Fire the press secretary

Sometimes, when I’m feeling dyspeptic, I wonder why psychologists have such a down on minds. Psychologists, of all people. In philosophy, ever since Plato, the mainstream opinion has been that the mind is the organ of thought; thinking is what the mind is for, and we act as we do because we think what we do. But psychologists, for the last hundred years or so, have mostly viewed that sort of ‘intellectualism’ as an embarrassing remnant of the Enlightenment: behaviourists say that the question of what minds are for doesn’t arise, since there aren’t any. Freudians say that the myth that we think is a sort of cover story that the mind tells itself to avoid having to confess to its libidinous urges. Associationists say that we don’t need a mind to think with (‘we don’t need an “executive”’ is how they put it) because ideas think themselves in virtue of the mechanical connections among them. And neuropsychologists say that since the mind is the brain, we don’t need the one because we have the other. That this bundle of muddle is recommended as the hard-headed, scientific way to do psychology is, I think, among the wonders of the age.

Anyhow, the version of anti-intellectualism that’s current is Psychological Darwinism, and Robert Kurzban’s new book is a typical instance. I’ll take a paragraph or two to sketch the connection between what Psychological Darwinists say and the anti-intellectualist thesis that there aren’t any minds. After that, I’ll focus on the latter, as does Kurzban.

Psychological Darwinism is the theory that some/many/all of the traits that constitute our ‘psychological phenotype’ (roughly, the catalogue of our innate mental traits) are adaptations to problems posed by the environments in which the mind evolved. There is plenty of disagreement about the details, but here’s the story to a first approximation. 1) The evolution of our psychological phenotype was largely driven by intra-specific competition for (viable) offspring. 2) Natural selection is the primary mechanism by which phenotypes evolve. This is true of psychological phenotypes inter alia. 3) The function of a phenotypic trait is determined by those of its properties in virtue of which it contributes to the adaptivity of creatures that have it. Thus, tigers have stripes because having stripes affords camouflage and, all else being equal, tigers that are camouflaged have more viable offspring than tigers that don’t. But bear in mind that the adaptivity of a phenotypic trait is its contribution to fecundity in the environment in which it was selected; not (or not necessarily) in the environment that it currently occupies. A penchant for a calorie-intensive diet was adaptive when we were running around in the veldt; but it probably isn’t adaptive now. (Followed by much minatory wagging of fingers at the obese.) 4) Our psychological phenotype is pretty much what you would expect it to be if it was selected for adaptivity to the environment that it evolved in. Therefore our psychological phenotypes are ‘massively modular’; therefore there is no such thing as the mind or the self; and (present company excepted) we are all self-deluding hypocrites.

I’ll concentrate on 4), because it is the burden of Kurzban’s book, but I don’t believe a word of 1) to 3) either; in particular, I don’t believe 2). Why I don’t is explained in What Darwin Got Wrong, which I wrote with Massimo Piattelli-Palmarini, a book which, unsurprisingly, is not in good odour among doctrinaire Darwinists.* A word about ‘massive modularity’ and we can then turn to Kurzban’s arguments and why they strike me as thoroughly unconvincing.

The ‘modularity theory’ entered the discussion of cognition around 1980. It was, in essence, a reaction to the then popular idea that how one perceives the world to be is in very large part determined by how one believes the world to be. You’ll find variants of the idea that beliefs saturate perception in the works of a number of iconic figures of the time and in several different fields: one thinks of Thomas Kuhn, Norwood Hanson and Paul Feyerabend in philosophy and of the New Look psychology of Jerome Bruner and Mitchell Ash. The art historian Erwin Panofsky made much of it, and the popular press swallowed it whole. The long and short is: one sees what one believes at least as much as the other way around. One views the world from the perspective of the ‘paradigms’ to which one is antecedently committed. So much for the objectivity of perception.

