1. Introduction
Research on the age factor in second language acquisition (SLA)Footnote 1 produces several publications every year. The four books to be examined in detail in this review are all written by leading experts in the field and – taken together – offer a good account of the issues under debate and current developments in the area. At the same time, they provide interesting points of comparison on central aspects of this often controversial topic.
Studies on the age factor in SLA can be broadly classified into two types. According to Carmen Muñoz (Reference Muñoz2008: 578), one type ‘aims to elucidate the existence and characteristics of maturational constraints on the human capacity for learning second languages’, the other ‘purports to identify age-related differences in foreign language learning, often with the aim of informing educational policy decisions’. The books chosen for this comparative review can also be conveniently categorised according to these two approaches.
The first line of research is principally concerned with the existence (or non-existence) of a critical period (CP) for SLA beyond which a second language cannot be learned, or language learning is said to be more difficult or incomplete. Much debate in the field has focused on the CP since Johnson & Newport (Reference Johnson and Newport1989) extended the hypothesis previously posed for L1 acquisition to L2 learning. After several years, defendants and detractors of the hypothesis continue to present arguments for and against the existence of this major constraint on learning languages. A clear example is David Birdsong's Second language acquisition and the Critical Period Hypothesis, which offers a range of studies supporting and opposing the Critical Period Hypothesis (CPH) from different points of view.
Closely linked to the possible existence of a CP, a second central concern in this line of research is that of ultimate attainment (UA): the final level of proficiency achieved in the L2. The argument goes that if maturational constraints operate on the human capacity for learning languages, starting to learn the L2 after the CP would mean not achieving levels of attainment comparable to learners who had started before the end of the CP. The notion of UA has been extensively discussed in the literature and systematically analysed from different angles (see Muñoz & Singleton, this issue). Of particular interest is the work presented by Alene Moyer in Age, accent and experience in second language acquisition, where she concludes that – apart from biological factors – cognitive, social and psychological elements do indeed contribute to explaining attainment. In her book she presents research in an immersion context on advanced learners who started learning the language after the CP. The book analyses one area – foreign accent – in detail and concentrates on those factors that can predict the incidence of accent in the process of learning a language.
The second line of research identified in age factor studies – that dealing with foreign language (FL) learning in instructional settings – has been pursued by more researchers over the last few years. The rise in the number of studies has partly been due to a growing number of governments having lowered the age at which schools first introduce foreign languages, especially English, which has steadily become a lingua franca for communication. Age and the rate of foreign language learning, edited by Muñoz, presents a collection of studies that focus on the learning rate and eventual outcomes of low-proficiency learners in a formal setting. It also offers a comparison of early and late starter groups of FL learners in different areas (e.g. pronunciation, morphology, fluency or learning strategies) and problematises the notion of ‘the younger, the better’ in instructional settings.
The two perspectives on research into the age factor outlined above are both taken into consideration in the book Language acquisition: The age factor by David Singleton and Lisa Ryan, which is a comprehensive investigation of studies on age in SLA. Two of its longest sections deal with L1- and L2-related evidence on the CPH, while the last part focuses on age studies relevant to educational settings.
Without doubt, many other books on the age factor also deserve the attention of the scientific community. However, I believe that, for the purpose of the present review, this selection is representative of a great deal of work being carried out in the field. Each book has its own objectives, although their contents and conclusions can be compared on several dimensions.
2. Overview
A first edition of Singleton & Ryan's book was published by Singleton in 1989, but this updated, co-authored second edition was published in 2004. Although the initial structure was maintained in this version, several sections were expanded, with the aim of being more informative and incorporating significant work from recent years. As the authors state, the book ‘sets out to explore impartially the pertinent data, proposals and speculations in all their diversity’ (p. 5). It is intended for a wide audience, as most of the issues related to age effects in SLA are covered, and provides the reader with sufficient information on each specific related topic. However, some background knowledge in language acquisition and linguistics is necessary to follow the discussions. Those with this initial knowledge but not yet fully acquainted with the intricacies of the age factor will get the most from their reading of the book. An ample bibliography of 49 pages with about 1,000 references – both classic and current – offers the reader the means to extend their understanding.
