Today there are perhaps four thousand spoken languages, about thirty times as many languages as there are countries in the world. Over half the world’s population is at least to some extent bilingual. These facts help to insure that bilingualism remains—as it has long been—an intellectually provocative and a socially charged issue. Why do human beings have so many mutually incomprehensible tongues? Should citizens be allowed to maintain their native language in an alien culture? Both questions have always been controversial. Thus, even before the United States came into existence, Benjamin Franklin complained about the prevalence of German in the province of Pennsylvania:

Few of their children in the country know English. They import many books from Germany…the signs in our streets have inscriptions in both languages and in some places only German…. I suppose in a few years [interpreters] will also be necessary in the Assembly, to tell one half of our legislators what the other half say. In short, unless the stream of their importation could be turned from this to other colonies…they will so soon outnumber us, that all the advantages we have, will, in my opinion, be not able to preserve our language and even our government will become precarious.1

Similar alarm has been expressed by many national officials including Chief Justice Marshall, the two presidents Adams, and the two Roosevelts. Most recently, Secretary of Education William Bennett called bilingual education a “failed path”; sharply criticizing the practice of providing non-English-speaking students with schoolwork in their native tongue, he urged local school systems to experiment with various methods that would help students master English as quickly as possible.

The same intellectual and social themes have recurred across the centuries. A rationalist strain, dating back to René Descartes and strongly expressed in the contemporary linguistic analyses of Noam Chomsky, has argued for the existence of universal grammar, while minimizing “surface” differences among languages. Against this “unitarian” perspective Giambattista Vico in the early eighteenth century, Johann Gottfried von Herder a century later, and Benjamin Lee Whorf in our own time celebrated the multiplicity of languages. In their romantically tinged accounts, each language has its own peculiar “genius” which strongly colors the thoughts and sentiments of its users.

As for the debate within the United States, for every spokesman who urged that immigrants absorb the American language and the dominant national culture, there have been others, usually (but not always) drawn from the immigrant population, who have argued for maintaining one’s foreign heritage and its distinctive linguistic and cultural values. Indeed, Thomas Jefferson declined an invitation to call for “national uniformity in Language.”2

In recent years, scholars have undertaken several lines of study in an effort to resolve controversial aspects of bilingualism. Biological and cognitive scientists have been investigating the nature of languages, the ways they are represented in the human brain, and the connections between linguistic and thought patterns. When considering the bilingual mind, such investigators have asked, In what ways are two languages different and what difference do these differences make? At the same time, educational and sociological researchers concerned with public policy have examined life in bilingual communities and the effects of programs that encourage or discourage the coexistence of two linguistic and cultural traditions. Typically these scholars ask, Is it useful and valuable to provide a significant amount of education in their native tongue to those who will live in an alien society?

While these two general approaches, cognitive and sociological, generally are pursued independently of each other, it is important to try to combine them. Decisions about the merits of bilingual education ought to be made in the light of the most reliable information about language, mind, and society. We can hope that cognitive studies will reveal whether the differences among languages are superficial or deep, and whether such differences affect the ways in which one thinks and learns. Sociological studies should indicate the ways in which bilingual people handle both their languages, the costs and benefits of maintaining a bilingual community, and the effects on children of enrollment in monolingual or bilingual programs.

No single study can by itself resolve the vexed issues surrounding bilingualism. However, it seems possible that explorations of the “degree of relatedness” between languages, or of the “transfer” of knowledge gained in one language to activities carried out in the second language, could provide links between usually disparate kinds of inquiry as well as offering suggestions about public policy. Suppose that it turns out that languages differ chiefly in superficial ways and have relatively little differential effect on basic thought patterns; the transition to a new language might then present relatively few problems. By contrast, suppose that both the differences among languages and imprints of each language on mental processes are deep ones. In that case, the risks of abandoning a native language would be higher and making a successful transition to an unfamiliar language would be more difficult.


