Applied Linguistics 29/4: 578–596 ß Oxford University Press 2008

doi:10.1093/applin/amm056 Advance Access published on 31 January 2008

Symmetries and Asymmetries of Age

Effects in Naturalistic and Instructed
L2 Learning

CARMEN MUÑOZ

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

University of Barcelona

The effects of age on second language acquisition constitute one of the most
frequently researched and debated topics in the field of Second Language
Acquisition. Two different orientations may be distinguished in age-related
research: one which aims to elucidate the existence and characteristics of
maturational constraints on the human capacity for learning second languages,
and another which purports to identify age-related differences in foreign
language learning, often with the aim of informing educational policy decisions.
Because of the dominant role of theoretically-oriented ultimate attainment
studies, it may be argued that research findings from naturalistic learning
contexts have somehow been hastily generalized to formal learning contexts.
This paper presents an analysis of the symmetries and asymmetries that exist
between a naturalistic learning setting and a foreign language learning setting
with respect to those variables that are crucial in the discussion of age effects in
second language acquisition. On the basis of the differences observed, it is
argued that the amount and quality of the input have a significant bearing on
the effects that age of initial learning has on second language learning. It is also
claimed that age-related studies in foreign language learning settings have
yielded significant findings that contribute to the development of an integrated
explanation of age effects on second language acquisition.

The distinction between naturalistic language learning and foreign language

learning is often ignored in discussions of maturational constraints in second
language acquisition. However, it is the contention of this paper that the
differences in the amount and quality of the respective input of the two
learning settings may have a significant influence on the effects that the age
of initial learning has on the outcome of second language learning.1 In
simple terms, naturalistic second language learning may be characterized as
learning through immersion in the second language environment, whereas
foreign language learning may be characterized as formal learning in the
classroom. To expand, a typical foreign language teaching/learning situation
in the schools of many countries may be characterized by means of some or
all of the following features: (i) instruction is limited to 2–4 sessions of
approximately 50 minutes per week;2 (ii) exposure to the target language
during these class periods may be limited in source (mainly the teacher),
quantity (not all teachers use the target language as the language of
CARMEN MUÑOZ 579

communication in the classroom) and quality (there is a large variability in

teachers’ oral fluency and general proficiency);3 (iii) the target language is
not the language of communication between peers;4 and (iv) the target
language is not spoken outside the classroom.5 While a binary distinction
serves the aim of this paper of pinpointing symmetries and asymmetries
between second language learning in the two different settings, it goes
without saying that the situation varies from country to country. Likewise,
although the paper will be focused on the two opposite ends of the

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

distinction, it is expected that the discussion will be relevant to second
language learning that takes place in school immersion situations, in which
input and use of the target language may be said to be limited but to a much
lesser extent.
In order to avoid confusion, in this paper the term second language learning
will be used in its general sense to refer to both learning that occurs in a
naturalistic setting and learning that takes place in a foreign, instructed or
formal setting. When referring to the former situation, the term naturalistic
learning will be used, in opposition to the terms foreign, formal, and instructed,
which will be used in reference to the latter situation. However, the terms
learning and acquisition will be used as synonyms.
Having drawn this first general distinction, we move now to a distinction
that is specific to the field of age effects on second language acquisition—that
between rate of learning and ultimate attainment—and which is drawn from an
extended review of research findings made almost three decades ago
(Krashen et al. 1979). These early research findings, which have been
corroborated by subsequent research, revealed that (i) older children,
adolescents, and adults generally make more rapid progress in the first stages
of the second language acquisition process than younger children,
particularly in morphosyntactic aspects, and that (ii) the younger a L2
learner is when the L2 acquisition process begins, the more successful that
process will be, that is, the more possibilities s/he will have of attaining
native-like proficiency. Therefore, the distinction identifies two different age-
related advantages: an initial rate advantage on the part of older learners
over younger learners, and an ultimate attainment advantage of younger
starters over older starters. Evidence from the latter comes from studies of
naturalistic second language acquisition, the so-called immigrant studies, in
which both phenomena are widely observed. Studies of formal learning in
foreign language settings to date have only provided evidence of the former
advantage, that is, the older learners’ initial rate advantage, in situations in
which older starters proceed faster than younger starters with the same
amount of instruction time (Garcı́a Mayo and Garcı́a Lecumberri 2003;
Muñoz 2006b). The initial rate advantage is also observed in situations in
which older learners begin instruction in the foreign language later than
younger learners. In these cases, as Krashen et al. (1979: 579) note, the
common finding is that children who start learning the foreign language later
eventually catch up to those who begin earlier.
580 SYMMETRIES AND ASYMMETRIES OF AGE EFFECTS

It has been claimed that, from a theoretical point of view, the crucial
notion in the distinction is ultimate attainment. The superior ultimate
attainment of younger starters is seen as evidence of the Critical Period
Hypothesis (CPH) (Lenneberg 1967) and, in general, for maturationally-based
explanations of the biological constraints on second language acquisition.
What is important for a theory of second language acquisition—according
to Long (2005: 291)—is not short-term differences in performance,
but rather long-term differences in capacity for acquisition, that is, ultimate

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

attainment.
Although the term ultimate attainment6 has often been used as a synonym
for nativelike proficiency, Birdsong (2004, 2006) points out that the term
properly refers to the final product of L2 acquisition, whether this is
nativelike attainment or any other outcome, that is, irrespective of the
degree of approximation to the native grammar.7 Although it is difficult to
dissociate the two terms because of their strong association in the literature
in this area, it is certainly desirable both methodologically and conceptually.
In fact, more and more voices in the field of SLA now argue against the
traditional view of nativelikeness and its characterization, claiming that a
bilingual or multilingual competence cannot properly be compared with a
monolingual competence (see Grosjean 1989, 1998; Cook 2002; Birdsong
2005, for a recent discussion of the nativelikeness standard in CPH studies).
It could certainly be claimed that people who have acquired two languages
simultaneously from birth constitute a better yardstick for comparison than
monolingual native speakers.
In the case of foreign language acquisition, where the proficiency levels
attained are known to fall very short of nativelike proficiency, the association
of ultimate attainment and nativelikeness is clearly inappropriate. More
importantly, not even the concept of final product can be transported or
adopted. The first reason for this is that the definition of the final product in
ultimate attainment studies may be said to entail the cessation of learning
(this is why we need a long period of residence to ascertain that L2 users
are not still learning, as discussed below). As in studies of fossilization, or
permanent stabilization (see Long 2003), a necessary methodological
requirement seems to be that cessation of learning appears in spite of
optimal learning conditions (see, among others, Selinker and Lamendella
1979; Han 2004; Han and Odlin 2005; Nakuma 2005). The requirement of
the presence of optimal learning conditions (including input that is neither
quantitatively nor qualitatively limited) is not met in foreign language
settings. Failing this requirement, in the case of classroom learning the
attribution of low achievement solely to starting age does not seem to be
justified. In the same vein, the case for the crucial influence of intensity of
exposure in second language learning has been recently reinforced by
neurolinguistic research that has obtained evidence of the crucial role played
by intensity of exposure in the neural representation of multiple languages
(Perani et al. 2003).
CARMEN MUÑOZ 581

