The first six chapters that follow are from the late 1960s, mostly based on talks for general university audiences, hence relatively informal. The final chapter is from 2004, based on a talk for a general audience. This recent essay reviews the “biolinguistic approach” that has guided this work from its origins half a century ago, some of the important developments of recent decades, and how the general approach looks today – to me at least.
The dominant approach to questions of language and mind in the 1950s was that of the behavioral sciences. As the term indicates, the object of inquiry was taken to be behavior, or, for linguistics, the products of behavior: perhaps a corpus obtained from informants by the elicitation techniques taught in field methods courses. Linguistic theory consisted of procedures of analysis, primarily segmentation and classification, designed to organize a body of linguistic material, guided by limited assumptions about structural properties and their arrangement. The prominent linguist Martin Joos hardly exaggerated in a 1955 exposition when he identified the “decisive direction” of contemporary structural linguistics as the decision that language can be “described without any preexistent scheme of what a language must be.”1 Prevailing approaches in the behavioral sciences generally were not very different. Of course, no one accepted the incoherent notion of a “blank slate.” But it was common to suppose that beyond some initial delimitation of properties detected in the environment (a “quality space,” in the framework of the highly influential philosopher W. V. O. Quine), general learning mechanisms of some kind should suffice to account for what organisms, including humans, know and do. Genetic endowment in these domains would not be expected to reach much beyond something like that.
The emerging biolinguistic approach adopted a different stance. It took the object of inquiry to be, not behavior and its products, but the internal cognitive systems that enter into action and interpretation, and, beyond that, the basis in our fixed biological nature for the growth and development of these internal systems. From this point of view, the central topic of concern is what Juan Huarte, in the sixteenth century, regarded as the essential property of human intelligence: the capacity of the human mind to “engender within itself, by its own power, the principles on which knowledge rests,”2 ideas that were developed in important ways in the philosophical–scientific traditions of later years. For language, “the principles on which knowledge rests” are those of the internalized language (I-language) that the person has acquired. Having acquired these principles, Jones has a wide range of knowledge, for example that glink but not glnik is a possible lexical item of English; that John is too angry to talk to (Mary) means that John is to be talked to (if Mary is missing) but John is to do the talking (if Mary is present); that him can be used to refer to John in the sentence I wonder who John expects to see him, but not if I wonder who is omitted; that if John painted the house brown then he put the paint on the exterior surface though he could paint the house brown on the inside; that when John climbed the mountain he went up although he can climb down the mountain; that books are in some sense simultaneously abstract and concrete as in John memorized and then burned the book; and so on over an unbounded range. “The power to engender” the I-language principles on which such particular cases of knowledge rest is understood to be the component of the genetic endowment that accounts for their growth and development.
Linguistics, so conceived, seeks to discover true theories of particular I-languages (grammars), and, at a deeper level, the theory of the genetic basis for language acquisition (universal grammar, UG, adapting a traditional term to a new usage). Other cognitive systems, it was assumed, should be conceived along similar lines, each with its own principles, and powers of engendering them.
Within this framework, cognitive systems are understood to be, in effect, organs of the body, primarily the brain, to be investigated in much the manner of other subcomponents with distinctive properties that interact in the life of the organism: the systems of vision, motor planning, circulation of the blood, etc. Along with their role in behavior, the “cognitive organs” enter into activities traditionally regarded as mental: thought, planning, interpretation, evaluation, and so on. The term “mental” here is informal and descriptive, pretty much on a par with such loose descriptive terms as “chemical,” “electrical,” “optical,” and others that are used to focus attention on particular aspects of the world that seem to have an integrated character and to be worth abstracting for special investigation, but without any illusion that they carve nature at the joints. Behavior and its products – such as texts – provide data that may be useful as evidence to determine the nature and origins of cognitive systems, but have no privileged status for such inquiries, just as in the case of other organs of the body.