It would be cheering to report that investigators of bilingualism have illuminated these questions. Unfortunately, as Kenji Hakuta writes in his new book,

What is remarkable about the…transfer of skills [from one language to another] is that despite its fundamental importance, almost no empirical studies have been conducted to understand the characteristics or even to demonstrate the existence of the transfer of skills.

Moreover, other scientific investigations have yielded meager results. Indeed, in the case of the most urgent issues—for example, the effectiveness of programs of bilingual education—there is a depressing consensus among critics that the research has been of poor quality. In a comprehensive review of the major studies of bilingual education, Ann C. Willig concludes:

A major result of the current synthesis has been the revelation that bilingual education has been badly served by a predominance of research that is inadequate in design and that makes inappropriate comparisons of children in bilingual programs to children who are dissimilar in many crucial respects.

…The overwhelming message from these data suggests that most research conclusions regarding the effectiveness of bilingual education reflect weaknesses of the research itself rather than effects of the actual programs.3

In Mirror of Language, Kenji Hakuta tries nonetheless to assess what we know about the cognitive and social aspects of bilingualism. Bilingual himself in Japanese and English, and currently associate professor of psychology at Yale University, Hakuta has in recent years emerged as one of the most thoughtful scholars of the subject. As an active researcher on bilingual education in the New Haven public schools and a national spokesman in favor of bilingual education, he has a personal stake in the issues.

When reviewing research of indifferent quality, one faces a number of unsatisfactory choices. One may simply overlook the large mass of inferior work and dwell on the few good studies. This sometimes leads to strong conclusions, though at the cost of oversimplification or misrepresentation. One may limit oneself to the disparate lines of inquiry that have yielded consistent findings, but this would give a very fragmented view of such a subject as bilingualism. Or one may survey the entire field of bilingual research, conceding that it is largely inadequate, and then report the modest results that have been achieved.

From appearances, it would seem that Hakuta has flirted with each of these approaches, but he has, for the most part, chosen to review all the available studies as best he can and to resist going beyond the data. Oddly, however, where the evidence is weakest—on the case for bilingual programs—he ends up drawing the strongest, and least warranted, conclusions. The result is a work that ends up by revealing rather less about the nature and consequences of bilingualism than one would like and much—both advertently and inadvertently—about the workings of contemporary American social science.


Hakuta’s mastery of the relevant literature is impressive, but the reader in search of a firm conclusion or at least a provocative hypothesis will be disappointed. Time and again, he builds up our hopes that research will lead somewhere only to undermine it with a sharp criticism or with an equally credible presentation of the opposing line of analysis.

Consider, for example, his study of the effects of bilingualism on measured intelligence (usually IQ). He describes how psychologists earlier in this century went to great lengths to document the intellectual inferiority of immigrants (particularly those who were neither Nordic nor Anglo-Saxon)—a familiar story. What is less well known is that the same prejudicial attitudes were reflected in comments on bilingualism. Ignoring the obvious fact that most standardized tests depended on mastery of English, leading researchers of the period like Carl Brigham, Florence Goodenough, H.H. Goddard, and Kimball Young consistently interpreted the poor performance of immigrants as evidence that they came from inferior stock. Their very failure to learn the language instantly was held as proof of their low IQs.

More recently scholars have provided evidence that bilingual people often outscore monolingual ones on intelligence tests, especially those that require the subjects being tested to reorganize patterns or concepts, rather than simply retain them. One might want to conclude that bilingualism contributes to a more flexible use of mental capacities, but Hakuta devotes several pages to a discussion of the weaknesses of even the best studies in this “liberal” tradition. The researchers have had difficulty matching groups according to crucial variables, such as the attitudes of parents toward bilingualism; they have found it virtually impossible to “control” for the respective influences of home and school environments; the tests used are inadequate; and so forth. Having called such investigations into question, Hakuta then concludes that, as it is now conceived, the question of the intellectual implications of bilingualism is largely a spurious one. It will, he believes, evaporate as deeper issues about language and cognition are explored. Having arrived at this cul-de-sac, one wonders whether the historical journey itself was worth making.