Returning to the generalization concerning the higher ultimate attainment

of younger learners in a naturalistic language setting, we can say that this
implies that the initial advantage of older learners must be progressively
reduced to allow the younger learners’ eventual advantage to emerge. In a
naturalistic immersion setting, research by Snow and Hoefnagel-Höhle
(e.g. 1978) provided evidence of the beginning of the process by which
differences in test results between older and younger learners were being
reduced in a period of approximately 1 year. When comparing children,

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

adolescents, and adults who had been in the L2 community for 6 months,
significant age differences were noticed in favour of the adolescent and adult
learners. However, by the second and third tests, after 4- to 5-month
intervals, the younger learners had begun to catch up with and, in some
tests, overtake the older learners. These findings reinforced the generalization
concerning a higher eventual attainment of younger starters in a naturalistic
learning setting but, as observed by Harley (1998) twenty years later, no
evidence has yet been found that the generalization of the higher eventual
attainment of younger starters in naturalistic settings can be automatically
extended to foreign language settings. This remains the case today. In other
words, no evidence has yet been found that the proficiency of younger
formal learners eventually becomes higher than that of older formal learners
after the same amount of instruction.
It has been suggested that the reason why younger beginners have not
been observed to overtake older beginners is that, in school conditions, there
is less exposure to the target language and the experience is less intensive
(Ellis 1994; Singleton 1995). That is to say, because of the differences in
input in both learning contexts, the period of time needed for younger
formal learners must necessarily be much longer than that needed in a
naturalistic learning setting. It then follows that the older starters’ initial
advantage as well as the younger starters’ eventual advantage must imply
longer periods than in a naturalistic situation. Following this line of
reasoning, Singleton and Ryan (2004) argue that the study by Florander
and Jensen (1969) offers a small amount of evidence that the same
phenomenon of progressive reduction of older L2 learners’ lead over younger
learners also occurs in formal instructional settings. In that study the older
learners’ lead diminishes after 320 hours, although it still remains signifi-
cant. This may indeed constitute evidence of the progressive reduction of
the older learners’ advantage, which has been observed recently with
respect to some language dimensions and after a longer period (more than
700 hours of instruction; see Muñoz 2006b), but it does not yet in itself
constitute proof that younger learners will necessarily outperform older
learners in the longer term. Indeed, if the older learners’ advantage is due
mainly to their superior cognitive development, as is claimed in Muñoz
(2006c), no differences in proficiency are to be expected when differences in
cognitive development also disappear with age. To summarize, there is still
a lack of empirical evidence to date confirming that, after the initial stages
582 SYMMETRIES AND ASYMMETRIES OF AGE EFFECTS

of foreign language learning, younger starters overtake older starters in

school settings.
On the other hand, even in studies where instructional time is not
constant and younger starters have had greater exposure than older starters,
the former do not succeed in showing a significant advantage. In this respect,
Harley (1998: 27) notes that no explanation has yet been provided for why
in school settings ‘the additional time associated with an early headstart has
not been found to provide more substantial long-term proficiency benefits’.

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

In contrast, studies comparing early and late immersion students in Canadian
immersion programmes have observed long-term benefits, particularly in oral
communication. While it is true that late immersion students have been
observed to catch up in aspects of proficiency such as reading comprehension
and writing skills (Lapkin et al. 1980; Swain 1981; Cummins 1983; Harley
1986; Turnbull et al. 1998) and that early learners have sometimes failed to
demonstrate the dramatic long-term advantage that was expected of them,8 it
can be argued that the massive amount of exposure provided by school
immersion has had a significant influence on the results. That is, an early
starting age produces long-term benefits when associated with greater time
and massive exposure, as in immersion programmes, but not when
associated with limited time and exposure, as in typical foreign language
learning classrooms.

THE REQUIRED LENGTH OF EXPOSURE

In ultimate attainment studies, the length of time of residence (LoR) is the
operationalization of the learners’ amount of exposure between age of
acquisition (AoA) or initial age of learning and testing time. In foreign language
learning studies, amount of exposure is operationalized either as the number of
hours of instruction or as courses of instruction, although equating hours of
exposure with hours of instruction is a gross generalization.
Researchers have speculated about the length of time of residence
or immersion in the L2-speaking environment that is required for a learner
to have reached ultimate attainment in naturalistic learning settings.
Importantly, the effects of the LoR variable appear to be limited to an initial
period, after which they diminish or disappear. Krashen et al. (1979) suggest
that length of residence plays little or no role after the first 5–10 years. This is
in line with DeKeyser’s (2000: 503) argumentation that a minimum of 10
years should be required in ultimate attainment studies in order to ensure
that it is ultimate attainment rather than rate effects that are being picked
up. Likewise, using self-estimations of L2 proficiency from an extremely large
sample of census data in the United States, Hakuta et al. (2003) found no
evidence of the effect of LoR on English in individuals who had 10 or more
years of US residence. Stevens (2006) claims, however, that large-scale
empirical analyses have uniformly shown that the positive relationship
between LoR and measures of English language proficiency among adult
CARMEN MUÑOZ 583

immigrants in the United States extends well beyond 5 or 10 years of

residence. This period of 10 (or even more) years would be a proxy—using
Birdsong’s (2004) words—for ultimate attainment, or L2 end state, given the
difficulty of determining when the end state has been reached.
Because the issue of reaching the end state, as discussed in the previous
section, loses its relevance in second language learning in a foreign language
setting, deprived by definition of unlimited input, no conceptual symmetry
can be traced here. In contrast, the length of instructional time is an

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

important matter of contention in relation to the alleged long-term benefits
of an early start, namely the higher eventual attainment of younger starters.
The issue then is to identify the long-term in foreign language learning, that
is, the amount of instructional time that may be considered sufficient for the
study of long-term benefits. Unsurprisingly, studies that have aimed to
observe the long-term effects of starting age in a foreign language classroom
setting have looked for them at the end or close to the end of secondary
education, although the long-term benefits sought have not been found (e.g.
Oller and Nagato 1974; Burstall 1975). More recently, the BAF (Barcelona
Age Factor) project has compared groups of learners after approximately 200,
400, 700, and 800 hours (Muñoz 2006b; Navés 2006). Other recent studies of
foreign language learning have used similar periods (from 600 to 800 hours)
in their longest-term comparisons (Garcı́a Mayo and Garcı́a Lecumberri
2003). Identifying the long term at the end of secondary education,
the natural limit of most students’ foreign language learning process, has
external validity because it represents the longest possible term for the
majority of the population and enables the determination of relevant
educational implications. In addition, it marks the point beyond which
keeping a tight control of variables becomes more and more difficult. In this
respect, studies whose informants are college students who began foreign
language education in elementary or infant school might be credited with
studying longer-term effects. In fact, the amount of exposure time observed
may be more equivalent to the amount known to be required for younger
starters to overtake older starters in naturalistic language acquisition
(18 years, according to Singleton’s (1989) estimations on the basis of the
findings made by Snow and Hoefnagel-Höhle (1978, 1979)). However, the
drawback of these studies is that they may depend too heavily on self-
reports, and they are therefore deprived of reliable information about the
individual learning trajectories. They also inevitably contain significant inter-
individual variability in exposure time in and out of the classroom.
To complete this account of time requirements and age effects, it is also
pertinent to note that experts in the area of L2 speech acquisition have
suggested that simple estimations of LoR may not always constitute a valid
index of L2 input. For example, the study by Flege and Liu (2001) shows
that L2 proficiency increases with LoR only if the L2 speaker is participating
in social settings, such as schools, in which they receive a substantial amount
of input from native speakers of the L2. Thus, very young children who do
584 SYMMETRIES AND ASYMMETRIES OF AGE EFFECTS

