspects of steppe assimilation in the Bronze Age. Pp. 249–263 in: Boyle, Renfrew, & Levine 2002.
Zeller, Gabriele. 1990. Die vedischen Zwillingsgötter: Untersuchungen zur Genese ihres Kultes. (Freiburger Beiträge
zur Indologie 24.) Wiesbaden: Otto Harrassowitz.
Zide, Norman H. 1969. Munda and non-Munda Austroasiatic languages. Pp. 411–430 in: Thomas A. Sebeok (ed.),
Linguistics in South Asia. (Current trends in linguistics 5.) The Hague and Paris: Mouton.
Zvelebil, Kamil V. 1970. Comparative Dravidian phonology. (Janua Linguarum, Series practica 80.) The Hague: Mouton.
_____
. 1973. The smile of Murugan: On Tamil literature of South India. Leiden: Brill.
_____. 1974. Tamil literature. (A History of Indian Literature 10:1.) Wiesbaden: Otto Harrassowitz.
_____. 1990. Dravidian linguistics: An introduction. (Pondicherry Institute of Linguistics and Culture publication 3.)
Pondicherry: Pondicherry Institute of Linguistics and Culture.
Index

For reasons of space modern authors and Vedic texts (listed in References and on pages xv–
xvi) have been excluded from this index. Attributes defining phases (Proto-, Old, Modern) or
dialectal groups (e.g., East, North) of languages and language families, if included in the
index, do not constitute separate lemmas but appear under the respective language and
language group (e.g., Indo-European).

Abashevo culture xii, 54–57, 59, 66–67, 297–298

abhicāra-, ābhicārika- 132
abhigara- 138
abhiṣeka-, punar-abhiṣeka 128
Abydos 222
Achaemenid 3, (6), (29), 55, 69, 97, 112, 149, 217, 257–258. See also Darius the Great, Xerxes
adhvaryu- 88, 124, 127, 133, 141, 193, 252, 302
Aditi 288
Āditya 108, 112, 301
Afanasyevo culture 47–48, 69
Agastya 153
aghnyā- 178
agni-, Agni x, 89, 92, 94, 96, 107, 113–115, 133, 143, 187–188, 196, 198, 206, 231, 253, 284–285
agniciti- (fire altar); agnicayana- x, 136, 139–140, 142–143, 171, 187–188, 197, 204, 206–208, 249, 303
agnihotra- 113, 124, 134, 232
āhanasya- 135, 137
Ahura 100;
Ahura Mazdā 98, 100, 103, 112, 114–116
Aihole 243
air root 177, 201–204, 206, 278–281, 284, 310–311
airyān m vaējō 98
aj- 40
aja-, ajā- 245, 249
Ajitā 252
Akanānūru 150
Akkad, Akkadian 164, 182, 211–216, 218, 220, 226–227, 229–230, 269, 307–308, 313
Akkadian language 85–87, 89, 163, 215, 229, 231, 238–239; gubāru/gupāru229
akṣara- 170
Alakar 157
Alakul' Andronovo culture 59, 62
Alalaḥ, Alalakh 85–88
Alanian language 102
Alašia (Cyprus) 85–86
Albania 16, 49
Albanian language 13–14, 16
Alcor 201, 280–281
Aldebaran 195, 198–199, 278, 310
Alexander the Great 29, 55, 94, 97–98, 153
Allahdino 20–21
alphabet, alphabetic script 14, 29, 172, 266, 274
Altai Mountains 59, 64, 68, 103–104, 297, 299
Altyn Depe 69–70, 211
ām 170
Amarna (Tell el Amarna) 85–86
Amenophis III (Nimmureya) 86
Am ša Sp ṇta 112
Ammianus Marcellinus 93
Amorite (West Semitic) language and people 213
Amri 17–18, 184–185, 200, 211, 272, 306
Aṁśa 112
Amu Darya (Oxus) River 69, 93
Amurru 216
An (Anu) 226, 238, 272
Ananta (Śeṣa) 158
Anarta 18
anas- 170
Anat 236
Anatolia 8, 14–15, 37–38, 49, 55, 76, 82–83, 85, 236–237, 295, 307
Anatolian languages 14–16, 38, 46, 48, 295
Andhaka 146
Andhra 259
Andronovo cultures 54, 59, 62, 68, 71, 79, 81, 109,120, 297–298, 300–301
Aṅga 150
aṅgāra-, aṅgāraka- 283, 315
Aṅgiras, Āṅgirasa 92, 115, 128–133, 141
aṇil, aṇil-piḷḷai 291
añj- 174
antennae-hilted sword 80
Anu tribe 93
Anurādhapura 150–151
anuṣṭubh- 245
apagara- 138
Apāṃ Napāt 114
apāṃ pati- 114
apāna- 127, 209
Aparājitā 252
apratiratha- 111
apsaras- 171, 180, 204, 317
Arabian Gulf. See Gulf
Arabic language 172
Arachosia 93, 97–98, 101, 103
Aramaic script 29
araṇya- 248–249
Aratta, Āraṭṭa 149, 211, 216, 307
Arbuda 94
Arcadian dialect of Greek 13
archaeology and linguistics viii–ix, xii, 35, 39
Areia 101
argha- 66
Arghandab River 93, 97
arghya- 174
Argolid 49
arhat- 134, 139
Aria, Ariana 93, 103
Ārjīkīyā 93
arjuna-, Arjuna 96, 111–112, 146, 154–155, 303
Arkaim 58, 110
arktos 197, 200
Armenian language 11–16, 40, 49, 51, 296
Arrian 97, 153–155, 159
Arsaces, Arsakēs 55, 101
Arsacids 55, 101
Arśasāna 95
Arthaśāstra 153
aru-mī n 274
arya-, ārya- 6, 100
Aryaman 112
Aryan languages. See Indo-Iranian languages
“Aryan problem / debate” vii–viii, 8–9
Arzawa 85–86
āsana- 209
āsandī- 115, 136, 138, 141
Ashkun Kāfirs 263–264
Asiknī River 93
asiktaretas- 141
aśoka tree 179
Asoka, his inscriptions 27, 145
Assam 265
Assur 82, 112
Assurbanipal 100
Assyria, Assyrian 49, 54, 82–86, 86, 100, 110, 112, 220, 298, 301
Aštar, Aštart 248
asura-, Asura 66, 82, 90, 100, 108–109, 114, 150, 241–242, 244, 252–254, 257, 309. See also Ahura, demon,
mahiṣa-.
aśva-, aśvā- 87–88, 90, 245, 249
aśvamedha- (horse sacrifice) 36, 76, 122, 139, 141, 239, 244–246, 248–249, 253–254
Aśvapati 155
aśvattha- 127
aśvayujau (aśvinī) 113, 198
Aśvin 88–89, 102, 107–115, 117–124, 126–129, 133–134, 136–137, 145, 154–156, 159, 300–304, 309, 317. See also
nāsatya-
āśvina-śastra- 113, 118–119
aśvinī. See aśvayujau
ātar- 114
Atharvan 115, 132
atirātra- 118
atithi- 90, 136
Atithigva 94, 97, 101
ātman- 155, 245
Atri, Atri clan 92, 118, 120–121, 124, 126–127, 302
Attis 237
Aupapātika-Sūtra 179
Austro-Asiatic languages 165, 259
Avadānaśataka 204
Avesta, Avestan language 6, 11, 13–15, 39–40, 54, 66, 85, 88, 91, 93, 97, 100, 103, 105–107, 114–116, 129, 131,
143, 147, 216, 259–261
avi- 249
awl 48, 53, 63
Ayodhyā 150
Ayyappan 248
Azerbaijan 83