The book begins with an introduction to the theoretical and practical aspects of research into the age factor. In particular, it emphasises the close connections with central themes such as language innateness and the investigation of a possible CP for language learning. It also highlights the direct implications of the research for L2 teaching and elementary curriculum planning. Before concentrating on the L1- and L2-related evidence on the CPH (chapters 3 and 4), chapter 2 provides a well-informed description of L1 speech acquisition milestones, concentrating on phonology. The authors compare stages in speech development across a number of studies by different researchers, and different explanatory tables are provided. Chapter 5 offers an accurate description of various theoretical explanations for age effects in language learning, such as sensory acuity, cognitive and affective factors, nativism or quality input factors. This theoretical section is followed by a chapter devoted to the role of age in the learning of languages other than the mother tongue in formal contexts.
Clear end-of-chapter summaries are used to highlight what is already known from the literature and what remains uncertain. These summaries, often in the form of lists, enable the reader to review the main trends, research findings and prospects in the areas dealt with in each of the sections. The book closes with a summary of the main ideas from the previous chapters and an emphasis on the still incomplete nature of research in the area.
Birdsong's volume is rather different from Singleton and Ryan's book. Second language acquisition and the Critical Period Hypothesis offers a series of articles selected from a larger number of papers delivered at the symposium New perspectives on the Critical Period for second language acquisition, celebrated in Jyväskyla, Finland, in 1996. It was published three years after this conference and contains chapters written both by proponents and opponents of the CPH in SLA.
The editor mentions in his introduction that he felt it was ‘important to compile recent and largely unpublished research’ and that the compilation offers a ‘balance of competing views’ (p. x). The selection of papers is consistent with these objectives, as half are written by proponents of the CPH and the other half by opponents of the hypothesis. The other aspect he emphasises is diversity because, apart from presenting different views on the CPH, the authors work in a wide range of disciplines and offer data of very different kinds.
The book is especially useful for researchers and applied linguistics graduates, in particular for those already familiar with the CP debate. However, it is not only those investigating the age factor who will be interested, as several research approaches and methodologies are used in the studies included. Among the empirical studies, we find Christine Weber-Fox and Helen Neville, who work in the neurophysiological field and use event-related brain potentials (ERPs) to test their hypotheses. James Hurford and Simon Kirby, working in the framework of evolution theory, base their argumentation on computer simulations. The study conducted by Ellen Bialystok and Kenji Hakuta examines linearity and asymptotic effects of the ages of arrival of immigrants on their ultimate proficiency. The data they use come from the New York State census and from self-reports on the English proficiency of these residents.
Other chapters involve re-examination of previous findings or meta-analyses of major studies in the field. James Flege's paper revisits the findings of previous empirical studies to show that the idea of a CP is inconsistent with certain research results. Similarly, Theo Bongaerts builds on three previous studies to defend the idea that native-like attainment is possible. One of the most theoretically-oriented chapters is by Lynn Eubank and Kevin Gregg, who revise and discuss a number of studies on L2 competence to prove that there are maturational constraints.
Moyer's Age, accent and experience in Second Language Acquisition sets out to explore ‘the significance of socio-cultural and psychological factors, such as identity and milieu, in relation to frequently cited neurobiological and cognitive influences in SLA’ (p. 50). In order to do so, the author studies an immigrant language learning situation, which is said to involve questions of a social, cultural and political nature that affect the language acquisition process. As the study is concerned with the UA of immigrants in naturalistic settings, she starts by discussing the universal and individual factors that may condition this attainment. She also justifies the choice of phonology as the main aspect to be examined on a number of grounds. For example, phonology relies on motor and higher analytical skills and can reveal neurological constraints and individual ability. It is also the element most often relied on when judging nativeness and identity.
In the research conducted, data on linguistic and non-linguistic performance was collected from 25 immigrants in Berlin, who were quite advanced in the study of the language. However, they had started learning German in their home countries at different ages. The mean age of onset (AO), i.e., the age at which the learners started learning the language, was 12. Four production tasks and an interview provided the main data sources for the study, and native raters assessed learners' performance. The outcomes are presented in two different chapters: chapter 3 on the quantitative, chapter 4 on the qualitative results. The quantitative analyses (mainly correlations and multiple regressions) provide evidence that age is closely tied to social and psychological variables, although they also show that no participant with an AO beyond 15 reaches a native level of proficiency, and that AO before age 12 is a variable with a strong predictive power. It would have been appropriate for this chapter to provide more complete information on the scores obtained by individual participants and on the AO of each learner, especially when the book aims at determining the actual function of this variable.