not yet attend school may not be exposed to the L2 immediately after arrival,
while some immigrants may reside in L1 ghettos, or may not have a job outside
the home that requires use of the L2. In addition, immigrants are likely to hear
different dialects of the L2, compatriots from the home country who speak the
L2 with a similar foreign accent, and individuals from other L1 backgrounds
with different L1-inspired foreign accents (Flege, in press). From his review
of previous research that has revealed small or non-existent LoR effects
(and significant age of arrival (AoA) effects), Flege infers that LoR may provide

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

a valid index of L2 input only for immigrants who receive a substantial amount
of native-speaker input. Hence, the quality in the input to which all learners are
exposed must also be examined alongside the quantity of that input or LoR.
In a foreign language learning setting, learners do not usually receive
native-speaker input and, when they do, it is not substantial. Moreover,
learners are not exposed to the target language during all hours of
instruction, and the exact proportion of lesson time varies according to the
different educational systems, schools, and teachers. As a result, it may be
argued that the length of instruction alone may not be a valid index of L2
input either, and that the quantity and the quality of the input to which
instructed learners are exposed must also be examined.

INITIAL AGE OF LEARNING AS THE BEGINNING OF

SIGNIFICANT EXPOSURE
Further asymmetries are revealed when considering the initial point of
learning. It has been argued that estimations of LoR should take as the initial
point the beginning of significant exposure.9 In naturalistic L2 acquisition
studies, this corresponds to the age of acquisition, also referred to as age of
onset. Birdsong (2006: 11) notes that age of acquisition is understood as the
age at which learners are immersed in the L2 context, typically as
immigrants.10 This landmark is separated from the age of first exposure,
which can occur in a formal schooling environment, visits to the L2 country,
extended contact with relatives who are L2 speakers and so forth. Although
age at immigration is taken as a proxy for age at onset of second language
learning in studies of L2 acquisition among immigrants (see Stevens 2006),
the age of acquisition may not coincide with the age upon arrival, because
residence does not guarantee exposure to and use of the L2, as seen above.
Therefore, precise estimations of the amount of L2 input are crucial because
it is often observed to be confounded with age of acquisition, since early
learners have typically used the L2 longer than late learners (Flege et al.
1999). The development of detailed questionnaires that measure amount of
L2 (and L1) input (e.g. Flege and MacKay 2004) is an important recent
contribution.
In studies of instructed foreign language acquisition, initial age of learning
is equated with the age at which instruction begins. In order to decide
whether this is ever significant exposure it is wise to examine what is meant by
CARMEN MUÑOZ 585

the term: significant exposure is widely conceived as immersion in the L2

context, which provides learners with a variety of contexts of use and
interaction. In other words, it is considered that learners have significant
exposure when they are able to carry out a variety of speech acts over a wide
range of situations and topics, and to participate in social settings effectively
dominated by the L2 (see Stevens 2006: 681). Following this argument, and
because foreign language learners are deprived of significant exposure of this
sort, initial age of foreign language learning may be equated with the age at

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

which insignificant exposure begins. As in studies in naturalistic settings,
however, it is also methodologically important to measure precisely the amount
of L2 input because, in this type of setting, initial age of learning may be
confounded with amount of exposure when groups of early and late learners
with different amounts of accumulated instruction hours are compared.
When learners move from one kind of language learning setting to the
other, it may be possible to identify an age of first significant exposure that is
different to the age of first insignificant exposure. As noted above, naturalistic
learners may have had some instruction in the target language before
moving to the L2-speaking community. In ultimate attainment studies there
has usually been no role for prior insignificant exposure, and the age of
initial exposure in classroom contexts has not usually been found to be
strongly predictive. For example, in the study by Johnson and Newport
(1989) the learners’ experience of formal learning before arrival in the
L2-country did not correlate significantly with the test scores; see also
Birdsong and Molis (2001). In contrast, the study by Urponen (2004) found
that the age at which learning of English as a foreign language began prior to
arrival in the L2-context, a significant predictor of ultimate attainment, was
in fact better than age upon arrival.
A different type of change of language learning setting occurs when
learners in regular school programmes with limited exposure later join
immersion programmes, as in the study by White and Genesee (1996) in
which French-speaking Canadians who had studied English as a compulsory
school subject for at least the period of secondary education joined an
English-medium university. White and Genesee (1996: 242–3), arguably
(see Long 2005), took this latter point in time as the age of onset, on the
basis that this was the first significant exposure to the target language.
A change of language learning setting may also take place for only a
limited period of time, as when foreign language learners spend a few weeks
in the target language community, usually attending language classes and
sometimes living with a host family. In such circumstances, the stay-abroad
experience may provide significant exposure even if only for a short period of
time, and in some cases the experience is repeated periodically over the
years. Clearly, research is needed that looks into the impact that these short
stays make on learners’ L2 proficiency in relation to their initial age of
learning (with limited exposure) and their age at the beginning of significant
exposure.
586 SYMMETRIES AND ASYMMETRIES OF AGE EFFECTS

Turning now to examine the relationship between age of acquisition and

second language outcomes, in naturalistic studies there is a wide consensus
about the importance of initial age of learning on second language achieve-
ment (for recent reviews of experimental studies, see Birdsong (2005) and
DeKeyser and Larson-Hall (2005)). For many researchers who defend the
existence of a maturationally defined critical or sensitive period, a necessary
though not sufficient condition for nativelikeness (see Hyltenstam and
Abrahamsson 2003) is to begin learning the second language during the