Ba'al 236
babhru- 87
Babino III culture (KMK culture) 54
Babylon, Babylonia, Babylonian 85–86, 89, 112, 208, 211–212, 217–218, 238, 256, 315
Babylonian language, Old 172
Bactria 69, 71, 77, 80–83, 97–98, 103, 211, 216, 255–256, 258, 307
Bactria and Margiana Archaeological Complex. See BMAC
Bactrian language (Middle Iranian) 255, 264
Badakhshan 216, 221
Baden-Pécel culture 42, 124
baga- 106
Bāhīka, Bāhlīka 216. See also Balhika, Vāhīka
bahiṣ 216
Bahrain (Bahrein) 211, 213–214
Bākhdī, Bākhtrī 216
Bala-kot 211
Balanovo culture 47–48, 55–56, 297
Bala-Rāma (Bala-Deva) 145, 153–159, 303–304, 317
Balbūtha Tarukṣa 82, 105–106
Balhika, Balhīka, Bālhīka 216. See also Bāhīka, Bāhlīka
bali- 152, 174, 285
Balkano-Carpathian (Carpatho-Balkan) metallurgical domain, 36–37, 46
Balkans 14, 37–38, 46, 49, 124, 295
Balkh 69, 216
Baltic languages 12–16, 41, 66, 109, 301
Baltics 109, 113, 118, 301
Balto-Slavic languages 13, 16, 47, 55, 296–297
Baluchi language 55, 163, 166
Bampur 79, 211
banana 165
Banawali 20–21, 241
bangle (bracelet) 284–286, 288
banyan tree 173, 177, 180, 201–204, 206–207, 235, 278, 280–281, 283–285, 306, 310–311
Barsaentēs 97
Bashkiria 68
basta- 245, 249
Bastar 186
bata- 143
Battle Axe (Corded Ware) cultures 47–49, 55–56, 295–297
BAULT xii
Bāveru, Bāveru-Jātaka 217
bekuri-, bhekuri- 171, 199, 204
Belaya River 55–56
Belo-Russia 222
Belo-Russian language 16
Bengal 141, 159, 167, 174, 182, 306
Bengali language 6–7
Beshkent 81
bevel-rimmed bowls 19, 210
Bhadra-Kāḷī 313
Bhaga 112
Bhagavad-Gītā 207, 311
Bhāgavata-Purāṇa 157
Bhāgavata religion 145, 159
Bhairava 150, 251
bhakti- 243
Bhalāna 93
Bharadvāja 92
Bharata 93–94, 106
bhargas- 115
Bhava 136, 140
bheṣaja- 114, 132
bhrātṛvya- 133, 136
Bhṛgu 115
bhūta- 246
Big Dipper. See Ursa Major
Bihar 166–167
Bindu- 277–278
Bīr-kōṭ-ghwaṇḍai 80
Birobā 150
Black Land 217–219
Black Sea 13, 35, 51, 222. See also Pontic/Pontic-Caspian steppes
Black-and-Red Ware (BRW) 145, 149, 151
BMAC viii, xii, 23, 69–84, 97–98, 103–105, 108–109, 112, 120–122, 150, 164, 210, 220, 255–260, 298–302, 309, 316
BMAC language 81–82, 197, 263
bodhi tree 180
Bodhisattva 180
Boğazköy 87, 89, 237
Bolan Pass and River 17, 79, 93, 103
brahmabandhu- 135–136, 139, 142
brahmacārin-, brahmacaryā- 138, 248, 252
brahma-loka- 205
brahman- 111, 200; Brahman priest 170; god (Brahmā) god. 142, 154, 156–157, 193–194, 203, 242, 311; priesthood
111, 116, 135–137
Brahmaṇaspati. See Bṛhaspati
Brahmāṇḍa-Purāṇa 158, 204
Brahmanism 4
brahmarandhra- 204–205, 208, 311
Brāhmī script 27, 29, 151
Brahui language 30, 163, 165–167
Bṛbu 81–82, 105
bṛbūka- 81
bṛhad- 128, 136–137
Bṛhaddevatā 127
Bṛhaspati (Brahmaṇaspati) 107, 111, 114–115, 136, 141, 188, 193, 198, 204, 250, 303, 310
bṛhaspatisava- 193
Bṛsaya 81, 97–98, 105
bṛsī- 81
BRW. See Black-and-Red Ware
Budakalász 42
Buddha 111, 143, 151–153, 204, 264
Budha 188, 204
Bug River 46
Bukhtarma 65
Burushaski 82, 163, 165
Burzahom 18, 178–179
Bustān 71, 81

caitya- (cēïya-) 151, 178, 180

cakra- 39, 63, 204, 297
cakra-pūjā- 251–252, 264
Cakra-Saṃvara 262
cakravartin- 111
Cappadocia 82
Caṇḍikā 254
Carcā, Carcikā 256
Carchemish 237
Carpatho-Balkan (Balkano-Carpathian) metallurgical province 36–37, 46
caṣāla- 134
Caspian Gates 54
Castor, Castores 110. See also Kastor
Catacomb Grave culture 49, 53–55, 59
Çatal Höyük 236–237
caturmukha- 194
Caucasian languages 83. See also Kartvelian languages
Caucasus, Caucasian 6, 16, 46, 53–54, 102, 149, 296, 300
cēl 315
Celtic languages 7, 11, 15, 36, 47, 66; Continental 15; Gaulish 36; Insular 15
Cemetery H of Harappa, Cemetery H culture 22–23, 80, 119, 185
cem-mī n 315
centum 13; centum languages 13
Cernavoda culture 37
Cēy, Ceyyavan, Ceyyōn 254, 312, 315
Chagar Bazar 80
Chanhu-daro 20–21, 79, 211, 213, 239, 266, 308
Chapli 37
Chelyabinsk 65
Chenab River 93
Cheremiss (Mari) 59, 86
Cherkaskul' culture 59, 62, 68, 301
Chinese script 30, 283
Chinese Turkestan. See Xinjiang
Chitral 80
Cholistan 20, 23–24
Chorasmia (Khwarezm) 103, 256
Chuvash Republic 55
Cicero 154, 303
Cilappatikāram 158, 201, 244
Cilicia 86
Citra 90
Cola, Cōla 152, 157
Confucius 190
Copper Hoards, Gangetic 23–24, 80, 150, 261
copper tablets of Mohenjo-daro 282–283
Corded Ware (Battle Axe) cultures 47–49, 55–56, 295–297
Corinthia 49
Cornish language 15
Crete 240
Cucuteni-Tripolye culture. See Tripolye culture
Cumuri 96
Curtius Rufus, Q. 98, 101
Cyavana, Cyavāna 118, 119
Cyprus (Alašia) 85–86
Cyrus the Great 54
Czech language 16

Dāai 101
Dabhīti 96
Dadhyañc Ātharvaṇa 122
daha-, Daha, Dahā, Dahae 100–103, 106, 299
Daimabad 22–23, 166, 168, 175–176, 308
Dakṣa 112
dakṣiṇāyana- 249
Damb Sadaat 17–18
dānu- ‘river’ 54
Dānu 95
Danube River and Valley 47, 49, 51, 54, 124, 296
Dāoi 101
Dardic languages 82, 163, 265
Dari language 163
Darius the Great 3, 6, 102, 149, 214
dasa-, Dasa 102
dāsa-, Dāsa xii, 81, 92, 94–98, 100–103, 105–106, 257, 259–263, 265, 314
dāśa- 261
daśahrā 249
daśamī- 249
Dasaratha-Jātaka 157
daśarātra- 250
dāsera-, dāśera-, dāseya-, dāśeya- 261
Dashly 70, 80, 97, 120, 257–260, 262, 316
Dasht-e Nawur 255
dāsī- 138
dasra-, Dasra 102, 111–112
dasyu-, Dasyu xii, 81, 92, 94–96, 101–102, 105–106, 259
Daulatabad 70
Deccan 18, 23, 149–150, 166
dehī- 264
Dēr 216
Dereivka culture 47
deva- 97, 106, 108, 136, 140, 150, 154, 156, 252–253, 300
Devī-Bhāgavata-Purāṇa 243
Devī-Māhātmya 242–243, 247
Devī-Purāṇa 243
Devlimadi 185
Dhammapada Atthakathā 180
dharma- 3, 112, 132, 312
dharma-rāja- 144
Dhenuka 304, 317
Dhiṣaṇā 264
dhiṣṇya- 197, 231, 310
Dholavira 20–22, 182, 195, 211, 306
Dhṛtarāṣṭra 146, 148
Dhṛti 156
dhruva- 200, 281, 311
Dhuni 96
dhuvana- 142
diacritic 269, 276–277, 315
Dieva dēli, Dievo sūneliai 109
digvijaya- 193
Dilmun 210–211, 213–215, 217, 219
Dionysos 159
Dioskouroi 49, 109–113, 117–118, 120, 301
Dīpavaṁsa 151
Disani (Dizani) 264–265
divo napātā 109–110, 154
Divodāsa 94–95, 87–98, 101, 299
dizā 264
Djokha (Umma) 211–212, 270
Dnieper-Donets culture 36
Dnieper River 36, 52–54, 296
Dniester River 54
Don River 36, 46, 54
Donets River 36, 54
Dorian dialects of Greek 49, 115
Dorian Greeks 49, 111, 301
Draco 200–201
Drangiana 97, 103
Draupadī 148
draviḍa- 165
Dravidian language(s) vii, x–xi, 30, 82, 132, 163, 165–172, 175, 261, 274, 286, 300, 305, 311; Central Dravidian
166–167, 186, 278, 291; North Dravidian 166–167, 170, 207; Proto-Dravidian 149, 169–172, 174, 199,
202–204, 206–207, 217–218, 229, 244, 256, 263, 267, 271, 276, 279–281, 286, 289, 291, 305, 310, 312–315,
317; South Dravidian 166–167, 217, 244, 278, 283–284, 315
Dṛbhīka 81
Druhyu 93
Drupada 148
Drutsy 38
Dumuzi (Tammuz) 239
dundubhi-, Dundubhi 137, 243
durga-, Durga 246, 257; durga-ghna-257
Durgā 158, 241–244, 246, 248–251, 253, 255–257, 264–265, 309, 313, 317
Durgā-pūjā 248–251
Dussehra 249–250
Dvārakā, Dvāravatī 153
Dyaus 108
dyumat- 66, 300
Dzarkutan 70

Eanna 224; Eannatum 227–228

Early Dynastic period 210, 212, 214, 224–228, 230, 232, 234, 307
Early Harappan culture 17–20, 69, 190–194, 305–306, 317
Early West Iranian Gray Ware 83
Easter Island 30
Ecbatana 54
Egypt vii, 8, 27, 55, 82–86, 90, 112, 187–188, 220–223, 230, 307
Egyptian hieroglyphic script 25, 27, 29–30, 221–222, 267–268, 305
Egyptian language 85, 172, 221, 298
eka- 87, 91
Elam, Elamite 100, 211–212, 215–216, 218–219, 221. See also Proto-Elamite culture, Trans-Elamite
Elamite language 218
El-Amarna 85–86
Elburz Mountains 83
Elephanta 243
Ellora 243
e lu mī n 275
Elunino 64
eṇī- 245
Enki (Ea) 181–182, 197, 217, 219, 229–230, 267, 273
Enkidu 226
Enmerkar 216
Eratosthenes 101
ē ru 263
erumai-ma ram 244
Eshnunna (Tell Asmar) 182, 211–212
Estonian language 59
Ethiopia 219
Euphrates River 49, 221
Eve 267
Ezero culture 37