Chapter 4 – on the qualitative findings – is of special significance, as it is devoted to an analysis of how learners identify themselves with the community both in linguistic and in cultural terms. This is an aspect which is not usually surveyed in such detail in studies on the age factor, and it gives interesting insights into topics such as opportunities for L2 contact, attitudes towards the community, and sense of self in the L2, as well as perceptions of foreignness and belonging. The book ends with a call for a comprehensive theory of SLA that accounts for both individual variation and possible universal constraints. This book is a must for students and researchers of age effects, especially for those interested in phonology, but it will also be of interest to researchers working on SLA and individual differences such as motivation, attitudes and identity issues.
Muñoz's Age and the rate of foreign language learning is the result of a nine-year research project conducted in public schools in Barcelona, where bilingual Catalan/Spanish students learn English as a FL. The main objective of the Barcelona Age Factor Project (BAF) is to compare the learning rates and ultimate proficiency of four groups of learners with different AO (8, 11, 14 and older than 18). Several data collections were carried out after 200, 416 and 726 hours of exposure at different ages at testing (AT), that is, the students' ages at the time the data was collected. Therefore, this research design, which combines different AOs and ATs as well as differing amounts of exposure, allows for two comparisons: one between Early Starters (ES) and Late Starters (LS) after receiving fixed amounts of exposure and another between groups with the same cognitive maturity (operationalised as AT) that differ, for instance, in AO or in amount of exposure to the FL. Furthermore, by keeping some variables constant, such as extra exposure to the language, the design makes it possible to throw light on the final product, that is, on the FL level attained by the end of secondary education and after learning English throughout the whole schooling period.
The book is structured into chapters according to the domains or skills analysed: phonetics, morphosyntax, lexis, writing, etc. Learning strategies and individual differences such as motivation are also taken into account and both oral and written language is assessed. Each author focuses on how the different groups of learners in the study perform in one specific area, thus contributing to fulfilling the main objective set, i.e. the comparison of the groups' of learning rates and attainment.
Apart from exploring questions of rate and attainment in low-proficiency learners, this volume also presents data and findings relevant to SLA in terms of methodology, interlanguage states or learners' perception of their own development. For this reason, it will be helpful not only for professionals in linguistics and FL curriculum planners, but also for researchers and language teachers interested in the description and measurement of the early stages of FL acquisition. For example, Joan Carles Mora, in chapter 3, shows that some fluency measures may not be indicative of the real competence of low level learners. Esther Álvarez (chapter 6) identifies and characterises a number of stages in the development of oral narrative abilities, which show how discourse components and morphosyntactic elements interact. Finally, Elsa Tragant & Mia Victori (chapter 10) find that the perceived strategic knowledge that learners develop when learning an L2 may not actually develop homogeneously for all skills. The tests and the interview used in the project are provided in the appendix. Indeed, Moyer also includes these materials at the end of her book. In this way, a set of tests implemented in two different language learning settings (formal and naturalistic) are made available to teachers and other researchers, to facilitate the replication of the studies.
3. Analysis and evaluation
This section focuses on three main critical issues that are debated in studies on the age factor. They appear to different extents in the books examined and are treated in different ways. The perspectives on these topics provided in these volumes can also help the reader appreciate the current state of affairs in age-related research. These topics are: (1) the CP in SLA, (2) the construct of UA and the role of AO and (3) the age effects in natural and formal settings.
3.1 The Critical Period in SLA
Birdsong's volume and Singleton & Ryan's book both deal extensively with this question and provide evidence and counter-evidence for the existence of a CP in SLA, thus enriching the debate and presenting the results and interpretations of several findings for the reader to judge.
