Even so, it is becoming apparent that there is deep discord within the vindicationist camp: discord over the nature of the body of facts in virtue of which folk psychology stands or falls. There are those who take these facts to involve specific claims about the functional organization of our cognitive architecture and so, apparently, to involve facts about the future of cognitive scientific inquiry. According to these vindicationists, if folk psychology is a true theory, then there must be a level of scientific psychological description and explanation that bears certain important relations to our common-sense psychologizing. There is, however, a second and increasingly influential school of vindicationism that regards what we can term "scientific vindication" as an error. According to such vindicationists, the facts in virtue of which folk-psychology stands or falls are just the facts as we know them to be. So, although folk-psychology is a theory, its truth is something of which we can now be assured. Its falsity is only barely conceivable; to imagine it is to imagine a radical skeptical hypothesis akin to Descartes' genie malin. The vindication of folk psychology, then, does not hinge upon future developments in the cognitive sciences; or, if we insist that it must, we can be certain, given what we already know to be the case, that the cognitive sciences will develop along certain rough lines. It follows that the elimination of folk psychological states and events is not a serious prospect.
I term this second school of thought "erstwhile vindicationism," since I shall argue that such accounts do not succeed in vindicating folk psychological explanation in a robust enough fashion to salvage our common-sense conception of ourselves as creatures who reason and deliberate-creatures who, at least sometimes, act and believe as they do because they have reason to act or to believe so. What must not be forgotten is that the central question of the vindication/elimination debate is this: can our common-sense self-image of ourselves be sustained in the face of the advance of the cognitive sciences? Thus, it may well be that there are ways of vindicating the bare existential claim that there are beliefs and desires that do not vindicate (or take seriously) our common-sense conception of ourselves as reasoners and deliberators. Yet, if folk psychology is to be vindicated as theory, then reasons-explanation must be taken seriously as theory, and so, the causal explanatory goodness of reasons-explanation must be vindicated.
This essay is organized as follows: In (1) I briefly characterize a version of scientific vindicationism defended by Jerry Fodor. In (2) I argue that, whatever else its merits, Fodor's view ably vindicates our common-sense self-image. Then, in (3) 1 characterize the competing account of vindication, erstwhile vindication. Such accounts, I shall argue, aim to insulate vernacular psychology from the burdensome demands of scientific vindicationism by claiming that its explanations do not offer information specific and determinate enough to put them at odds with scientific psychological inquiry. I will argue that this maneuver responds to the letter of the threat of eliminativism without responding to its spirit. In (4), I take a closer look at Frank Jackson and Philip Pettit's version of erstwhile vindicationism and show that it must confront the dilemma: either folk psychology (as understood by the erstwhile vindicationist) does not offer informative causal explanations and so cannot make good sense of reasons-explanation, or vernacular psychology does offer informative causal explanations, and so, scientific vindicationism must be defended. Still, I grant that there is much with which one can rightly sympathize in erstwhile vindicationism. I conclude by suggesting that, contrary to their stated aims, a good deal of what erstwhile vindicationists say is best understood as militating against the view that common-sense intentional psychology is an empirical theory of mind and behavior. If this is so, the survival of our folk conception of ourselves as reasoners need not hinge solely upon the causal explanatory goodness of folk intentional psychology.
I. THE "DEEP" VIEW
OF FOLK PSYCHOLOGY
Central to the rationale of scientific vindicationism is its conception of the nature of folk psychological explanation. Such explanations do offer relatively informative causal explanations of our behavior and cognition. It is for this reason that the vindication of folk psychology hinges upon its relation to cognitive science. Jerry Fodor writes:
When [folk psychological] explanations are
made explicit, they demonstrate the `deductive
structure' that is so characteristic of explanation
in real science. There are two parts to
this: the theory's underlying generalizations
are defined over unobservables and they lead
to its predictions by iterating and interacting
rather than by being directly instantiated ...
It is a deep and powerful fact about the world
that the most powerful etiological generalizations
hold of unobservable causes. Such facts
shape our sciences (they'd better). It is thus a
test of the depth of a theory that many of its
generalizations subsume interactions among
unobservables. By this test, our implicit common-sense
meteorology is presumably not a
deep theory since it consists largely of rule-of-thumb
generalizations of the `red sky at night,
sailor's delight,' variety. Correspondingly, the
reasoning that mediates application of common-sense
meteorology probably involves not
a lot more than instantiation and modus ponens.
(All this being so, it is perhaps not surprising
that common-sense meteorology
doesn't work very well.) common-sense psychology,
by contrast, passes the test.
More precisely, a scientific psychology will count as vindicating common-sense intentional psychology, "just in case it postulates states (entities, events, whatever) satisfying the following conditions":
i. They are semantically evaluable.
ii. They have causal powers.
iii. The implicit generalizations of common-sense
belief/desire psychology are largely
true of them.
This will mean that the explanations of folk psychology are "deep" in the sense that they provide causal explanations that are "structurally informative" with respect to the causal relations between stimuli, behavior, and unobservable internal states. And, like any informative functional theory, folk psychology offers a componential account of the system whose behavior it aims to explain. In particular, it offers an account of the causal relations between the components (beliefs, desires, intentions, etc.) of the system that are posited to explain observed behavior. If this is so, the structure of these causal relations must be mirrored in the explanations of a scientific psychology if folk psychological explanation is to be vindicated.
To see this conception of vindication in practice, consider an example of common-sense psychologizing. We ask, Why did Silvio leave in such a hurry? And we may answer, Well, he wanted some cigarettes, and he realized that the store was closing in a short while. The common-sense explanation has it that Silvio left as he did because he wanted some cigarettes and he believed that the store was soon closing. Presumably, this is an instance of the common-sense generalization:
(PI) Whenever a subject desires that p, and believes
that he must q in order to secure that
p, then, ceteris paribus, the agent q's.
According to the deep view, folk psychological explanations and generalizations can be vindicated by the science of psychology only if the explanations of that science take seriously the structural features of the causal explanations offered by the vernacular. The explanations of such a science will, then, appeal to states (belief-like and desire-like states) that are causally relevant to the production of the behavior and that are determinately semantically evaluable. Vindication hinges upon there being some level of scientific psychological description and explanation that takes folk psychology seriously as an informative causal account of mind and behavior. No doubt, the common-sense explanation is incomplete and will need much tidying up. What is crucial, in any case, is that our common-sense generalizations are more or less preserved by such an account. This is secured by the rough structural isomorphism between folk explanations and their scientific counterparts.
Not surprisingly, this conception of folk psychology and the associated constraints on its vindication are shared by eliminativists. For, if folk psychology is, as it is on this model, an effort to give some fairly detailed information concerning causal processes, and it does this by positing unobservables in order to give a deeper account of observed regularities, then the eliminativist's suspicion that it would be a miracle if folk psychology turned out to be fundamentally sound is hardly unreasonable.
The "deep" view of common-sense psychology amounts, then, to the following: folk explanations offer information concerning the nature, structure, etc., of causal relations between stimuli, behavior, and unobserved mental states; so, if folk psychology is to be vindicated, these claims must be more or less correct. This will mean that there is some level of cognitive organization, and so some respectable scientific psychology, that carves things up more or less as common-sense psychology does. Of course, there can be little doubt that there are many legitimate and informative levels at which psychological inquiry may be usefully pursued. Fodor's claim is not that every explanatory enterprise that aims, in some broad sense, to understand, e.g., human inference, must satisfy his constraints. What is important to Fodor's vindicationism is that there can be a respectable psychology that individuates states and posits causal relations between them in ways that more or less mimic the folk conception. It is for this reason that we can view folk psychology, as Fodor notes, as having depth that it would not possess were it on par with, say, common-sense meteorology. It is deep because it makes claims about underlying cognitive organization. This kind of depth is crucial if the vernacular is to make the causally informative claims in virtue of which its vindication depends upon its relation to cognitive science.