Failaka Island 211, 213–214

Fara 225–226
Farukabad 70
Fatehgarh 80
Fatyanovo culture 47–48, 55–56, 297. See also Balanovo culture
Fëdorovo Andronovo culture 59, 62, 68, 71, 81, 109, 298, 300–301
Finland, Finnish ix–xii, 47, 57, 59, 297
Finnic languages and folk epics 59, 66, 122, 300
Finnish language 41, 59, 66, 100
Finno-Ugric languages. See Uralic languages

Gadaba language 167

gājan 182
gaṇa- 140–142
gaṇapati-, Gaṇapati, Gaṇeśa 140–141, 291
Gandhāra 103, 120–121, 149, 211, 214, 216, 255, 261–262
Gandhāra Grave culture xii, 80–81, 120–121, 124–127, 301–302
Gāndhārī Prakrit 261
Gandharva- 141, 171, 180, 204, 252, 317
Ganges River (Gaṅgā-) and Valley 18, 20, 23–24, 80, 93, 150, 152, 261, 285, 310
Ganges-Yamuna doab 23–24, 146, 255
gardabha- 169–170
gārhapatya- 124
Garino-Bor culture 59, 297
gāthā- 129, 141, 170
gaura-, gaurī- 171, 245, 249, 304
gavaya-, gavayī- 245, 249, 250
gāyatra-, gāyatrī- 128–129, 156, 302
Gebel el-Arak knife 220, 222–223
Geoksyur 69, 79
Germanic languages 7, 11–16, 39–41, 47, 66; East 15–16; North 15, 41; Proto-Germanic 12, 39–40; West 16
Ghaggar river 22, 93, 97
Ghālegay 80–81
gharma- 88, 113, 120–121, 124, 126–128, 136–137, 302. See also pravargya-
ghāṭ 169, 175, 180
ghora- 132, 137
Gilgamesh 220, 226
Gilund 79
Girikṣit 94
Girsu (Tello, Telloh) 211–213, 218, 228
Gish 264
Gomal River 93
Gomatī 93
Gond people (Maria Gonds) 186
Gondi language 167
Gonur xii, 69–71, 76–77, 80, 122, 211
Göpekli Tepe 307
Gorakṣaśataka 204
Gorgan 83, 211, 298
Gorgan Gray Ware 83
Gorodsk culture 47
Gothic language 7, 11, 13–16, 40, 117, 126
gotra- 168
grabh-, grah-, graha- 281–283, 285
gṛda- 81
Greater Indus Valley 21, 69, 163, 167, 213–215, 236, 307
Greece, Greek 3, 15, 37, 49, 51, 69, 109, 111, 113, 155, 159, 236–237, 255, 300–301
Greek language 3, 7, 13, 36, 39–40, 41, 49, 51–52, 66, 69, 100, 115, 172, 197, 231, 255, 264
Greek (alphabetic) script 29–30. See also Linear B
gṛhapati- 135
Gṛtsamada 92
Guabba 218
Gudea 213, 218
Gujarat 18, 20–24, 79, 145, 149, 151–153, 163, 166, 168–169, 175, 182, 185–186, 195, 200, 211, 213, 217, 299–300,
305–306
Gujarati language 151, 163, 166
Gulf (Arabian/Persian Gulf) 21, 76, 163, 211–214, 219, 221, 307
Gumla 178
Gupin 215
Gurjara tribe 166
guru- ix–x, 173
Gutium 216

Habuba Kabira 221

Haētumant 97
Hakra culture 17
Hakra River 22, 93
hala- 157; Halabhṛt 156; Halāyudha 158
Hamadan 54
Hamvarti 260
Hanti (Khanty, Ostyak) language 59
haoma 103, 107, 128, 143, 302
Harahuvati 93, 97
Haraiva 93
Harappa vii, 17–18, 20–21, 23–24, 27, 29, 31, 79, 96, 123, 173, 182–186, 192, 203, 207, 211, 214, 220, 240, 271,
274–276, 282, 284, 286–287, 289–290, 306
Harappan culture 17–24, 173–291, 305–317. See also Early Harappan, Late Harappan, Mature Harappan = Indus
civilization, Post-Harappan, Pre-Harappan culture
Haraxvaitī 93, 97
Hari 277, 317
Hari-Naigameṣin 285
Hariścandra 199–200, 306, 317
Haroiva 93
Haryana 17–18
Hāstinapura 146
haṭha-yoga- 204–205, 311
Haṭhayogapradīpikā 204
Hatti. See Hittite empire
Hattuša 87
haumavarga- 103
Hebrew language 172, 267
Helen, Helenē 109, 155
Helmand River 93, 97
Hemacandra 172, 217
Heracles 153–156, 303
Herat 93
Herodotus ix, 20, 29, 53–54, 101, 103, 111, 148–149, 164, 263, 300
heruka-, Heruka 263
Hili 211–212
Himachal Pradesh 256
Himalaya(s) 152, 165, 265
Hindi language 6, 163, 170, 175, 217, 249, 263, 265, 277
Hindu Kush mountains 21, 71, 94, 97, 102, 163, 231
hiṅgulāja-, Hinglāj 256
hippodamos, hippota 110, 117
Hittite empire (Hatti, 85–86) and people 83, 85–87, 89, 112, 237, 300; (Luvian) script 30
Hittite language 14, 38, 40, 87, 172, 236
holī 176–177, 251
Homer ix, 14–15, 39, 110, 117–118, 197, 301
Hooghly River 182
Horodistea-Foltesti culture 47
Hsiungnu 255
Hsüan-Tsang (Xuan-Zang) 152
Hungary, Hungarian 30, 42, 59, 68, 124, 260
Hungarian language 11, 59, 68
Hurrian language and people 83, 85–86, 88, 298
Hyksos 90

Ibbi-Sīn 216
Ibnišarrum 230
Iceland viii
Icelandic language 11, 15
iḍā- 133
IE. See Indo-European
ignis 107, 284
Ilībiśa 82
Illyrian language 14
ilma, Ilmari/Inmar, and Indra 66, 109, 300
Imra (< *Yama-rāja) 143, 264
Inanna 216, 218, 224, 238–239, 256, 265, 309, 312–313. See also Ishtar
Indo-Aryan languages 6–8, 53, 55, 92–93; Middle IA 6, 94. See Pali, Prakrit; Modern/New/Neo-IA 6, 265; Old IA. See
Sanskrit; Proto-IA viii, 13–14, 55, 63, 66–68, 76, 79, 81–91, 97, 102, 105, 108–109, 112, 114, 121
Indo-European (IE) languages and linguistics ix, xii, 8, 10–16, 35, 39–41, 49–50; Proto-Indo-European (PIE) language
11, 35–50, 102, 105, 127, 302, Early 11, 35–38, 51, Late 11, 38–40, 43–47, 49, 51; Early Post-PIE: Northwest
IE 47, Southeast IE 47
Indo-Iranian borderlands 17–18, 97, 210, 299, 307–308
Indo-Iranian (Aryan) languages viii, xii, 6–7, 11, 13–15, 35, 47–48, 52–53, 55, 71, 76, 81–82, 89, 102, 117, 163–165,
173, 305, 314; Proto-Indo-Iranian 7, 13–14, 36, 55, 63, 93, 105, 127, 296–297, 302. See also Indo-Aryan
languages; Iranian languages; Nuristani languages
Indra 22, 66, 89–90, 92, 94–97, 102–103, 107–109, 111–112, 114–116, 121–122, 124, 128, 136, 155, 203, 206,
252–254, 259–260, 264, 300, 302–304, 309
Indrota 90
Indus civilization (Mature Harappan) 8, 17–31, 161–291, 305–317
Indus River and Indus Valley vii, 3, 17–18, 20, 22–23, 55, 79, 93, estuary 267
Indus script x–xi, 19, 25–31, 70, 266–291, 314–315
Inmar. See ilma
Iran xii, 6, 15, 25, 47, 53–55, 71, 76, 80, 82–83, 107, 114, 122, 149, 210, 212–215, 220, 225, 234, 298, 300, 306–307
Iranian (Irano-Aryan) languages 6–7, 53–55, 63, 67, 76; East Iranian 54, 98, 101–102, 106; Middle Iranian 55, 255;
Modern Iranian 6, 55; Old Iranian 6, 15, 36, 54–55, 106, 131, 143; Proto-Iranian 13–15, 49, 53, 67, 100, 102,
105–106, 115, 263; West Iranian 54–55, 149
Iranian plateau 17–20, 54, 69
Iraq 55, 224
Irāvatī River 93
Irtysh River 65
īśāna- 136, 140, 249
Ishim River 54, 59, 62
Ishtar, Ištar 238, 257, 265, 309, 312–313. See also Inanna
Ishqi-Mari 227
Isin-Larsa dynasty 219
iṣṭi- 134. See also kāmya-iṣṭi-
iṣu- 134
iti 168
itihāsa- 146
Iustinus 101

