Chomsky in retrospect — 2
The debacle of Chomsky’s foray into historiography is in keeping with the fact that after 1960 he became complacent and gradually stopped doing serious academic work. Notably, his Aspects of the theory of Syntax (MIT Press, 1965), though still regarded as a seminal work by his followers, is in fact a sloppy, not well thought-through and immature work. Its main inadequacy lies in the fact that the kind of deep structure version proposed in that work does not represent meaning adequately, whereas it is claimed—a claim he gave up soon afterwards—that it does. In particular, as was observed by a few linguists, including Paul Postal, Jerrold Katz, Jim McCawley and also myself, Chomsky’s 1965 version of deep structure was unable to express scope differences, such as exist between sentence pairs like No Dutchman knows two languages and Two languages are known by no Dutchman, of which the first is false and the second probably true in the world as it is.
The standard way to account for scope differences is to treat the crucial elements as (logical) operators, whose placement with respect to each other in logical structures in most cases causes semantic differences. In the cases at hand, no and two are such operators and their scope with regard to each other determines which interpretation applies in either case. The first sentence would then be read as something like ‘there is no Dutchman such that there are two languages such that the Dutchman knows these languages’, while the second sentence would read as ‘there are two languages such that there is no Dutchman such that the Dutchman knows these languages’. Such scope differences are very widespread in all languages. Take, for example, the three sentences Because of the rain I did not go out, Not because of the rain did I go out and the ambiguous (pace intonation) I did not go out because of the rain, where the (logico-)semantic difference is naturally described in terms of scope differences between the two operators because of and not. Such differences exist in all languages in analogous ways (at least, no counterexamples are known), which means that speakers do not have to learn them. Since the version of deep structure proposed in Aspects lacks the means to express such scope differences, it was felt to be necessary to assume a more abstract and more logic-oriented version of deep structure. In fact, as was stressed by Jim McCawley and, independently, by me in my PhD dissertation Operators and Nucleus of 1969, the formal language of predicate logic, as developed by Peano, Whitehead and Russell, would serve well as an approximate model for the kind of deep structure we were after. This then gave rise to the idea that a grammar of a language L is an algorithmic, meaning-preserving mechanism transforming logico-semantic ‘deep’ structures into the surface structures of L.
This newly found structural relation between grammar and logic came as an enormously liberating experience to those who were working in this area, as it was felt that we had found the way back from a purely form-based method of description of language to one that would integrate meaning—a notion that had hitherto been considered out of bounds to science. (For me, this discovery has been a very fruitful main driving force behind my work ever since.) But when, around 1968, young linguists, mainly those mentioned above, actually started to develop a more abstract model of grammar, which they called Generative Semantics, in terms of a modular algorithm transforming a logico-semantic input into a phonological output, Chomsky, having first gone along with this development, suddenly turned round and set out to destroy it, while claiming, against obvious evidence, that his own views had been consistent all along. He reverted to his earlier, pre-1965, view of autonomous syntax, adding the notion that the structures randomly generated by an ‘autonomous’ generative algorithm would be open to (i) a phonological and (ii) a semantic ‘interpretation’—the latter in terms of an ill-defined ‘logical form’ (LF). This got him into a mess, as this notion of grammar did not sit well with his newly acquired realism: speakers obviously do not randomly produce well-formed strings of symbols and then ‘interpret’ them.
Why did he do that? There are good reasons to think that it had everything to do with his amour propre and nothing with academic argument. The initiative was being taken over by younger linguists, who were engaging in sophisticated syntactic arguments and were attracting many more students than he did at the time—as I saw with my own eyes when I visited MIT in December 1969. During those few years, the vast majority were doing Generative Semantics, not Chomsky’s ‘Autonomous Grammar’, which was clearly on the wane. When this happened, he pulled out all his reserves, in fact issuing a fatwa on the new movement. He started a war of defamation and isolation, compelling his students to ignore anything written by the ‘enemies’, misrepresenting the generative semanticists’ positions in his classes (I was there: they were a sham), and posturing as a guru leading his congregation on the strength of his alleged supernatural insights—all this without even the shadow of an academic argument and without ever answering those who did present arguments. And he won, at least for the eyes of the world and for the time being. (In later years, in his popular but, significantly, not in his more technical talks and writings, he dishonestly took to presenting the Generative Semantics view of sentences as schemata for the expression of thoughts as his own view.)