II. THE DEEP VIEW AND OUR FOLK
What bears special emphasis is that, in virtue of the fact that it must take very seriously the folk conception of the structure of reasoning, the deep view seems especially well-suited to vindicating our conception of ourselves as reasoners and deliberators. For consider: we said of Silvio that he left for the store because he wanted some cigarettes and believed that the store was soon closing. If this is correct, it is to say, at least, that what moved Silvio to leave was this particular reason in the form of the relevant belief/desire pair. What moved him was the significance of the thought that the store would soon be closed and that this would result in a cigarette-less evening. As we know, Silvio may have had many other desires, which, together with his beliefs, could serve to rationalize his action. He may have wanted some English muffins for breakfast, he may have wanted some brisket for Friday's supper, etc. Yet these desires - even if satisfied at the store while purchasing cigarettes - were, in the case we are considering, no part of the reason that moved him to act. He reasoned to the intention to go to the store straightaway because he wanted some cigarettes and believed the store was soon to close. And this is why he went to the store. The same point can easily be made with respect to theoretical reasoning. Reasons-explanation, if regarded as an expression of an informative causal explanatory theory, points to the causal relevance of particular states/properties of an agent.
I am taking my cue here from a suggestion offered by Ramsey, Stich, and Garon. Folk psychology is, they argue, committed to propositional modularity: propositional attitudes, on the folk conception, are functionally discrete, semantically interpretable states that play a causal role in the production of other mental states and behavior. This, they argue, is a supposition that will turn out to be groundless if certain connectionist models of our cognitive architecture are correct. Now, it may be an open question as to whether or not the commitment to propositional modularity is essential to the vindication of the existential claim that there are beliefs and desires. Nonetheless, this commitment is essential to our folk conception of ourselves and our fellows as deliberators and reasoners. When we reason, whether to intentions or to other beliefs, we are moved by our consideration of, our grasp of the meaning and significance of, particular intentional contents. Silvio may, for example, answer us, when we express surprise at his running to the store Wednesday night to buy brisket for Friday's supper, with "But that's not why I went; I needed some cigarettes for the evening." In reasoning, we focus upon the contents of discrete intentional states and the relations between their contents. So, if our conception of ourselves as reasoners and our associated explanations of our behaviors is taken seriously as an informative theory, then propositional modularity is built into vernacular psychological theory. If there is no making sense of our acting on some particular reason rather than some other, then there is no making sense of our acting on reasons at all.
This result is all the more telling if we consider matters from our own first personal perspective as deliberators and thinkers. In deliberation and reasoning we think about what to do and what to believe; we ponder our alternatives, rank our preferences, evaluate evidence. What moves us to the formulation of intentions or the acquisition of beliefs is our appreciation of the significance of various propositions believed or desired. And, of course, retrospectively, we explain our behavior (to ourselves and to our fellows) in this same way: we advert to the reasoning that moved us to action or belief. If this self-explanatory framework expresses a causal explanatory theory, if it is true that we are moved by reasons, then these reasons must be causally relevant states of the reasoner.
Thus, the claim that in reasoning and deliberation we are causally influenced by particular mental states amounts to the following: there are mental states that possess the causal profile belief and desire are thought to have in folk psychology. And, again, what this means is that there must be some level of psychological description of the sort that Fodor and Ramsey, Stich, and Garon describe if our folk conception. of ourselves is to be vindicated as psychological theory; there must be some level of psychological inquiry that takes seriously the structure of causal relations posited by vernacular psychology. And you can not take this structure of causal relations seriously without propositional modularity.
John Heil suggests, however, that there is no good reason to regard folk psychology as committed to the "crude `billiard-ball' model" of causation that, apparently, undergirds propositional modularity. Surely, no one can imagine that belief/desire properties are basic causal properties; that beliefs and desires somehow "bump" into each other. Yet, the fundamental claim made by the scientific vindicationist is that our cognitive architecture is organized in such a way as to mirror the structure of folk psychology. No doubt, what realizes our mental states are complex neural states; but the point is that these neural (ultimately, micro-physical) states are organized in such a way as to have causal powers in virtue of their complex organization. And this complex organization is such as to realize something like the causal structure posited by common-sense. In this sense, folk psychology does seem to be committed to a crude billiard ball model of causation.
It is, however, not easy to see why this is an objection. Any deep causal explanatory theory earns its keep by making informative claims about the structure of causal relations between its theoretical posits. So any informative causal explanatory theory is, in this sense, committed to the crude billiard-ball model of causation.
Heil's central reason for rejecting propositional modularity is expressed in the following passage:
Psychological events (my believing that p on
the basis of my believing that q, for instance)
could well turn out to have a finer-grained underlying
psychological - perhaps connectionist - basis.
Consider an example. Suppose I
decide to give myself the disposition to form
beliefs in accord with modus ponens. There
are at least two ways I might accomplish this.
I might set out to instill the principle itself,
perhaps by memorizing it. Or I might undertake
to imbue some other, perhaps very different,
principle, one that, once acquired, would
result in my forming beliefs in accord with modus
ponens.... Similarly, it could easily be
imagined that an efficient way for Mother Nature
to ensure that we satisfy folk-theoretical
principles is by endowing us with a connectionist
The passage is ambiguous. If all that is meant is that there are levels of psychological description and explanation finer-grained than Fodor's, there need be no dispute. As we saw above, all Fodor need claim is that there is a level of scientific psychological explanation and description that respects his constraints. Even so, the first sentence of the cited passage might reasonably be taken to suggest that it might turn out that there is no level of psychological explanation at which such constraints are respected; but this would appear to have the result that, e.g., our believing that p on the basis of believing that q is merely emergent out of its implementation. Moreover, what Heil says about modus ponens deepens this suspicion. When we say, as we sometimes do, that we believe that q because we believe that p and also believe that p -> q, (i.e., that the reason for our coming to believe that q is that we believed that p and that p -> q) we mean to say not just that we came to believe in accord with the rule, but hat we actually followed it - that our grip upon the rule guided our arrival at the belief. But in the last sentence of the quoted passage, it is suggested that it might always be true that we come to "satisfy" modus ponens not by following the rule per se but rather by virtue of some "connectionist psychological architecture." And this does no justice to our vernacular conception of ourselves as reasoners - in this case, to our claim that our reason for believing as we do is our belief in modus ponens.
Now it may be that, as a matter of fact, no one (save, perhaps, students of logic) does explicitly represent and so is explicitly guided by modus ponens. Even so, the point of principle remains; for, if, like Silvio, I run off to the store because I believe that it is soon closing and want some cigarettes, then what moves me is something like the thought: Given that I want cigarettes, the thing for me to do is to leave for the store now. Thus we need not have the explicit representation of or guidance by modus ponens. Nonetheless, we do have explicit rule-following behavior. In such a case, the agent is moved by particular considerations - her reasons.
One of the chief issues raised by Ramsey, Stich, and Garon is that folk psychology is committed to the thesis that a belief may be caused by one reason, while other reasons (beliefs) might be causally inert with respect to the bringing about of that belief. In response, Heil protests:
It might plausibly be supposed that even when
S fails to hold p on the basis of r, r might play
some role in S's holding p: S may not have held
p on the basis of q had he not held r. Even
when we take so-called basing relations to be
causal, we need not suppose that other beliefs
are causally inert, that the psychological or
epistemic basis of a given belief is functionally
isolated in some absolute sense from an
agent's doxastic background.
This remark does not count against propositional modularity nor against the deep view of folk psychology. All these demand is that there be some level of psychological description and explanation that is committed to properties such as believes that p and desires that q, and that the explanations and relevant generalizations at that level be framed in ways that display these properties as causally relevant. In short, it may well be that but for the fact that S believed that q, that r, that s, . . . , she wouldn't have believed that t and so wouldn't have come to believe that p (since she has come to believe that p because she believes that t). Even so, it is proper to explain S's coming to believe that p by appeal to the functionally isolable and discrete belief that t. Thus, the above does not challenge the claim that there is a level of psychological description and explanation which mirrors the vernacular.