Jaffna 170, 291

Jain, Jaina, Jainism 3, 142, 178, 203, 205, 276, 285, 311
Jaipur 149
Jamdat Nasr 227
Janaka 157, 303
Janamejaya 146
Japan ix, 75
Jaṭ 260
Jātaka 157, 217, 281
Jayā, Jayantī 251–252
Jebel Aruda 221
Jew, Jewish 3
Jhelum River 93–94
Jhukar, Jhukar culture 22–23, 79
Jiroft 71, 210–211, 216–217, 307
juma, jumala 66, 300
Jumna (Yamuna) river 93
Jupiter 108, 188, 198, 204–205, 314–315
jyeṣṭhā-, Jyeṣṭhā, jyeṣṭhaghnī- 199, 250

kabara-, kabarī- 229

Kabul River and Valley 93, 98, 120, 255
kadala-, kadalī- 165
Kafir, Kafiri. See Nuristani
Kailāsa 157
Kaiṭabha 154
Kakṣīvant 118
kāla-, kaḷa-, kāẓ, Kāla 317
kalar, kaḷar, kaḷu 218
Kalevala 122
Kālī 158, 261, 317
Kalibangan 18–21, 198, 229, 240, 265, 309
Kālidāsa's Meghadūta 265, 317
Kālikā-Purāṇa 243–244, 251, 253, 265, 277
Kalittokai 241
Kāliyā 158
Kalpasūtra (Jaina) 285
kalutai 170
Kāma (god of love) 176, 181, 198, 251, 285
Kama River 47, 54–56, 59, 66–67
kampam 175
kāmya-iṣṭi, kāmyeṣṭi- 133
kaṇ, kaṇṇu, kāṇ 158, 267
kāṇa- 169
Kandahar (Qandahar) 17, 93, 98, 299
Kanesh 82
Kaniš 216
Kānīta 105
Kanmer 21, 169
Kannada language 165–168, 170, 172, 207, 217, 315
Kannapuram 175
Kaṇva, Kāṇva 90, 92, 120–121, 124, 126, 128–129, 131–133, 141
kapāla-, kapāla-pa- 263
Kāpālika 262
kaparda- 165, 229
Karachi 196, 269
Karagash near Karaganda 48
Karelian language 59
Karṇa 111
Karnataka ix, 218, 250, 278, 305
karpāsa- 165
Kārttikeya 198, 248. See also Skanda
Kartvelian languages, *grgar 46
karu-nāṭu 218
Kashmir 130, 165, 178–179, 255, 257, 278
Kashmiri language 82, 265
Kasperovtsy culture 47
Kassite language and empire 85, 89, 216
Kastor 49, 110, 155. See also Castor
Kātelai 80
Kathāsaritsāgara 172
Kati language and people 14, 264–265
kaṭuka- 169
kaulāla- 82
kaunakes garment 71, 257
Kaurava 146, 148, 299. See also Kuru
kavar, kavaram, kavari 229
kavi- 92
kayal 158, 315
Kazakhstan viii, xii, 54, 59, 68, 81, 297–298
Kelleli 70
Kelteminar culture 69
kem-mīn 314–315
Kerala ix–x, 167, 207, 248, 313
Kerār tribe 260
Kerman 71, 210, 213, 215–216, 306–307
ketu-, ketavaḥ 206, 310
keṭubha- 154
Khabur Valley 83
Khafajeh 211, 232, 234–235
khala- 169
Khaṇḍobā 150
Khanty (Hanti, Ostyak) language 59
Khapa 149
Khapuz Depe 69–70
Kharoṣṭhī script 27, 29
Kherai 80
khila- 93, 200, 285
Khorasan 54
Khotan 102
Khotanese Saka language 102–103, 106, 143, 299
Khowar language 163
Khurgak Govi 48
Khvalynsk culture 36, 38, 51, 295
Khwarezm (Chorasmia) 103, 256
Ki 238
Kikkuli 87
kīlāla- 82
kilbiṣa- 82
kiṃpuruṣa-. See kinnara-
kinnara-, kiṃpuruṣa- 141, 171–172, 248–250, 312
Kirta 88
Kish 211–213, 216
kiyāmbu- 165, 169
Kizil’ 65
KMK culture (Babino III culture) 54
Knossos 266
Kodagu language 165, 167
Kohistan 80
Kohistani language 163
kol 244
koḷ, kōḷ, kōḷi, koḷḷi 281, 283–284
Koldyri 46
Komi (Zyryan, Zyryene) language 59, 63, 66
Konar Sandal 216
Konkani language 7
Kopet Dagh 69–71
Korku language 165
Kot Diji 18, 20–21, 178–179
Kot Diji culture 17–20, 165, 178–179, 190, 192, 194, 312
ko ttam,ko tti, Kotti, Kottavai 243–244
kṛ- 132
Krivoe Ozero 61
kṛṣṇa- 154, 245
Kṛṣṇa, Krishna (Vāsudeva) 111–112, 145–146, 148–149, 153–156, 158, 303–304, 317
kṛta- 89
kṛttikāḥ. See Pleiades
kṛtyā- 132–133
Krumu 93
kṣatra- 111, 116, 136, 248
kṣatriya- 128, 155, 196, 316. See also rājanya-
Kubāba 236. See also Kybele
Kuban river and steppe 46
Kubera 152, 180
Kubhā 93
kulāla- 82
Kulli culture 17–18, 20, 211
Kültepe 82
Kulu Valley 256
Kumaon 177, 256
Kumaoni language 265
kumāra-, Kumāra 141, 232, 248, 254, 286, 310, 312, 317. See also Rudra; Skanda
kumārī-, Kumārī 248, 252–253, 317. See also Durgā
kuṇḍalinī-yoga- 204–205, 311
kuntāpa- 139
Kura-Araxes culture 46
Kurdish language 55
Kurgan culture 35
kūrma- 206–207
Kurram River 93
Kuru 92, 94, 145–148, 250, 299. See also Kaurava
Kurukh (Oraon) language 166–167, 172
Kurukṣetra 94, 146–147, 149
Kurum River 93, 121
Kuruśravaṇa 94, 146
Kuṣāṇa 241, 243, 255–256, 260, 309
Kutch 18, 20–22, 170, 182, 195, 211
Kutlug Tepe 97, 258
Kutsa 94–96, 133
Kuṭṭiccāttan x, 304
Kuvaṇṇā / Kuveṇī 152, 157
Kuvi language 167, 170, 172
Kuzali 71
Kvitanska culture 47
*k wek wlo- 39, 46, (> *kekra-) 63
Kybele 237, 246, 257. See also Kubāba

Lagash 211–212, 218, 227–228

Lahnda language 163, 260
Lahore 17
Lakṣmī 199
lāṅgala- 157, 165, 169
Lāṅgaleśvara 159
Laṅkā. See Sri Lanka
Las Bela 256
Late Harappan culture 22–24, 185
Late Helladic culture. See Greece
Late Kot Diji culture. See Kot Diji culture
Late Tripolye culture. See Tripolye
Late Uruk culture. See Uruk
Late West Iranian Buff Ware 54
Latin language ix, 3, 7, 11–15, 39–41, 100, 107–108, 110, 115, 131, 170, 200, 267, 284, 313
Latium 14
Latvian language 13, 16, 40,109
Leo 230
Levant 221
Linear B (syllabic) script x, 15, 266
Linear Elamite script 213
liṅga- 177, 194–195, 198, 275
Liṅga-Purāṇa 204
lip- 174
Lithuanian language 12–14, 16, 109, 155
Livonian language 59
Loebanr 125
logographic script, logogram 26, 172
logosyllabic script 221, 266–268, 305
lohita- 136, 204–205
loka- 131, 205
loṣṭa-, loṣṭa-citi- 119
Lothal 20–21, 23, 211, 213–214, 268, 282
Luristan 89, 211, 221
Luvian language 14, 38–39
Luvian (hieroglyphic) script 14, 30
Lycian language 14
Lydian language 14