The general shift of perspective is sometimes called the “cognitive revolution” of the 1950s. However, for reasons discussed in the early essays that follow, I think it might more properly be considered a renewal and further development of the cognitive revolution of the seventeenth century. From the 1950s, many traditional questions were revived – regrettably, without acquaintance with the tradition, which had been largely forgotten or misrepresented. Also revived was the view that had been crystallizing through the eighteenth century that properties “termed mental” are the result of “such an organical structure as that of the brain” (chemist–philosopher Joseph Priestley). This development of “Locke’s suggestion,” as it is called in the scholarly literature, was a natural, virtually inevitable, concomitant of the Newtonian revolution, which effectively dismantled the only significant notion of “body” or “physical.” The basic conclusion was well understood by the nineteenth-century. Darwin asked rhetorically why “thought, being a secretion of the brain,” should be considered “more wonderful than gravity, a property of matter.” In his classic nineteenth-century history of materialism, Friedrich Lange observes that scientists have “accustomed ourselves to the abstract notion of forces, or rather to a notion hovering in a mystic obscurity between abstraction and concrete comprehension,” a “turning-point” in the history of materialism that removes the surviving remnants of the doctrine far from the ideas and concerns of the “genuine Materialists” of the seventeenth century, and deprives them of significance. They need be of no special concern in the study of aspects of the world “termed mental.”
It is perhaps worth noting that this traditional understanding is still regarded as highly contentious, and repetition of it, almost in virtually the same words, is regularly proposed as a “bold hypothesis” or “radical new idea” in the study of the domains “termed mental.”3
Another significant feature of the original cognitive revolution was the recognition that properties of the world termed mental may involve unbounded capacities of a limited finite organ, the “infinite use of finite means,” in Wilhelm von Humboldt’s phrase. The doctrine was at the heart of the Cartesian concept of mind. It provided the basic criterion to deal with the problem of “other minds” – to determine whether some creature has a mind like ours. Descartes and his followers focused on use of language as the clearest illustration. In a rather similar vein, Hume later recognized that our moral judgments are unbounded in scope, and must be founded on general principles that are part of our nature – genetically determined, in modern terms. That observation poses Huarte’s problem in a different domain, and is, by now, the topic of intriguing empirical research and conceptual analysis.
By the mid twentieth century, it had become possible to face such problems as these in a more substantive way than in earlier periods. There was, by then, a clear general understanding of finite generative systems with unbounded scope, which could be readily adapted to the reframing and investigation of traditional questions that had necessarily been left obscure. Another influential factor in the renewal of the cognitive revolution was the work of ethologists and comparative psychologists, then just coming to be more readily accessible, with its concern for “the innate working hypotheses present in subhuman organisms,” and the “human a priori,” which should have much the same character.4 That framework too could be adapted to the study of human cognitive organs and their genetically determined nature, which constructs experience – the organism’s Umwelt, in ethological terminology – and guides the general path of development, just as in all other aspects of growth of organisms.
Meanwhile, efforts to sharpen and refine procedural approaches ran into serious difficulties, revealing what appear to be intrinsic inadequacies. A basic problem is that even the most simple elements of discourse are not detectable by procedures of segmentation and classification. They do not have the required “beads on a string” property for such procedures to operate, and often cannot be located in some identifiable part of the physical event that corresponds to the mind-internal expression in which these elements function. It became increasingly clear that even the simplest units – morphemes, elementary lexical items, for that matter even phonological segments – can be identified only by their role in generative procedures that form linguistic expressions. These expressions, in turn, can be regarded as “instructions” to other systems of the mind/body that are used for mental operations, as well as for production of utterances and interpretation of external signals. More generally, study of the postulated mechanisms of learning and control of behavior in the behavioral sciences revealed fundamental inadequacies, and even at the core of the disciplines serious doubts were arising as to whether the entire enterprise was viable, apart from its utility for design of experiments that might be useful for some other purpose.