I have lately been emphasizing what it takes to vindicate our folk conception of ourselves as reasoners and deliberators. There is, no one should deny, much that our folk conception does not tell us about the nature of reasoning, intention formation, etc. But if this common-sense conception expresses an empirical theory, then its vindication hinges upon its causal explanatory goodness. When we offer explanations of the behavior of ourselves and others in terms of beliefs and desires, intentions, and the like, we offer reasons-explanations. Any vindication of folk psychology must take reasons-explanation seriously; and taking reasons-explanation seriously is just a matter of taking seriously the view that we act because of certain reasons rather than others; that what moved us on an occasion- was some consideration (in the form of beliefs, desires, values, etc.) in particular. Again, the proper conclusion to draw is that, if reasons-explanations are to be taken seriously as part of an informative casual explanatory theory, propositional modularity is part of the picture. If it turns out to be true that there is no level of scientific psychological description and explanation that is committed to such an architecture, then our folk conception of ourselves is in jeopardy.
On this "deep model," folk psychology is at risk. It is eminently falsifiable. Still, its being at risk in this way is a good thing. As with any other informative empirical theory, the vernacular's being at risk is just a matter of its making specific claims about the structure of the causal relations between and the nature of the states and events causally responsible for behavior; and so its vindication hinges upon demonstrating the truth of these claims. I have argued that there is good reason for regarding folk psychology as making such specific claims, since this makes for a clear way in which we can regard our common-sense self-image as vindicated. The eliminativist, of course, argues that such claims are likely to be false. No respectable psychology will find a role for such discrete states and events.
Notice that, in the face of such a worry, we might seek to save the common-sense conception by understanding it as less informative than it is on Fodor's view. For, if folk psychology did not possess the above sort of depth, it would not be at risk from developments in the cognitive sciences. On this view, folk psychology does not make claims precise or determinate enough to be at odds with cognitive scientific inquiry. This is the maneuver of erstwhile vindicationism. There is, however, a different sort of risk here: the more we eviscerate folk psychology in order to insulate it from science, the more our vindication" becomes virtual elimination. Erstwhile vindicationism offers no vindication worth having.
III. AN ALTERNATIVE MODEL
In short, the rationale for scientific vindicationism makes obvious a certain response: simply deny that folk psychology does make such specific and informative claims about the causal architecture of our cognitive processes. If this is done there is no reason to expect a scientific psychology to mirror the relevant claims of common-sense; there are no such claims to mirror. Consequently (as Terence Horgan and George Graham write), there is no basis for demanding that the following be a necessary condition of "true believer-hood":
(SA) Humans are true believers only if folk
psychology is absorbable into mature
Rather, they argue, were it to become clear that folk psychology is not scientifically absorbable, this would be evidence, not that human beings are not true believers, but that SA is not a necessarv condition of true believer-hood. Such a view must be founded upon a different conception of the claims of common-sense.
Perhaps the best way to begin to characterize this alternative conception of common-sense psychology is to consider some of the analogies mustered by its advocates. Horgan and James Woodward offer the following:
Both we and our ancestors judge that the impact
of the rock caused the shattering of the
pot, that the lack of water caused the camel to
die ... These examples serve to remind us of
the fact that not all folk theory is now regarded
as radically false.
In a similar spirit, Charles Chastain writes of "folk physics" that:
It contained folk terms like `fire,' `burn,' `hot,'
`cold,' `soft,' `slimy,' `slippery,' and many others.
It contained generalizations like `Fire
burns things,' `Hot things make cold things
warmer,' `Slimy surfaces are slippery,' and so
on. These terms and generalizations are still
with us, and the inspiring success story of modern
physics does not include a chapter describing
the abandonment of this folk vocabulary .
... Nothing that has happened or would happen
in physics could demonstrate that there
are no soft objects, or that nothing is slimy.
Frank Jackson and Philip Pettit offer the following:
The hypothesis that metals are good conductors
of heat and electricity was established
much earlier and much more securely than the
explanation of the fact in terms of atomic theory,
and this was precisely because being a
good conductor is functionally defined in
terms of playing a certain role between observable
inputs and outputs. Provided, but
only provided, that we insist that the folk-hypothesis
be understood in purely functional
terms, we can regard it as peculiarly well confirmed.
And, finally, Horgan and Graham write:
[Consider the concept] being able to fly.
There is little doubt that the very concept of
being able to fly renders it correctly applicable
to ordinary, proto-typical birds - that is birds
who (among other things) behave in ways that
meet the behavior-based epistemic standards
we employ when attributing to some creature
the ability to fly. Suppose that someone advances - possibly
on the basis of plausible and
scientifically sophisticated reasoning - a hypothesis
to the effect that any creature capable
of flying must satisfy a condition C. And suppose
that thereafter it is discovered that some
birds, although they do not satisfy condition C,
nonetheless do meet all the standards we normally
employ in attributing the ability to fly,
should we then conclude that, appearances to
the contrary, these birds cannot fly because
they do not satisfy condition C? Surely not;
rather the proper conclusion would be that
condition C is not after all a genuine prerequisite
for being able to fly.
Analogies such as these are exceedingly common in the literature. They aim to demonstrate that many folk explanations and folk "theoretical" terms do survive the advance <)f science. Indeed, it is hard to imagine that we would come to regard the above explanations as false, and the associated terms as empty. And even if we can get ourselves into philosophical moods in which we are able to take such possibilities seriously, there seems little consolation for the eliminativist in the thought that there are no beliefs and desires in the way that nothing is slippery. (Or in the thought that the explanations of folk psychology are false in the way that it is false that the pot was smashed because it was struck by the rock.)
Still, in so far as one is tempted by the "deep" conception of folk psychology one will find scant solace in the vindication promised by such analogies. I want to consider briefly some of the characteristics of the above folk explanations and predicates in virtue of which they are alleged to be peculiarly insulated from going the way of phlogiston and caloric. Not surprisingly, such insulation is secured by denying essential elements of the deep conception.
First, many of the above explanations do not posit unobservables in aid of the explanation of observed phenomena. Re folksy "the impact of the rock smashed the pot" is, to use Sellarsian terminology, at the "manifest" level. As such, this folk theory (perhaps best expressed by the regularity "fragile objects are smashed when struck by very hard objects") amounts to the systematization of manifest regularities. More controversially, our grip on the relevant concepts or predicates is secured - at least in part - independently of their causal/theoretical role.
The claim that such concepts are phenomenologically grounded and that the regularities in which they figure are manifest amounts to this: one notes that, e.g., the rock is very hard and that it is sailing towards the fragile pot. One further notes the correlation between such events and the smashing of the fragile object. Finally, one concludes that there is a causal connection between such events. A tedious story, no doubt. Still, there is an important lesson here. Such explanations do not pretend to tell us anything informative about the causal relations between the events responsible for observed phenomena; i.e., there is no effort to say anything about the structure of causal relations or processes beyond the manifest level. Such explanations are not "deep" in the way theoretical explanations typically are, and the consequence of this lack of depth is a lack of informativeness. As a result, we are able to regard such folksy explanations as true without worrying about scientific absorbability.
