I would like to share as much of this book as I can so I'll keep the commentary to a minimum. I know far more about Chomsky's theories than Piaget's,I would like to share as much of this book as I can so I'll keep the commentary to a minimum. I know far more about Chomsky's theories than Piaget's, so it was useful to be given a helpful comparison in the Foreword, by Howard Gardner:
Although both Piaget and Chomsky pay homage to models provided by and logic, they are fundamentally interested n quite kinds of examples and explanations. Piaget is fascinated by the behaviors children emit and, more specifically the errors they make when solving the challenging puzzles he poses. He has developed an elaborate technical vocabulary, rooted in biology, to describe thees phenomena, a rich description of the stages through which children pass in each of these realms of achievement, and his own logical formalism to describe the affinities that obtain across discrete mental stages. [...] At most, Piaget's adventures into technical vocabulary and formal models offer a convenient way of synthesizing the enormous amount of data he has accumulated. In the end, it is his overall vision of how capacities relate and of how knowledge in its varied forms develops that inspires workers in the field (of psychology). [Chomsky's work] is of a fundamentally different order. Rather than being struck by behavioral phenomena that he feels compelled to describe, Chomsky is driven by a powerful vision of how linguistic science should be pursued[...] In his view the student of linguistics should construct models of human linguistic competence and thereby specific the "universals" of langage. Foreword xxix-xxx
After the introductions by Gardner and the editor, Piattelli-Palmarini, we have two introductory papers by Piaget and Chomsky. It's probably a function of my lack of familiarity and strong familiarity with each respectively, but I felt as if Piaget's writing was a lot less clear than Chomsky's. I got the impression that he thought that, in the end, theories only shift investigative emphasis this way or another, and so prefers liberal use of biological metaphor, as Gardner has just criticised, for example using the term "fixed nucleus" rather than the technical term "universal grammar":
I believe I understand why Chomsky proposed this hypothesis ["fixed nucleus"]: simply because it is a very common opinion to presuppose that a behavior is more stable if it is firmly rooted, that is, if it is hereditary and not simply a product of auto-regulation. In other words, Chomsky's fixed nucleus would appear more stable, more important, and thus of higher value if it were hereditarily fixed. Piaget, p.57
Piaget is concerned with the actually developing individual, and so it's right that he prefers the term "nucleus" which denotes a broadly defined area of origin for genetic information, however locating only a small piece of the journey from genetics to behaviour. On the other hand, Chomsky is interested in the uniquely black and white nature of language and favours evidence that shows the fine line between sense and nonsense:
If there are two copies of this book [that John wrote] before us, I can point to either one and say, "John wrote this book," but I cannot conclude that John wrote two books. Chomsky, p.46
As the general reader, we are well equipped by the editor to understand what is spoken about, but that did not prevent certain parts of the discussion from flying hopelessly over my head. For example, this part in which, as per Piaget's borrowings from biology, the debate crucially hangs upon whether language development is "more like morphogenesis" or "like philogenesis." Hmm:
To say that language acquisition is "more like morphogenesis" is equivalent to invoking a problem solving or specialized inferential mechanism that is innately determined and that gives rise subsequently to a mental organ (a "body of beliefs") that is innately determined. To say that it is "more like phylogenesis" is to affirm that the mechanism is multipurpose and innately determined. However its output is not innately determined, since a new factor intervenes in its generation (random variation, in the case of Darwinian evolution). [...]The alternative becomes: is the mechanism a composer of programs (phylogenesis) or merely an executor (morphogenesis)? [...] If we adopt Chomsky's morphogenetic view, man becomes the passive plaything of genetic predestination and environmental determinism, whereas in Piaget's view a strong measure of free will and creative autonomy is maintained. Guy Cellérier 84
I understood it in the end, roughly, and it's to be expected that a book with contributors from so many areas (Cellerier is a biologist) should wander into the tall grass at times. Roughly, then, morphogenesis is the change of form that occurs within the same species, whereas philogenesis is the change of forms that occur in the course of speciation. In which case, the idea that Piaget prefers is that intelligence, including language intelligence, can branch out infinitely and endlessly specialize, just as the family tree of life (apparently) can.
On the other hand, Chomsky quite clearly shows examples where - with never a mistake made - language takes a limited form with black and white rules of correctness.
Before presenting his examples, he first stipulates that a "theory of learning" must be assigned some domain and organism (e.g. Theory of Learning for the domain of language in humans: LT(H,L)) unless it is to be an impossibly broad characterisation of intelligence for "all domains." Chomsky clearly resents that the system he advocates, which successfully explains much of what it ought to, is criticized in favour of a competing theory that doesn't exist. He proudly points out that he has eaten such attempts to define mechanisms for "general intelligence" for breakfast before:
Chomsky: [...] If anyone thinks there is such a general developmental mechanism, fine, propose it, make it explicit, then I and others will investigate it to see whether there is any relation whatsoever, be it direct or metaphorical, to the concrete problem of attaining the final state.