For every action, there is an equal and opposite reaction; the putative continuity of perception with thought was always in jeopardy from well-known counter-examples. The most obvious is the persistence of many perceptual illusions even when one is fully aware that they are illusions. The classic example is the Müller-Lyer. Line A looks longer than line B even though one knows perfectly well that, in fact, they are equally long. Even measuring the lines doesn’t get rid of the illusion; try it and you’ll see. So why, if belief penetrates perception, doesn’t knowing about the Müller-Lyer make it go away?

The Müller-Lyer illusion

Enter the modularity theory: some mental processes are performed by special-purpose computational modules, i.e. by mechanisms that are largely encapsulated from beliefs and from one another. The perceptual mechanisms that determine apparent length are among these; they are, as one says, ‘cognitively impenetrable’. Indeed, it turns out that quite a lot of perception works this way. ‘You see what you believe’ may be true; but it can’t be the whole truth. Accordingly, a couple of decades of cognitive psychology were invested in a search for ‘mental modules’. Enter the ‘massive’ modularity thesis that pretty much all cognitive processes are performed by encapsulated, special-purpose computational modules; in effect, by little homunculi who don’t much talk to one another. Kurzban says that the mind is a ‘bundle’ of domain-specific ‘software’; and these days a lot of other psychologists say much the same. The professional journals are up to their ears in massive modularity.

The question of what is right and wrong about the massive modularity thesis is deeply interesting; much turns on it, and it has prompted a lot of very fruitful empirical research. But suffice it, for present purposes, to note how well Psychological Darwinism comports with massively modular minds. If one assumes that a creature’s executive evolved to solve quite specific problems of adaptation to its ecology, it’s entirely natural to think of evolved cognitive processes as correspondingly specific solutions of such problems. If you’re a fish, a module for estimating depth might be a good thing for you to have; less so if you’re a bird. So it isn’t entirely surprising if fish evolve depth finders and birds don’t. Conversely, if you’re a bird, a module for stellar navigation might be a good thing to have; less so if you’re a fish. Iterations of this line of thought might well lead a committed Darwinist to predict that the mind is a bundle of modules. Enter the massive modularity theory: a mind just is a totality of available heuristics; there isn’t anyone in charge, least of all me.

Enter, finally, Kurzban. Kurzban is an enthusiast both for Psychological Darwinism and for massive modularity, and he thinks that the one predicts the other. He also thinks that, Darwinism aside, there is a lot of independent evidence for massive modularity; in particular, that it yields plausible answers to such traditional questions as ‘Who am I?’ and ‘How is it possible for me to have a divided mind?’ Thus: ‘Who am I?’ Nobody in particular; I’m not whoever it is my mind belongs to. That’s because, strictly speaking, there is no such thing as my mind; there is only a bundle of heuristics with no one in charge. Kurzban does admit (what he calls) a ‘press secretary’; it functions to announce ‘my’ views to the world. But he thinks that’s mostly agitprop since, strictly speaking, if there is nothing that corresponds to ‘me’, there is likewise nothing that corresponds to ‘my’ views. Strictly speaking, it’s modules (not minds and not persons) that ‘believe’ things. (Kurzban has unlimited faith that shudder-quotes can solve hard problems.) The modules compete for control of what the press secretary says that ‘I’ think; sometimes one of them wins, sometimes others do. The press secretary does what it can to harmonise contradictions between what the various modules believe. But, since the modules are largely encapsulated from one another, attempts to reconcile their ‘beliefs’ are usually post hoc and unconvincing.

The question about ‘divided minds’ is answered accordingly. How is it possible that someone should hold contradictory beliefs? There is a whole field of psychology (‘dissonance theory’) devoted to such matters, but massive modularity can do without it. I can’t believe P and not-P because there isn’t any me to believe them. At worst, some of my modules ‘believe’ P and some of them ‘believe’ not-P, but none of my modules believes (or ‘believes’ ) both. You might think this view would invite a worry about why it is that someone who believes P and believes Q is quite likely also to believe P&Q. A theory of mind that prohibits beliefs that are self-contradictory is likely also to prohibit lots of beliefs that aren’t. But Kurzban doesn’t talk about that.