This last point deserves some comment, since it is important that we be clear about just what SA, in this context, entails. It may be suggested that such folk explanations do, in fact, meet the demand of SA, at least in the sense that, even though "pot" or "rock" will figure in no law, any particular instantiation of one of the properties will be token identical to physical states and events that will, presumably, be describable in the vocabulary of the sciences. And, since it will be claimed that this is the situation with regard to the special sciences quite generally, we should conclude that such explanations and such predicates are, in fact, scientifically absorbable. But this is to misconstrue SAkd, at least in this context. For SA is a matter of horizontal absorbability (or intra-theoretic relations) and not vertical absorbability (or inter-theoretic relations). Vertical absorbability is a matter of the relation between theories at different explanatory levels; as such, it is a synchronous relation. Horizontal absorbability is a matter of the relation between theories at the same explanatory level; as such, it is a diachronic or successional relation. This will raise the obvious question: Why should the vindication of folk psychology hinge upon the horizontal scientific absorption of folk psychology?
But there is an obvious answer: the deep conception regards folk psychology as making specific structurally informative causal claims, and, since it does, these claims must be vindicated by science. Just because the scientific vindicationist takes common-sense psychology to make specific claims about the structure of causal relations between the events postulated to explain observed behavior, those claims must be vindicated by science. On the other hand, when I noted above that folk explanations of the pot/rock variety lack informativeness by virtue of their lack of depth, I meant something quite specific. Such explanations are silent about the structure of causal relations between the unobservables posited in aid of the explanation of observed phenomena. Thus, there are no such causal claims to be vindicated by a science. We need not worry about the satisfaction of the relevant version of SA just because of the relative uninformativeness of such explanations. It is true that no one expects a science in terms of pots and rocks, etc.; no science will carve states and events up in this way. Yet, of course, we regard many of the explanations in which such predicates figure as true. What permits this unitarian (and sensible) conclusion is just the fact that such explanations are not terribly informative. Again, you generate a rationale for horizontal SA only if, in aid of the explanation of manifest phenomena, the theory in question makes informative claims about the structure of causal relations between unobservables. In such circumstances you then have two causal explanations (the scientific and the suspect) of the same phenomenon. The suspect causal explanation must either be absorbed or jettisoned. And on Fodor's account, this is indeed the situation, since folk psychology offers structurally informative causal explanations.
To see the vertical/horizontal SA distinction at work, consider the ill-fated phlogiston. One of the important differences between this case (and folk psychology on the deep conception) and the pot/rock case is that the Phlogiston theory makes specific claims about the underlying causal processes responsible for the manifest regularities concerning combustion. It was claimed, for example, that the residue remaining when lead was burned weighed more than the original sample; thus, it was concluded that Phlogiston possessed a "positive lightness." In short, the theory made claims about the discrete underlying causally relevant properties responsible for the observed phenomena. Lavoisier's work, which had it that combustion was to be accounted for by appeal to combustible substances and oxygen, was rightly perceived as a threat to - because in competition with - the phlogiston theory precisely because it found no explanatory role for such events as the release of phlogiston. Nothing in the oxygen theory had the rough causal role played by phlogiston. It is notable that various of Schele's and Priestley's defenses of phlogiston were informed by this worry; for example, in the face of certain recalcitrant data, it was asserted that what had been thought to be phlogiston was really a compound of phlogiston and water. Again, the aim here was to vindicate the Phlogiston theory by demonstrating that the discrete causal properties attributed to phlogiston played an essential role in explaining phenomena at the manifest level.
My point here is that the demise of phlogiston was a matter of its failure to be horizontally absorbable by a better confirmed, more powerful theoretical account of the very phenomena it had been invoked to explain. Robert McCauley remarks of theories in such intra-level conflicts, "since they offer thoroughly incompatible accounts of at least much the `same' phenomena, science cannot abide both for long."(26)
As we have seen already, these are the issues at play in the dispute between eliminativists and scientific vindicationists. Thus, Ramsey, Stich, and Garon emphasize that on their reading connectionist models "are intended as cognitive models and not merely as implementations of cognitive models."(27) For were one to construe such models as mere implementations, one would not have competing explanations of the same phenomena. They make the point directly: "So viewed, however, connectionist models are not psychological or cognitive models at all, any more than a story of how our cognitive processes are implemented at the quantum mechanical level is a psychological story."(28) But, of course, if connectionist models are psychological models, then they are in competition with folk psychological accounts. They offer competing explanations at the same explanatory level.(29)
Thus the demand for horizontal scientific absorbability follows directly from the scientific vindicationist's view that many of the generalizations of folk psychology "subsume interactions among unobservables."(30) And it is just this that makes for the view that folk psychology offers an informative account of our psychological processes (and so explanations of our behavior) by appeal to discrete, causally potent folk psychological states and events. This is why the scientific vindicationist - but not the erstwhile vindicationist - demands that there be a science, if not in terms of belief and desire, then at least in terms which carve up the causal pie in ways that mirror belief/desire explanation.(31) The erstwhile vindicationist suggests, to the contrary, that common-sense does not make such informative claims, and that, as a result, the vindication of the vernacular does not hinge upon the possibility of such a science. Furthermore, the erstwhile vindicationist will note, our regard for other folk explanations makes this possibility immensely attractive.
Let us, then, turn to these other folk explanations in aid of finding another model for the vindication of folk psychology. Consider first the explanation: "S fell because the floor was so slippery." It is apparent that "...is slippery" will figure in no science. "Slipperiness" is not a scientific kind. So, we know in advance that this folk explanation is not horizontally absorbable. This hardly undermines our faith in the explanation or in the relevant existential claims. In this case we are confident that the explanation is true and that there are slippery things.
Can this be a model for the vindication of folk psychology? No. What this example illustrates is that when we offer folk explanations of the "slippery" sort, we both recognize and don't much care about the fact that such explanations are not structurally informative - and this is not true of vernacular psychology. For consider the fashion in which "slippery" functions in the above explanation. I submit that it functions in such a way as to be of little real importance to the strictly causal explanation of the phenomenon we pick out as "a falling by S."
We, as folk explainers, perfectly well recognize that terms like "slippery" function as causal place-holders. Even in the offering of such causal explanations, we do not commit ourselves to regarding "slipperiness" as somehow deeply implicated in the bringing about of the event that is S's failing. Still, individuating states and events in such terms is convenient and useful. We are familiar with slippery things. When we pick out something as "slippery" and deploy this term in the making of causal explanations, we nonetheless recognize that it is not the property of being slippery that is crucial to bringing about S's falling. Rather, when we deploy such terms we mean for them to function in something like the following way: there's something about the event we characterize as "slippery" that causes the event we characterize as "a falling by S." Presumably, in so doing we mean to be pointing to more basic physical (which is not to say basic microphysical) properties that do the real causing and the real informative explaining. It is for this reason that we may regard such terms as figuring as place-holders in legitimately informative causal explanations.
So, perhaps, we can take some solace from the fact that these folk explanations will survive. Sadly, there is no similar solace to be had in the case of our common-sense psychologizing. The analogous thought would have to be something like: when we explain by appeal to beliefs and desires we are doing no more than claiming that there's something presumably, some neural/physical states) being picked out by the attribution of propositional attitudes that causes the state we pick out as, e.g., "S's intentionally A-ing." But if beliefs and desires serve as causal place-holders in this way, then folk psychology is a similarly uninformative "theory." Belief and desire are doing no serious explanatory work - so much so that it is hard to see what all the bother is about. As Brian Loar puts it: "If mental properties explain merely by pointing to what really explains, why resist the eliminativist idea that they are just manners of speaking-that what points are the predicates without the properties? Puzzling."(32)
It will, however, certainly be suggested that I have, with the above, chosen too handy an example in the form of a dispositional term such as "slippery." And it will be argued that many folk predicates are not plausibly viewed as empty causal place-holders. In particular, the appeal to, e.g., sliminess or liquidity in the explanation of manifest phenomena would appear not to be so understood. Rather, such terms would appear to point to something relatively specific and not to just anything possessing certain dispositional properties. It is, we say, the property sliminess or liquidity, per se, that causes some particular event. Of course, we must note that the folk explanations in which such terms figure can go awry. So, for example, many folk would regard the following as uncontroversially true:
It is because leaves are so light that they make their way to the earth so slowly in the fall.
