"Folk psychology" is, to a first approximation, the intuitively appealing hypothesis that mental life is composed of beliefs and desires that really exist and, perhaps even more important, have the causal powers necessary to initiate and control other beliefs, desires, and, ultimately, behavior itself. Even the most ardent defenders of folk psychology (e.g., Fodor, Pylyshyn) would attach a number of caveats to this rough formulation (what about, e.g., hopes and fears? feelings? consciousness? unconscious motivations? involuntary behavior?), but it is a reasonable first approximation. To formalize things a bit, folk psychology endorses the general claim: "If person P desires X and believes s/he can get X by engaging in behavior B, then P will do B in order to obtain X."
It would only be a slight exaggeration to say that the status of this generalization, as a scientific psychological law, has been the central issue of cognitive science since its inception. In the early days, it was the behaviorists who opposed belief/desire accounts of psychology, arguing that they were in some sense unscientific, a priori. Behaviorism was unable to make good on its promises to be able ultimately to account for complex human activity, even after decades of research. Having given up so much to behaviorism for so little return, psychology began a tactical retreat to the once seemingly-obvious possibility that beliefs and desires might really exist and, moreover, might be crucial elements in the explanation of behavior. As is well known by everyone now, the innovation that allowed beliefs and desires to be treated scientifically was the invention of the digital computer; a machine that can manipulate physical tokens that bear representations of things in the world (and even things not in the world) according to inferential rules. In short, it is a machine the operation of which is reminiscent of the way in which folk psychologists believe the mind to operate. The hope was that the best of behaviorism-its methodological scrupulousness-could be integrated with the best of mentalism-its content-to produce a scientific mentalist psychology.
After a quarter century, however, the computational approach to folk psychology seems stalled, much like the behaviorism it replaced was a quarter century ago, and there has now appeared on the horizon a new opponent to folk psychology: eliminativism. Eliminativists (e.g., the Churchlands, Stich) endorse the claim that a detailed account of brain function, once it is achieved, will simply leave nothing behind for folk psychology to explain. Consequently, terms like "belief" and "desire" will then simply be eliminated from the vocabulary of scientific psychology. Many eliminativists think that connectionist networks hold the greatest promise for modeling the activities of the brain in their entirety.
In this paper, I will argue that even if the eliminativist dream of describing brain function in its totality succeeds, it will have a great deal of trouble dislodging folk psychology. This is because folk psychology is not only a scientific hypothesis about the mind and its relation to behavior. It is also the basis of a great many non-scientific social institutions that are deeply entrenched in our culture, not least of which are the governmental system, the legal system, the educational system, and the social service systems.
This paper is entitled "Foucault and folk psychology," and at least some of you are probably wondering by now where Michel Foucault is going to fit in to this, thus far, very un-Foucauldian account of the history of cognitive science. His role derives mainly from the inspiration provided by his histories of madness, medicine, the human sciences, the prison, and sexuality. One of the key things he taught us in those works is that old ideas that form the bases of deeply entrenched institutional practices are not easily displaced, sometimes even long after they are discredited, unless institutional practices can be reformed, or their existing functions reinterpreted, on the basis of the new ideas.
To take what is perhaps his classic example, we have for over a century now complained that prisons do nothing to reform criminals, and that they probably make criminals worse, yet we still employ them, now more than ever, because we literally can't think of anything else to do that does not violate some constraint (e.g., ethical, economic, etc.) elsewhere in the network of ideas that, in part, defines our civilization. As a result, the prison has come to form a kind of state-within-a-state, developing internal ways of "handling" (and I use that term advisedly) its own perpetual failure as an institution. This "handling" has less to do with "correcting" its failure and converting it into success through reform or somesuch, than it does with learning to live with such failure-figuring out what to do in the face of such failure. "If we can't succeed in actually reforming criminals and returning them to society," the crucial but implicit question goes, "then what can we succeed at?" Perhaps at "controlling" the criminal, and control comes of "knowing" the criminal intimately.