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

sensitive period (see also Flege et al. 1999). For example, in Long’s (1990)
formulation of the maturational state hypothesis, it is explicitly stated that
even beginning to learn (my emphasis) the second language during that period
is sufficient.
In contrast, in foreign language learning studies the consensus concerning
the relationship between age of acquisition and second language outcomes is
that older starters have a faster rate of learning (e.g. Cenoz 2002, 2003;
Garcı́a Lecumberri and Gallardo 2003; Garcı́a Mayo 2003; Lasagabaster and
Doiz 2003; Muñoz 2003, 2006c; Navés et al. 2003; Perales et al. 2004; Álvarez
2006; Miralpeix 2006; Mora 2006; Torras et al. 2006). As seen above, no
evidence exists that an early start in foreign language learning leads to higher
proficiency levels after the same amount of instructional time, and even
younger starters with more instructional time have often failed to show a
particularly substantial advantage in terms of long-term proficiency benefits.
To discuss this asymmetry one can turn to the formulation of the
maturational state hypothesis above. It can certainly be argued that the
condition that it is sufficient to begin learning the second language during
the sensitive period could only be relevant to foreign language learning if,
during the privileged period, children have access to significant exposure,
which marks the beginning of acquisition. Similar considerations have led
researchers to argue that the CPH is irrelevant to formal language acquisition
(Patkowski 1994, 2003; Lightbown 2000; DeKeyser and Larson-Hall
2005: 101). DeKeyser’s argument (e.g. 2000, 2003) is based on an
interpretation of the CPH according to which it applies only to implicit
language acquisition (acquisition from mere exposure to the language).
Children necessarily learn implicitly, DeKeyser argues, and implicit learning
requires massive input. But in regular school programmes children are not
provided with the massive amounts of input which constitute a necessary
condition for triggering the implicit learning mechanisms that characterize
their language learning capacity (2000: 520). Further, while a lack of
significant exposure in the classroom deprives young learners of the
possibility of using implicit learning mechanisms, the older learners’ superior
cognitive development allows them to make more efficient use of faster,
explicit learning mechanisms (Muñoz 2001, 2006c).
To conclude, turning again to the maturational state hypothesis and its
applicability to foreign language learning settings, it is arguable that age at
the initial point of learning has as much influence in that context as in
CARMEN MUÑOZ 587

naturalistic learning settings. In fact, with low intensity of exposure, foreign

language learning proceeds in a slow cumulative way through multiple and
discontinuous short amounts of limited exposure. Therefore, in a low-
intensity or distributed programme, the whole age range that extends during
the period in which foreign language learning takes place may have a
stronger influence on the process and the outcome of second language
learning than the initial point of learning.

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

LEARNERS’ CHRONOLOGICAL AGE
Learners’ chronological age at time of testing (AT) is a biographical variable
that has drawn the attention of a number of researchers, most of whom have
been critical of the methodologies used in studies of the effects of age on
second language acquisition. It has been argued that AT may impact on
ultimate attainment by confounding with cognitive factors, education, and
other background variables (Bialystok and Hakuta 1999). Hakuta et al. (2003)
attribute the steady decline across the life-span (from age 5 to 60) in their
census data to cognitive declines of older learners, which ‘gradually erodes
some of the mechanisms necessary for learning a complex body of
knowledge, such as a new language’ (2003: 31). The learning abilities that
are relevant to language learning include, according to Hakuta et al. (2003),
the ability to learn paired associates (Salthouse 1992), the ability to encode
new information (Craik and Jennings 1992) and to recall detail as opposed
to gist (Hultsch and Dixon 1990). In a similar vein, Birdsong (2006) has
highlighted the difference between the effects of aging on L2 learning
(i.e. effects in the temporal-associative areas of the brain, particularly) from
the effects of maturation on L2 learning, that is, the probability of nativelike
ultimate attainment if learning begins before the end of the sensitive period.
Other researchers, however, have argued against the claim that the
confound between age of onset or age on arrival and age at testing may
partly explain the inverse correlations found between AoA and test scores in
some ultimate attainment studies. Such a claim would diminish the strength
of the effects of maturational constraints. Specifically, DeKeyser (2000)
reports a study in which the correlation between test scores and AoA is still
significant with age at test partialled out, whereas that between test scores
and AT is non-significant.
A new perspective on the interplay of these variables has been brought to
the field of SLA by Stevens’s (2006) discussion of the linear dependency of
age at immigration, length of residence, and age at testing. The linear dependency
among trios of variables has been extensively discussed by demographers
in relation to what is known as the age–period–cohort problem or conundrum
(see e.g. Wilson and Gove 1999). This linear dependency is expressed by
Stevens (2006: 672) by means of an equation, which can be rewritten as
AT ¼ LoR þ AoA. A methodological consequence is that the age–length–onset
problem cannot be unambiguously addressed by correlational analyses,
588 SYMMETRIES AND ASYMMETRIES OF AGE EFFECTS

and hence it requires a more careful consideration of the concepts and

conceptual processes that these three variables index. Stevens (2006) uses
Johnson and Newport’s (1989) landmark study to illustrate how difficult it is
to reach unambiguous conclusions when investigating the relationship
between age at arrival and L2 proficiency among adult immigrants. Although
that classic study considers only age at arrival and length of residence,
Stevens argues that the interdependence between age, length of residence,
and age at onset entails a negative relationship between chronological age

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

and L2 proficiency, which provides grounds for some ambiguity about the
conclusion concerning the importance of age at onset, interpreted as the only
predictive factor. In addition, Stevens claims that the negative relationship
may be explained by the fact that the participants in Johnson and Newport’s
study were faculty and students at a university; the older participants in the
study were more likely to be faculty and the younger ones to be students.
The types of experience the two groups have with the language may be very
different and much more intense and demanding in the case of students than
for the faculty, which may explain, at least in part, the relationship between
age at immigration and success in L2 acquisition found in the study (see also
Bialystok and Hakuta 1994, 1999).
In instructed foreign language learning, AT may also be seen as
confounding with age of onset, but in this case it has a negative effect on
the performance of the youngest learners in comparison with older learners
in school settings, and thus contributes to the positive relationship between
L2 proficiency and older age of learning. Specifically, AT may have an impact
on test-taking skills, favouring older children, adolescents, and adults over
younger children when taking language tests. In consequence, the older
learners’ higher scores in cognitively-demanding tests might be attributable
not only to a language proficiency advantage over younger learners, but also
to their superior cognitive development, which helps them achieve a better
understanding of the task in comparison with younger learners as well as
choose better strategies for accomplishing the learning task. This highlights
the importance of the type of tasks that can be used with different-age
learners in order to ensure that the effects of AT are not confounded with
AoA effects (see Muñoz 2006c).
Chronological age is not just an indicator of biological processes associated
with senescence; it is also, as illustrated in the discussion above, an excellent
indicator of life-cycle stage, strongly associated with motivations and oppor-
tunities to speak and to maintain or improve proficiency in an L2 (Stevens
2006: 684). In instructed foreign language learning chronological age may
also be seen as associated with attitudes towards learning the language, the
teacher, and the classroom. A confound between initial age of learning and
chronological age may also be suspected in studies that analyse learner
motivation in relation to age. To illustrate, it may be argued that adult
beginners who choose to enrol on a foreign language course may possess
stronger motivation than younger learners following compulsory education
CARMEN MUÑOZ 589

courses, as it has been argued that students in late immersion programmes

may be more strongly motivated than students in early immersion
programmes (Turnbull et al. 1998). On the other hand, it has often been
reported that younger school learners have a more positive attitude towards
a foreign language than older school learners, as shown by their answers to
questionnaires on motivation, and that this is a definite advantage of an early
start (e.g. Hawkins 1996; Blondin et al. 1998). However, the young learners’
positive attitude may be seen as associated with chronological age rather