This explanation is false; in this case, the error is that causal potency is attributed to the wrong property. But even so, the fact that this particular folk explanation is false does not tempt us to the view that there is no such thing as lightness. I take it that not only does lightness figure in many true folk explanations, but that, in addition, it is phenomenologically grounded.
What is central here is the manifest nature of such folk explanations as "hot things make cold things warmer," or "the red wine spilled off the table and onto the rug because it is a liquid." The property liquidity is robust; it is something the sciences can certainly account for - in terms, let us say, of van der Waals molecular interaction properties. And liquidity is, at least roughly, vertically absorbable. But, not to put too fine a point on the matter, the manifest behavior of substances we folk pick out as liquids is part of the data for the sciences; such manifest characterizations are not in competition with those of the sciences. The crucial matter is that there is no question of some version of "propositional modularity" being put to use in such a case. We do not expect there to be, at any level below the manifest, some discrete property invoked in explanations that plays a structurally isomorphic role to that played by liquidity at the manifest level. Indeed, it would be something of a miracle if we folk had stumbled onto an informative account of the processes that underlie manifest regularities.(33)
Of course, such folk explanations, just by virtue of the fact that they are at the manifest level, do not aim to offer information about the causal processes underlying those very manifest regularities. Now one might at this juncture, quite reasonably wonder how such a model could sustain a theoretical conception of folk psychology; moreover, one might question the propriety of viewing folk psychological properties on the model of manifest properties such as liquidity. These are, however, not worries I am inclined to pursue.
Rather, the chief issue is that such a conception of folk psychology does, after all, aim to defend folk psychology as a theory. It is a relatively easy matter to see that the price of this sort of defense is, again, uninformativeness. The view on offer has it that folk psychology is not hostage to developments in the sciences because explanations at the manifest level do not presume to make informative claims about causal processes below the manifest level. Anything like the constraint of propositional modularity (along with the demand for horizontal absorbability) is to be rejected, since these strictures are clearly inappropriate in manifest/scientific inter-level contexts.
I want to emphasize that this is a perfectly sensible move in the case of many of the folk explanations we have been considering - and just because explanations at the manifest level aim to offer no informative account about the relevant causal processes. What information about causal processes does "the wine spilled because it was a liquid" provide? Rather, such an explanation points the way to a deeper inquiry concerning such causal processes. And this is why the availability of such more informative scientific accounts - far from undermining the folk explanation - serves as a vindication (a spelling out) of our folk account.
I have urged throughout that our commonsense conception of ourselves as reasoners and deliberators hinges upon taking very seriously folk psychology as an informative account of mental processes. We act as we do and reason to the conclusions that we reach because we appreciate the significance of the particular reasons that we possess. A folk psychological account of even the most mundane of inferences involves us in talk of a raft of our reasons, and in talk of a myriad of the particular causal consequences of those reasons. Yet folk psychology on the manifest conception does not even aim to offer an informative account of cognitive processes. If we are attempting to defend folk psychology as a causal explanatory theory, this maneuver secures vindication simply by claiming that folk psychology does not offer an account of mental processes. Yet if folk psychology does not express an informative account of mental processes, our commonsense self-image is no better than a fiction.
It may be claimed that we have, in any case, succeeded in showing both that there are beliefs and desires (in the way that there are, e.g., slippery things, and liquids) and that folk psychological explanations are true in the way that many other folk explanations are true. Surely, this puts our eliminativist worries to rest. But all we have shown is that one may respond to the letter of the threat of eliminativism while failing to respond to its spirit. For example, if beliefs and desires play no serious deep explanatory role, then our common-sense conception of ourselves as reasoners cannot be sustained. Such a vindication" does nothing to secure our common-sense self-image; it leaves worries about this wholly unaddressed. But if we regard such a conception as expressing a theory, then we must somehow make sense of the claim that we believe and behave as we do for reasons; that in deliberation and reasoning we are moved to intend and to believe for particular reasons. Were intentional states to play no more than the explanatory role secured above, this folk conception of ourselves must crumble.
Erstwhile vindicationism, if it cleaves to a manifest conception of folk psychology, fails to vindicate our common-sense self-image. It fails to take reasons-explanation seriously as causal explanation. It is, then, insufficient to say as Horgan and Graham do that we can vindicate reasons-explanation simply by noting that beliefs and desires figure in a "broad-ranging pattern of counterfactual dependencies."(34) After all," they note:
There are numerous singular causal explanations
involving non-scientific language - e.g.,
Sam's SAT score improved because he took
the SAT preparation course,' or 'the grass
turned green because Tom spread fertilizer on
it ..... Predicates like [these] presumably ...
do not figure in a body of mature scientific
theories.... Causal/explanatory truth is one
thing; absorbability into science is quite another.
Again, I do not disagree. We do not regard it as a fiction that things are liquid and slippery; nor do we regard as fictive the explanation that the wine spilled because it was a liquid. But again this sort of explanation does not pretend to offer structurally informative causal explanations. Relative invulnerability to fictionalization can be bought; but the price is relative uninformativeness. So, on this model we can agree to the bare claim that there are beliefs and desires and that, for example, but for the fact that we believe and desire we would not behave as we do. What I have argued is that such vindication fails to take our conception of ourselves as reasoners and deliberators seriously. What we would have to regard as a fiction is not that there are beliefs and desires, but that we deliberate and reason.
Perhaps the issue is best appreciated by noting its relation to what Jaegwon Kim terms "the problem of explanatory exclusion."(36) Criticizing those who would seek to understand mental causation and intentional causal explanation solely in terms of counterfactual dependencies, he writes:
If the demonstration of the possibility of mental
causation consisted merely in showing that
there are true counterfactuals of the form, If
S hadn't thought that p, S wouldn't have done
A,' who could disagree? What makes mental
causation ... a deep philosophical problem is,
among other things, the issue of compatibility,
or exclusion between rationalizing causes of
behavior and its neurobiological causes.... (37)
The issue is just how seriously we are to take mental causation and intentional causal-explanation. I have argued that if we are to make good our common-sense self-image as psychological theory, we must take it very seriously indeed, Kim's concerns and my own dovetail in their similar regard for the counterfactual test for intentional causal explanatory adequacy. The causal explanation "the wine spilled because it is a liquid" can surely be true, and its truth is supported by the truth of the counterfactual "If the wine hadn't been in liquid form, it wouldn't have spilled off the table." Yet we have seen that these causal explanations are exceptionally uninformative. In particular they aim to say nothing about the relevant causal processes. If this is as deep as belief/desire explanation cuts, then our common-sense self-image is vanquished. Vindicating our image of ourselves as reasoners and deliberators as causal/ explanatory theory hinges upon securing a more serious causal role for mental states.
IV Erstwhile Vindicationism
Thus the dilemma: You can insulate folk psychological theory from elimination, but only by rendering it so uninformative as to leave our common-sense self-image behind. It may be alleged, however, that I have taken the proffered analogies too seriously. And it may be suggested that folk psychology need not be so uninformative as I have made it out to be. Thus, while Jackson and Pettit do indeed think that common-sense psychology is insulated from elimination, at least in part, because of its uninformativeness, they do not think it is quite so uninformative as I have made it out to be.
According to them, "it is sufficient for having beliefs and desires that we are in states that satisfy the folk roles."(38) And they take it to be more or less obvious that we are sometimes in states which satisfy these folk roles. The basis of this last claim is their assertion that vernacular psychology is a purely functional theory. As a result, "to behave as if we had beliefs and desires is to provide overwhelming evidence that we in fact have beliefs and desires."39 Here is where the uninformativeness of folk psychology comes into play: "By being non-committal about the nature of what realizes the functional roles [folk psychology] minimizes the extent to which it goes beyond its observational base, and so is made peculiarly secure by that observational base."(40)
If the sole point here is that folk psychology says nothing about the nature of the state (presumably, a highly complex neural state) that realizes the role state, then who could object? This, after all, is a familiar feature of functional analyses of psychological states. More is, however, implied by this passage. For here Jackson and Pettit are relying upon the earlier cited analogy of being a good conductor.
