Vol. 4. No. 2 R-14 November 1999
Return to Table of Contents Return to Main Page

Second Language Acquisition and the Critical Period Hypothesis

Birdsong, David (Ed.) (1999)
Mahwah, NJ: Lawrence Erlbaum Associates
Pp x + 191
ISBN 0-8058-3084-7 (cloth)
US $45.00 (special prepaid price, US $19.95)

Is there a single key issue in the field of second language acquisition/learning, an as yet unresolved matter on which all else depends? A good case could be made for the question of whether or not there is a critical period for second language learning being just such a key issue. In other words, does the nature of second language acquisition change if the first exposure to the new language comes after a certain age? This question is closely linked to the question of whether first language (L1) acquisition and second language (L2) acquisition are essentially the same process, or very similar processes, and if so whether this is the case for some learners, or for all. In practical terms, it could be central not only to such issues as the optimal age at which children should start learning foreign languages, but also to the best teaching/learning approach for adults. Krashen's Input Hypothesis (Krashen, 1985) is totally undermined if a critical period does indeed exist, since the hypothesis assumes not only that L2 acquisition is similar in nature to L1 acquisition, but also that this is the case for learners of any age. Alhough many would claim that Krashen's theories are seriously flawed in any case, their influence in the field of second language teaching can hardly be denied. Issues such as the relative importance of lexis and syntax in teaching materials must ultimately link back to the way in which second language knowledge is organised in the brain. If that organisation is different in learners who have first been exposed to L2 after a certain age, then this has a bearing on choice of teaching approach. Yes, I believe there is a strong prima facie case for regarding the debate over the Critical Period Hypothesis (CPH) as a central issue.

The concept of a critical period is well known in nature. One example is imprinting in ducks and geese, where it is claimed that ducklings and goslings can be induced to adopt chickens, people, or even mechanical objects as their mothers if they encounter them within a certain short period after hatching. (Note, however, that the exact nature of even this apparently well-documented instance of a critical period is now coming under fire; see Hoffmann, 1996). In humans, on the basis of extant evidence, it seems that there is a critical period for first language acquisition; those unfortunate persons who are not exposed to any language before puberty seem unable to properly acquire the syntax of their first language later in life. (Inevitably, our knowledge in this area is sketchy and unreliable, being based solely on a very few cases, of which that of "Genie" is the most celebrated and best known; see Eubank and Gregg's article in Birdsong for a discussion.) [-1-]

Provided that a person learns a first language in the normal way, the question is then whether there is a certain biologically-determined critical period during which that person can acquire further languages using one mental mechanism, probably resulting in a high level of achievement if learning continues, and after which the learning process for new languages changes, so that the learning outcome will not be as good. Note that we are not talking here about the commonly-observed and widely-accepted generalisation that learning gets harder as one gets older; nor is the question one of whether changes in attitudes or situation alter the learning process as one gets older. The issue is whether a fundamental change in the learning process and thus in potential learning outcomes related to second languages occurs in the brain at a fairly fixed age, closing a biological "window of opportunity" (although as Birdsong points out in his introduction to this book, there is no single formulation of the Critical Period Hypothesis, but a number of different versions of the theory).

This book contributes to the debate by juxtaposing a number of papers which consider the CPH from a variety of points of view, and which arrive at a variety of conclusions. Most of the papers in the book are based on talks given at an AILA symposium on the CPH which took place in Finland in 1996. It must have been quite a conference; the names of the contributors to this book make up a Who's Who of researchers in fields related to the CPH issue, and the diversity of the opinions held by the contributors must have made for some sharp exchanges. The book contains research papers by both proponents and opponents of a CPH for SLA, thus drawing the reader into the controversy.

What you get in the book is what you might expect from the above description. First, it must be said that it is a fairly tough read. Some of the writers are easier to follow than others, but these are research papers, and anyone unfamiliar with the fields covered--and there is a considerable range of fields--is likely to have to work quite hard at some of the texts at least. Second, there is no overall conclusion, even though the editor does have his own clearly-expressed view. This is not because of differences in the interpretation of data; it is because the various writers operate in different areas of research, each casting a different light on the central issue. These varied areas of research produce conclusions which point in different directions, and because of the lack of common ground on which to debate, the differences cannot easily be resolved. Third, there is an unevenness about the book. Some writers report on tentative conclusions from ongoing research; others simply reproduce material on completed projects which can be found in almost identical form elsewhere. The relevance of the research presented to the central issue also varies. These points might be regarded as drawbacks. But the compensation comes in having so much relevant and fairly up-to-date material on the issue collected together in one volume, providing insights into current knowledge and thinking from a variety of angles.