Singleton & Ryan concentrate first on the possibility of a CP for the L1 (chapter 3) and then make a similar consideration of a CP in SLA (chapter 4). After providing basic information about the CP, they synthesise what research has established about cut-off points for language learning. The studies included come from the fields of neurology, deaf children, feral children and children with Down's syndrome. From the references reviewed, they conclude that the literature does not clearly establish cut-off points and that there are no conclusive data confirming that language acquisition is absolutely impossible after puberty, the most frequently cited period where capacity for learning is said to start becoming less effective. This seems to be supported by studies showing that learning development continues well into adulthood. Therefore, they cast doubt on the existence of a CP as defined in biological terms, that is, one with a clear discontinuity between phases.
Five positions are analysed in relation to a possible CP in SLA: (1) ‘the younger, the better’, (2) ‘the older, the better’, (3) ‘the younger, the better in some respects’, (4) ‘the younger, the better in the long run’ and (5) the ‘qualitative change’ position, which claims that after a certain age, qualitative changes occur in the learning process. The studies summarised in these five sections help the reader understand the complexity of the elements involved in the analysis and the number of factors that need to be taken into consideration when testing any theory in this field. For instance, information is provided on the function of learning contexts and length of exposure, the existence of talented learners, competence in each language domain and its developmental pattern and the role (if any) of possible innate rules.
The information presented on the CP in Birdsong's book is structured in a different way. Apart from Birdsong's opening chapter, which offers a detailed account of arguments both in support of and against the CPH, three chapters argue in favour of a CP for SLA, while three argue against it. Among those in favour we find Weber-Fox & Neville, who use a combined behavioural-electrophysiological approach; Hurford & Kirby, who model the relationship between two dimensions of human language (language size and speed of language learning); and Eubank & Gregg, who establish a comparison between the CP for language acquisition and other CPs such as those in nonhumans. In the first case, Weber-Fox & Neville show that age of immersion in an L2 affects the language-processing neural subsystems in different ways. They examine ERPs from early and late bilinguals when noticing phrase structure violations and when processing open and closed class words. When trying to model the relationship between language size and speed of acquisition, Hurford & Kirby demonstrate that the two dimensions do not develop independently of one another and that there will inevitably be a cost–benefit effect at some point that favours one at the expense of the other. That would suggest that the idea of a CP is feasible, because at a certain time an external feature may impact on the developing elements. Eubank & Gregg choose to focus on very specific aspects related to CPs (both linguistic and non-linguistic) and discuss the plausibility of a CP for SLA, pointing out that Universal Grammar (UG) certainly has a role in late SLA.
Chapters by Bongaerts and Flege both deal with pronunciation. Bongaerts presents studies in which late learners pass for native speakers, which is interpreted as calling into question an ‘absolute biological barrier to the attainment of nativelike accent in a FL’ (p. 154). Flege explains that the findings that accuracy in L2 pronunciation declines linearly with age are not consistent with the existence of a CP, which should represent a significant turning point in the acquisition process. The same ideas presented by these two authors are shared by Bialystok & Hakuta in the last chapter, although they address syntax rather than phonology. Their conclusion is more in line with the arguments offered in Moyer's book: a contextualisation of the CP enquiry is seen as necessary because experience – and not maturation alone – is thought to be crucial in language acquisition.
3.2 The construct of Ultimate Attainment and the role of Age of Onset
According to Moyer, the CP in SLA research has been ‘operationalized’ as a simple measure of AO. She argues that this ‘operationalization’ is not correct and that AO should not be given the preeminent role that it still has in studies on the CP, or at least it should not be decontextualised. Variables such as AO or Length of Residence are indirect measures of L2 experience and as such they do not account for linguistic outcomes on their own. Thus the idea that UA is primarily a function of age needs a re-evaluation, as UA is not just a function of maturation but also of psychological and social influences. She also points out that experience is unique to each learner and is relevant to UA. However, there are no studies on social and psychological variables related to late attainment in phonology.
From the study conducted with the German learners on phonological acquisition, she proposes a model including all the influences on L2 phonological attainment that have been shown to be significant in the analyses performed. The model consists of a set of ‘processing clusters’ (cognitive, psychological and social) that help to process L2 input and feedback. These clusters are said to be ‘universally significant’, although they depend on individual orientation. Apart from the fact that each of them reflects cumulative experience, they are linked together by means of flexibility/plasticity, style/preference, intention/behaviour and access/opportunity. AO is certainly present in the model, but it is linked by different mechanisms to other clusters and is definitely not the only factor that leads to UA.