The folk theory that metals are good conductors of heat really is made peculiarly secure by its observational base. But this is by virtue of the fact that the explanatory appeal to being a good conductor is about as uninformative as functional explanation can get; as such, "being a good conductor" is more a dispositional than a functional term. Being a good conductor just is defined in terms of certain observable inputs and outputs. So, in this case, it is true that "having the functional roles wrong is a very hard matter to conceal, precisely because they are functional roles."(41) Adverting to such properties (fragility, dormitive powers, etc.) is, of course, explanatory. Like all functional explanation, the appeal to being a good conductor aims to explain the capacities of a system by pointing to the capacities of its parts. But in the case of explanation by appeal to dispositional terms all we learn is that there is some categorical state of, e.g., the metal rod responsible for its rapid rise in temperature.(42) We already know that the rod's exposure to, say, a source of heat is responsible for its rapid rise in temperature. What we learn is that there is something about the rod that explains this. And it is true that this folk theory is unlikely to appear on the reasonable eliminativist's hit-list.
Yet there are more. and less informative functional theories and functional explanations. In particular, the more informative a functional theory is, the more specific are its claims about the structure of causal relations between the components of the system that the theory posits in order to explain observed behavior. This must be so, since functional explanation in general seeks, as we noted above, to explain the behavior of a system by appealing to the capacities of it constituent parts. And the more informative this becomes, the finer-grained and more specific is the account of the structure of causal relations between those constituent components or parts.
Moreover, folk psychology certainly seems to say a lot more than that some internal state or other causes behavior. As a result, folk psychology, unlike our above explanation by appeal to the notion of something's being a good conductor, does go significantly beyond its observational base. As I emphasized above, it is the hypothesized internal causal links between mental states that seem to be crucial to the vindication of our vernacular view of ourselves. For example, we say that I came to intend to A because I was so impressed by reason r. Jackson and Pettit are rather cavalier in their treatment of this worry. Common-sense psychology, they grant, does posit internal causal links; but it is, according to them, an easy matter to vindicate such claims. They appeal to the following example:
Well, suppose I punch 4 followed by 7 on my
pocket calculator, and that an I I then appears
on the screen. I discard the chance hypothesis
and adopt one which includes the clause: my
calculator stores for subsequent utilization the
number n when n is punched. This is a functional
hypothesis that goes beyond its observational
base both in respect of internal causal
links and in respect to hypothetical inputs and
outputs. Nevertheless, after a relatively few
trials, the hypothesis would become overwhelmingly
credible, although there might still
be much doubt about how the calculator does
What are we to make of the claim that the calculator stores n, when n is punched? What are the conditions under which we might regard this as the basis of a correct explanation of the machine's behavior? Are we to assume that we will have such a basis only if the rule has engineering or computational reality? If so, then even though there will still be much that we do not know about how the calculator does the storing, we will certainly have a process explanation, an account in terms of mechanisms at the computational level, of the calculator's behavior.
We know from Dennett's famous discussion of these matters that the rule "when n is punched, store n" need not be explicitly represented, need not have any computational reality. Dennett writes:
In a recent conversation with the designer of a
chess-playing computer program I heard the
following criticism of a rival program: 'It
thinks it should get its queen out early.' This
ascribes a propositional attitude in a very useful
and predictive way, for as the designer went
on to say, one can usually count on chasing that
queen around the board. But for all the many
levels of explicit representation to be found in
that program, nowhere is anything roughly
synonymous with `I should get my queen out
early' explicitly tokened. The level of analysis
to which the designer's remark belongs describes
features of the program that are, in an
entirely innocent way, emergent properties of
the computational processes that have 'engineering
reality.' I see no reason to believe that
the relation between belief-talk and Psychological-process
talk will be any more direct.(44)
Are Jackson and Pettit to be understood as arguing in a way which is consistent or at odds with the upshot of Dennett's remarks? The answer is unclear. Consider the first possibility. That is, imagine that they agree that there need be no computationally real level at which there is some discrete, causally relevant state that is the storing of n, in order for us to view as vindicated explanations which appeal to the rule (viz., that there is no level of computational reality at which the calculator operates upon the explicit rule "when n is punched store n"). If this is so, then the explanation of the calculator's behavior - and presumably the same will hold of folk psychological explanation - is to be understood in something like the following manner: There are internal causal processes going on which can be understood as if they carried out the rule: "when n is punched store n." Now it is true that this is a claim about internal causal processes; but it is not the sort of informative claim about causal processes that we might have expected given what Jackson and Pettit say in the above passage. That sort of claim is the following: When n is punched the machine enters a particular computational state which is the storage of n in its memory. Such a claim posits functionally discrete causally salient internal states which mirror the structure of the rule-based explanation. Again, there are, no doubt, many levels of analysis of the goings-on in the calculator at which such a state will not appear (most obviously at the micro-physical level of analysis, but there are, presumably, finer grained computational levels where this is true as well). Of course, the point of Dennett's example is that there may be no computational reality to such rules or algorithms. Taking the analogy seriously has the result that, notwithstanding this possibility, there are beliefs and desires and true folk psychological explanations.
If we are to understand Jackson and Pettit in this way, then it is true that folk psychology is made peculiarly secure by its observational base. Yet it is made so secure not just by remaining non-committal about the nature or composition of what realizes the functional-role states but also by being non-committal about specific internal causal links. According to the erstwhile vindicationist, belief/desire explanation says nothing particular about internal states or the causal relations between internal states. But, as I have urged, if you are non-committal about such internal causal links (at a certain level of analysis) then you can not expect to vindicate our common-sense conception of ourselves - even if you have vindicated the existential claim that there are beliefs and desires.
Again, this meager conception of what it takes to make common-sense psychology true leaves our conception of ourselves as reasoners behind. On this view, the truth of the explanation that, for example, Silvio went to the store because he wanted some cigarettes and believed that the store was soon closing does not demand that there be functionally discrete causally relevant states that mirror the structure of the folk explanation. Yet Jackson and Pettit insist that it is obvious that the folk roles are satisfied. Thus, the claim that we are sometimes in states that satisfy the folk roles amounts to no more than the following: we are sometimes in states (which perhaps vary quite typically both inter-and intra-personally) that are such as to make it possible for us to use the folk psychological idiom in explanation and prediction. What this makes apparent is that the distinction between being such as to behave as if one had beliefs and desires and in fact having beliefs and desires has been made to disappear by fiat alone.(45)
On the other hand, if I am wrong in suggesting that Jackson and Pettit argue in a way consistent with Dennett's remarks, then folk psychology is not made peculiarly secure by its observational base; and the vindication of our common-sense conception of ourselves would seem to depend upon our adopting the "deep" view of folk psychology and all the associated burdens.
And so the dilemma: If folk psychology does make the informative causal claims Fodor's deep view suggests, then to imagine its falsity does not entail that we must imagine radical skeptical hypotheses. It does mean that we must defend a version of scientific vindicationism. Contrarily, if it does not make such claims, then it is presumably insulated from such elimination, but only at the price of uninformativeness. And not just uninformativeness about the composition or nature of the realizer states, but uninformativeness about the structure of causal relations between the role states. Such uninformativeness means that we have empty vindication, vindication of the existential claim, without vindication of our common-sense conception of ourselves as deliberators and reasoners.
The erstwhile vindicationist has it that folk psychology is a theory whose its falsity is only barely conceivable. I have argued that there is considerable tension in this view. In seeking to reduce this tension, the erstwhile vindicationist insists that folk psychology need not be understood to make the sort of specific and determinate claims that are falsifiable. But in relieving the tension in this way, such a theorist fails to vindicate our common-sense self-image.