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

than, or as well as, age at onset. It is well known that young children are
more eager to please the adult than pubescent and adolescent learners and,
consequently, the former may be more inclined than the latter to respond
positively to motivation-related questions. Thus, the relatively common
finding that younger starters have a more positive attitude towards learning a
second language than older starters may be a result of their chronological age
(they are younger) when they answer the motivation questionnaire rather
than or in addition to their earlier start. This is clearly an empirical question
since evidence could be provided by studies that asked attitudinal questions
to learners with different AoA when they were the same chronological age.
It was mentioned above that the amount of formal education at AT has
been observed to be a very good predictor of second language learning
among immigrants. Strong evidence has been provided by the large-scale
study using census data in US conducted by Hakuta et al. (2003). Flege and
coworkers have also considered the amount of education in the host country
in their studies of L2 speech learning. For example, in the study conducted
by Flege et al. (1999), AoA effects disappeared when education was
controlled. In a more recent study, Flege et al. (2005) observe the influence
on foreign accent of variables that were confounded with AoA and find that
AoA was correlated with length of residence in Canada, which was correlated
with years of education in the host country.
In instructed foreign language learning, the type and characteristics of
formal education may have a strong influence on learners’ proficiency levels.
For one thing, in many countries, socioeconomic background may determine
a child’s initial age of foreign language learning through the choice of schools
(privately- or state-funded) as well as the amount of extracurricular exposure
learners can enjoy (in the form of extracurricular lessons, technological
devices in the home, and studies abroad, mainly). More importantly, the
educational level of parents has a significant influence on children’s foreign
language learning success (see for example the PISA report in Europe 2003),
which suggests a parallel with the role played by the formal education of
immigrant learners in studies of naturalistic L2 acquisition. Because children
of parents from different socioeconomic backgrounds and with different
educational levels are not homogeneously distributed in all types of school,
the effects of AoA may be confounded with the effects of socioeconomic and
sociocultural variables which result in varying availability of curricular and
extracurricular exposure to the foreign language (Muñoz 2006a).
590 SYMMETRIES AND ASYMMETRIES OF AGE EFFECTS

CONCLUSION
This paper has presented an analysis of symmetries and asymmetries that
exist between a naturalistic learning setting and a foreign language learning
setting with respect to those variables that are crucial in the discussion of age
effects in second language acquisition. The analysis has highlighted important
differences that have largely been ignored in such discussion; the reason
being that due to the higher theoretical relevance granted to ultimate

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

attainment studies in the field of SLA, the findings of classroom studies have
been interpreted in the light of the assumptions and priorities of the former.
One notable asymmetry is derived from the different status of ultimate
attainment or final product in the two types of learning setting. Studies of
age effects on naturalistic second language learning compare younger and
older starters in terms of the final product of their second language learning
process, bounded by a minimum length of time (i.e. at least 10 years of
residence). In contrast, age-related studies in a foreign language setting, in
which the conditions for reaching the end state are not met, compare the
gains of younger and older learners after different lengths of time with an
emphasis on long-term benefits. In such studies, long-term benefits are often
temporally bounded by the educational system itself, which does not usually
require foreign language learning beyond secondary education. In con-
sequence, the amount that can be learned in such a limited period and with
such limited exposure—that is, rate of learning—becomes a crucial concern.
In relation to the long-term benefits of an early start, it is important to
underline the absence of empirical evidence to date confirming that younger
starters overtake older starters in school settings with the same amount of
input. Nevertheless, the generalization has been accepted in the field of SLA
and extended beyond the field that, as occurs in a naturalistic learning
setting, younger starters will overtake older starters in the long run. Likewise,
it has been tacitly agreed that when younger starters show better outcomes
than older starters in situations in which the former have had a greater
amount of exposure or instruction, this is due to the learners’ earlier starting
age and not to the greater instruction time to which they have had access
due to their earlier start, ignoring this methodological flaw in variable
control.
Noticeable asymmetries are also found in the amount of exposure between
the two types of learning setting. While in naturalistic second language learning
settings input is generally unlimited, in foreign language settings input is, by
definition, limited and it is usually distributed in very small doses. Therefore, it
may be hard to find equivalent amounts of exposure, and research that takes
into account such long periods of time (i.e. 18 years of uninterrupted
instruction) is difficult to conduct with methodological rigour. Similarly, it has
been pointed out in this paper that while in naturalistic second language
learning studies initial age of learning is taken to be the age of first significant
exposure, foreign language learning studies irremediably take as the initial age
CARMEN MUÑOZ 591

of learning the age of first insignificant exposure. A consequence of these

considerable differences in amount and type of exposure is that a foreign
language learning setting cannot provide the optimal learning conditions (input
that is neither quantitatively nor qualitatively limited) that constitute the
methodological requirement for establishing the cessation of learning.
Notwithstanding, the generalization has again been made that the compara-
tively lower levels of language proficiency in classroom settings are due to
starting age, and that they will automatically be improved by lowering the

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

initial age of learning, disregarding the crucial role played by intensity of
exposure in language learning.
In summary, for a number of years now in the field of SLA there has been
a tacit acceptance of sweeping generalizations of findings from natural
settings to classroom settings that have not been upheld by research into the
latter.11 An inferential leap has been made in the assumption that learning
age will have the same effect on learners in an immersion setting as on
students of a foreign language, when the latter are exposed to only one
speaker of that language (the teacher) in only one setting (the classroom)
and for only limited amounts of time (Muñoz 2006b). However, recent
studies conducted in foreign language settings have clearly illustrated the role
of input and exposure in the equation: an early start leads to success but only
provided that it is associated with enough significant exposure.
Not only have findings been generalized; the aims and priorities of research in
naturalistic settings have also been extended to research in classroom settings.
As has been discussed in this paper, the aim of age-related classroom-based
research cannot be the same as that of ultimate attainment studies, namely, to
provide ‘evidence for the existence, scope or timing of maturational constraints
on the human capacity for learning second (including foreign) languages’
(Long 2005: 288). This recognition frees age-related classroom research from
endeavours which strictly belong to the area of CPH studies and provides it with
a clearer focus. Similarly, legitimate priorities for research in foreign language
learning settings that should have implications for guiding age-related
classroom research may be: (i) to determine the amount of input that is
required for an early start to be effective in promoting language learning; (ii) to
focus on the relative gains of different-age pupils with different types of time
distribution, that is the advantages of intensive learning programmes that may
be plausible in schools; (iii) the distinction between short-term and long-term
benefits of starting at different ages; and (iv) the comparative study of the
learning rate of different-age learners, which should inform educators about
what to expect after 4, 8, or n years of foreign language instruction from
different-age learners.
To conclude, the field of SLA should now be mature enough to recognize
that it is ultimately the findings of studies in actual classrooms that are most
relevant to decisions concerning the time and timing of second language
instruction. It is these findings that have yielded a more complete picture of
the factors that lead to success in second language learning, providing
592 SYMMETRIES AND ASYMMETRIES OF AGE EFFECTS

evidence of the essential role of input, and thus contributing to the

development of an integrated explanation of age effects on second language
acquisition.