Generative Semantics was thus buried without a tombstone in the dry sands of the desert, meant to be forgotten forever. Its practitioners mostly went on to do other things. Ross turned to the teaching of poetry; Lakoff took to posing as a cognitivist guru; McCawley (who died in 1999) gave up his brilliant strain of thought and started doing less cutting-edge work, incorporating a great deal of traditional grammar with a strong shot of pragmatics. Postal went on to develop an alternative theory of grammar, Relational Grammar, and has stayed a productive linguist till the present day. Less prominent generative semanticists just toed the line, hiding their youthful sins. The only one who went on doing pure Generative Semantics, under the more appropriate label of ‘Semantic Syntax’, was me. But when I published my book Semantic Syntax (Blackwell, Oxford, 1996), it just fell on the dry sand the movement had been buried in. There was not a single reaction from any theoretical linguist. The only reactions came from computational linguists, who appreciated that this book presented the central elements of fully algorithmical and crucially correct grammars converting well-motivated logico-semantic structures into well-formed surface structures of English, French, Dutch, German, and to some extent also of Turkish. But theoretical linguistics buried the book in desert sand along with the movement it instantiated. Fortunately, desert sand is just about the best material for preserving treasures….
The period between 1970 and 1980 is now known as that of the ‘linguistics wars’ (the term was introduced by Randy Harris in his 1993 book The Linguistics Wars). And it is here that the rot in Chomsky’s work, which had been there all along, came to the surface. It became clear that the ultimate driving force behind his work was the urge to be seen and revered as a superhuman being, not to advance science by developing better theories. In fact, as from as early as 1960, when he was barely thirty years old, Chomsky increasingly refused to submit to the normal discipline of academic work, subordinating the norms and rules of academia to his own interests. And he was bewilderingly successful in this—which, of course, also reflects badly on those he fooled.
This raises the question of how Chomsky managed to achieve this. The answer lies, I believe, to a large extent in his way of exploiting and abusing people’s reasonable expectations of fellow humans’ standards of honesty, politeness and civil behaviour, always exploring the limits of how far he could go. In this respect he is indeed exceptionally gifted. He has always been extremely fast on the uptake, which still makes him a dangerous and slippery opponent in discussion. But a little study and observation will make one see through his ploys. A typical feature of his written and spoken texts is the use of value judgements reflecting only his own assessment, without any further argument, such as “the only serious analysis” or “outlandish arguments” or “the best work, in my opinion” and many others of a similar nature, meant to browbeat the reader or discussant into believing that what they took to be a valid counterargument must be inferior and based on bad thinking or lack of knowledge and, at the same time, making it impolite on the part of the discussant to raise or maintain the objection they may have had in mind.
Or not mentioning essential literature when discussing opponents, as in his lengthy interview with the French politically engaged linguist Mitsou Ronat, held in 1976, and recently republished in Noam Chomsky, On Language (The New Press, New York, 2007), where, on pp. 148–162, the school of Generative Semantics is discussed. A great deal of attention is paid there to Jerrold Katz and Jerry Fodor’s 1963 article ‘The structure of a semantic theory’ (Language 39.2, pp. 170-210), where a view of semantic interpretation is defended that Chomsky embraced at the time of the interview. But the seminal book An Integrated Theory of Linguistic Descriptions (MIT Press, 1964), by the philosopher Jerrold Katz and the linguist Paul Postal, which actually gave rise to Generative Semantics on the strength of highly persuasive syntactic, not semantic, arguments, is never mentioned.
Or, in discussing a movement or school of thought he disapproves of, he will mention a few names, including one or two that are not typical at all of the school concerned, and then focus on those non-typical names. Thus, in the same interview with Mitsou Ronat on the same topic, Chomsky speaks at length about Charles Fillmore, who did not belong to the group of Generative Semanticists but had gone his own way, without making any mention at all of James McCawley, who made the most centrally important and brilliant contributions to that theory. Outsiders will not notice, but insiders do. Another ploy, often used in discussions, consists in interrupting the discussant in mid-sentence and finishing the sentence in a way not intended by the speaker, and then replying to the part of the sentence supplied by himself. Or he will argue on the strength of examples that have little to do with the topic at issue, or ask questions in return he knows his opponent cannot answer, or play on the fact that the opponent has very little time to phrase his point while he has all the time in the world. And a few more such ploys could be mentioned. The only counterstrategy that has proved effective is to lay bare such tricks in public as the discussion is going on. But that means breaking the rules of polite discussion—a necessary step if one wants to maintain them. … to be continued