Even so, I do think that we should have considerable sympathy for the erstwhile vindicationist. It is very hard to take seriously the thought that we might not be deliberators and reasoners, a result which would be forced upon us by the falsity of folk psychology. I have in mind here not the charge that such an eliminativism would be pragmatically self-defeating,(46) but rather the more mundane point that - baldly stated - we know that we deliberate and reason. Still, the difficulty is that giving good sense to our assurance that we are deliberators, while insisting that folk psychology is an informative causal explanatory theory, forces upon us the demands of scientific vindicationism. And scientific vindicationism, as we have seen, places our folk conception of ourselves at risk.
Our assurance - on the basis of the facts as we know them to be - that we are deliberators and reasoners is, then, at odds with a robust theoretical conception of folk psychology.(47) I have been arguing that if we regard our common-sense conception of ourselves as expressing a causal explanatory theory, then its vindication hinges upon scientific vindication. We can, however, seek to avoid these difficulties by opting for a non-theoretical conception of common-sense.
This is not the time to consider what is to be said in favor of simulation accounts of the folk psychology.(48) I want only to note that much that the erstwhile vindicationist says indicates congeniality to simulationism. For example, Jackson and Pettit make much of the fact that we move with much confidence back and forth "between behavior, situations, and beliefs and desires."(49) What this makes crucial is the nature and status of our everyday common-sense psychologizing. They wish to understand our explanatory practices as serving to limn the common-sense functional roles of belief, desire, and the like. But we are not forced to such a view. Indeed, the centrality of the explanatory and self-explanatory practices of this picture is suggestive not so much of a common-sense functionalist account of folk psychology, as of simulation accounts.(50) According to such views, our folk psychologizing is not undergirded by an implicit psychological theory, but is carried out by an kind of empathy. As Simon Blackburn puts it, I "`recentre' my gaze as yours, or change my `egocentric map' and think about the world as it appears from that point of view."(51) I make adjustments for your cognitive/motivational/affective constitution, and, then, once in the imagined state, I project the thinking I do to you (or to myself retrospectively). As Stephen Stich and Shaun Nichols note, such a view "pulls the rug out from under eliminativists."(52) If our commonsense psychologizing is not undergirded by theory, then it cannot be a "radically false theory." Of course, such a view pulls the rug out from under familiar forms of vindicationism as well, for if folk psychology is not a theory, neither can it be a true psychological theory.
The simulation account has it that beliefs, desires, intentions, and the like are not theoretical posits. Rather, intentional mental states just are the states we inhabit when we engage in deliberation and reasoning - the states that, for example, Silvio is in when he occurrently thinks to himself, "I really would like some cigarettes this evening" and "the store closes within the next half hour." These are the states that we project onto ourselves in retrospective self-understanding, and they are the states we project onto others when we seek to explain their behavior. No advances in the cognitive sciences will demonstrate that we do not reason and deliberate, but all this means is that we do reflexively and occurrently consider such sentences, or - in the psychological mode - think such thoughts as Silvio's.
Yet if this is so, folk psychology does not offer a "deep" account of our cognitive architecture. What secures reference for the terms "belief," "desire," etc., is just such experiences of self-conscious reasoning and deliberation. There are, to be sure, many difficulties to which the simulationist must respond.(53) For my current purposes, what bears note is that if, at the end of the day, simulation can be defended, much that the erstwhile vindicationist claims will turn out to be correct. There can be no serious question as to the existence of beliefs and desires, nor can there be much worry about whether we do deliberate and reason. The error of the erstwhile vindicationist was to defend folk psychology as a causal explanatory theory. For if you defend vernacular psychology as a causal explanatory theory, and you opt for erstwhile vindicationism, our common-sense intentional psychological explanations become so uninformative that our common-sense self-image is left behind.
My aim in this essay has been to show that if folk psychology is to be meaningfully vindicated as a causal explanatory of mind and behavior, the formidable (and Fodorian) burdens of scientific vindicationism must be taken up. I cannot here set out in detail an alternative conception of the explanatory status of folk psychology. It is apparent, however, that this alternative conception must show that the primary explanatory contribution of folk psychology is not causal explanatory. I do think that the simulationist perspective offers us a clue as to the nature of this explanatory contribution. For just as our understanding of our own actions and belief acquisitions - in deliberation and reasoning - is not primarily causal/predictive, so will the explanation of the behavior of our fellows, in so far as this is accomplished by simulation, not be causal/predictive. The explanation of the behavior of others will amount to the projection of simulated self-understanding.(54,55)
(1.) See, for example, Terence Horgan and James Woodward, "Folk Psychology is Here to Stay," The Philosophical Review, vol. 94 (1985); George Graham and Terence Horgan, "How to be Realistic About Folk Psychology," Philosophical Psychology, vol. 1 (1988); Terence Horgan and George Graham, "In Defense of Southern Fundamentalism," Philosophical Studies, vol. 62 (1991); Charles Chastain, "Reply to Baker," in Contents of Thought, Robert H. Grim and Daniel D. Merrill, eds., (Tucson: University of Arizona Press, 1988); Frank Jackson and Philip Pettit, "In Defense of Folk Psychology," Philosophical Studies, vol.59, (1990); and, Lynn Rudder Baker,"The Attitudes as Nonentities," Philosophical Studies, vol. 76 (1994). (2.) This is emphasized by Horgan and Graham, Chastain, and Jackson and Pettit in the papers cited above. (3.) Psychosemantics, (Cambridge, MA: MIT Press, 1987) p. 8. (4.) Ibid., p. 10. (5.) This is explicit in Paul Churchland's "Eliminative Materialism and the Propositional Attitudes," Journal of Philosophy, vol.78(1981),and Stephen Stich's From Folk Psychology To Cognitive Science, (Cambridge, MA: MIT Press, 1983), chapter 11. (6.) Horgan and Graham evince puzzlement over Fodor's claim that "vindicating folk psychology means demonstrating the mere possibility of scientific absorption" ("Southern Fundamentalism, p. 131, n. 6). But there ought to be nothing surprising about this claim. (7.) "Connectionism, Eliminativism, and the Future of Folk Psychology," in The Future of Folk Psychology John D. Greenwood, ed. (Cambridge: Cambridge University Press, 1991). (8.) Ibid., p. 97. (9.) The assumption here is, of course, that connectionist models are psychological models and not implementational ones. See Jerry Fodor and Zenon Pylyshyn, "Connectionism and Cognitive Architecture" in Connections and Symbols, Steven Pinker and Jacques Mehler (Cambridge, MA: MIT Press, 1988), pp. 64-66. (10.) "Being Indiscrete," in Greenwood, p. 132. (11.) Ibid . p. 131. (12.) In any case, not all the inferential rules of a system can be explicitly represented in that system. This is the upshot of Lewis Carroll's "What the Tortoise Said to Achilles," Mind, vol. 4 (1985). (13.) Fodor makes the related point: [It is sometimes alleged] that intentionalist methodology permits inference from 'x's behavior complies with rule r,' to 'R is a rule that x explicitly represents.' The embarrassment is supposed to be that this allows the inference from 'The movement of the planets comply with Kepler's Law,' to some astronomical version of LOT.
