Jerry Fodor 1935-2017
04 December 2017
In his introduction to a public lecture Jerry Fodor gave in Cambridge about 15 years ago Ross Harrison referred to him as 'one of the great philosophers'. The only person I could see showing surprise at this was the man himself. His position in philosophy is secure. But Fodor was also a theoretical psychologist; indeed, an experimental psychologist.
For my money, Jerry Fodor was also one of the great psychologists, so great indeed that one can, within cognitive developmental psychology at least, talk about BF (before Fodor) and AF (after him). 'BF'… an unfortunate shorthand surely? No, serendipitous. For some of us were indeed 'bloody fools' to have been as happy as we were to theorise as we did BF. Most broadly, we conflated development and learning (more on this later). We speculated merrily about how children might bootstrap themselves from a lower developmental level to a higher one armed only with the representational format of the lower one. We tended to think that children could acquire a new concept from scratch simply by learning a word, or by way of a bit of judicious social interaction.
The Fodor works best known to psychologists are probably The Language of Thought (1974), with its stark message for developmental psychology, and The Modularity of Mind (1983). The first was clearly a work of philosophy while the second looks more like theoretical cognitive psychology. The Modularity of Mind can be imaged: at the top, a horizontal cloud (the Central Systems); at the bottom a horizontal line of matchboxes (the Input Systems). The former are, among other things, domain-general and global; the latter are modular, for being which Fodor gave nine criteria (e.g., mandatory, informationally encapsulated). Arrows travel in isolation from each matchbox up to the cloud and no arrows travel downward. Sentential parsing, he argued, is modular; and he quotes at the start of the book Merrill Garrett’s view that basically parsing is a reflex.
Nowadays, this looks rather quaint and too contentious for its own good; but at the time it was just what the doctor ordered. We don’t always like our medicine; and many cognitive psychologist gagged on 'Fodor’s First Law of the Non-Existence of the Cognitive Sciences' which states that the less modular a system the less we understand it, and that when it comes to thinking we hardly understand it at all.
Empiricism is the philosophical position that all knowledge derives from perceptual experience. Fodor was the anti-empiricist, the heir of Plato and Descartes, and so it was not surprising that when connectionist modelling (empiricist to the core) came on the scene in the mid-1980s Fodor was strapping on his sword. These models, he pointed out, can categorise but they cannot form concepts; concepts are what we think in, thinking is systematic; and associative models with no symbolic content cannot model systematicity. Of course, much of our learning is associative/Pavlovian/Humean (e.g. German children learning the six forms of the definite article) but cognition itself is not associative. Fodor has been proved correct at least insofar as connectionism as a model of cognition has not delivered on its promise.
This brings me back to what I said earlier about our confusing development with learning in the BF era. Connectionist developmental theories turned back this particular clock. To expand. Mental development involves a lot of learning; but mental development itself is not learning. Do we learn to think ('first pick a subject, then add a predicate..')? Indeed, do we learn to walk ('put one foot forward then…')? So back to development and The Language of Thought. Fodor argued that knowledge acquisition is a matter of hypothesis formation and testing, hypotheses that can only come from a pool of innate hypotheses that must have a symbolic representational format… the famous Language of Thought (LOT). LOT is strongly analogous to the machine code of a digital computer. I hated this up until the mid-‘80s. Many of us did. 'Fodor is a pest,' Robin Campbell said to me around that time, but added 'an interesting pest'. As time went on, the pest-element withered and the interest filled the frame. There were some strong arguments here. Critics tended to argue that the LOT idea invited an infinite regress, and so Fodor had to spend a lot of time explaining how bedrocks do not themselves need bedrocks: that’s what a bedrock is.
At this point I must say loud and clear that we were not all BFs in the BF era. Certainly not Peter Bryant. In the lectures I attended in 1968 Peter discussed the behaviouristic view that the reason younger children cannot do a kind of relational learning that older children can do is that the latter have learned the word for the relevant property (e.g. 'bigger'). This, said Peter at that time and in a later book, 'leaves unanswered the awkward question of how they learned the word in the first place'. The passage in The Language of Thought in which Fodor cites Peter’s insight with gratitude, while acknowledging its uniqueness in psychology, I can recite like Invictus.
Before we get hoarse with cheering, it is well to bear in mind that Fodor would not be Fodor without the over-reaching, indeed without the bull-in-a-china-shop element. I’m sure many supporters watched in dismay as he took on Darwin in What Darwin Got Wrong (co-authored with Massimo Piatelli-Parlmarini). His main complaint was that the natural selection theory was a species of behaviourism, in which the environment selects between what organisms blindly offer up (his objection was to the 'blindly'). He argued too that Darwin tended to construe Nature teleologically. Second, I don’t know what philosophers thought of it, but I was underwhelmed by his LOT2 in which he defended his outrageous ideas about all concepts being innate. He seemed to be defending his position by way of a brilliant series of hair-splittings; and the dialogues became tiresome. In fact, many would say that Steven Pinker has proved to be a better spokesman for nativism than Fodor. His ideas about how children come to pass the false-belief task were conceptually unexciting and empirically wrong, as it turned out. While being just the person to take on the pretentions of cognitive neuroscience there was again over-reaching aplenty. To say that all we need to know is that these things go on 'North of the neck' would sound adolescent coming from anybody else. Finally in this flurry of negatives, even his writing on opera could come a cropper. 'Professor Fodor’s views on Die Meistersinger are informed by something other than by knowledge of the plot' wrote Michael Tanner to The London Review of Books.
But amid all this, Fodor’s arguments for nativism, LOT, and for other things, stand unblemished. Isn’t the point to do the right kind of thing whether or not we always do it well enough?
This brings me round finally to Jerry Fodor the man. Here I can only write from personal experience, and not much of that. I first saw him in a Q&A session that preceded the lecture I mentioned at the beginning. I was expecting some arrogance and not a few lacerating put-downs. In fact there was modesty and good humour. He began by saying how much he enjoyed the lunch Steven Butterfill had given him, looking round at the small group and saying 'One feels a bit of a fraud in circumstances like this,' before removing his suit jacket not by undoing the buttons but by tugging it over his head like a sweater.
I had come prepared with what I thought was a decent question: Why did he espouse globalism when it came to the Central Systems while being a sworn enemy of holistic theories of word meaning? (He held that it was possible to know the meaning of only a single word… could there be only a single thought too?). He gently explained that global and holistic are not at all the same idea and that in any case epistemology (global) and semantics (atomistic) are two domains never to be confused. He added 'I would like it to be emblazoned on my gravestone in neon "He never muddled his epistemology with his semantics."'
A year or so later I was writing a book on theories of syntactic development and emailed him for advice about his positions on this and that. He proved to be a chatty correspondent; and we discovered a shared love of Fred Astaire as a vocalist. Astaire and Fodor. Fred and Jerry. What pair could be more different? Maybe not. There was flair, speed, a surface of devil-may-care above a fierce grinding of detail, and a capacity to delight. I think delight is the word. Mostly, Jerry Fodor was a delight to read, whether encapsulating the contrast between two philosophies of mind within a single sentence of jazzy aplomb, teasing 'Granny' (who embodied those sunk in dogmatic empiricist or Wittgenstinian slumbers) or reviewing the Elton John-Tim Rice version of Aida… 'like processed cheese.'
- James Russell is Emeritus Professor of Cognitive Development at Cambridge. His Psychology as King of the Ghosts: A Personal Critique appeared last year, and was reviewed for The Psychologist online by Patrick Rabbitt.