Kurzban has other arguments that the mind is massively modular. Here’s one: everybody knows that beliefs, wants, preferences and the like are heavily sensitive to context. You might like drinking wine, but not at breakfast; you might like drinking coffee, but not while you are drinking tea … and so forth. If one thinks of one’s beliefs as a sort of list that the mind consults when trying to figure out how to act, there are going to be an awful lot of items in the list: not just ‘things to do when offered wine’ and ‘things to do when offered coffee’ but also ‘things to do when offered coffee and wine’; ‘things to do when offered wine in the morning’; ‘things to do when offered coffee and wine in the morning’ and so on, pretty much without end. Hence massively many modules: there is a coffee module and a wine module and there is a coffee-in-the-morning module, and so forth. Any one of these can be turned off while any of the others are on. Kurzban thinks that this simplifies the problem of contextualising beliefs, desires, preferences and the like. But, on second thoughts, he’s made a bookkeeping error since, for every item on the intellectualist’s list of preferences, Kurzban will require a corresponding module in one’s head: the more list-items you need, the more modules Kurzban needs, with a total gain of no yardage.

Similarly for Kurzban’s thesis that we can do without anybody who’s in charge. Something has to ensure that, in many, many cases, if you believe P and you believe Q, you also believe P&Q. In traditional intellectualist models, this is part of what executives do. The executive is an inference-making organ: it is structured so that when it finds P is on the list of your beliefs, and Q on the list of your beliefs, it adds P&Q to the list of your beliefs. Very roughly, it allows for the construction of relatively complex mental states from relatively simple mental states; so it can (maybe) explain how it is possible for a creature with a finite head to have indefinitely many beliefs. On this sort of view, it is not an accident that the belief P is a constituent of the belief P&Q; and it is not an accident that the sentence ‘John prefers coffee’ is a constituent of the sentence ‘John prefers coffee in the morning.’ If you have an executive, you can (maybe) make sense of all that. If not, then – so far as anyone knows – you can’t. Intellectualism suggests the possibility of a unified treatment of logic, language and thought. I think that’s a terrific idea: the best idea any psychologist has ever had. If it’s true that our Darwinism seriously threatens our intellectualist psychology, then maybe we’d better reconsider our Darwinism.

Quite aside from its very impressive provenance, the idea that we are, to a first approximation, rational animals, and that our rationality is built into how our minds work, has a certain nobility; it would be nice if it turned out to be true; it is not, in any case, to be lightly dismissed. Kurzban says repeatedly that he knows of no reason why psychological science should acknowledge minds, selves and the like. Well, here’s one in a nutshell: selves are the agents of inference and of behaviour; you need executives to account for the rationality of our inferences; you need the rationality of our inferences to account for the coherence of our behaviour; and you need the coherence of our behaviour to explain the successes of our actions. To be sure, our inferences aren’t rational all of the time; and, God knows, our actions aren’t always successful. But I think we are very often rational and successful when push comes to shove: when there’s a lot at issue, we do, quite reliably, infer P&Q from our beliefs P and Q; and we do, quite reliably, act on the basis of such inferences; and actions that are based on such inferences are, very often, successful. These facts need to be explained. The notions that there are minds and selves, and that the behavioural successes of our selves are very often explained by the rationality of the inferences our minds draw, is the traditional intellectualist view of the relation between cognition and behaviour. If it is a left-over from the Enlightenment, so much the better for the Enlightenment. If it doesn’t comport with Psychological Darwinism, so much the worse for Psychological Darwinism. Intellectualism may well be the only tradition in cognitive psychology that’s worth the trouble of saving. If, like Kurzban, you have it in mind to write a popular, introductory tract that tells the reader where our theories of the mind have got to, it would be a crying shame to slight it.