Macedonian language 16
madhu- (< *medhu-) 63, 66, 109, 122, 154, 297, 302, 317
Madhu, Madhusūdana 154, 317
madhu-śiṣṭa- 63, 297
Madhurā 152–154, 157–158, 300. See also Mathurā
Madhya Pradesh 24, 167
Madra 152, 155–156
Mādrī 156
Madurai 152, 157–158, 300
madya- 251
Magadha 135, 139, 150, 255, 261
māgadha- 136, 250, 252
Māgadhī Prakrit 150, 252, 259, 261
Magan (Makan) 210–211, 213–217, 219
Magar Talão 269
magha- 86
Māgha's Śiśupālavadha 171
maghavat- 122
magilum boat 217
Mahābhārata 4, 111–112, 114, 145–146, 148–150, 152, 154–155, 157–158, 180, 188, 193, 197–198, 206–207, 216,
231, 259, 262, 299, 303
mahādeva- 136
mahānagnī- 252
Maharashtra 24, 149, 151, 166–168, 175–177, 305, 308
Mahāvaṁsa 151, 158
mahāvīra- 121, 124, 126, 302; M. Jina 285
mahāvrata- 137, 188, 241, 246, 249–251, 303, 309, 312–313
mahiṣa-, Mahiṣa Asura 241–243, 249–250, 254, 309; Mahiṣa-Saṃvara 263; Mahiṣāsuramardinī 241, 244, 253, 255
mahiṣī- 248, 250, 253
Mahurjhari 149
mai-m-mī n 276
maithuna- 246, 251. See also mithuna-
Majdanets’ke 42, 44–45
Majkop culture 46
Maka 214
Makan 255. See also Magan
makara- 181
makha-, Makha, makhavat- 111, 121–122, 124, 126, 128, 204
makhē, makhesasthai 111, 122
Makran 55, 103, 210, 214–215, 306
Malayalam language x–xi, 165–167, 172, 217, 313, 315
Malto language 165–168, 171–172
Māmallapuram 241–242
māṁsa- 251
man- 102
Manchar Lake 210
mañci 217
maṇḍala-, of Rigveda 92, Tantric 259, 261–262, 316
maṅga-, maṅginī- 217
maṇi- 86
Mansi (Vogul) language 59
mantra- 6, 8, 124, 126, 137–138, 140, 155, 200–201, 204, 207, 249, 253–254, 277, 301
manu-, Manu 95–96, 102, 143, 197, 261, 272
Māra 143, 264
Marathi language 6, 151, 166, 172
Marduk 112, 256
Margiana viii, 69–71, 77, 79, 82–83, 98, 101, 104, 164, 211, 258, 260, 298
Margush 69, 101
Marhaši 211, 215–216, 219
Mari of Syria 110, 227, 230, 237, 257
Mari (Cheremiss) language and people 59, 63, 86, 102
Maria Gonds 186
marīca- 165
Mariupol' 37
Mārkaṇḍeya-Purāṇa 242
markhor 171, 231, 240, 265
Mars 204–205, 254, 283, 285, 312, 314–315, 317
martiya- 86
Martu 218
Maru 261
Marudvṛdhā River 93
Marut 107, 127
Marwar 261
Marwari language 163
marya-, maryanni, maryannu 83, 85–86, 90, 102, 298
Massagetae 148–149
Mātali, Mātalī, Mātariśvan 114–115
math- / manth- 115
Mathurā 145, 149, 153–154, 158, 255, 284–285, 299–300, 303. See also Madhurā
matsya-, Matsya 149, 181, 206, 251, 315
Matsya-Purāṇa 206
Mature Harappan. See Indus civilization
Mauryan kingdom 153
māyā- 254
Mayan language 11, 274
Mayan script 30, 274
Maysar 211–212
mazdā- 114. See also Ahura Mazdā
medhā- 90–91, 114
Medhātithi, Medhyātithi 90, 121
Media, Median empire and people 54–55, 90, 149
Median language 100–101
Medinet Habu 84
Mediterranean 240, 307
Megalithic culture of India 27, 145, 149–159, 299–300
megaron 37
Megasthenes 153–155, 157, 159, 303
Mehatnū 93
Mehrgarh 17–18, 79, 173, 190, 193, 211, 312
mēl, mēlu 203–204
Meluhha 21, 163, 210–211, 213, 215–219, 227, 229, 307
Meluhhan language 163–164, 215
Mercury 188, 204–205, 277, 314–315, 317
Meru 203, 205–206, 281, 311
meru-daṇḍa- 205
Merv 69
meš 214, 217, 219
meṣa- 249, 284
Meskalamdug 227
Messapic language 14
Middle Danubian culture 124
Middle Dnieper culture 47–48, 297
mīḍha- 87, 91
Mikhajlovka 47, 54
min, mīn 271, 288–289. See also mai-m-mīn; poṭṭu-mīṇ; vaṭa-mīn; veḷ-mīn; veḷḷi-mīn
Mīnākṣī 157–158
Minoan 240
mī n-ūrti 181
Minussinsk basin 47
Minyan Ware 49
Mir-i Qalat 19–20, 210–211
Mitanni (Mittani, Mittanni) 8, 15, 69, 71, 76, 79, 81, 83–91, 111–112, 121, 131, 222, 298, 300, 302
Mithilā 157
Mithra 112, 114, 116
Mithridates 55
mithuna- 252. See also maithuna-
mitra-, Mitra 90, 92, 107–108, 111, 116, 122, 249, 302
Mitra-and-Varuṇa 89, 108–109, 112–114, 117, 128, 134, 300–303
Mizar 201
Mleccha 259
Mohana 210
Mohenjo-daro vii, 8, 17, 20–23, 25
Moldavia 36, 295
Mollali 71
Mongolia 47–48, 54, 76
Moon (lunar) 113, 119, 134, 158, 171, 187–189, 198–201, 204, 206, 226, 232, 234–235, 250, 256, 278, 281, 303,
308–311, 317
Mordvin language 59, 63
Mot 236
mṛga- 81
Mṛtyu 142–144
mudrā- 251
Mughals 3, 260
muhūrta- 188
mukha- 126, 169, 196
mūlādhāra- 204
multiroller ceramics 54
Mumbai vii, 22, 151, 308
Munda languages 165
Mundigak 17–18, 69, 79, 211
Murghab River 79, 101
muruku, Muruku, Murukan (Murugan) 254, 285–286, 288–289, 291, 312, 314–315
Muski 49
Muslim (Islam) vii, 3, 6, 55, 166, 189, 259, 269, 288
Mūtiba 259
Mycenae, Mycenaean civilization 49, 53, 237. See also Linear B script
Mycenaean Greek language 15, 39

nāḍī- 204–205
nāga-, Nāga, Nāgarāja 158, 173, 180
Nāgapaṭṭinam 158
Naigameṣa 124, 284–285. See also Nejameṣa
Nainā, Nainital 256
nakṣatra- 171, 187–189, 196–199, 204, 250, 254, 278, 310
Nakula 156
Nal 17–18, 211
Namazga Depe 69–71, 79, 211
Namuci 96
Nana, Nanaya 256–257, 309
Naqada 221–222
nara- 102, 171
Nārada 156
Narām-Sīn 215–216, 218, 307
Narmer 222
nas- 127, 302
nāsatya-, Nāsatya 89, 108–112, 117–118, 120, 127, 300–302. See also Aśvin
nāsika- 126
naudhasa- 137–138
Nausharo 17–18, 20–21, 79, 278, 280, 290
nava- 87
nava-graha- 285
navarātra-, navarātri- 173, 249, 253, 259, 309, 313
NBPW. See Northern Black Polished Ware
Nejameṣa 284–285. See also Naigameṣa
Neo-Assyrian 54, 100
Neo-Sumerian period 218, 227, 232
Nepal 107, 167, 249, 253, 259, 263
Nepali language 265
nestōr, Nestor 110, 117–118, 301
Nestor's Chronicle 222
Netherlands 47, 296
Nidānakathā 180
nidhana- (finale) 133, 137
nij- 285
nīla- 136, 205
nīlakaṇṭha- 186
Nile River and Valley 220, 222
Nimmureya (Amenophis III) 86
Nindowari 211, 247, 290–291
Ninhursag 219
Nin-ti 267
Nippur 212–213, 216, 224–225
Nizhne-Mikhailovka culture 47
Nonnos 231
Northern Black Polished Ware (NBPW) 145, 147, 150
Northern Neolithic culture 18, 165
nostos 117, 301
Novodanilovka 37
Novotitarovskaya culture 46
Nubia 219
Nuristan 6, 143, 263–265
Nuristani (Kafiri) languages 6, 14, 82, 105, 163, 263, 265
Nuzi 87, 89

Ochre Coloured Pottery (OCP) culture 23–24, 261

odana- 246
Oka River 47, 54
Okhos 101
Old Church Slavic 13, bogǔ106
Old Norse. See Icelandic
Old Persian cuneiform script 29
Old Persian language 6–7, 11, 13, 55, 88, 90, 93, 97, 100, 102–103, 105–106, 131, 215, 217, 264, 299
Old Prussian language 16
oṁ 170, 205, 311
Oman 210–211, 213–215
Omsk 64–65
Onesicritus 153
Oraon (Kurukh) language 166–167, 172
Ordos Plateau 76
Orissa 130, 155, 167, 169, 243, 259
orja 100
Oscan language 14
Ossetic language 6, 102, 106
Ostyak (Hanti, Khanty) language 59
Oxus (Amu Darya) 69, 93, 101
Oxus civilization viii, xii, 69. See BMAC

