Linguistics / Book Review

Vol. 2, NO. 3 / September 2016

Reinventing Linguistics – Again

W. Tecumseh Fitch

Letters to the Editors

In response to “Reinventing Linguistics – Again


What Kind of Creatures Are We?
by Noam Chomsky
Columbia University Press, 200 pp., $19.95.

Noam Chomsky is one of the most influential—and controversial—intellectuals of our time, and his writings have had enormous impact in linguistics, cognitive science, philosophy of mind, and political discourse. If the breadth of these contributions is remarkable, the depth of his insights in each of these fields is formidable and daunting. This scope, combined with the fact that much of his scholarly writing is highly technical, sometimes makes the unity of his thinking difficult to discern. Indeed, until I read this short and accessible volume, I always found it difficult to reconcile Chomsky the linguist/philosopher with Chomsky the political critic—his focus on innate limitations in language and mind seemed at odds with his steadfast political championing of individual freedom against oppressive forces. This recent book, in refreshingly clear if sometimes still challenging prose, provides the answer, and reveals an underlying unity in Chomsky’s perspective and thought on these topics.

In the four essays that make up the chapters of this recent book, What Kind of Creatures Are We?, he clearly leads us through his thinking in the domains of language (“What is Language?”), philosophy of mind (“What Can We Understand?”), and politics and ethics (“What is the Common Good?”), returning in the final essay to the general question of human knowledge and the history of science (“The Mysteries of Nature: How Deeply Hidden?”). Of course, Chomsky is first and foremost a linguist, and it is natural that his own historical trajectory and that of this book start with language—in my opinion the most biologically special and cognitively precious capacity our species possesses. I will thus start this review with some historical context concerning Chomsky’s contributions to linguistics and cognitive science before turning to the contents of individual chapters.

Chomsky famously reinvented linguistics as a young man, by reconceptualizing our language capacity in terms of both rigorous mathematical formalism, on the one hand, and strong innate biological predispositions, on the other. At the time, in the late 1950s, linguistics was strongly focused on speech phonetics (the “hard science” part of the language sciences, dominated by engineers and psychologists) and historical linguistics (clearly a branch of the humanities, with its roots in classical philology), with work on semantics based mainly on anthropological and sociological methodology. Syntax was barely discussed, if only because the flexibility of sentence structure and the potentially infinite variety of sentences seemed to put it beyond the empirical methods customary at the time. This potential for creativity in sentence structure, available to every ordinary language user and not just great orators or savants, posed a serious puzzle—so challenging as to remain basically ignored.

The young Chomsky, with a background in mathematics, realized that this must not be so. Some of the greatest mathematicians of the first half of the twentieth century, with Alan Turing, Alonzo Church, Emil Post, and Kurt Gödel most prominent among them, had already developed a rigorous conceptual framework for analyzing infinite sets, and Chomsky recognized that with some modifications this framework could be put to use analyzing linguistic syntax as well. In particular, this formal framework enabled Chomsky to put some analytic meat on the bones of Wilhelm von Humboldt’s famous observation that “language makes infinite use of finite means.” To Chomsky, “grammar” was reinterpreted as the finite set of syntactic rules for combining words and phrases, possessed by each language user, and a “language” was the potentially infinite set of sentences that could be created using these tools. Thus, the obviously limited nature of our knowledge of language could be reconciled, in a rigorous fashion, with the seemingly unlimited freedom and creativity we evidence in using language.

Chomsky’s second great early insight was that syntax could not be adequately captured by the notion of “word order” with which syntax was often conflated (especially in the English-speaking world, due the relatively impoverished nature of English inflection and morphology). Rather, it is the tree-like syntactic structure of sentences that must be the object of syntactic inquiry. It is only because we can capture the phrasal hierarchical structure of a sentence like “the boy who fed the dog chased the girl” that we can understand that semantically it is the boy who chased the girl (despite the word sequence “the dog chased the girl” being present in the sentence).

These two insights—the structure-dependent nature of syntax and the value of rigorous mathematics in capturing the creative syntactic capacity—were almost instantly recognized as revolutionary. The timing of these insights seemed particularly opportune, because computers were at that time just becoming powerful enough to engage with human language and large-scale machine translation was seen as a major and quickly attainable goal at the time. Before Chomsky had even defended his PhD at Penn, he was invited in 1951 to join the prestigious Society of Fellows at Harvard, and after completing his PhD in 1955 he was invited to join the faculty of MIT, where he has remained ever since.