ACKNOWLEDGEMENTS
This work was supported by research grants HUM2004–05167 and 2005SGR00778. The author

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

thanks Teresa Navés, two anonymous reviewers, and the editors for their comments on the
manuscript.

NOTES 5 Variation among countries is large

and in the case of English the
1 The term second language will be used situation is more correctly depicted
in this paper in a general sense to as a gradient or a cline (Berns 1990).
include all its different aspects See also Bonnet (2002).
(from phonology to discourse) not- 6 Other terms used more or less
withstanding the differences in the interchangeably in the literature on
way they are learned and the speed naturalistic L2 learning are end state,
and degree of success with which they final state, steady state, and asymptote
are learned. (see Birdsong 2006: 11).
2 For example, in Europe the average 7 On the other hand, and on the basis
number of foreign language classes is of their recent research findings,
between 30 and 50 in primary educa- Abrahamsson and Hyltenstam (2006)
tion, and a little over 90 hours in argue that nativelikeness alone may
secondary education (see Eurydice not be enough to assess the CPH, and
2005). This amounts in many coun- claim that language aptitude seems
tries to around 800 hours by the end crucial for reaching nativelikeness.
of compulsory education. 8 Cummins (1980, 1981) and Cummins
3 For example, in a study assessing and Swain (1986) argued that older
pupils’ skills in English in eight learners show greater mastery of L2
European countries (Bonnet 2002) it syntax, morphology and other lit-
is observed that ‘in most countries it is eracy-related skills, such as vocabu-
basically the teacher who does the lary and reading comprehension, due
talking’ (p. 93) and that teachers’ use to their greater cognitive maturity.
of the target language varies among 9 MacWhinney (2005: 136) qualifies it
countries: only 15% of the Spanish as serious exposure.
teachers state that they always speak 10 In that respect, Long (2005) argues
in English in their lessons, whereas that estimations of LoR should take
40% of teachers from Norway, as the initial point the beginning of
Sweden, or Denmark usually do so. significant exposure even in cases
4 For example, in the study referred to in which periods spent in the
in note 4 (Bonnet 2002) only 8.5% of L2 environment are interrupted
pupils state that they work in groups or discontinuous, such as when
and talk in the target language most of learners spend different periods of
the time. residence in the L2 community.
CARMEN MUÑOZ 593

11 This is not exactly parallel to direct use of results from basic

the divide between pure and applied theoretical L2 research in dealing
research in SLA (see, for example, with applied questions generated
Sharwood-Smith 2004: 5), although in instructional language learning
some of the features of the latter may situations. I thank an anonymous
be reminiscent of it, such as the reviewer for the suggestion.

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

REFERENCES
Abrahamsson, N. and K. Hyltenstam. 2006. T. Taeschner. 1998. Foreign Languages in Primary
‘The role of language learning aptitude for and Pre-School Education. London: CILT.
near-native L2 attainment.’ Paper presented in Bonnet, G. (ed.) 2002. The Assessment of Pupils’ Skills
the Joint AAAL and ACLA/CAAL Conference, in English in Eight European Countries. Project
Montréal, 17–20 June. commissioned by the European Network of
Álvarez, E. 2006. ‘Rate and route of acquisition in Policy Makers for the Evaluation of Education
EFL narrative development at different ages’ in Systems. http://cisad.adc.education.fr/reva/pdf/
C. Muñoz (ed.): Age and the Rate of Foreign assessmentofenglish.pdf.
Language Learning. Clevedon: Multilingual Mat- Burstall, C. 1975. ‘Primary French in the balance,’
ters, pp.127–55. Foreign Language Annals 10/3: 245–52.
Berns, M. 1990. ‘ ‘‘Second’’ and ‘‘foreign’’ in Cenoz, J. 2002. ‘Age differences in foreign lan-
second language acquisition/foreign language guage learning,’ ITL Review of Applied Linguistics
learning: a sociolinguistic perspective’ in B. Van 125–42, 135–6.
Patten and J. F. Lee (eds): Second Language Cenoz, J. 2003. ‘The influence of age on the
Acquisition–Foreign Language Learning. Clevedon: acquisition of English’ in M. P. Garcı́a Mayo
Multilingual Matters. and M. L. Garcı́a Lecumberri (eds): Age and the
Bialystok, E. and K. Hakuta. 1994. In Other Words: Acquisition of English as a Foreign Language.
The Science and Psychology of Second-language Acqui- Clevedon: Multilingual Matters, pp. 77–93.
sition. New York: Basic Books. Cook, V. 2002. ‘Background to the L2 user’ in
Bialystok, E. and K. Hakuta. 1999. ‘Confounded V. Cook (ed.): Portraits of the L2 User. Clevedon:
age: Linguistic and cognitive factors in age Multilingual Matters, pp. 1–28.
differences for second language acquisition’ in Craik, F. I. M. and J. M. Jennings. 1992. ‘Human
D. Birdsong (ed.): Second Language Acquisition and memory’ in F. I. M. Craik and T. A. Salthouse
the Critical Period Hypothesis. Mahwah, NJ: (eds): The Handbook of Aging and Cognition.
Erlbaum, pp. 161–81. Hillsdale, NJ: Erlbaum, pp. 51–110.
Birdsong, D. 2004. ‘Second language acquisition Cummins, J. 1980. ‘The cross-lingual dimensions
and ultimate attainment’ in A. Davies and of language proficiency: Implications for bilin-
C. Elder (eds): The Handbook of Applied Linguistics. gual education and the optimal age issue,’ TESOL
Malden, MA: Blackwell Publishers, pp. 82–105. Quarterly 14/2: 175–87.
Birdsong, D. 2005. ‘Nativelikeness and non- Cummins, J. 1981. ‘Age on arrival and immigrant
nativelikeness in L2A research,’ International second language learning in Canada. A reassess-
Review of Applied Linguistics 43/4: 319–28. ment,’ Applied Linguistics 11/2: 132–49.
Birdsong, D. 2006. ‘Age and second language Cummins, J. 1983. ‘Language proficiency, biliter-
acquisition and processing: A selective over- acy and French immersion,’ Canadian Journal of
view,’ Language Learning 56/1: 9–49. Education 8/2: 117–38.
Birdsong, D. and M. Molis. 2001. ‘On the Cummins, J. and M. Swain. 1986. ‘Bilingualism in
evidence for maturational constraints on Education’. London: Longman.
second-language acquisition,’ Journal of Memory DeKeyser, R. 2000. ‘The robustness of critical
and Language 44: 235–49. period effects in second language acquisition,’
Blondin, C., M. Candelier, P. Edelenbos, Studies in Second Language Acquisition 22:
R. Johnstone, A. Kubanek-German, and 499–533.
594 SYMMETRIES AND ASYMMETRIES OF AGE EFFECTS