Patrick Suppes at Stanford has presented a very explicit and interesting proposal[...] It was clear and explicit enough to investigate, and he showed that this system could attain in the limit a finite-state Markovian system which produces symbols from left to right. That is interesting because we know that such a system is inadequate for language, that is, we know that hte systems that can in principle be attain within this model are not the system that are attained in fact by humans [...]this theory was presented in a clear enough way so that it was possible to determine whether or not it is correct. It is a merit of a theory to be proved false. Chomsky, p.111
The main piece of evidence in the first paper is that we can say "they liked each other," but not "They thought John liked each other." "Each other" requires a plural as the corresponding subject, even though "they thought John liked each other" could conceivably, in another world, be interpreted as "they thought John liked each of others":
It cannot be imagined that the language learner is taught these facts[...]. No one ever makes mistakes to be corrected. [..]Passive observation of a person's total performance might not enable us to determine whether the principles are in fact being observed though [thought] experiment will quickly reveal that it is so. The only rational conclusion is that the SSC and the relevant abstract notion of "subject" and "bound anaphor" are properties of S0 [the initial state of the language faculty], that is part of the Learning Theory in the Domain of Human Language, LT(H,L).
This comes in for criticism from the Piagetian, Seymour Papert:
Of course the child does not learn the SSC by being taught in any direct sense; this is not the way any fundamental structures or skills are ever acquired. And of course the child does not learn the SSC by passive observation. Piaget has removed any lingering tendencies people might have had to see the child as ever engaging in "passive observation." So I would have thought that the only rational conclusion from Chomsky's set of statements is that the child discovered through "experiments." If asked where the ability to do experiments came from, I would be a little more inclined to grant that this is a property of S0. (Seymour Papert, p.97)
On top of misinterpreting Chomsky's use of the word "passive" (our passive observation of children, not the child's observation of language) this strikes me as turning a blind eye to the obvious empirical test: do children in fact "experiment" with the SSC (the "each other" problem) property or not? I.e., do they ever make the mistake "they thought I liked each other but I didn't" and learn by trial and error that it is wrong?
What's more, as Chomsky hammers home, the point does not pivot upon the child's behaviour, it's that all of us with healthy language development agree that "they thought john loved each other" is wrong but "they thought he loved all the others" is fine:
Notice that there is nothing logically necessary about that principle; English would be a perfectly fine language if it didn't have that principle, only in English you would then say "John seems to the men to like each other." There would be no problems about communication; all the properties would be acceptable. It just wouldn't be a human language; it could be the language of another organism, namely an organism whose initial state would not include this principle (Specified Subject Condition). One could conceive of an organism exactly like humans, but minus the specified subject condition, and it would talk with a fine language which it could use for all possible purposes. In fact, we could observe that organism for a very long time and not even know that certain physically interesting properties have ever been realized. Likewise, if you were simply to observe the flow of experience, to make a movie of people talking for instance, you might not know whether thery are using the SSC at all, yet the fact is that everyone, without ever committing an error, uses that condition even if they have had no relevant experience whatsoever. Chomsky, p.123
Fodor makes the point even more categorically when he says that it's nonsense, since the prevailing computational theory of mind is the metaphor that the mind is a manipulator of symbolic logic, to say, as Piaget does, that the stages of the mind, in maturation, develop STRONGER logics that are capable of things erstwhile impossible is plain wrong:
One way of reading the Piagetian position is to say that if you [characterized the computational capacities of the organism] for several different time slices[...], what you would get is a fundamentally different galaxy of constraints on the organism's concepts. Moreover, this difference would have the following important characteristic: the logic instantiated by the system of concept at any i^th stage is weaker than the logic instantiated by the i-1^th stage. [...]In short, if you look at the organism as a succession of logics, and powerful in some fairly rigorous sense: as for example that the set of truths that could be expressed by using the concepts availabe at i is a subset of the truths that could be expressed by using the concepts available at i+1. (Fodor, p.147)
(Btw, is there a mistake here - shouldn't it read "stronger than the logic instantiated...."?)