paccai, pacu-mī n 276–277

Pahari language 265
Painted Grey Ware (PGW) culture 145–147, 299
Paippalāda xv, 130, 132, 135
pakhsa 98–99
Palaic language 14, 38
palālī-, palāla- 82
Pali, Pāli language 151, 317
palita- 87
Pallava 157, 241
paḷḷi, palli 166
Pamir Mountains 71
pañca- 87, pañca makārāḥ251
Pañcajana 317
Pañcāla 146, 148, 250
Pañcavīra 153
Pandaíē (Pandaea) 153–157, 303
Paṇḍu, Pāṇḍu, Pāṇḍava 145–149, 152–157, 299–300
Paṇḍu-Vāsudeva 152–153
Pāṇḍya 152, 300
Paṇi 81, 92, 94, 96, 101, 105–106
Pāṇini 10, 145–146, 152, 199, 299
Panjab. See Punjab
Parachi language 82
Parahšum 215–217
parasvata- 249
parihasta- 288
Parikṣit 146
pāriplava- 141
pariṣkanda- 136
Parji language 167, 172
Parna 55, Parṇaya, Parnoi 101
Parsatatar 88–89
Parsee 3, 6
Parśu (Sparśu) 149
Parsuva 54
Parthia, Parthian empire 55, 98, 100–101, 103
Parthian language 101, 106
Paruṣṇī River 93–94
Pārvatī 173, 176
pāśa- 144
Pashai language 82
Pashto language 6, 97–98, 163, 299
paśu- 248
paśupati- 136, 140
pat- 245
Pātāla 267
Patañjali's Mahābhāṣya 152; Yogasūtra 207, 209
pati-, patnī- 245
Patroclus 118
Paurava 94
pavamāna- 128
peace 232, 249
Pechora River 59
Pengo language 167, 278
Pepkino 55
Permic languages 59, 66
Persepolis 103
Persia, Persian. See Achaemenid; Old Persian
Persian Gulf. See Gulf
Peshawar 125, 191–192
Petrovka culture 57, 59, 68, 297
peya 66
PGW. See Painted Grey Ware
phala- 149, 169
phāla- 157
Phrygian language 14–15
PIE. See Indo-European languages
piḷḷai, Piḷḷaiyār 291
piṅgala- 87, 205
Pipru Mṛgaya 81, 94–96
Pirak 17, 103–104, 298
Pit Grave cultures. See Yamnaya cultures
pīta- 205
Pitṛmedhasūtra 142
Pleiades (kṛttikāḥ) 196–198, 250, 265, 274–275, 284–286, 310, 312
Polaris 200–201
Pole star 190, 200–201, 203, 206–209, 235, 281, 306, 310–311, 314, 316
Polish language 16
Pollux 110. See also Poludeukes
Poltavka culture 55–56, 297
Poludeukes 49, 101, 155. See also Pollux
Pomponius Mela 101
po n, po n-mī n, po nna n, po nnava n 314–315
Pontic/Pontic-Caspian steppes 35, 38, 40–41, 46–47, 49, 51–52, 54–55, 67, 105, 295–298
Poros 94
Portuguese language 14
Post-Harappan cultures 18, 23, 79
Post-Tripolye cultures. See Tripolye
Post-“urban” / Late BMAC. See BMAC
Potapovka culture 56, 122–123
poṭṭu, poṭṭu-mīn 278
Pozdnyakovo culture 54
Prabhas Patan 79
prācya- 135
pragātha- 128–129, 302
Prajāpati 119, 136, 142, 156–157, 177, 188, 196, 198, 206–207, 209, 244–248, 250, 252, 254, 303, 309, 312–313,
316
Prakrit 6, 101, 131–132, 140, 150, 252, 259, 261. See also Pāli
pramātha- 115
prāṇa- 127, 207–209
prāṇāyāma- 209
prastotar- 128, 139
Prasun Kāfirs 264
pratihartar- 128
pratiprasthātar- 88, 124, 133, 302
pratoda- 136
pratyāhāra- 207, 311
pravargya- 121–122, 124, 126–128, 133–134, 136–137, 204, 302. See also gharma-
pravarta- 136
Pre-Harappan cultures 17
Priyamedha 90
Prometheus 114–115
Proto-Elamite culture 19, 29, 69, 209–210, 212, 221, 224, 226, 231, 237–238, 241–242, 306. See also Elam, Trans-
Elamite
Proto-Elamite script 19, 25, 29–30, 69, 238, 306
“Proto-Śiva” 25, 143, 178, 194–195, 209–210, 212, 231, 308, 312
proto-writing 26, 221
pṛṣṭha- 127, 137
Pṛthivī 108
Prut River 46
Ptolemy's Geography 101
pūjā- 169, 173–174. See also cakra-pūjā-, Durgā-pūjā-
Pulinda 259
puṁsavana- 127, 177, 284
puṁścalī-, puṁścalū- 136, 252
Puṇḍra 259
Punjab (Panjab) 17–18, 211, 262
Punjabi (Panjabi) language 6, 163, 265
pur- 246, 257
puraḥsara- 136
puraḥsthātar- 88
Puranānūru 154, 315
Puri 155
purohita- (royal priest, court priest) 111, 114–116, 130, 136, 141, 196, 301, 303
Pūru 93–94. See also Paurava, Poros
purudaṁsa(s)- 112
Purukutsa 94–95
puruṣa-, Puruṣa 171, 205–206, 209, 245–246, 248–249, 254, 316
puruṣamedha- (human sacrifice) 140, 172, 206, 241, 248, 254, 303, 312, 316

qala 98–100
Qala'at al-Bahrain 213–214
Qandahar (Kandahar) 17, 93, 98, 299
Quetta 17, 79, 211

Rahman Dheri 190–192

rājanya- 141, 196. See also kṣatriya-
Rajasthan 18, 20, 23–24, 79, 149, 163, 299, 305
rājasūya- (royal consecration) 115, 128, 139, 170, 193, 231, 316
Rajput 168
Rakhigarhi 18, 20–21
Rāma 150, 156–157, 246. See also Bala-Rāma
Rāmāyaṇa 4, 145, 150, 152, 156–157, 168, 300, 303
Ramses 84
Raŋhā 93
Ra's al-Hadd 211–212
Ra's al-Junayz 211–213
Rasā River 93
raśmi- 204, 206, 209, 310
ratha-, rathya- 88, 128, 136
rathantara- 128, 136–138
rathīr adhvarāṇām 115
Rāvaṇa 152
Ravi River 17–18, 93–94
Ṛbhu 113
ṛbīsa- 81–82, 126
Red Sea 221
Reshnoe 57
Rhā 93
Rheto-Romanian language 14
Rigveda xii, 6, 8, 39, 81–82, 90–98, 106–108, 120–122, 128–134, 146, 299
Rimuš 215–217
Ṛjiśvan 94–95
ṛkṣa- 197, 204, 272
rohiṇī-, Rohiṇī 198–199, 250, 254, 278, 313, 317
rohit- 246
rohita-, Rohita 198–200, 278, 312
rohita-matsya- 315
roller pottery 49
Roller Pottery cultures 54, 105, 298
Roman script 30
Romance languages 14
Rome, Romans 14–15, 83, 101, 108, 117, 200, 237, 300
rongorongo tablets and signs 30
Rosetta stone 27
Rostovka 36, 57, 64, 67
ṛṣi- 81, 92, 197. See also Seven Sages
ṛśya- 246
ṛta- 144, 312
Ṛtadhāman 88
Ṛtamna 90
Ṛtasmara 88
rud- 141, 254
rudhira- 254, 283, 312
Rudra 4, 108, 122, 135–136, 139–143, 156, 158, 186, 188, 209, 229, 231, 247–252, 254, 283–286, 291, 303, 312,
317
rukma- 136
Russia, Russian viii, xii, 35, 59–60, 69, 86, 122–123, 164, 222, 296–297
Russian language xii, 13–14, 16, 39–40, 51, 54, 107, 131

