A Chomsky Reader

Oct 01 2010 Published by under Quotable

"Questions of fact cannot be resolved on the basis of ideological commitment." --Noam Chomsky

Not familiar? Not a problem. Here, in much abridged form, are some of the main ideas from "Language and Problems of Knowledge: The Managua Lectures," one of the most accessible of Chomsky's texts on language, and a clear, cogent and articulate elaboration of his views.

Why post this? I frequently receive comments on my posts accusing me of caricaturing Chomsky's position.  "Chomsky didn't say that!" the standard line goes (and sometimes, more vehemently: "Chomsky wouldn't say that").  There's really no good response to this except to say, "yes he did, it says so right here, on page..." which strikes me as annoyingly pedantic.  While I would encourage all my readers who have more than a passing interest in the debate to read the original texts, this compilation of quotes should serve as a helpful introduction (or refresher) on Chomsky's ideas.  Of course, The Managua Lectures are a concise elaboration of Chomsky's theoretical stance, rather than arguments for or against it -- but this is actually quite helpful, since it allows me to present short, well-formulated excerpts, without doing a hack-job on an extended piece of reasoning.

Of the excerpts that follow, there are some that I think are reasonable, some absurd, and some simply amusing. In the upcoming months, I hope to provide a similar treatment for some of Chomsky's other books and for Wittgenstein and Skinner.

On innate concepts:

"The speed and precision of vocabulary acquisition leaves no real alternative to the conclusion that the child somehow has the concepts available before experience with language and is basically learning labels for concepts that are already part of his or her conceptual apparatus. This is why dictionary definitions can be sufficient for their purpose, though they are so imprecise. The rough approximation suffices because the basic principles of the word meaning (whatever they are) are known to the dictionary user, as they are to the language learner, independent of any instruction or experience."

"Now that can only mean one thing. Namely, human nature gives us the concept "climb" for free. That is, the concept "climb" is just part of the way in which we are able to interpret experience available to us before we even have the experience. That is probably true for most concepts that have words for them in language. This is the way we learn language. We simply learn the label that goes with the preexisting concept. So in other words, it is as if the child, prior to any experience, has a long list of concepts like "climb," and then the child is looking at the world to figure out which sound goes with which concept."

On how little is actually determined by experience:

"...there must be possibilities of variation permitted by the fixed biological endowment, these possibilities being resolved by experience; the same must be true of the variety of languages more generally. But a great deal is constant, determined quite independently of experience. Analogy [i.e., learning by analogy or generalization] seems to be a useless concept, invoked simply as an expression of ignorance as to what the operative principles and processes really are."

On how the innate grammar structures the linguistic input:

"Suppose that a child with the human language faculty as a part of its innate endowment is place in a social environment in which Spanish is spoken. The language faculty selects relevant data from the events taking place in the environment, making use of these data in a manner determined by its internal structure, the child constructs a language, Spanish, or more properly the variety of Spanish to which it is exposed. This language is now incorporated in the mind. When the process is completed, the language constitutes the mature state attained by the language faculty. The person now speaks and understands this language."

On what the child "knows" in advance:

"The child learning Spanish or any other human language knows, in advance of experience, that the results will be structure dependent. The child does not consider the simple linear rule R, then discard it in favor of the more complex rule R-Q, in the manner of the rational scientist inquiring into language. Rather, the child knows without experience or instruction that the linear rule R is not a candidate and that the structure-dependent rule R-Q is the only possibility. This knowledge is part of the child's biological endowment, part of the structure of the language faculty. It forms part of the mental equipment with which the child faces the world of experience."

On the import (or lack thereof) of general learning mechanisms in language learning:

"Evidently, the language faculty incorporates quite specific principles that lie well beyond any "general learning mechanisms," and there is good reason to suppose that it is only one of a number of such special faculties of mind. It is, in fact, doubtful that "general learning mechanisms," if they exist, play a major part in the growth of our systems of knowledge and belief about the world in which we live--our cognitive systems. ...It is fair to say that in any domain in which we have any understanding about the matter, specific and often highly structured capacities enter into the acquisition and use of belief and knowledge."

