Behaviourism- as pioneered by Pavlov, Thorndike & Watson- was the dominant 'scientific' school of psychology in post War America. B.F Skinner, tenured Professor at Harvard from 1948-1973 was its high priest. Chomsky first gained fame, in 1959, by attacking Skinner's 1957 book 'Verbal behaviour'. Chomsky's 'nativism' (the idea that language is 'innate') was seen as championing freedom & creativity & the qualitative difference between human behaviour and that of 'meat-machines' (animals). However, the great utility & increasing sophistication of large language models- based on Reinforcement Learning from Human Feedback (RLHF)- seems to support Skinner, not Chomsky.
On the other hand, Chomsky's first mentor, Zellig Harris- who was critical of Skinner's book- did foundational work on distributional structures in the 1950s thus providing the theoretical underpinning for modern LLMs and vector-space models of meaning. His "distributional hypothesis"—the idea that the meaning of a word is defined by its environment and the company it keeps—is the core principle behind how LLMs learn to process and generate language. By contrast, Chomsky produced exactly zero 'Generative Grammars' of any language. This was a linguist who started to forget languages he had learnt and whose naive 'computational' metaphor was repudiated by neuroscience. In other words, he specialism was in ultracrepidarianism- i.e. pontificating about things of which he had very little knowledge and doing it in a paranoid style.
To be fair, there was an aspect of 'moral panic' & 'straw-manning' to the polemics associated with the anti-Skinnerian crusades. Everybody could claim that they had been misrepresented. This only left the question as to which approach would yield more useful results. Suppose there was an 'i-language' based machine translator rendering English into Russian with high fidelity, then Chomsky would be on the winning side. If the translator is statistical and 'e-language' based, then Harris & Skinner had been on the right track all along. We might say Skinner should have sugar-coated things but, au fond, the guy was doing good science'.
Chris Knight has suggested that Chomsky- afraid his work might have military value- deliberately made it so abstract as to be useless. The problem here is that abstraction would be a good thing if the program itself was viable. It wasn't. It was nonsense.
I think, Chomsky's value to his colleagues at MIT is that he made it appear as though his profession was 'self-regulating'- i.e. had its own internal watchdogs equipped with a liberal sensibility and a moral conscience. Indeed, Chomsky was a member of the Pound Panel in 1969 which sought to defuse anti-War protesters from attacking Pentagon funded laboratories.
It did not harm that Chomsky's political tirades were naive and counter-productive. All that was required of him was the appearance, not the reality, of liberal or humanistic values and deep sympathy for Left wing socio-economic policies.
I suppose even Chomsky must have finally come to understood that everything he believed or advocated for was nonsense. Perhaps he hoped Nowak- whose Evolutionary dynamics research Epstein was financing- or some other such math maven could find a way to vindicate him. What is certain is that the Epstein scandal had destroyed Chomsky's legacy. It seems his son, who was in charge of his deceased mother's trust, unwittingly contributed to the destruction of his own most valuable inheritance- viz. the moral halo associated with the Chomsky brand-name. This was because the father was left short of money and thus had to turn to Epstein for help. Interestingly, Ehud Barak- a Mathsy Israeli general and senior politician- was another interlocutor Chomsky gained through Epstein. But this undermines his anti-establishment, anti-Zionist, image. The one good thing is that, because of his age, Chomsky can't be accused of sexual impropriety.
The most influential opponent of Skinner back in the Sixties and Seventies was Arthur Koestler who advocated a holistic philosophy. Like Chomsky, Koestler was anti-Stalinist and a great believer in human freedom & creativity. Koestler's reputation declined as genuine biologists, like Medawar, criticized his Lamarckian evolutionary theory. But that wasn't what was utterly fatal to his reputation. After his death, it was revealed that he had a predatory attitude towards women- he raped Michael Foot's wife!- and thus he was dismissed as a sociopathic charlatan who must be 'cancelled' .
Interestingly, Skinner, in his 1948 Utopian novel, Walden 2, had supported many 'progressive policies' under the rubric of what we might term 'mechanism design' or even 'nudge' theory. The problem is that this compromises human freedom & dignity. It may be that we can be indeed manipulated into being good but is would that actually be a good outcome? This does not mean that we shouldn't try to make social mechanisms 'incentive compatible'- i.e. ensuring there are penalties for doing harmful things and rewards for doing helpful things. It just means we should not consider the outcome Utopian or universalizable.
Had Chomsky done something useful with his life- rather than frittered away his energies writing stupid ultracrepdiarian shite, he wouldn't, in old age, have had to turn to Epstein for help. Money matters. In the short run, you can make enough of it by bullshitting. Long term, utility is what gives you liquidity. Lack of liquidity drove Chomsky into the sticky hands of Epstein. The Pavlovian reflex of the brain-dead Left is to eviscerate an old man who was never guilty of any type of sexual misconduct or financial irregularity in his whole life.
Thus Skinner's victory has been total.
No comments:
Post a Comment