For the study of language, a natural conclusion seemed to be that the I-language attained has roughly the character of a scientific theory: an integrated system of rules and principles from which the expressions of the language can be derived, each of them a collection of instructions for thought and action. The child must somehow select the I-language from the flux of experience. The problem appeared to be similar to what Charles Sanders Peirce had called abduction, in considering the problem of scientific discovery.5 And as in the case of the sciences, the task is impossible without what Peirce called a “limit on admissible hypotheses” that permits only certain theories to be entertained, but not infinitely many others compatible with relevant data. In the language case, it appeared that the genetic endowment of the language faculty must impose a format for rule systems that is sufficiently restrictive so that candidate I-languages are “scattered,” and only a small number can even be considered in the course of language acquisition. In later work in the cognitive sciences, such approaches are often called “theory theory” conceptions.6 Like abduction, and for that matter every aspect of growth and development, language acquisition faces a problem of poverty of stimulus. The general observation is transparent, so much so that outside of the cognitive sciences the ubiquitous phenomenon is not even dignified with a name: no one speaks of the problem of poverty of stimulus for an embryo that has somehow to become a worm or a cat, given the nutritional environment, or in any aspect of post-natal development, say undergoing puberty.
In the essays reprinted below from the 1960s, the nature and acquisition of language presented and discussed adopts the general framework just outlined. “The most challenging theoretical problem in linguistics” was therefore taken to be “that of discovering the principles of universal grammar,” which “determine the choice of hypotheses” – that is, restrict the accessible I-languages. It was also recognized, however, that for language, as for other biological organisms, a still more challenging problem lies on the horizon: to discover “the laws that determine possible successful mutation and the nature of complex organisms,” quite apart from the particular cognitive organs or other organic systems under investigation.7 As the same point was made a few years earlier: “there is surely no reason today for taking seriously a position that attributes a complex human achievement entirely to months (or at most years) of experience [as in the behavioral sciences], rather than to millions of years of evolution [as in the study of the specific biological endowment, UG in the language case], or to principles of neural organization that may be more deeply grounded in physical law”8 – a “third factor” in growth and development, organ- and possibly organism-independent. Investigation of the third factor seemed too remote from inquiry to merit much attention, and was therefore barely mentioned, though, in fact, even some of the earliest work – for example, on elimination of redundancy in rule systems – was implicitly guided by such concerns.
In the years that followed, the topics under investigation were substantially extended, not only in language-related areas but in the cognitive sciences generally. By the early 1980s, a substantial shift of perspective within linguistics reframed the basic questions considerably, abandoning entirely the format conception of linguistic theory in favor of an approach that sought to limit attainable I-languages to a finite set, aside from lexical choices (these too highly restricted). This Principles and Parameters approach may or may not turn out to be justified; one can never know. But as a research program, it has been highly successful, yielding an explosion of empirical inquiry into a very wide range of typologically varied languages, posing new theoretical questions that could scarcely have been formulated before, often providing at least partial answers as well, while also revitalizing related areas of language acquisition and processing. Another consequence is that it removed some basic conceptual barriers to the serious inquiry into the deeper “third factor” issues. These topics are reviewed in the lecture that closes this collection. They raise possibilities that, in my personal view at least, suggest novel and exciting challenges for the study of language in particular and problems of mind more generally.
1 Chapter 3, note 12. Joos was referring explicitly to the “Boasian tradition” of American structuralism, and had only a few – rather disparaging – remarks about European structuralism. But the observations carry over without too much change.
3 For examples and discussion, see my New Horizons in the Study of Language and Mind (Cambridge, 2000).
6 Advocates of these approaches disagree, but mistakenly, I believe. See L. Antony and N. Hornstein, Chomsky and his Critics (Blackwell, 2003), chapter 10, and reply.
7 Pp. 47, 85f., below.
8 Chomsky, Aspects of the Theory of Syntax (Cambridge, Mass: MIT Press, 1965), p. 59.