Yet another strong point of Chomsky's is one raised by another contributor earlier: that since Piaget predicates exploration and development of language with exploration and manipulation of the physical world, why do the severely handicapped not suffer from handicapped language faculties?:
What about a weaker [version of Piaget's theory]? Could it be, for example, that the constructions of sensorimotor intelligence are a necessary condition for the emergence of language? [...]If in fact the constructions of senorimotor intelligence are a necessary condition for the development of language, then it should turn out that insofar as those sensorimotor constructions are impeded, the intelligence that leads to the acquisition of language should also be impeded; and if they are drastically reduced, language ought to be virtually eliminated. [....] blind children, who have [...] a significant reduction in their capacity to develop constructions of sensorimotor intelligence [...]acquire language more rapidly than sighted children, which isn't surprising because they are more dependent on it. [...] If a child were paralyzed, for example, my prediction would be that it would have no noticeable effect on his language development. Chomsky, 171
Glass, Premack and Gazzaniga have found something that I would have certainly predicted [...] that severe global aphasics (people who apparently have almost total destruction of the physical basis for language capacity) they were able to induce a system very much like the one that the chimpanzee acquired. Some other studies that I know of have suggested the same thing. I think this is the kind of result one would expect, because what it means is that a a chimpanzee is very smart and has all kinds of sensorimotor constructions (causality, representational functions, semiotic functions, and so forth), but one thing is missing: that little part of the left hemisphere that is responsible for the very specific structures of human language. (Chomsky, 181-2)
Chomsky does make one compromise to the programatic conception of language development, that I noticed. Namely, that the "stable state" that comes after the language acquisition state of growth may not be equal in all adults, even within the same language. He cautiously cites some evidence:
We talk about a [homogenous, adult] fixed steady state, which is of course idealized, it may very well be that the steady state attained is rather different among people of difficult education levels, even if there is no reason to believe that there is a difference in intelligence. Carol Chomsky did some work on the acquisition of moderately complex linguistic structures, as in sentences like "I asked him what to read " and "I told him what to read." If I say "...asked..." then I am doing the reading; if I say "...told..." then he is doing the reading. [The ability to distinguish the two develops] between ages 5 and 10. [...]Zella Lyria did a study on adults to find out whether they were [...]making these distinctions, and it turned out that there were a remarkable number of adults, at least on the basis of the most sophisticated tests the ycould devise, who were not making these distinctions. This is very hard for me to believe, and I must say there may be some data that are inconsistent with it. Chomsky p.175-6
A saving grace for the aging Piaget, who generally appears to have little new to say in the face of new evidence and theoretical frameworks, is that the biologists (or at least Changeux) sympathise with his belief that theorizing should be undertaken in close observation, and that we should be suspicious about inductive theories about how brain matter "must" organize itself. In my view this is overly cautious: I see nothing wrong with building a program for observation from a theory inductively sketched out. Without some guiding beliefs about the new territory of discovery, the morass of firing brain cells is a blank canvas for interpretation. As Chomsky likes to say, who's to authoritatively say that the brain is not to cool the blood?:
In the past decades inductive methods have appeared to be rather unsuccessful in the neurosciences. The reasons for this lie basically in the anatomical organization of the nervous system. Apparently simple operations such as the movements of the eye or the killing of a mouse by a cat, in fact involve the recruitment of a large number of neurons (thousands, even millions) from many different areas of the brain. In addition, no simple rule appears to exist in the macroscopic and histological organization of such centers in the brain. Why should there be any logic, for instance, in the presence and the role of subcortical structures in high "corticalized" mammals, except for the very fact that these structures existed in the brain of more primitive animals from which they evolved? A given behavioral act may indeed engage, simultaneously and necessarily, groups of neurons which appeared at different periods in the evolution of vertebrates. The stabilization or selection of these centers had its own logic at the time they were formed, but becomes masked by millions of years of history. [...] Anatomy cannot be inferred from anything other than its direct investigation. Changeux, p.185
Towards the end, Sperber makes a welcome attack against the vagueness of the semitoic/structuralist conception of a "grammar" of symbols and the hard-line "environmentalism" (belief that the environment and tradition shapes man's concepts) that it is inevitably accompanied by. He makes a good point against environmentalism in the form of a discussion of humour:
Every human being living in society knows how to recognize, understand, and appreciate an open repertory of witticisms. Now, now only are adults' instructions on grasping witticisms very vague, but in addition, a witticism is really recognized as such only by a person who does not need to see others laughing in order to find it humorous himself, and it is appreciated only by a person who does not need any explanation of it in the first place. [...] One cannot assert that acquired knowledge is deducible from heard utterances through the sole operations of a general intelligence deprived of innate specializations. Dan Sperber p.247
The thesis most generally accepted [...] is that semantic and symbolic represenations together would come under one semiotic function and would bear on signs, utterances, or symbols which the mind would decipher, in contrast to other objects of perception which the mind would just describe. [...] Either the semiological conception is without any precise empirical significance [...] or it must assert that there exists a grammar of symbolism comparable to that of language. A grammar is a device that enumerates (in the mathematical sense of the term) the sentences of the language. In this respect, a grammar differs from other mental devices the inputs or outputs of which cannot be enumerated. For example, no one could conceive of making a grammar of possible visual stimuli. The question, therefore, is to know whether the set of phenomenon susceptible of being symbolically interpreted is enumerable and dependent on a grammar, or whether it can be defined only by input conditions. Once expressed in these terms, the answer is obvious: [...]every object of thought can elicit a symbolic evocation. Dan Sperber p.248
This is a book that is filled with ideas and features a meeting of minds of diverse fields that is exciting for its rareness. It is patiently and considerately edited to maximize the ease of reading for the non-expert reader, although it is a rather heavy read and by it's nature it is wont to be rather dull in some areas, depending on where you're coming from. ...more