Of course, sixty years later, we know that machine translation (or any type of computational linguistics at the semantic level) is a daunting challenge, and that the hopes of early engineers were naïve if not quixotic. But at that time, there was a degree of hubris among US scientists that Chomsky found distasteful. Chomsky tells of his MIT job interview with Jerome Wiesner (presidential science advisor and future Provost and President of MIT), who was considering hiring Chomsky to join the faculty at the Research Lab of Electronics (RLE). There was some linguistic work going on at the RLE at the time, much of it aiming at automatic computer translation. Most attempts at machine translation in those days were using what were then the state-of-the-art methods, like Markov chain analysis, that implicitly assume that a sentence is nothing more than a sequence of words. Wiesner was looking for linguists to help with the project, and asked Chomsky what he thought of this work. Chomsky replied that he thought it was fundamentally misguided and “a complete waste of time,” and Wiesner immediately offered him the job.1

Chomsky’s first short book, Syntactic Structures, published in 1957 (based on lecture notes to MIT undergraduates), summarized his insights into structure-dependence and infinite use of finite means, and immediately sent shock waves through linguistics (including computational linguistics and engineering).2 But his scathing 1959 critique of B. F. Skinner’s latest tome Verbal Behavior, a book attempting to apply Skinner’s behaviorist principles to language acquisition, was perhaps more important in establishing Chomsky’s early reputation outside of linguistics.3 While framed as a book review, this essay was in fact a trenchant critique of the entire intellectual edifice of behaviorism and related associationist approaches to language. In those days, behaviorism reigned without challenge in English-speaking psychology departments on both sides of the Atlantic, and the very notions of “mind” or “mental” were seen as unscientific. But the creativity and structure-dependence of language seemed to cut to the heart of naïve associationist conceptions of language. How could the child learn, from repeated occurrences and gradual association, to understand and produce entirely novel sentences? Chomsky’s critique, mercilessly underscoring the utter inadequacy of Skinner’s ideas to even approach this problem, was seen by many young psychologists as a welcome breath of fresh air, and is now widely seen as a key founding event in the birth and flowering of cognitive science, as well as the beginning of the end for behaviorism.4 In reinventing linguistics, Chomsky had also served as midwife to a whole new science of the mind.

But Syntactic Structures was just the warm-up act: the true bombshell came with Aspects of the Theory of Syntax which is widely seen as the founding document of modern linguistics.5 While it continues the themes of his first book, Aspects (as it is widely known) fleshes out another crucial component of Chomsky’s linguistics: its biological basis. Chomsky argued that the only if there are strong pre-formed constraints about the types of rules that are allowable, and the types of meanings that are conceivable, can the child successfully acquire the syntactic rules and word meanings of her native language, based on the rather random and limited evidence provided by input words and sentences she happens to encounter. Chomsky dubbed this set of biologically-given constraints “Universal Grammar,” adopting an old term to a new usage (often abbreviated to “UG”). This again sent shock waves through many intellectual circles, if only because the freedom and creativity of language that Chomsky was championing seemed at odds with the notion that the human language capacity was heavily constrained by our biology. How could humans be so free and flexible, and at the same time so strongly limited by our innate predispositions and abilities?

This tensions between freedom and limitation is, I think, at the heart of Chomsky’s thought and helps to explain why his ideas, while almost universally recognized as important, have been so controversial from 1965 to the present day sixty years later. Indeed, his notion of Universal Grammar has repeatedly been ridiculed, dismissed, or set up as a straw man and gleefully destroyed by many a young critic as almost obviously wrong. The most recent incarnation of this persistent theme is the vehement denunciation by Tom Wolfe just published in Harper’s, which criticizes both Chomsky’s scholarly and political writings, but so clearly fails to come to grips with either the depth or breadth of Chomsky’s thought that it is, even to one not well acquainted with Chomsky, self-evidently a vindictive personal attack.6 Fortunately, the book under review provides an up-to-date and accessible antidote to any suspicions that Chomsky’s thinking can be so easily dismissed.

More importantly, the essays in this book illustrate how Chomsky’s superficially incompatible ideas about human limitation and human freedom are not just sensible, but may be correct by logical necessity: that our limitations go hand in hand with our freedom. But the first rule when reading a short summary of a lifetime of thought should be to “check your intuitive certainties at the door.” If something seems obvious to you (about language, mind or politics) expect to have it and other implicit assumptions and dogmas challenged, and (often) found wanting.