Foucault argued that this will-to-knowledge about the criminal led to a new kind of combined surveillance-observation that serves a double function. On the one hand it keeps the criminal "frozen" in space, so to speak, fearful of being caught doing anything wrong. On the other hand it serves as a continuous supply of information on the criminal and, supposedly, on criminality itself. Ironically, this constant stream of information has done little to improve our ability-as per our original intent-to reform the criminal. But this very failure is continuously put forward as justification to continue and even intensify this observation in the never-dying hope that useable knowledge will eventually be obtained. What is more, Foucault argued, though efforts at reform have been failures, efforts at control have been much more successful-so much so that they have been exported to society at large when additional "control" has been deemed necessary. Thus, expressed intentions and theories aside, prison practices have led to it becoming a sort of laboratory in which more effective forms of discipline are developed and exported for use in other venues. Let me emphasize that this outcome was not the intention of any particular group of people with the power to make it come about. It is more like a point of stability within a dynamic system (though Foucault never uses this exact analogy); prison practices, as opposed to penal theory or policy, and the interaction of those practices with the practices of society as a whole, opened up a niche which made prison very nearly indispensable, while simultaneously failing at its stated goals.
Now what has all this to do with cognitive science and the apparent rise of eliminativism in the past few years? The key point of comparison is the relation between theory and practice. Practices are often much more difficult to dislodge than theories, primarily because they are often bound up-spatio-temporally, not just inferentially-with a whole network of habits, rituals, and performances-in short, with other practices-that will have to be disturbed and reformulated if the offending practice is to be changed.
Before I look explicitly at this relation with respect to folk psychology and eliminativism, consider the case of the relation between the quantum-relativistic and classical physical theories, and their relations, in turn, to our socio-political institutional practices. Imagine the following scenario: a woman is charged with murder; specifically with shooting another person to death with a gun. In court, she pleads not guilty. Her lawyer argues that she cannot have committed the crime of which she is accused for, after all, under current theories of physics, there really are no such objects as guns, bullets, and the like; only swarms of minuscule particles that randomly pop in and out of existence, energy fields, wave functions, and the like. In fact, under the very same theories, the same can be said for the defendant, the victim, the judge, and the court house as well. There is room for none of them in any current physical ontology. (The lawyer, strangely, leaves lawyers out of this list.) Beliefs about guns, people, buildings, and the like are only the residue of naive and outdated "folk theories" of physics that have long since been discarded by serious scientists. Moreover, since time itself has been shown to be relative, even if we were to accept these fictional objects as real, any effort to establish simultaneity of the defendant, the victim, and the gunshot would be hopeless unless we also accepted Newton's archaic belief that time "flows equably without regard to anything external." And surely indulgence in such a misguided ontology would make a laughing stock of the whole justice system. As Lord Rutherford himself once remarked, "All science is physics, or it is stamp collecting," and surely the court would not want to lower itself to mere philately.
The prosecutor is at first thrown off balance, but then retorts that even quantum physicists concede, in the main, that the traditional ontology of guns, people, etc. is still approximately captured by the aggregate activity of quantum particles, on average, anyway, so the charges should stand. The wiley defense lawyer is prepared for this move, however, and howls to the court that if his client is to be convicted for approximately shooting an approximate gun and, only on average, killing the approximate victim that she should be sentenced to only go approximately to prison; that perhaps her average location, taken over many years, should correspond to that of the prison.
One should be able to see the analogy to eliminativism immediately. Suppose that eliminativism becomes the dominant theoretical position in psychology at some point in the relatively near future. Scholars all over the world stop speaking of beliefs, desires, and the like. They begin to explain events like Clarence's going to the store in terms of the firings of neurons in Clarence's brain, organized in such a way that they implement connectionist networks of specified types. What impact will this have the world outside the academy? Virtually none in the remotely foreseeable future, I would predict.
Consider the possibility of the thoroughly eliminativist lawyer. "Your honor, my client could not possibly have believed that her husband was having and affair, and then desired him dead for, as modern psychological science tells us, there are no such things as beliefs and desires. Such ideas (such whats?) are nothing but the residue of naive and outdated folk theories of psychology that have no place in the modern court house." In fact, there would be no reason for there being a court, or even a justice system, if eliminativism were to be accepted in the society at large, for the justice system is utterly dependent on the assumption the people are moral agents; i.e., that they can tell the difference between-that is, have beliefs about-right and wrong.