Saami languages 11, 59

śabara-, Śabara 249, 259
śabarotsava-, śābarotsava-, śābara-utsava- 246, 249, 251, 259, 262, 313
sabhā- 252
śad- 118
Sāhā 252
Saha-Deva 156
sahasra- (66)
Śaiva. See Śiva
śak-, śaka-, Saka people 54, 101–106, 259–260, 299, 314
Saka language 101–102, 106, 143, 259
śakaṭa-, śakaṭī- 170–172
Śākta 3, 241, 252, 255, 262, 314
Śākya 152
Śalya 111, 156
sāman-, Sāmaveda ix–x, 128–139, 141–142, 170, 302–303
Samara 122–123
śambara-, Śambara 94–98, 106, 259–260, 262–264
saṃdaṁśa- 134
saṃgrahītar- 110
Saṃkarṣaṇa 157–158
Samoyed languages 59, 68
Sampo 66, 300
samrāj- 115–116, 121, 138
saṃvara-, Saṃvara 259, 262–263
śanaiścara-, śani- 204, 276
saṅgha- 142
Sañjaya 148
Sanskrit (Old Indo-Aryan) vii, x, 6–11, 13–15, 36, 39–40, 63, 66, 81–82, 85–91, 102, 105–106, 110, 114–115,
128–133, 140, 149, 165, 168–172, 174–175, 197, 199–200, 203–204, 207, 217, 229, 249–250, 255–257,
259–265, 278, 280–281, 283, 285, 297, 300, 304, 310–313, 317
Sanur 27
Saora 259
Sapalli 70, 76, 120
sapatna-, sapatnī- 133; sapatna-sāhī- 253
sapta- 87; sapta hotrāḥ 197; sapta rṣayaḥ 197; sapta sindhavaḥ 93
śara- 217
śarabha-, śarabhī- 243–245, 248–249, 265
Śāradā, śāradī-; Śār(a)dī 257–259
Saraṇyū 127
Sarasvatī River 93, 97, 101, 119, 299; goddess 156, 248, 252, 313
sārathi- 110, 136
Sarayū River 93
Sarazm 48, 69, 79
Sarcostemma 108
Sargary (Alekseevka) culture 54, 105
Sargon the Great, Sargonid 215–216, 226–227, 229–230, 240, 307
Šarkališarri 216, 230
Sarmatians 54
sarpa-bali- 173–174
sārparājñī- 240
Sarsuti River 93, 97
Śarva 136, 140, 143
Śaryāta Mānava 119
Sasanids, Sasanian 55, 98, 229
śata- 13, 66
śatarudriya- 140, 188
satem languages, satemization 13, 48, 52, 55, 296
Satī 156
Sattagydia 103
Šattivaza = Sātivāja 88–89
Saturn 204–205, 208, 276, 312, 314–315, 317
Satyavat 156
Satyga 57
Śaunaka xv, 130, 132, 150, 288, 303
Saurashtra 168, 211
Sauštat(t)ar 88–89
sautrāmaṇī- 128, 302
śava- 134
Sāvatthi 180
Savitar 108, 119, 198
Sāvitrī 155–158, 198
savyaṣṭha-, savyaṣṭhā-, savyeṣṭha- 88, 110
Sazagan 76
Scandinavia, Scandinavian xi, 48, 222
script (writing system). See alphabetic, logographic, logosyllabic, syllabic script; Aramaic, Brāhmī, Chinese,
cuneiform, Egyptian, Greek, Indus, Kharoṣṭhī, Luvian, Mayan, Old Persian, Proto-Elamite, Roman, runic,
Semitic, Sumerian script
Scythian language 101, 105, 164, 299
Scythian people 36, 53–54, 101, 103, 263
Seistan 69, 97, 210, 306
Sejma 57, 64
Sejma-Turbino trader network 36, 57, 59, 63–65, 67–68, 76, 297
Seleucus, Seleucid empire 55
Semipalatinsk 65
Semitic languages 35; (Amorite) 213; East Semitic (Akkadian) 85, 215; West Semitic 85, 163, 248
Semitic (consonantal alphabetic) script 29
senāpati- 140
Serbo-Croatian language 16
serekh 222
Śeṣa, Ādi-Śeṣa 158, 317
Seven Sages 196–198, 209, 265, 275, 285, 288, 311. See also Ursa Major
Shahr-i Sokhta 69, 79, 211, 306
Shalmaneser 54, 110
Sheikh Hassan 221
Shimashki 70
Shina 163
Shipunovo 65
Shortugai 21, 120, 211, 216
Siberia 47, 55–56, 59, 64–65, 67–68
Sibri 17, 79
Sikh, Sikhism 3
siṁha- 220; siṁhī- 253
Siṁhala (Sinhala) 151–153
Siṁhapura (Sīhapura) 151–152
śimyu- 102
Sindes 101
Sindh 3, 17–18, 22, 149, 163, 166, 169, 171, 178, 185, 207, 211, 216, 218–219, 239, 260–261, 299, 306
Sindhi language 82, 163, 172, 218, 260–261, 265
Sindhu 3, 93
sindūra-, sindura-dāna- 278
Sinhalese language 151, 265
Sinitic languages 165
Sinkiang. See Xinjiang
Sintashta 58, 60, 63
Sintashta culture 49, 56–61, 63, 68, 76, 78, 103, 109–110, 118, 257, 297–298, 301
Sippar 182
Siraiki language 163
śiśumāra- 199–200
Siswal 17–18
Sītā 156–158, 246, 303
Śiva (Śaiva) 4, 108, 124, 139–141, 150, 154, 156–159, 173, 176–177, 182, 186, 194–195, 198, 209, 242–243, 251,
257, 259, 262, 275, 285, 304, 312, 317. Seealso “Proto-Śiva”
Siwalik Mountains 97
skambha- 175
skand-, caskanda 285
Skanda 140, 155–156, 186, 198, 231, 248, 264, 285–286, 288, 291, 310, 312, 315
Skanda-Purāṇa 155
Skelya culture 37–38, 49, 51, 53, 295–296
Slav, Slavic people 100
Slavic (Slavonic) languages 13–14, 15–16, 39–40, 106; East Slavic 16; South Slavic 16; West Slavic 16
śloka- 119, 131, 141, 204, 209
Slovacian language 16
Slovenian language 16
Sofievka culture 47
Sogdiana 103
Sogdian language 106
Sohan River 93
soma-, Soma 4, 93–94, 103, 107–109, 118–119, 121–122, 124, 128, 130, 133, 142, 188, 252–253, 300, 302, 304; as
the moon 119, 188, 198, 310
Somnath 79
Sonkh 158
Sora 259
Sothi 17–18
Sotka-koh 210–211
Southeast Asia 165
Spain 15, 240
Spanish language 14
Sparta 111
Spica 188
śrauta- 130, 134–136, 146
Sṛbinda 81–82, 105
Srednij Stog II culture 36–38, 47, 51, 295
śṛgāla- 81
Sri Lanka (Laṅkā) 145, 149–153, 157–158, 167, 170, 176, 201, 300
Śrīvidyārṇava-Tantra 158
śru- 131
Srubnaya (Timber Grave) culture 54, 62, 67, 105
stambha- 66, 300
Starokorsunskaya 46
Stogovska culture 47
stoma- 133, 135
Strabo 101
Stril’cha Skelya 37
stūpa- 150, 284
Subandhu 90
Sub-Carpathian culture 47–48
Subhadrā 156
Subir 218
Sudās 94, 97–98
śūdra- 82, 138, 316
Šu-ilišu 164, 215
Sukanyā 119
śukla- 205
śukra- 122, 204
Šulgi 218
Śulvasūtra 187, 195–196
Sumakha 110–111, 122
Sumer, Sumerian 25, 43, 46, 71, 181–182, 197, 210–232, 269, 296, 307. See also (Late) Uruk
Sumerian language 163, 215; abba 219; abgal 197; dar 217; gigir 46; gú.bar 229; haia 217; kaunakes 71; ki 238; kur
217–218; magilum 217; meš 214, 217, 219; mul 230; ti 267; ú 219; ušū 218. See also An(u), Aratta, Enki,
Inanna, Marhaši, Meluhha
Sumerian script, archaic, 25, 29–30, 43, 230, 238, 272, 305
Sun (solar) 39, 89, 95–96, 108–109, 112–114, 117–119, 121–122, 127, 134, 136, 154–156, 188–190, 193–201,
204–208, 226, 231–232, 235, 249–251, 254, 256, 267, 278, 281, 303, 308–312, 315–317
Śunaḥśepa 119, 170, 177, 197, 199–201
sundara-, Sundareśvara, Sundara-Pāṇḍya 157
suparṇa- 119
Suppiluliuma I 89
surā 66, 109, 128, 302
Śūrasena 154, 157
Surkotada 18, 20–21
Sūrya 39, 89–90, 108, 113
Sūryā 119
sūrya-raśmi- 204
Susa 70, 79, 100, 102, 211–213, 220–221, 225, 227, 237–238, 256, 306
Susartu 93
Śuṣṇa 94–96
Suṣomā 93
suṣumṇa-, suṣumṇā-, suṣumnā- 204, 311
sūta- 115, 141, 254
Sutkagen-dor 20–21, 210–211
Sutlej River 93
Sūtra texts 8
Śutudrī River 93
Suvāstu 93, 120
Suvorovo 37–38
svadhiti- 122
Śvetyā 93
Swat River and Valley 80, 93, 98, 125, 165, 255, 299. See also Suvāstu
Swedish language 15, 41
śyaita- 137–138
śyena- 134
syllabic script 221, 266–267
Syria 8, 55, 69, 71, 76, 79–91, 112, 121, 227, 231, 236–237, 257, 273, 298
Syriac language 172

Tacitus 101
Taidu 83, 85
Taip-depe 70, 82
Tajik language 6
Tajikistan 48, 69, 76,78, 81
Takhirbaj 71, 98, 103
Takla Makan 255
Taliqan 69
Tall Fakhariya 83, 85
Tall Hamidiya 83, 85
Tall Leilan 83
Tall-i Malyan 211
Tal’yanky 42
Tambapaṇṇi 151, 153
Tamil language xi, 30, 82, 150, 153, 157–158, 165–167, 169–172, 175, 181, 187, 201, 203, 207, 217, 241, 243–244,
254, 265, 271, 274–278, 280–283, 285, 288–289, 291, 310–315
Tamil Nadu ix, 145, 149–150, 152–154, 157–158, 167, 175, 241–242, 285, 288, 300
Tammuz (Dumuzi) 239
Tāmraparṇī 153
Tantra, Tantric 209, 241, 251–252, 255, 257, 259, 261–265, 314, 316
Taprobane 153
Tārā 188, 198, 310
Taraqai Qila 192
Tarim basin 255
tārpya- 193, 199, 231, 310
Tarut 211
tathā, tathāstu 170
Taurus 230, 278
Taxila, Taxiles 98, 152
Tedzen / Tejend River 69, 101
teîkhos 264
Tell Abraq 211
Tell Asmar (Eshnunna) 182, 211–212
Tell Brak 89
Tell el-Amarna 85–86
Tell es-Sulema 211–212
Tello, Telloh (Girsu) 211–213, 218, 228
Telugu language 165–168, 171–172
Tepe Hissar 79–80, 83–84, 211, 298
Tepe Yahya 79, 211, 213
Thor 63
Thracian language 14–15
Thuban 200–201, 281
ti 267
Tibet, Tibetan 165, 259, 263
Tibeto-Burman languages 165
tigraxauda- 103
Tillya Tepe 98,101, 299
Timber Grave culture. See Srubnaya culture
tīrtha- 175
Tirukkuraḷ 207
Tiruviḷaiyātar-Purāṇam 157
Tobol River 54–56, 59, 62
Tocharian (Tokharian) languages 11–16, 39–41, 48
Togolok 70–71
Transcaucasia 46
Trans-Elamite 71, 76, 210, 298, 306
Trasadasyu 95, 146
tri- 87
Tri Brata 52
Tripolye (Cucuteni-Tripolye) culture 36–39, 42–49, 51, 85, 295–296, 298
tripura- 257–259, 262
Tripurā, Tripura-Sundarī 257
triṣṭubh- 129
Tṛkṣi 94
Troy, Trojans 14, 37
Tṛṣṭāmā 93
Tryambaka 140, 251
Tulu language 167, 171–172, 207, 217
Tumburu 251
tuṇaṅkai 243–244
Turbino 57. See also Sejma-Turbino
Tureng Tepe 79–80, 83, 211
Turkestan, Chinese. See Xinjiang
Turkey 14, 16, 55, 87, 236, 274
Turkmenistan xii, 19, 54, 59, 66, 68–70, 79, 297. See also BMAC, Margiana
Turvaśa 93
Tušratta 86, 88
Tvaṣṭar 108, 245, 250, 288
tveṣaratha- 88