Hakuta turns to such “deeper issues” in a section on the bilingual mind. First he reviews the major theoretical debates between those, like Jean Piaget and Noam Chomsky, who minimize the importance of differences among languages, and those, like Lev Vygotsky and Benjamin Lee Whorf, who stress the effects of language on thought as well as the ways that specific languages mold perceptions. Reviewing the different theories he reaches a conclusion that sounds weak:

If one’s orientation is toward the view that bilingualism influences thought, one will tend to believe that there are general capacities common to language and thought…. On the other hand, if one believes that language and thought are autonomously structured…one will find the influence of bilingualism on thought trivial.

It should be possible to go well beyond this bland conclusion. For the last few decades, researchers in various disciplines have been attracted by the intriguing hypothesis of Benjamin Whorf:

We dissect nature along lines laid down by our native languages. The categories and types that we isolate from the world of phenomena we do not find there because they stare every observer in the face; on the contrary the world is presented in a kaleidoscopic flux of impressions which has to be organized by our minds—and this means largely by the linguistic systems in our minds.4

However, despite dozens of valiant empirical attempts to support this hypothesis, there is no evidence in favor of any interesting version of it. In fact, in those cases where Whorf’s hypothesis apparently had support, subsequent research has countered the provisional claims made for it. During the 1950s, some researchers believed that names used for colors affected the perception and classification of colors, but this was undermined in the 1970s by convincing evidence that people everywhere perceive colors in essentially the same fashion. More recently, Alfred Bloom noted that the Chinese language does not include the linguistic form called the counterfactual conditional (“If Reagan had been a politician when he was young, he would now be a consummate actor”). He raised the possibility that, as a consequence, monolingual Chinese speakers would be unable to appreciate counterfactual lines of reasoning. While some early studies supported Bloom’s intriguing speculation, more careful investigations have documented that counterfactual reasoning presents equally few problems for Chinese as for English speakers.5

On the basis of this and much other evidence, I would myself conclude that most differences among languages (even those as remote as English and Chinese) are not deep ones and that thinkers from different linguistic backgrounds perceive, classify, and reason about the world in fundamentally similar ways. The differences that exist among speakers of disparate languages seem to reflect the different circumstances (including cultural conditions) under which languages are acquired rather than the structure or the lexicon of the languages themselves. What troubles me is that Hakuta is unwilling to take any stand whatever on this central question of cognitive science. At the very least, if he should decide that a conclusion is premature, he might offer his readers a hypothesis or two of his own.

Hakuta is similarly uncertain when he turns to studies of the effects of bilingualism on patterns of thought. He considers three different kinds of research: by neurologists, by cognitive psychologists, and by educational psychologists. Some of the studies and findings are fascinating. Neurologists tell us that one’s primary language is more concentrated in one part of the brain than a secondary language; for some reason as yet not understood, a first language is located almost entirely in the left (or dominant) hemisphere, while a second language has significant representation in the right (or nondominant) hemisphere. Cognitive psychologists report that bilinguals remember lists of words from two languages much as monolinguals recall lists from a single language; in this sense, they have but a single store of words. At the same time, however, bilinguals can typically remember the language in which a word was first presented, thus suggesting that they have an auxiliary code for the source, or the physical form, of the word. Educational psychologists have found that skills in certain school subjects (such as arithmetic) taught in one’s first language can readily be transferred (and are therefore retained) when one begins to speak a new language.

Hakuta reviews all of this research thoughtfully. But once again his summaries are undercut by his own critical spirit. So disturbed is he by problems arising from the design of various research studies, that he seems unable to reach any strong conclusions. Once more, he finds a bland middle ground, declaring, “I would venture the guess that the language variable would be important in tasks where language occupied a central role.” He goes on to deplore the primitiveness of recent research: it is characterized by “no conclusive results,” “confusion,” “deprived state of knowledge,” “crude research procedures.” It is also “misleadingly simplistic,” uses “gross overall measures,” has “many gaping holes,” etc. One is tempted to invoke the biblical question, “Why did you lead us into this desert?”