DeKeyser, R. 2003. ‘Implicit and explicit learning’ Language. Clevedon: Multilingual Matters,
in C. J. Doughty and M. H. Long (eds): Hand- pp. 94–114.
book of Second Language Acquisition. London: Garcı́a Mayo, M. P. and M. L. Garcı́a Lecum-
Blackwell, pp. 313–48. berri (eds). 2003. Age and the Acquisition of English
DeKeyser, R. and J. Larson-Hall. 2005. ‘What as a Foreign Language. Clevedon: Multilingual
does the critical period really mean?’ in J. F. Kroll Matters.
and A. M. B. de Groot (eds): Handbook of Bilin- Grosjean, F. 1989. ‘Neurolinguists, beware! The
gualism: Psycholinguistic Approaches. Oxford: bilingual is not two monolinguals in one person,’
Oxford University Press, pp. 88–108. Brain and Language 36: 3–15.
Ellis, R. 1994. The Study of Second Language Acquisi- Grosjean, F. 1998. ‘Studying bilinguals: Methodo-

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

tion. Oxford: Oxford University Press. logical and conceptual issues,’ Bilingualism:
Eurydice. 2005. Key Data on Teaching Languages at Language and Cognition 1: 131–49.
School in Europe. European Commission. Hakuta, K., E. Bialystok, and E. Wiley. 2003.
Flege, J. in press. ‘Give input a chance!’ in ‘Critical evidence: A test of the Critical Period
M. Young-Scholten and T. Piske (eds): Input Hypothesis for second language acquisition,’
Matters. Clevedon: Multilingual Matters. Psychological Science 14: 31–8.
Flege, J. and S. Liu. 2001. ‘The effect of experience Han, Z.-H. 2004. Fossilization in Adult Second
on adults’ acquisition of a second language,’ Language Acquisition. Clevedon: Multilingual
Studies in Second Language Acquisition 23/4: Matters.
527–52. Han, Z.-H. and T. Odlin. 2005. ‘Introduction’ in
Flege, J. and I. MacKay. 2004. ‘Perceiving vowels Z.-H. Han and T. Odlin (eds): Studies of Fossiliza-
in a second language,’ Studies in Second Language tion in Second Language Acquisition. Clevedon:
Acquisition 26: 1–34. Multilingual Matters, pp. 1–20.
Flege, J., G. Yeni-Komshian, and S. Liu. 1999. Harley, B. 1986. Age in Second Language Acquisition.
‘Age constraints on second language learning,’ Clevedon: Multilingual Matters.
Journal of Memory and Language 41: 78–104. Harley, B. 1998. ‘The outcomes of early and later
Flege, J., I. MacKay, and S. Imai. 2005. ‘What language learning’ in M. Med (ed.): Critical
accounts for ‘‘age’’ effects on overall degree Issues in Early Second Language Learning. Scott
foreign accent?’ Manuscript submitted for pub- Foresman-Addison Wesley, pp. 26–31.
lication. Hawkins, E. 1996. ‘The early teaching of modern
Florander, J. and M. Jensen. 1969. Skoleforsøg i languages. A pilot scheme’ in E. Hawkins (ed.):
engelsk 1959–1965. (School experiments in English Thirty Years of Language Teaching. London: CILT,
1959–1965). Stencilled. Denmark’s Institute of pp. 155–64.
Education. Cited in Ekstrand (1976) ‘Age and Hultsch, D. F. and R. A. Dixon. 1990. ‘Learning
length of residence as variables related to the and memory in aging’ in J. E. Birren and
adjustment of migrant children with special K. W. Schaie (eds): Handbook of the Psychology of
reference to second language learning’ in Aging 3rd edn. San Diego, CA: Academic Press,
G. Nickel (ed.) Proceedings of the Fourth Interna- pp. 259–74.
tional Congress of Applied Linguistics. Volume 3. Hyltenstam, K. and N. Abrahamsson. 2003.
Stuttgart: Hochschulverlag. Reprinted in ‘Maturational constraints in SLA’ in C. Doughty
S. Krashen, R. Scarcella and M. Long (eds). and M. Long (eds): The Handbook of Second
1982. Child–Adult Differences in Second Language Language Acquisition. Malden, MA and Oxford,
Acquisition. Rowley, MA: Newbury House. UK: Blackwell, pp. 539–88.
Garcı́a Lecumberri, M. L. and F. Gallardo. 2003. Johnson, J. and E. Newport. 1989. ‘Critical
‘English FL sounds in school learners of different period effects in second language learning: The
ages’ in M. P. Garcı́a Mayo and M. L. Garcı́a influence of maturational state on the acquisi-
Lecumberri (eds): Age and the Acquisition of tion of English as a second language,’ Cognitive
English as a Foreign Language. Clevedon: Multi- Psychology 21: 60–99.
lingual Matters, pp. 115–35. Krashen, S. 1981. Second Language Acquisition and
Garcı́a Mayo, M. P. 2003. ‘Age, length of exposure Second Language Learning. Oxford: Pergamon.
and grammaticality judgements in the acqui- Krashen, S., M. Long, and R. Scarcella. 1979.
sition of English as a foreign language’ in ‘Age, rate, and eventual attainment in second
M. P. Garcı́a Mayo and M. L. Garcı́a Lecumberri language acquisition,’ TESOL Quarterly 13/4:
(eds): Age and the Acquisition of English as a Foreign 573–82.
CARMEN MUÑOZ 595

Lapkin, S., M. Swain, J. Kamin, and G. Hanna. Muñoz, C. (ed.) 2006b. Age and the Rate of
1980. Report on the 1979 evaluation of the Peel Foreign Language Learning. Clevedon: Multilin-
County late French immersion program, grades gual Matters.
8, 10, 11 and 12. Unpublished report, University Muñoz, C. 2006c. ‘The effects of age on foreign
of Toronto, OISE. language learning: The BAF Project’ in C. Muñoz
Lasagabaster, D. and A. Doiz. 2003. ‘Matura- (ed.): Age and the Rate of Foreign Language Learn-
tional constraints on foreign-language written ing. Clevedon: Multilingual Matters, pp. 1–40.
production’ in M. P. Garcı́a Mayo and Nakuma, C. K. 2005. ‘Attrition: easy
M. L. Garcı́a Lecumberri (eds): Age and questions, difficult answers’ in Z.-H. Han and
the Acquisition of English as a Foreign Language. T. Odlin (eds): Studies of Fossilization in Second