This claim provided the impetus for the post "The question is: are you dumber than a rat?" and the follow-up "Intelligent Nihilism." For an informative (and clarifying) discussion of what learning models can and cannot do, see: Rescorla (1988). For some recent applications of learning theoretic models to language, see e.g., Ramscar et al. (2010), Koo & Collins (2010), Baayen (2010)and Singh-Miller & Collins (2007). For some recent work in information theory, suggesting that people are sensitive to how words distribute probabilistically in language see e.g., Jaeger (2010), Levy & Jaeger (2007). Both the Jaeger and Ramscar et al. (2010) papers provide comprehensive citations of other theoretical work in this vein.

On the empiricist tradition:

"In this empiricist tradition it was held that the constructions of the mind result from a few simple operations of association on the basis of contiguity, phenomenal similarity, and so on, perhaps extended by a capacity for induction from a limited class of cases to a larger class of the same type. These resources must then suffice for all intellectual achievements, including language learning, and much else."

On behaviorism:

"Attention to the facts quickly demonstrates that these ideas are not simply in error but entirely beyond any hope of repair. They must be abandoned, as essentially worthless. One has to turn to the domain of ideology to find comparable instances of a collection of ideas accepted so widely and with so little question, and so utterly divorced from the real world. And in fact that is the direction in which we should turn if we are interested in finding out how and why these myths achieved the respectability accorded to them, how they came to dominate such a large part of intellectual life and discourse."

Compare this to R. Roediger's take on Chomsky and Kenneth MacCorquodale's review of Chomsky's review of Skinner.

On what the language faculty and universal grammar amount to:

"A theory of the language faculty is sometimes called universal grammar... universal grammar attempts to formulate the principles that enter into the operation of the language faculty. The grammar of a particular language is an account of the state of the language faculty after it has been presented with data of experience; universal grammar is an account of the initial state of the language faculty before any experience. It would include, for example, the principle that rules are structure-dependent, that a pronoun must be free in its domain, that there is a subject-object asymmetry, [etc]... Universal grammar provides a genuine explanation of observed phenomena. From its principle we can deduce that the phenomena must be of a certain character, not some different character, given the initial data that the language faculty used to achieve its current state."

On parameters:

"The principles of universal grammar are exceptionless, because they constitute the language faculty itself, a framework for any particular human language, the basis for the acquisition of language. ...The observed facts...follow from the principles of the language faculty in combination with the data presented to the language learner, which have determined various options left unsettled by universal grammar. The principles of universal grammar have certain parameters, which can be fixed by experience in one way or the other. We may think of the language faculty as a complex and intricate network of some sort associated with a switch box consisting of an array of switches that can be in one of two positions. Unless the switches are set one way or another, the system does not function. When they are set in one of the permissible ways, then the system functions in accordance with its nature, but differently, depending on how the switches are set. The fixed network is the system of principles of universal grammar; the switches are the parameters to be fixed by experience."

On what a fully-described generative grammar is supposed to achieve:

"At this stage, the linguist is attempting to construct a grammar of a particular language, that is, a theory of that language. If the grammar is sufficiently explicit--what is called a generative grammar--it will predict an unbounded range of structured expressions and can be tested for empirical adequacy by investigation the accuracy of these predictions."

"The task of description is difficult enough, but the task of explanation, of developing universal grammar, is far harder and more challenging. At the descriptive level the linguist is presented with an array of phenomena and seeks to discover a computational system that will account for these phenomena and others that are predicted. At the explanatory level it is necessary to show how the phenomena can be derived from invariant principles once parameters are set. This is a far more difficult task..."

On how comprehension works:

"The computational principles of the mind now carry out a series of operations to yield the actual sentence."

"...the mind makes use of general principles of universal grammar and certain values for parameters and, of course, the meanings of particular words. These resources should suffice to determine the form and meaning of any sentence."

"For a person to understand a linguistic expression, the mind/brain must determine its phonetic form and its words and then use the principles of universal grammar and the values of the parameters to project a structured representation of this expression and determine how its parts are associated."

"The specific interpretation assigned... is determined by a series of mental computations carried out by the language faculty in accordance with its fixed principles, making use of information provided by the choice of parameters and the lexical properties that are specific to [the language]... The computations involved in determining the meaning... are moderately complex. At several points in the computation, more than one option is available, but only one is selected because others lead to the violation of general principles of universal grammar, and the path to the correct analysis involves quite a few steps."