The justice system would not be the only victim. Consider the educational system. How would one describe what goes on in school if not that certain people who have certain socially-sanctioned beliefs about the world, called "teachers," attempt to cause similar beliefs to arise in other people, called "students"? Eliminativists governments, always eager to cut costs, would undoubtedly be anxious to eliminate this extravagance. After all, if there is no such things as beliefs, we're paying thousands of people millions of dollars for being in possession of, and transmitting, entities that patently don't exist. "My God," cries out President Churchland, "it's no better than paying state alchemists!" and poof go the schools. Actually, Paul Churchland (1981) has fantasized in print about worlds in which spoken language disappears because the brains of the whole population have been linked together by artificial commissures, and libraries contain not books, but "long recordings of exemplary bouts of neural activity.... [which] do not consist of sentences or arguments" but, rather, something more like pure information, in the Shanon-and-Weaver sense. I think it safe to say that such eventualities are extremely remote, at best.
Governmental structures, too, would be in danger if eliminativism were to take hold. It would undermine virtually every single reason for having a liberal, democratic government. Governments, insofar as such institutions continued to exist, would look much more like giant laboratories which, rather than exhorting citizens to taking personal responsibility (i.e., instilling certain beliefs in them), would be arranging for maximally-useful patterns of neural firings, or somesuch. Please take care to note, here, that I am not whining that eliminativism could not be right because it would be anti-democratic. Rather, I am pointing out the radical lack of fit between what we now consider a government to be, and what would be minimally rational under an eliminativist "regime."
From an historical perspective, one of the really interesting things about this little set of thought experiments is that the implications of having strongly eliminativist socio-political institutions are well-nigh indistinguishable from those of having strongly behaviorist ones. If one goes back to look at the criticisms of B. F. Skinner's project for a behaviorist society, one sees fears expressed of having institutions not very different from those that I presented above. One may also notice that behaviorism, in the final analysis, did not have anything like the impact on our socio-political institutions that radical behaviorists of the time would have had us expect. This leads to the interesting question of just what differences there are between eliminativism and radical behaviorism. Though his advocates seem to have recently launched a campaign of telling us what a decent, even humanistic, guy old Fred Skinner was-that he didn't deny phenomenal experience, personal values, or human dignity-it is undeniable that he sometimes wrote much like what might be termed a "behavioral eliminativist." Consider the following passage from Behavior of organisms (1938, p. 7) : "We must not take over without careful consideration the schemes which underlie popular speech. The vernacular is clumsy and obese; its terms overlap each other, draw unnecessary or unreal distinctions, and are far from being the most convenient for dealing with data." Now one can only guess whether this sentiment was to have led to a mere "cleaning-up" of ordinary language (as, say, the logical positivists worked to achieve) or whether it was to have resulted in the elimination of ordinary psychological terms in favor of good behavioral ones. I suspect the latter, though many of Skinner's current apologists seem to insist that he had nothing so radical in mind.
In any case, the relation of behaviorism to eliminativism is a digression from my main point, this being that the theory of eliminativism stands so contrary to the practices of our culture-practices that typically prove quite recalcitrant to change-that I suspect that even if it were to become dominant in the academy, even if it were indeed TRUE, it would be doomed to have little immediate impact in the world at large. And this brings us to an interesting question about truth itself. Why is it that we're able to get along so well day-to-day if our understanding of ourselves is, potentially at least, so completely misguided? Is it possible that what counts as truth is somehow relative to the projects we undertake; not, perhaps, wildly and unconstrainedly relative in that sense that we've all learned to refute with a deft application of petitio principii, but relative in the sense that the entities assumed in stating what is true must, in some sense be "proportional" to the aim we have in mind. This would mean that many different descriptions, including some that have unclear relationships to each other, but seem incompatible, might all be true nevertheless. But this line of thought leads us into deep metaphysical space that we do not have the time to explore at present, and so I will conclude.
I have not, it should be noted, argued that eliminativism is correct, nor that it is incorrect. Nor have I argued that if it becomes the dominant psychological theory that it would be impossible to rearrange our institutional practices to conform with it. What I have presented is more on the order of a prediction, viz., that even if the hopes of eliminativists were to be realized in the academy, folk psychology would put up far more resistance to "elimination" than its opponents are, I think, wont to believe.