Hakuta may be well advised in his reluctance to draw conclusions. At the very least, however, he owes us a more penetrating discussion of how progress might be forthcoming. I am not a researcher on bilingualism but it occurs to me that the question of “transfer of skills or knowledge” across languages would be one promising path to pursue. One could look for the degree of transfer, in learning a second language, of aspects of syntax, phonology, semantics, and the denotative and connotative features from the native tongue. Turning to the transmission of knowledge, it should be possible to examine different groups of skills or knowledge, including those where language seems clearly incidental (the visual arts or music), those where language seems clearly central (history, literary criticism), as well as those where the role of language is less easy to determine (how important is language in learning logic, physics, algebra?). By examining the performances of people who learn in one language but who are tested subsequently in a second language, one could gain some idea of the extent to which various forms of knowledge move readily across languages, or are stored in a more abstract, nonlinguistic form, or are associated with a particular linguistic format.

Of course my point here is not to endorse a particular kind of research in bilingualism. It is not enough to say, as Hakuta lamely concludes, that “the most important task facing research is the identification of the conditions under which transfer occurs and the conditions under which it does not.” Such a remark should be followed by more concrete suggestions about how to proceed and what one might find.

For a while it appears that much stronger conclusions will emerge from the chapter on adult (or “later”) learning of second languages. Most of us believe that the earlier one is exposed to a second language, the better one will learn it. However, Hakuta reviews evidence that older children (at least through adolescence) learn second languages more quickly and more effectively than do young ones. Indeed, only in the acquiring of phonology is there no difference between the younger and the older learner of a language. But once again Hakuta is reluctant to endorse the apparent conclusion that “getting older means getting smarter, and the smarter you are, the better you should be at learning most things, second language included.” Not surprisingly, he is overwhelmed by the methodological problems presented by these studies.

I, too, am skeptical about these conclusions, but for other reasons. To my mind the question remains why younger children appear to pick up languages in a qualitatively different, and perhaps more natural way. Perhaps most observers tend to contrast children who learn a second language through immersion in an alien environment with older people who receive instruction in school. The first method may simply prove most effective for most of us. To test this hypothesis, one would need to compare people of different ages as they acquire language under different conditions; and one would as well have to examine the means by which scientists and lay observers draw their conclusions (e.g., formal tests versus observations of casual interactions in daily life). Hakuta’s constant strictures on method make me painfully aware, however, of how difficult it would be to carry out such a study in a reliable way. One wonders: Is a social science of bilingualism possible?

In the final chapters of the book, one finds a change of voice, if not of perspective. As Hakuta turns away from cognitive issues to the controversy surrounding bilingual education in America today, he becomes an advocate who seeks to mobilize “popular support for bilingual education.” These final pages have a passion and a sense of purpose that have not been evident before. Unfortunately, the data on the value of bilingual education are even less convincing than those on the cognitive issues I have discussed; so Hakuta is left in the uncomfortable position of arguing with the greatest enthusiasm on the basis of his weakest evidence.

Not that Hakuta completely abandons the objectivity of the rest of his book. He helps us to understand some of the reasons for American resistance toward bilingualism. In his view (which echoes that of many other authorities), there have always been considerable pressures on American immigrants to abandon their original language and culture; it has been pretty much left to the immigrant group itself to withstand these pressures. In general, the immigrant group has low social status and a poor image of itself; both become exacerbated to the extent that the group looks or sounds different from the majority community. Most immigrants to this country have sought economic advancement as well as a larger part in the political process; in nearly all cases these goals combine to dictate that they assimilate to English-speaking culture.