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

Clevedon: Multilingual Matters, pp. 136–60. Language Acquisition. Clevedon: Multilingual
Lenneberg, E. H. 1967. Biological Foundations of Matters, pp. 21–34.
Language. New York: Wiley. Navés, T. 2006. ‘The long-term effects of an early
Lightbown, P. 2000. ‘Classroom second language start on writing in foreign language learning.’
acquisition research and second language teach- Unpublished doctoral disertation. University of
ing,’ Applied Linguistics 21/4: 431–82. Barcelona.
Long, M. 1990. ‘Maturational constraints on lan- Navés, T., M. R. Torras, and M. L. Celaya. 2003.
guage development,’ Studies in Second Language ‘Long-term effects of an earlier start. An Analysis
Acquisition 12: 251–85. of EFL written production’ in S. Foster-Cohen
Long, M. 2003. ‘Stabilization and fossilization in and S. Pekarek (eds): Eurosla-Yearbook, vol. 3.
interlanguage development’ in C. Doughty and Amsterdam: John Benjamins, pp. 103–30.
M. Long (eds): The Handbook of Second Language Oller, J. and N. Nagato. 1974. ‘The long-term
Acquisition. Oxford: Blackwell, pp. 487–536. effect of FLES: An experiment,’ Modern Language
Long, M. 2005. ‘Problems with supposed counter- Journal 58: 15–19.
evidence to the Critical Period Hypothesis,’ Patkowski, M. S. 1994. ‘The critical period
International Review of Applied Linguistics 43/4: hypothesis and interlanguage phonology’ in
287–317. M. Yavas (ed.): First and Second Language Phonol-
MacWhinney, B. 2005. ‘Emergent fossilization’ in ogy. San Diego: Singular, pp. 205–21.
Z.-H. Han and T. Odlin (eds): Studies of Fossiliza- Patkowski, M. S. 2003. ‘Laterality effects in multi-
tion in Second Language Acquisition. Clevedon: linguals during speech production under the
Multilingual Matters, pp. 134–56. concurrent task paradigm: Another test of the
Miralpeix, I. 2006. ‘Age and vocabulary acquisi- age of acquisition,’ International Review of Applied
tion in English as a foreign language’ in C. Muñoz Linguistics 41/3: 175–200.
(ed.): Age and the Rate of Foreign Language Learn- Perales, S., M. P. Garcı́a Mayo, and J. M. Liceras.
ing. Clevedon: Multilingual Matters, pp. 89–106. 2004. ‘The acquisition of English sentential
Mora, J. C. 2006. ‘Age effects on oral fluency negation by bilingual (Basque/Spanish) chil-
development’ in C. Muñoz (ed.): Age and the dren: The age factor in an institutional setting.’
Rate of Foreign Language Learning. Clevedon: Paper presented at the Annual Meeting of the
Multilingual Matters, pp. 65–88. American Association of Applied Linguistics,
Muñoz, C. 2001. ‘Factores escolares e individuales Portland, OR, May 1–4.
en el aprendizaje formal de un idioma extranjero’ Perani, D., J. Abulatebi, E. Paulesu, S. Bram-
in S. Pastor and V. Salazar (eds): Estudios de bati, P. Scifo, S. F. Cappa, and F. Fazio. 2003.
Lingüı́stica. Universidad de Alicante, Anexo 1: ‘The role of age of acquisition and language usage
249–270. in early, high-proficient bilinguals: An fMRI
Muñoz, C. 2003. ‘Variation in oral skills develop- study during verbal fluency,’ Human Brain Map-
ment and age of onset’ in M. P. Garcı́a Mayo ping 19: 170–82.
and M. L. Garcı́a Lecumberri (eds): Age and PISA. 2003. Learning for Tomorrow’s World—First
the Acquisition of English as a Foreign Language. Results from Pisa 2003. OECD Publishing.
Clevedon: Multilingual Matters, pp. 161–81. Salthouse, T. A. 1992. Mechanisms of Age–Cognition
Muñoz, C. 2006a. ‘Investigating age-related differ- Relations in Adulthood. Hillsdale, NJ: Erlbaum.
ences in SLA: Methodological issues’. Paper pre- Selinker, L. and J. Lamendella. 1979. ‘The role of
sented in the Symposium Doing SLA Research: extrinsic feedback in interlanguage fossilization:
Theoretical, Methodological and Ethical Issues. A discussion of ‘‘rule fossilization: A tentative
AESLA Conference, Madrid, 31 March. model’’’, Language Learning 29/2: 363–75.
596 SYMMETRIES AND ASYMMETRIES OF AGE EFFECTS

Sharwood-Smith, M. 2004. Second Language Learn- Swain, M. 1981. ‘Time and timing in bilingual
ing: Theoretical Foundations. London: Longman. education,’ Language Learning 31: 1–16.
Singleton, D. 1989. Language Acquisition: The Age Torras, M. R., T. Navés, M. L. Celaya, and
Factor. Clevedon: Multilingual Matters. C. Pérez-Vidal. 2006. ‘Age and IL development
Singleton, D. 1995. ‘A critical look at the Critical in writing’ in C. Muñoz (ed.): Age and the Rate of
Period Hypothesis in second language Foreign Language Learning. Clevedon: Multilin-
acquisition research’ in D. Singleton and Z. gual Matters, pp. 156–82.
Lengyel (eds): The Age Factor in Second Language Turnbull, M., S. Lapkin, D. Hart, and M. Swain.
Acquisition. Clevedon: Multilingual Matters, 1998. ‘Time on task and immersion graduates’
pp. 1–29. French proficiency’ in S. Lapkin (ed.): French

Downloaded from http://applij.oxfordjournals.org/ at Columbia University Libraries on November 14, 2014

Singleton, D. and L. Ryan. 2004. Language Acqui- Second Language Education in Canada: Empirical
sition: The Age Factor 2nd edn. Clevedon: Multi- Studies. Toronto, ON: University of Toronto Press,
lingual Matters. pp. 31–55.
Snow, C. and M. Hoefnagel-Höhle. 1978. ‘The Urponen, M. I. 2004. Ultimate Attainment in
critical period for language acquisition: Evidence Postpuberty Second Language Acquisition. Boston
from second language learning,’ Child Develop- University: Boston.
ment 49: 1114–28. White, L. and F. Genesee. 1996. ‘How native is
Snow, C. and M. Hoefnagel-Höhle. 1979. ‘Indi- near-native? The issue of ultimate attainment
vidual differences in second language ability: in adult second language acquisition,’ Second
A factor analytic study,’ Language and Speech 22: Language Research 12/3: 233–65.
151–62. Wilson, J. A. and W. R. Gove. 1999. ‘The
Stevens, G. 2006. ‘The age-length-onset problems age–period–cohort conundrum and verbal
in research on second language acquisition ability: Empirical relationships and their
among immigrants,’ Language Learning 56/4: interpretation,’ American Sociological Review 64:
671–92. 287–302.