But in fact no such principle of inference is assumed. What warrants the hypothesis that r is explicitly represented is not mere behavior in compliance with K; it's an etiology according to which r figures as the content of one of the intentional states whose tokenings are causally responsible for x's behavior. And, of course, it's not part of the etiological story about the motions of the planets that Kepler's Law occurs to them as they proceed upon their occasions. (Psychosemantics, p. 156, n. 9.) (14.) "Connectionism, Eliminativism," pp. 99-100. (15.) "Being Indiscrete," p.132. (16.) It should be emphasized, however, that Heil closes his discussion with the remark: "Connectionism, then, poses no special threat to ordinary appeals to beliefs and desires. It remains to be seen whether intentional states can reasonably be regarded as purely theoretical posits" (p. 132). This last sentence anticipates the central thrust of this essay's conclusion. (17.) "Southern Fundamentalism," p. 108. (18.) In "How to be Realistic About Folk Psychology," Graham and Horgan come close to denying that scientific absorbability is a constraint on vindication. In "In Defense on Southern Fundamentalism," they defend the weaker claim that the vindication of the vernacular is neutral with respect to such constraints as SA. My claim is that vindicationism cannot afford such neutrality. (19.) "Folk Psychology Is Here to Stay," pp. 202-3. (20.) "Reply to Baker," pp. 22-23. (21.) "In Defense of Folk Psychology," p. 42. (22.) "Southern Fundamentalism," p. 116. (23.) See John Greenwood's "Reasons to Believe" in The Future of Folk Psychology, pp.71-80. (24.) It is clear that this is Fodor's view of the matter; he is, after all, a vehement foe of the notion that vertical absorbability (where this means smooth reduction) is a test of adequacy of the special sciences. See, in addition, Robert McCauley's "Intertheoretic Relations and the Future of Psychology," Philosophy of Science, vol.53 (1986). (25.) See Jaegwon Kim's, "Mechanism, Purpose, and Explanatory Exclusion," in Philosophical Perspectives 3, James Tomberlin, ed. (Atascadero, CA: Ridgeview Publishing Co.,1989). (26.) "Intertheoretic Relations and the Future of Psychology," p.192. (27.) "Connectionism and the Future of Folk Psychology," p.102. (28.) Ibid. p. 105. (29.) It is to be emphasized, of course, that not all intra-level conflicts result in the outright elimination of the ontology of one of the theories. McCauley remarks: "In intralevel contexts with little conceptual friction (and a relatively straightforward mapping between theories) although new theories replace old ones, features of the older theories persist .... In addition, a reasonably faithful image of its ontology endures in that of its successor" ("Intertheoretic Relations," p. 192). (30.) Fodor, Psychosemantics, p. 8. (31.) Brian Loar makes the same point: "If it were to turn out that the physical mechanisms that completely explain human behavior at no level exhibited the structure of beliefs and desires, then something that we had all along believed, viz., that beliefs and desires were among the causes of behavior, would turn out to be false," Mind and Meaning, (Cambridge: Cambridge University Press,1981). (32.) "Elimination Versus Non-Reductive Physicalism" in Reduction, Explanation, and Realism, David Charles and Kathleen Lennon, eds. (Oxford: Clarendon Press,1992), p.251. (33.) This is an important intuition pump of eliminativism. Thus Stephen writes: "The very fact that it is a folk theory should make us suspicious. For in just about every other domain one can think of, the ancient shepherds and camel drivers whose speculations were woven into folk theory have a notoriously bad track recond" From Folk Psychology to Cognitive Science, p. 229. Stich is, of course here assuming that folk psychological explanations are not at the manifest level. (34.) Something emphasized by Baker in her "The Attitudes as Nonentities." (35.) "Southern Fundamentalism," pp. 127-28. (36.) "Mechanism, Purpose, and Explanatory Exclusion," and "Explanatory Exclusion and the Problem of Mental Causation," in Information, Semantics, and Epistemology, Enrique Villanueva, ed. (Oxford: Basil Blackwell, 1990 (37.) "Explanatory Exclusion and the Problem of Mental Causation," p.45. (38.) "In Defense of Folk Psychology," p. 36. (39.) Ibid., p. 43. (40.) Ibid., p. 42. (41.) Ibid., p. 41. (42.) Ryle would not agree: "to possess a dispositional property is not to be in a particular state...." The Concept of Mind (London: Hutchinson,1943), p.43. (43.) "In Defense of Folk Psychology," p.42. (44.) "A Cure for the Common Code," Brainstorms, (Cambridge,MA: MIT Press,1978), p. 107. (45.) In fact, these difficulties are very near to the surface of Jackson and Pettit's account. For they claim that although folk psychological explanations are causally explanatory, the properties to which such explanations advert are not causally efficacious. Folk psychological explanations are a species of "program explanation"; such explanations are not "process" explanations. (See their "Program Explanation: A General Perspective," Analysis, vol. 50 .) My claim amounts to the following: One cannot vindicate our common-sense conception of ourselves as a theory, unless one defends folk psychological properties as causally efficacious, and so folk psychological explanations as process explanations. (46.) Notably, Lynn Rudder Bakers' "Cognitive Suicide," in Contents of Thought. (47.) For an argument to this effect see Simon Blackburn's "Theory, Observation, and Drama," Mind and Language, vol. 7 (1992), pp.197-202. (48.) I discuss simulation accounts in "Simulation, Rationality, and the Status of Vernacular Psychology," (in preparation). (49.) "In Defense of Folk Psychology," p.51. (50.) The same can be said of Horgan and Graham's "conceptual/semantic competence" argument for their "austere" conception of vernacular psychology ("Southern Fundamentalism," pp.116-20). They write: When competent speakers routinely . . . think it intuitively obvious that certain kinds of descriptions are correctly applicable in various situations, normally the most plausible explanation will be that these intuitive judgments are the direct and reliable product of the speakers' conceptual/semantic competence. In such cases, these patterns among the judgments will provide empirical evidence that under an adequate account of the relevant concepts and the terms expressing them, the judgments will usually be correct (p.118).
It seems clear that this sort of argument when applied to folk psychology is neutral with respect to the theoretical or simulationist account. Moreover, it has what is an odd consequence if one cleaves to the theoretical conception: it is only because we are folk psychologists ourselves that we can be certain the common-sense psychology is true. It is only because we, as a matter of fact, do make judgments whereby we attribute folk psychological states to ourselves and others and that there is wide-spread inter-subjective agreement about such judgments that we can afford to be unconcerned about conditions like SA.
But imagine a collection of non-linguistic creatures; they have a complex behavioral repertoire such that we must ask ourselves whether intentional characterization is justified. Yet there is nothing like robust inter-subjective agreement as to whether the behavior of such creatures demands appeals to beliefs, desires, etc. (perhaps for reasons adduced by Donald Davidson or Stephen Stich). The result is that a determination of whether the behavior of such creatures is to be explained by appeal to beliefs and desires is radically discontinuous from the way in which we are meant to understand the vindication of the belief/desire framework with respect to ourselves. What's more, if we are to remain non-instrumentalist, it seems apparent that with respect to such creatures we must resort to scientific vindicationism. Our grip upon common-sense psychological concepts is secured in our own case by appeal to one body of evidence, and, in the case of these other creatures, by a very different body of evidence. All of this suggests that the concept of belief applied to us, on the one hand, and to these creatures, on the other, are different concepts. If we are conceiving of folk psychology on the theoretical model, this is certainly an untoward result. (51.) "Theory, Observation, and Drama," p. 191. (52.) "Folk Psychology: Simulation or Tacit Theory? "Mind and Language, vol.7 (1992),p.37. It should be noted that Stich now believes that this judgment was too quick. See Stich and Ian Ravenscroft, What Is Folk-Psychology?" Cognition, vol. 50 (1994). (53.) The current debate over the standing of simulation accounts is well represented in Mind & Language. op. cit. (54.) I spell this out in significantly more detail in "Simulation, Rationality. and the Status of Vernacular Psychology."
I owe thanks to Paul Hurley and Charles Young, both of whom commented upon earlier drafts The comments of an anonymous referee for this journal much improved this essay.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||folk psychology|
|Publication:||American Philosophical Quarterly|
|Date:||Jul 1, 1995|
|Previous Article:||Philosophical anarchism and legal indifference.|
|Next Article:||Dreams and the coherence of experience: an anti-idealist critique from classical Indian philosophy.|