Reminding us that things don’t have to be this way, Hakuta describes four communities where relatively stable bilingualism exists: the northeast Amazon, a community near the Austro-Hungarian border, a Puerto Rican section of Jersey City, and Eastern Canada. The facts are quite different in each of these cases: bilingualism is promoted in the Amazon because Indians must marry outside their tribe, while it is likely to disappear in Austria once the older generation dies out. Examples from Jersey City and Canada are more relevant to the current situation in the United States, of course. Hakuta convinces us that if the circumstances are right, people can use two languages productively in their society. The Puerto Ricans in Jersey City told researchers that they tended to use Spanish at home, in talking with friends, and in their religious lives, while they would use English at school and at work. But Hakuta’s account of the research does not bring us close to the language people actually use in various settings. Although he seems to suggest that the Puerto Ricans are at ease with bilingualism in Jersey City, he does not indicate whether their experience is applicable to that of Indo-Chinese in St. Paul, Cubans in Miami, or Chicanos in California. We are thus left not with a blueprint but rather with four isolated models, of varying degrees of suggestiveness.

It is precisely the quest for such a blueprint that animates discussions in the United States. During the late 1960s, as a result of the civil rights movement and pressure from the growing Hispanic population, federal legislation was passed to address “the special educational needs of children of limited English-speaking ability in schools having a high concentration of such children from families…with incomes below $3,000.00 per year.”6 Then in 1974 the Supreme Court, in the case of Lau v. Nichols, ruled that “there is no equality of treatment merely by providing students with the same facilities, textbooks, teachers, and curriculum; for students who do not understand English are effectively foreclosed from any meaningful education.”

While not imposing any particular model of bilingual education, the court’s ruling was widely interpreted as mandating some form of bilingual education wherever it was needed. And in fact, while bilingual education is still decidedly the exception rather than the rule in the United States, several hundred programs have been set up throughout the country. Moreover, during the past decade, there has also been support for activities ranging from fellowships for graduate students who are preparing to become bilingual teacher trainers, to funds for the development of teaching materials to be used in bilingual education programs.

These measures have been undertaken in the hope that they may change the depressing pattern of poor school performance, high dropout rates, and difficulty in finding employment that has characterized many youngsters who do not speak fluent English, and most particularly those of Hispanic origins. It hardly needs to be said that bilingual programs are difficult to carry out properly and that there are disturbing differences between the few successful programs and teachers and the many that reflect the malaise of contemporary American education. Still, it is now nearly twenty years since these ideas became national policy and over ten years since many programs were put into effect; it is not too early to consider their effectiveness and, more broadly, to consider the arguments for and against some form of education in which children receive at least part of their instruction in their native language.

Hakuta undertakes such an appraisal. First he dispassionately reviews the various arguments surrounding bilingual education, giving equal attention to those who claim it improves academic performance and to those who claim that there is little need for it, and that it has even less demonstrated effectiveness. Then the very author who was reluctant to draw any strong conclusions in the main part of his book makes a revealing shift in his approach. Rather than attempting to justify the need for or the effectiveness of bilingual education by research findings, he instead considers each of the arguments against bilingual education and finds them wanting. He documents the historical precedents in this country for bilingual education; during the nineteenth century there was non-English-language instruction, mainly for people of German, French, and Scandinavian extraction.

Next, he points out the considerable popular support for such programs (interestingly, his own studies in New Haven suggest the greatest resistance comes from older men who were themselves raised in non-English-speaking homes). Countering the widespread belief that bilingual programs are mere fences to maintain a foreign culture, he provides evidence that in general such programs are intended, at least, to facilitate a smooth transition to instruction in English. He also rejects the claim that young children can learn a second language without effort or that time spent in a non-English-speaking class has small value for learning the skills needed to survive in American society. Hakuta does not always make his own case persuasively. For example, the evidence on transfer of learning from one language to another is simply too sparse to allow one to make any conclusions. But he effectively undercuts the often careless or inadequately documented arguments of “the opposition.”