On why, actually, the language faculty can't account for comprehension:

"The actual use of language involves elements of the mind/brain that go beyond the language faculty, so what the speaker perceives or produces may not precisely reflect the properties of the language faculty taken in isolation. In cases such as these, where speakers of a language have no clear idea of what an expression means or are informed that their interpretation is not the correct one, the speakers "think about the expression" (whatever this means), and after a period of reflection a conclusion springs to mind about the meaning of the expression."

On the "reality" of his theory:

"The computations just reviewed and the representations they form and modify have the same claim to reality as other constructs of science: chemical elements, valence, molecules, atoms, and so on. They enter into the explanation of curious and complex phenomena, and we can look forward to the discovery of physical mechanisms that have the properties brought to light in this inquiry into the functioning of the human mind/brain."

On the take home message:

"So, I think the answer to your question is, I don't think modern linguistics can tell you very much of practical utility."

9 responses so far

  • Marcus says:

    This whole thing with Chomsky seems very odd and dogmatic to me. I can't think of any other person that researchers get so riled up about, for or against.

  • Avery Andrews says:

    Well one thing that needs to be said is that Chomsky made linguistics much more fun than it was previously. My first exposure was Gleason's (1957) _Introduction to Descriptive Linguistics_, which was in my high school library, and later at college I got started on transformational grammar, which was really much more entertaining, because you could integrate much more different things into the analysis. As in my early work on 'non nominative subjects in Icelandic', and other work by other people on all sorts of things.

    The basic ideas of recursion etc are considerably older than Chomsky, being quite evident in Bloomfield and Harris, and let's not forget Otto Jesperson. What transformations let you do is combine your descriptions of different constructions in a much wider variety of ways than Bloomfield or Harris could, as I might try to discuss at greater length when I'm not on the road and can upload stuff more easily,

    I think Melodye is greatly overestimating the extent to which recursion and systematicity have been refuted, for example in the last 1000 years of English there have been various changes in the structure of NP, which appear in all positions where NPs occur now, indicating that NPs really do exist after all. So once upon a time you could say things like 'my the beloved son' in lots of grammatical positions, and now its wrong in all of them, supporting the idea that there really is an NP phrase-type, which has various possible internal structures, and various places it can appear, with each of these able to change independently.

  • GuessHandsOn says:

    Thanks for these maelstrom of Chomsky quotes.

    I'm now once again conviced that Chomsky has the right theory about language acquisition and its innate aspects.


    I'm looking forward to your posts in which you'll be discussing some of the corroborating evidence.

    Thanks for all your hardwork. Once again...


  • daedalus2u says:

    Is Chompsky a dualist? That is, is he arguing from the position that the "mind" and the "brain" are two separate things? and that "language" is something "sprinkled" on to that dualism?

    (by "sprinkled" on, I mean is quasi-independent and added later and to some degree separable from both mind and brain).

  • Fred says:

    @daedalus2u Several people have claimed that Chomsky is a dualist; an interesting review if you can track it down is a book chapter by Robert Kluender called "In Search of the Golden Slash" (referencing C's frequent use of the "mind/brain" collocation, and what it might mean).

    In fact, however, if you read enough of Chomsky's stuff, the ontological position he stakes out is more or less monadic, but actually seems to end up closer to idealism than physicalism/materialism. You can find several places where he says something to the effect of (paraphrasing) "The mind/body distinction was collapsed when Newton disposed of the concept of 'body'" (I'm probably not remembering that exactly right).

  • Avery Andrews says:

    I've heard it just the way you remember it, several times, so I think you are remembering it right.

  • daedalus2u says:

    The idea that all or even most, or even many concepts are "innate" is just wrong. There is no way that a brain can be specified to already have "innate" concepts.

    Most concepts are known to be not innate, they take active learning, either to learn them from someone else, or to generate them de novo. Parsimony suggests that the mechanism for acquiring most concepts is very likely the mechanism by which essentially all of them are acquired.

    Even if the brain could be specified to have innate concepts, that doesn't explain how those concepts can be understood, it just posits a homunculus inside that does the innate understanding. How does that homunculus have an innate understanding? Another homunculus inside that one?

  • >>>NOTE:- Links to other interests replaced with a link to relevant material at human-rights dot org<<<

  • […] about humans -- whether they come with language acquisition devices or no; whether they are stocked with innate concepts or merely with the desire to think they are; whether there is truly a hardwired "tool kit" (as […]