The concept of UA is defined in Muñoz's volume as the optimal level of language proficiency realistically attained. However, for this author, the concept cannot be applied when dealing with instructional settings, mainly because this context does not meet the criteria for optimal learning conditions (see also Muñoz Reference Muñoz2008). One of the most frequently repeated conclusions reached in the studies presented in the book is that, all conditions being equal, an early AO does not lead to higher proficiency in formal contexts in the long run. In this case, the ‘long run’ amounts to nine years, which coincides with the end of secondary education and comprises all the curricular exposure that learners get from their years at school. Regarding phonology, the same area studied by Moyer, Natalia Fullana (chapter 2) concludes that AO is not a decisive variable for perceiving and producing English sounds in a native-like manner. Similar results are obtained in other domains analysed in the different chapters. For example, as far as oral performance is concerned, LS groups tend to outperform ES in oral fluency and in narrative development. The same happens in vocabulary, grammar or writing performance. Starting to learn a language at a young age does not significantly alter the levels of motivation of learners either. Additionally, the non-primary role of AO is accompanied by the idea that success in SLA in a formal setting can be as much a function of exposure as of age. Therefore, here, as in Moyer's study, albeit in a different context, AO is not considered to be the only determinant factor for successful language acquisition.
Nevertheless, the results do not necessarily mean that positive benefits for the Early Starters would not have emerged had the study been conducted over a longer period of time, as in formal settings the total amount of exposure is low. A longer period would probably have been necessary in order to confirm or disprove the long-term advantage of the Late Starters group described in the book.
In a similar vein, the results in the book are not interpreted as running counter to an early AO to start learning a FL. From the meta-analysis conducted by Singleton & Ryan in chapter 4, it is also clear that young L2 learners are not always as successful as older learners in all skills and in all phases of learning, but this does not necessarily mean that considerations other than the age factor would not support an early start.
3.3 Age effects in natural and formal settings
It is not only AO that can affect the learning process; biological age has also been shown to play a role in SLA. The influence it may exert, however, depends on whether the language is acquired in a naturalistic setting or in a classroom.
In Moyer's book, immigrants learned the language in a naturalistic setting as adults, although they also received formal instruction in the target language (German) and studied at various advanced learning institutions in Berlin. She emphasises that socio-psychological adaptation to the new country is a determining factor and that assimilation concerns are of paramount importance. The study shows that personal and professional motivation of adult learners is correlated with phonological attainment. The author also suggests that language fluency is a key issue in adaptation to the new environment and that biological age may influence attitudes as well, and that this may also, in turn, influence experience with the new culture.
The studies in Muñoz's book, all of which have been conducted in formal settings, put forward some interesting ideas regarding FL acquisition and the biological age of the learners. Some of the findings are not new, while others have not been sufficiently reflected in the literature or prominent enough to be taken into account in classroom practice. For instance, the superior cognitive ability that adolescent learners possess in comparison to children assists them in taking greater advantage of explicit teaching processes in the classroom. Nevertheless, the implicit learning mechanisms typically used by young learners cannot be fully and properly exploited in formal contexts, due to the limited amount of input provided in this environment.
The book presents valuable evidence on specific areas in language acquisition with respect to age and development in formal settings. As regards writing development, M. Rosa Torras, Teresa Navés, M. Luz Celaya and Carmen Pérez-Vidal (chapter 7) show that there are differences in the distribution of gains that affect the rate of acquisition in learners of different ages. In addition, there also seem to be different developmental patterns for the various areas of writing competence. In morphosyntax (chapter 5), Muñoz shows that learners of different ages learn L2 grammatical morphemes in qualitatively similar ways. In relation to learning strategy use, Tragant & Victori (chapter 9) saw a shift towards complex strategies at the onset of adolescence (around 11–13 years of age).