Hakuta concedes that he is unable to counter the major claim made by critics that “bilingual education is not effective.” Here, as he puts it, “there is a sober truth that even the ardent advocate of bilingual education would not deny.” Nonetheless, he remains convinced that the better programs work. In testimony he recently gave (with Catherine Snow) to the House Labor and Education Committee Hearings on Bilingual Education, Hakuta said that “evaluation research conducted with greater rigor would bear out the superiority of bilingual education as an instructional method in many educational contexts.” Clearly, this is a hope rather than a fact.

As an empirical scientist with a cause, Hakuta would like to be able to demonstrate to a skeptic the effectiveness of bilingual education. Yet he also despairs of doing so. He believes that while most observers pay deference to scientific research, few in fact seem to be persuaded by it. Those who feel threatened by a society in which a significant percentage of people speak an unfamiliar tongue, as well as those who believe that the strength of our society depends upon a rapid integration into the linguistic and cultural mainstream, will continue to call for the abolition of any “two-language” system, no matter what “the data say.”

Hakuta deplores this state of affairs and praises those societies and communities where bilingualism and biculturalism have been made to work and made to last. In his view survival of a multilingual perspective in America depends upon a number of factors: a demonstration of the pedagogical value of bilingual education, a dilution of its association with ethnicity, and the adoption of functional bilingualism as a goal for all students.

It is difficult to quarrel with Hakuta’s vision. Other things being equal, it would certainly be desirable if most Americans could express themselves in and appreciate the literary traditions of more than one language; less hostility among ethnic groups might be one result if they could do so and this would be as welcome in our land as elsewhere. After reading Hakuta’s book, however, I find it difficult to be optimistic about the prospects of a genuinely bilingual or multilingual culture in America. To begin with the cognitive perspective, the deep differences among languages that some theorists had anticipated seem not to exist. This conclusion invalidates one potential line of argument in favor of lengthy transitional courses or of continued instruction in two languages. Both the history of America and its current situation, moreover, sharply distinguish it from societies in which two or more language communities have lived together. Americans—whether old or new—tend to be pragmatic and it is simply not practical to sustain a variety of language communities in our country. Those from remote linguistic traditions (like Indo-Chinese) will attempt to integrate themselves into American society as rapidly as possible. For Spanish-speaking people who expect to return to Puerto Rico, Mexico, or other Latin countries, some kind of program to maintain the Spanish language appears likely; but I would expect that most Spanish-speaking American citizens, like most earlier immigrants, will gradually lose their linguistic ties. Cultural links to an ethnic or religious past are likely to remain longer but these will not for the most part be encouraged in the public school system.

In his concluding remarks, Hakuta discusses the nature of social-science research. He finds the same disdain for bilingualism within the universities as in the society at large. He claims that contemporary social scientists are biased against applied research in general and against research on bilingualism in particular. As a result, research on bilingualism is considered a “marginal” activity, in which scholars of indifferent quality produce mediocre work. To improve research he recommends the training of larger numbers of researchers from minority-language backgrounds and the involvement of students from “majority backgrounds” in “immersion programs” where they can acquire another language.

Again Hakuta’s vision here is appealing but, at a time when support for any social-scientific research is vanishing, his hopes are probably forlorn. No doubt the social scientists themselves deserve some of the blame for the current confusion over bilingualism. Too often, we social scientists exhibit excessive caution in interpreting results. Even if particular studies are flawed (as they inevitably will be), we should not desist from suggesting provocative hypotheses and drawing conclusions when the preponderance of evidence points in one direction. Even as we become beset by excessive caution in interpreting our own empirical work, or in linking disparate kinds of investigation, we often throw caution aside when asked to address a problem of social importance. Forgetting the concerns for method that have become obsessive in technical journals, we tend to claim, once we mount a public platform, too neat a fit between “the data” and our values. Unless we can find an acceptable way to reconcile our scientific consciences with the public demands for “answers” to pressing problems, we risk failure on both fronts.

This Issue

October 23, 1986