With this historical context established, I will now discuss the individual chapters in sequence. Starting with language, Chomsky lays out his perennial concern with our odd combination of biological limits and personal freedom by offering a very clear overview of his current conception of human language and its biological basis. This conception builds on empirically-driven insights about the world’s languages that first attained wide attention with his 1995 book The Minimalist Program, but which for Chomsky have their roots in his much earlier engagement with biologists like Eric Lenneberg in the 1970s in a movement that became known as “biolinguistics.”7 In the 1960s, Chomsky’s conception of the biological basis of language and Universal Grammar was that quite complex and human-specific rules must be, somehow, innately wired into the human brain. He had few specific hypotheses about what brain mechanisms are involved or how they might have evolved. But today, in another revolutionary reinvention of linguistics, he has reconceptualized the core of language as consisting of a single very powerful but abstract capacity that is unique to humans.

This fundamental rethinking, from the ground up, of central notions of linguistics is itself not without controversy, and many linguists who grew up in the earlier tradition, even Chomsky’s former students, have refused to follow this conceptual turn.8 Thus, although Chomsky is sometimes caricatured as the leader of some linguistic orthodoxy, this change in his thinking (along with earlier, less radical, rethinkings) illustrates that in fact Chomsky is a scholar who, while steadfastly focused on the same scientific problems, has been perfectly willing to change his mind (even when others in the discipline are not). In doing so, with the minimalist approach, Chomsky has reinvented linguistics once again.

In this new conception, most of the complexity of human languages, and the differences between languages, is relatively superficial and stems from accidents of history and culture (examples would include the reduction of English grammatical complexity, relative to German or Icelandic, that resulted from language contact subsequent to first the Norse and then the Norman invasions of England). Furthermore, most of the biological hardware that enables us to perceive and learn spoken language is shared with other species.9 What’s left—the uneliminable “minimalist” core—is a basic operation that can combine syntactic/semantic objects (words and phrases) into ever more complex hierarchical structures. In its recent incarnation, this operation is termed Merge and simply combines two items or sets into a larger set, with no notion of order built into this operation. But because language needs to be produced serially (whether as sequences of words in spoken languages, or signs in signed languages) conventions are needed to determine how these abstract hierarchical structures get expressed or “externalized” as strings. Such conventions are not determined by our biology or UG, but instead are rather arbitrary conventions that are one of the main source of differences between languages (e.g. between English, where the verb typically comes second, and Japanese, where the verb typically comes last).

This reconceptualization of the biological core of human language has important consequences. First, it brings discussions of language evolution into the realm of scientific discourse (instead of idle speculation and story-telling), partly because it means that what needed to evolve was rather simple (although very powerful computationally). Although this evolutionary issue is only briefly discussed in the book under review, it was a central topic of a much-cited 2002 paper in Science (that—full disclosure—included me as a co-author) and another recent book by Robert Berwick and Chomsky.10 But in raising the question of evolution (and the evolutionary advantages that language must have brought our forebears), it raises a more central issue in the relationship between language and thought, again a topic of perennial interest for Chomsky.11 Chomsky’s argument is that in fact this relationship is almost an identity, and that the same operation that provides the creative power of language underlies the creative power of human thought—the “language of thought.” He argues therefore that the original selective pressure driving the spread of the language capacity in our early forebears was not its utility for communication (an assumption so widely accepted as to be almost a dogma in some circles) but rather the vastly increased conceptual ability it granted its possessors. Put simply, “language,” in this rather specific recombinatory sense, evolved for thought rather than communication, and its use in communication came later in our evolutionary history. But just as limitations are central to our human ability to acquire language, Chomsky argues that similar limitations must exist in our capacity to reason. Thus, the apparent dichotomy between freedom and limitation, central to language, again arises in our capacity to understand: the human mind itself is similarly rich but bounded in scope.

I find most of what Chomsky argues in these two chapters both plausible and in many cases compelling, but there are two specific issues that troubled me. One is his contrast between the apparent complexity of linguistic rules that depend, invariably, on structural considerations (such that the relevant notion of “distance” in a sentence is structural distance, rather than the serial number of words intervening) with a hypothetical language capacity that relies simply on linear order. But is linear order really “far simpler” for brains, and if so how much simpler? It seems to me that if all of the basic conceptual and semantic structures implicit in thought and expressed in language have hierarchical structure, then hierarchically structured sentences may be in fact the simplest computational option. In general, it seems to me that caution is needed in our notions of simplicity, as they are applied in a “minimalist” program that seeks in Einstein’s words to make everything “as simple as possible, but not simpler.” To the extent that principles of “minimal computation” are supposed to play a central role in this enterprise, they deserve to be cashed out in as clear and biologically-based terms as possible: what is simple for a neurally-instantiated system might be complex for silicon computers, and vice versa.