ucci 203–204
udgātar- 128–129, 137–138, 141–142
udgītha- 128, 137, 141
Udmurt (Votyak) language 59, 63, 66
udumbara- 138
ugra- 102; ugro devaḥ 136, 140
Ugric languages 59
Ū a 100
Ujjain 317
ukha- 169
ukhā- 136, 171
Ukraine 35–36, 164, 222, 295
Ukrainian language 16
uloka- 131
Ulug Depe 69–70
Umbrian language 14
Umma (Djokha) 211–212, 270
Umm-an Nar 211, 214
upanayana- 130
Upaniṣad 150, 204, 315
Ur 211–213, 217–218, 232; Royal cemetery 212–213, 227; (Neo-Sumerian) Ur III dynasty 36, 216, 218–219, 227,
232, 256, 307
Ural Mountains and river 47–49, 51, 53–60, 63, 68, 76, 109–110, 118, 257, 296–297, 301–313.
Uralic (Finno-Ugric) languages xii, 11, 35, 41, 59, 66, 68, 81, 86, 102, 108–109, 114, 297, 300; Central Uralic 59,
300; East Uralic 59, 66, 68, 301; Proto-Uralic (PU) 15, 59, 63, 66–67, 114, 123, 297, 300; West Uralic 59, 63,
297, 300
Urmia Lake 54, 83
Ursa Major (Big Dipper) 190, 197, 201, 209, 274–275, 280, 311. See also Seven Sages
Ursa Minor 200–201
Uruk 211, 216, 224; Late Uruk culture 25, 43, 71, 221–224, 226–227, 231, 307; Uruk expansion 46, 221–222, 296
Uśanas Kāvya 95
Uṣas 108–109, 113, 155–156
Usatovo culture 47
uṣṇīṣa- 136
Ust’-Muta 64
uṣṭra-, uṣṭrī- 120, 245, 249–250
Uttarakhand 256
uttara-vedi- 253
uttarāyaṇa- 249
Uzbekistan 70, 81

vāc-, Vāc 137–138, 156, 241, 244–246, 250–253, 303, 309, 312–313; vāg-īśvarī-252, 313
vācaspati- 245
Vadhryaśva 97
vadhū-rā- 87
vahana- 87, 91
Vāhīka 216. See Bāhīka
vahni- 115
vaiku ru mī n 171
vairāja- 137–138
vairūpa- 137–138
Vaiśampāyana 146
Vaiṣṇava. See Viṣṇu
vaiśya- 196, 316
vājapeya- 109, 111, 114–115, 139, 193, 303
vājasāti-, vājasya sāti- 88
vajra 63, 66, 107, 114, 134, 136
Vajra-Varāhī 262
Vakhsh 81
vāla- / vāra- 128
valikovaya keramika 54. See also roller pottery
Vāmadeva 92
vāmadevya- 136–137
Vāmana 108
Vāmana-Purāṇa 243
Vandana 118, 126
vā n, vā n-veḷḷi 289
var- 259
var, vara- 98, 259–260
Varanasi 175
Varcin 96
varṇa- 316
vart-, vartana- 87
varuṇa- 112, Varuṇa 108–109, 114–116, 119–120, 128, 143–144, 174, 177–178, 181, 193, 199–201, 203, 206–207,
231, 235, 250, 253–254, 281, 302, 306, 309–310, 312. See also Mitra-and-Varuṇa
vasanta- 249
Vasiṣṭha 92, 113, 120
Vassukanni 83, 85
Vāsudeva (Kṛṣṇa) 146, 149, 153–155. See also Paṇḍu-Vāsudeva
vaṭa, vaṭam 201, 278, 280–281, 283, 310
vaṭa-maram 280
vaṭa-mīn 201, 206, 280–281, 291, 310
vaṭa-sāvitrī-vrata- 156, 173
vaṭa-veḷḷi 291
Vāta, Vāyu 108
vāta-raśmi- 206, 209
Vāyu-Purāṇa 204
Veḍḍa 151
veḷ, vēḷ, Vēḷ, veḷi, veḷḷi 288–289, 291, 313–314
veḷiccam, veḷiccappāṭu 313
veḷ-mī n(veṇ-mī n), veḷḷi-mī n 289, 314
Venetic language 14
Venus 187, 204, 238–239, 289, 313–315, 317
vibhindu- 165
Vidarbha 149
Vidathin 95
Vīdēvdād 98
vijaya-, Vijaya 151–153, 157, 300; vijaya-daśamī- 249; Vijayā 152, 158, 252
Vikings viii, 222
vi+kram-, vikrama- 205
viḷ, viḷakku 313
Vima Takhtu 255
viṇ, viṇ-mī n 289
vīṇā- 141–142
vipatha- 135–136, 140
vipra- 116
vīra- 121, 251. See also mahāvīra-
Virāṭanagara 149
Virgo 188
Viṣṇu (Vaiṣṇava) 4, 108, 121–122, 145, 148, 153–159, 205, 242, 277, 317
Viṣṇu-Purāṇa 201
viṣṭuti- 133
viṣu-, viṣuvat- 250, 312
Viśvāmitra 92
Vitastā 93
viṭi-mī n, viṭi-veḷḷi 291
Vivasvant, Vivasvat 113, 127, 143
Vogul (Mansi) language 59
Volga River 36, 47, 53–56, 62, 67, 86, 93, 122–123297
Votyak (Udmurt) language 59, 63, 66
Vrācaḍa 261
Vraja 317
vrata- 94, 136–137, 139, 141–142, 156, 173, 188, 241, 246, 249–251, 303, 309, 312–313
vrāta- 135
vrātīna- 134
vrātya- 130, 204, 209, 229, 241, 244, 249–252, 255, 261–262, 303, 309
vrātyastoma- 134, 249, 303, 309
vṛkṣa-śāyikā- 291
vṛṣabha- 249
vṛṣṇi-, Vṛṣṇi 146, 153, 249
Vṛtra 95–96, 107, 114, 116, 206, 253
vṛtrahan- 114
Vulgar Latin language 14
vyāghra- 248
vyāna- 127
Vyāsa 148

Wakhi language 102–103, 163, 299

Waziristan 98, 299
Welsh language 15, 39–40
writing 19, 25–26, 29, 163, 221, 266–267, 274–276, 305. See also script
Wusun tribe 255

Xerxes 101–103
Xinjiang (Chinese Turkestan, Sinkiang) 16, 48, 102, 255, 300
Xuan-Zang (Hsüan-Tsang) 152

Yadu, Yādava 93, 153, 155, 303

yajata- 106
yajña- 6, 134, 245, 252
yajñāyajñīya- 136–137
Yajurveda, Yajurvedic 8, 82, 126, 130, 133–135, 137–138, 142, 146, 170–171, 187–188, 196, 204, 302–303, 311
yakā- 215
yakṣa-, yakkha-, yakṣī-, yakṣiṇī-, yakkhinī 151–152, 157, 178, 180
yama (Russian) 51
Yama 118–119, 127, 143–144, 156, 208, 264, 301, 312. See also Imra, Yima
Yamī 143
Yamnaya (Pit Grave) cultures 37–38, 47–49, 51–53, 55, 295–298
Yamunā (Jumna River) 23–24, 93, 146, 154, 158, 255
Yāska 112
yasna- 6, 128, 134, 302
Yašt 129
yātu-, yātudhāna- 132
yaudha- 134
Yavyavatī 93
Yaz I -related cultures 54, 98, 103, 105–106, 298–299
Yaz II culture 54, 149
Yenissei River 59
Yima 98, 143, 259. See also Yama
yoga-, Yoga 25, 204–205, 207–210, 306, 311, 315
yogeśvara- 209
yoginī- 251
yoni- 198
Yudhiṣṭhira 146, 193
Yuezhi 255
yūpa- 177
Yurino 57
yuvarāja- 232

Zagros Mountains 89
Zahar 215
Zarathuštra 112, 114, 129
Zardcha Khalifa 76, 78
Zarif Karuna 125
Zeravshan Valley 48, 69, 76
Zeus 108–109, 117, 155, 300
Zhob River 93
Zincirli 237
Zoroastrian religion 6, 85, 98, 103, 106–107, 112, 116, 128, 134, 143, 298, 302. See also Zarathuštra, Parsee
Zyryan, Zyryene (Komi) language 59, 63, 66