As mentioned in the introduction to this review, Singleton & Ryan's book also takes into account the educational dimension of the age issue, in this case acquisition in instructional settings by learners of different ages. In chapter 6 they concentrate on two specific topics: advantages and disadvantages of an early start in a foreign setting and L2 programmes for adult learners. In relation to the latter, they present research suggesting that advanced biological age may cause problems in phonetic coding and in handling aural materials. The acquisition of an FL may also be affected by reduced durability of memory traces. Despite the inconveniences, the authors point out that learning FLs as an adult might not be as problematic as one may think, especially if some classroom recommendations are considered and an appropriate teaching approach is used. The suggestions and implications for the classroom presented in this chapter on the basis of research findings are thought-provoking and essential for language teaching practice. However, as pointed out before, they are reduced to just two well-defined areas.
4. Concluding remarks
All four books contribute to the debate on critical issues regarding the age factor. They present research from several linguistic sub-disciplines, offer new data, in-depth analyses and meta-analyses, and bring different perspectives and insights to the study of age effects in SLA.
Singleton & Ryan's book provides the reader with a good ‘state-of-the-art’ review of the field, although inevitably nowadays more studies would need to be incorporated and discussed. Birdsong's collection contains a balanced selection of articles, but seven papers cannot really cover all aspects of the debate on the CP in SLA.
It is obvious that there is considerable complexity involved in research on maturational constraints (see, for instance, Hyltenstam & Abrahamsson Reference Hyltenstam, Abrahamsson, Doughty and Long2003). The conflicting positions are clearly seen in Birdsong's work, which presents views for and against the CPH, and in the pertinent chapters by Singleton & Ryan. They are just a reflection of a wider debate that is going on today and will surely go on for a long time to come (e.g. Long Reference Long2005; Rothman Reference Rothman2008).
Among those sceptical of the existence of maturational constraints, there is the view that AO is perhaps confounded with other variables that do in fact regulate adult L2 learning. That is why, for example, Moyer investigates the effects of social and cultural concerns of immigrants when they acquire the L2. Previous work has usually studied age of arrival effects without much interest in other conditions that could account for different levels of attainment. Nevertheless, it does not present conclusive evidence for the non-existence of a CP, as only two out of the 25 learners in her study conclusively passed as native speakers. The lack of background information in the book on these particular learners prevents the reader from determining if other factors could also have accounted for their success. In addition, the findings might not be applicable to areas other than phonology (learners might not pass as natives in other domains) or to larger samples of participants who learn languages typologically different from the mother tongue.
If attainment depends on a combination of effects of AO and other specific variables, that is, if it is the interaction of multiple causes that shape the final outcome in SLA, as Singleton & Ryan suggest, then further research may throw light on which variables these are. As Moyer points out, more information is needed in order to determine the most typical combination of influential factors. However, taking other factors into consideration may also be a way to justify time constraints on SLA. For instance, Hurford & Kirby (chapter 3 in Birdsong's book) simulate rates at which ‘the biological’ (genetic contribution to speed of learning) interacts with ‘the social’ (language size, understood as a social process). These rates might lead to the conclusion that there is a CP at some point. In addition, analysing variables other than age that might affect UA may also result in findings that can support the CPH (e.g. the study on aptitude by DeKeyser Reference DeKeyser2000).
Complexity is not only present in the study of possible maturational constraints in SLA. The second line identified in the introduction, which aims at describing age-related differences in FL learning in order to inform official curricula, also poses several challenges. Muñoz's book has responded only to some of them. It offers data on a large cross-linguistic sample (more than 750 students) and follows some of these learners longitudinally, analysing their rate of development and final outcomes in various areas of performance after receiving extended periods of instruction. Nonetheless, the book does not attempt to translate the findings into practical implications for the classroom, and suggestions remain largely at a theoretical level. This should be taken into account in future publications on the topic, as books on age effects in FL settings do not abound (see, for instance, García Mayo & García Lecumberri Reference García Mayo and García Lecumberri2003). Age-related classroom-based research is necessary not just to inform language policies, but also to improve the quality of teaching offered to learners.
Acknowledgements
I would like to acknowledge the help of Dr Graeme Porte and David Hall. However, I alone remain responsible for any shortcomings.
Imma Miralpeix is currently a lecturer and researcher at the English Studies Department, University of Barcelona, where she is also a member of GRAL (Language Acquisition Research Group). Her research focuses on second language vocabulary acquisition, although she is also interested in the Age Factor and in testing. Her Ph.D. thesis dealt with the effects of age on vocabulary acquisition in EFL.