My second doubt is empirically grounded in my own research in animal cognition and communication. In the second half of chapter 2, Chomsky develops a detailed argument against referentialist assumptions about word meaning (in a nutshell, the intuitive idea that meaning reflects a direct, mind-independent linkage between words and things). Chomsky convincingly argues that this intuition is misleading, and that even such words as “water” or “tree” that seem to refer to real things in the world are in fact crucially dependent on our individual mental representations. That is, word meanings must always pass through the mind of language users in a way that is highly context- and goal-dependent. Although I strongly agree with this conception of linguistic word meaning, I do not agree with Chomsky’s claim that animal concepts are fundamentally different.12 Chomsky bases this claim on an apparent discontinuity between language and animal communication, and particularly the idea that the signals of animal communication are in some sense reflexively or automatically elicited by circumstances (whether external factors, like the appearance of food or a predator, or internal factors such as hunger, pleasure or fear). The degree to which reflexivity characterizes animal communication is itself debated, but even if we grant that animal calls are mostly limited in this way, it would not mean that animal concepts are so limited.13 In fact, I think a considerable body of work in animal cognition indicates the opposite: that animal categorization, navigation and motor control reveals the existence in other species of flexible, context- and goal-dependent, mind internal concepts. While I thus agree with Chomsky that there is a considerable gulf between human language and animal communication systems, I think he greatly overstates the difference between animal and human concepts, thus overestimating the degree to which our type of conceptual structure poses a significant evolutionary problem.

A recurrent central theme in these first two chapters is the great value of a scientist’s ability to be puzzled by a problem or phenomenon, and the key role that such puzzlement can play in progress. For Aristotle, nothing could be more natural than an apple falling to earth, for “that is where it naturally belongs.” It was precisely Newton’s puzzlement about this sort of obvious everyday occurrence that led to the greatest scientific breakthrough of the seventeenth century. In modern times, the equivalent would be the puzzlement we should feel when we see a young child, with little explicit instruction, effortlessly “absorb” the rules of their parent’s language(s): an everyday occurrence, but one that no other species can duplicate—thus a deserving object for wonderment. “Willingness to be puzzled is a valuable trait to cultivate, from childhood to advanced inquiry.”14 The value of a “willingness to be puzzled” is a central theme not just of this book, but of Chomsky’s entire career.

The third chapter, concerning the nature of humans as social and political creatures, was for me the most eye-opening (though I confess to being much less familiar with Chomsky’s political work than with his linguistic and philosophical oeuvre). Although this chapter initially seems like a break from the first two, which deal centrally with individual cognition, issues of personal responsibility still loom large here. Although he is sometimes labeled an anarchist, Chomsky’s political perspective defies such simple labels, and he sees his perspective as closer to the libertarian democratic ideals of John Dewey. In this chapter, Chomsky clearly lays out some of the fundamental precepts that have driven his lifelong political writing and activism. One of these is the fundamental right of all humans to dignity, independence, self-development and self-determination, rights that any political system or government should curtail only under specific circumstances, and with serious consideration and justification. That such curtailments occur everywhere and all the time, without any viable justification, is a virtual truism, and Chomsky argues that it should be our intellectual responsibility to observe, name and resist such curtailments and the coercion that drives them. Whether by governments or (increasingly) international corporations, such coercion should be everywhere and always resisted.

Whether it is called anarchism, libertarian socialism, or anarcho-syndicalism, the fundamental ethical precept that Chomsky defends here is one that characterizes many democratic and cooperative movements worldwide (e.g., the civil rights or feminist movements) which seek

to identify structures of hierarchy, authority, and domination that constrain human development, and then to subject them to a very reasonable challenge: justify yourself [emphasis added]. Demonstrate that you are legitimate, either in some special circumstances at a particular stage of society or in principle. And if they cannot meet that challenge, they should be dismantled.15

It is hard to see how one could disagree with this basic principle—and easy to see that it is very widely ignored in practice. For Chomsky, humans are again rich in potential but bounded in scope in the political and ethical domain, but in this case the restrictions are mostly imposed by repressive political and economic forces to be resisted, rather than our biology (which can only be humbly acknowledged).