In his introduction, Birdsong briefly surveys the background to the debate, outlining some of the arguments previously advanced for and against the existence of a critical period. He is particularly well suited to this task, having "changed sides" on the issue in the early 1990s. After the background section, Birdsong goes on to present a careful summary of each of the chapters in the book. While admitting his adherence to the "anti-CPH" camp, he makes no attempt to resolve the evidence presented in the various chapters, and concludes that in total the contributions to the book demonstrate "the richness, depth and breadth of the critical period enquiry" and that they "testify to the unmistakable centrality of the CPH in L2A research" (p. 18). The introduction is clearly written, and since it contains so much summarising material it can stand as a valuable survey of the field in itself. However, some of the chapters in the book do not lend themselves to brief summaries (the chapter by Eubank and Gregg, for example, is far too broad in scope) and thus reading the introduction is no substitute for reading the entire book. [-2-]

There are three chapters providing evidence for the existence of an SLA critical period. First, Weber-Fox and Neville take a frontal approach to the issue with an investigation of neural activity while performing L2 tasks in subjects whose first exposure to L2 was at different ages. Their paper has the rather daunting title of "Functional Neural Subsystems Are Differentially Affected by Delays in Second Language Immersion: ERP and Behavioural Evidence in Bilinguals." The findings do not point to the existence of a single critical period; the patterns of change vary for different language tasks. The authors claim, fairly circumspectly, that "our findings are consistent with the hypothesis that the development of at least some neural subsystems for language processing is constrained by maturational changes, even in early childhood. Additionally, our results are compatible, at least in part, with aspects of Lenneberg's . . . original hypothesis that puberty may mark a significant point in language learning capacity and neural reorganization capabilities" (pp. 35-36). Eubank and Gregg, in a wide-ranging paper entitled "Critical Periods and (Second) Language Acquisition: Divide et Impera," recognise the importance of neurological investigation in their consideration of whether second language learners retain access to Universal Grammar, and find Weber-Fox and Neville's line of research a promising one. Sandwiched between these two chapters comes a paper by Hurford and Kirby which takes a very different approach to the problem. The writers' argument is an evolutionary one; they produce computer models to suggest that a critical period for language acquisition finishing at puberty inevitably evolves in order to produce maximum language learning by the time a reproductive age is reached. However, their argument appears predicated on the assumption that the level of an individual's language knowledge governs the likelihood of his or her being able to reproduce. Alas, it has never been my personal experience that linguistic ability provides a crucial advantage in the competition for sexual partners, and I do not find myself convinced in any great measure by this application of the currently fashionable evolutionary approach to exploring the nature of the human mind.

Then come three chapters arguing against the existence of a critical period. James Flege provides research evidence to show that level of achievement in pronunciation is closely related to age of first exposure to the second language. He claims that, even for children, the later in life the first exposure to L2, the greater the degree of foreign accent, with no sudden discontinuity in the figures at a certain age to suggest that a critical [-3-] period has ended, a window of opportunity suddenly closed. Other hypotheses, he claims, can be advanced to explain the linear nature of the relationship between age of first exposure and L2 pronunciation, notably that pronunciation of L2 varies as a function of how well one pronounces L1. Theo Bongaerts takes a very different approach in his paper, which again focusses on pronunciation; his view is that people who begin learning L2 later in life can sometimes achieve native-like pronunciation. If such learners do indeed exist, and Bongaerts presents evidence to suggest that they do, then there can be no biological window of learning opportunity that closes at a fixed age; instead, there must be other explanations for the lack of success of the majority of learners. These two papers, then, while arriving at the same conclusion with regard to the existence of an L2 critical period, do so on the basis of more or less contradictory evidence! Finally, Bialystok and Hakuta, focussing primarily on syntax rather than pronunciation, again point to the lack of any age-related discontinuity in the nature of L2 acquisition. They also suggest that belief in a critical period may be the result of misattribution of causality in examining the evidence; even the neurological differences pointed out by Weber-Fox and Neville could be the result of differences in the learning experience, rather than causes of such differences.

This is all somewhat confusing, and the only conclusion that a reader can come to at the end of the book is that there are no easy answers on the CPH. What is clear is that the old notion that the nature of L2 acquisition changes suddenly and dramatically at around the age of 12-13 because of changes in the brain is much too simplistic (as has been generally recognised for some time). If there is any truth in the CPH, then there may be different critical periods for different language skills, different types of change at different ages. If on the other hand there is no physical change in the brain which can be directly related to language learning, other powerful explanations are needed to account for the dramatic decline in ultimate achievement generally seen in later second language learners compared to young children -- and such explanations are no more than tentative guesses at present. None of this is of much immediate help to the practising language teacher; it may even be in the long run that exact age of first L2 exposure and the CPH will not turn out to be such a central issue after all, at least not in a formal learning context. But whether it is itself a key field, or whether it simply takes us into other areas which are key fields, further research into the relationship between age and language learning is likely to help us delve deeper into the mysteries of the mechanisms of second language acquisition.

Birdsong's collection of papers describes a waystage--or rather, a series of different waystages on different roads--in research and speculation on the CPH issue. It is a significant publication, and important reading for those who need to keep up-to-date with second language acquisition research.


Hoffmann, H. S. (1997). Imprinting: A brief description. Available: http://www.animatedsoftware.com/family/howardsh/imprint.htm

Krashen, S. D. (1985). The input hypothesis: Issues and implications. London & New York: Longman.

Tim Caudery
University of Aarhus, Denmark

© Copyright rests with authors. Please cite TESL-EJ appropriately.

Editor's Note: Dashed numbers in square brackets indicate the end of each page for purposes of citation..

Return to Table of Contents Return to Top Return to Main Page