The bounded scope of human understanding is again the theme of the final chapter of this book, where Chomsky takes science as a model of our capacity to understand the world (returning to themes alluded to in the second chapter). He is particularly fascinated by the difficulties that Newton and his contemporaries had in accepting the notion of “action at a distance” that was central to Newton’s concept of gravity. For Newton, it was intuitively obvious that the only rational model for action involved mechanical contact between the actor and its object. The model for this was clockwork, where simple actions could be combined into highly complex behaviors in a seemingly unlimited way, but only because the gears and springs were in direct contact. But Newton was forced by his science to accept the notion of gravitational “action at a distance” while simultaneously finding this notion not just intuitively unsatisfying, but virtually unintelligible. During Newton’s era this intellectual conflict seemed every bit as great as that caused, more recently, by the “spooky” and non-intuitive aspects of quantum mechanics (which led physicist David Mermin’s to summarize the Copenhagen interpretation as an injunction to “shut up and calculate!”). There are two morals: one is that our intuitions provide a poor guide to (or yardstick of) scientific reality, and the second is that science often makes progress precisely by learning to overcome and ultimately ignore (or retool) intuitive conceptions and prejudices.

Despite the diversity of topics covered in these wide-ranging chapters, several unifying threads are clearly visible throughout the book (and I think reflect the unity of Chomsky’s thought more generally). These include the importance of the capacity for puzzlement in intellectual progress at either the individual or scientific levels, and a conception of the human mind as simultaneously bounded in scope (largely by our biology) but at the same time remarkably flexible, creative and free, at a level that dwarfs the cognition of any other species. This view of our species—as cognitively “bounded but rich”—encapsulates the answer to the title’s question “what kind of creatures are we?” We are both bounded, in ways that should keep us humble, and rich in creativity and potential, in ways that should make us hopeful as intellectuals and defiant against oppression and coercion in the political domain.

In summary, by making evident his capacity to be puzzled by things that others find trivial or obvious, his willingness to reject accepted dogma in pursuing such puzzles, and the often non-intuitive conclusions he reaches after decades of thinking, Noam Chomsky reveals himself in this book to be not just a scholar’s scholar, rich in insight and erudition, but also a surprisingly accessible communicator. With perhaps a bit of extra effort, any college-educated reader can absorb and profit from the wisdom distilled into this book. I think it is the clearest and most accessible summary available of this important thinker’s key insights, and recommend it without reservation.

Endmark

  1. This anecdote and other personal details stem from a series of interviews I conducted and recorded with Noam Chomsky at his MIT office in Dec 2010; I also thank him for comments on the current review. 
  2. Noam Chomsky, Syntactic Structures (Paris: Mouton & Co., 1957). 
  3. See Noam Chomsky, “A Review of ‘Verbal Behavior’ by B. F. Skinner,” Language 35, no. 1 (1959): 26–58; B. F. Skinner, Verbal Behavior (New York: Appleton-Century-Crofts, 1957). 
  4. Howard Gardner, The Mind’s New Science: A History of the Cognitive Revolution (New York: Basic Books, 1985). 
  5. Noam Chomsky, Aspects of the Theory of Syntax (Cambridge, MA: MIT Press, 1965). 
  6. Tom Wolfe, “The Origins of Speech: In the Beginning was Chomsky,” Harper’s Magazine (2016): 25-40. 
  7. Noam Chomsky, The Minimalist Program (Cambridge, MA: MIT Press, 1995). 
  8. See Ray Jackendoff, Foundations of Language. (Oxford: Oxford University Press, 2002); Steven Pinker and Ray Jackendoff, “The Faculty of Language: What’s Special About It?” Cognition 95 (2005): 201–36. 
  9. Marc Hauser, Noam Chomsky, and W. Tecumseh Fitch, “The Language Faculty: What Is It, Who Has It, and How Did It Evolve?” Science 298 (2002): 1,569–79. 
  10. See Marc Hauser et al., “The Language Faculty: What Is It, Who Has It, and How Did It Evolve?” Science 298 (2002): 1,569–79; Robert Berwick and Noam Chomsky, Why Only Us: Language and Evolution. (Cambridge, MA: MIT Press, 2016). 
  11. Noam Chomsky, Language and Mind (Mew York: Harcourt, Brace & World, 1968). 
  12. W. Tecumseh Fitch, The Evolution of Language (Cambridge: Cambridge University Press, 2010), 603. 
  13. W. Tecumseh Fitch and Klaus Zuberbühler, “Primate Precursors to Human Language: Beyond Discontinuity” in Evolution of Emotional Communication: From Sounds in Nonhuman Mammals to Speech and Music in Man, eds. Eckart Altenmüller, Sabine Schmidt, and Elke Zimmerman (Oxford: Oxford University Press, 2013), 26–48. 
  14. Noam Chomsky, What Kind of Creatures Are We? (New York: Columbia University Press, 2016), 12. 
  15. Chomsky, What Kind of Creatures Are We?, 63. 

W. Tecumseh Fitch is an American biologist and cognitive scientist at the University of Vienna.


More on Linguistics


Endmark

Copyright © Inference 2024

ISSN #2576–4403