Larry Hauser
Email: lshauser@aol.com
Introduction
It is interesting to compare the multiplicity of tools in language and of the ways they are used, the multiplicity of kinds of word and sentence, with what logicians have said about the structure of language.{1}
I wish to extend the comparison to theories or "conceptual networks." The network of terms, background knowledge, and generalizations which Paul Churchland and others have termed "folk psychology" is widely held to have same logical structure and cognitive functions as scientific theories generally, and it is a small step from this characterization of folk psychology as science (or proto-science) to the observation that it's bad (imprecise, theoretically stagnant) science; bad beyond repair; calling for its elimination and replacement. This diagnosis of the `trouble' with folk psychology rests on a misapprehension of the cognitive aims, and hence the logical form, of folk psychology's central pattern of explanation (of actions by motives or aims). Such explanations conform neither to Hempel's Deductive Nomological nor Inductive Statistical patterns: conformity to these scientific patterns would bar them from having the practical, evaluative uses they have.
Against the claim that folk psychology is a theory, I contend that folk psychology is not empirically vulnerable in the same way theories are, and has evaluative functions that make it irreplaceable by a scientific theory. It is neither would-be nor has-been science.
Folk Psychology
`Folk psychology' refers to the system of motive, cognitive, and other concepts that support our mutual understanding and, we think, explain what we do. A striking feature of this conceptual apparatus is the teleological character and, consequently, the patently unscientific look of folk psychology's central pattern of explanation and prediction (of actions by aims). Such explanations have the logical features they do, I maintain, to suit them to evaluative and deliberative functions (which seem extratheoretic or practical){2} while also suiting them to the explanatory and predictive functions they seem to have in common with science. These extratheoretic (evaluative and deliberative) functions suggest this "network of common-sense concepts" Churchland terms "folk psychology" is not "an empirical theory with all the functions, virtues, and perils entailed by that status".{3} It has functions scientific theories lack and -- to tell the truth -- few scientific virtues. Rather, it has a host of scientific vices. Or would-be vices, if folk psychology were a scientific theory. The "theory theory" (as I'll call the view that folk psychology is a science-like empirical theory){4} sins both against the `text' of our usage (since folk psychology is very unlike scientific theories on its face) and also against interpretive charity since folk psychology is notoriously lacking in scientific virtues. To construe it as a science-like theory is by that very fact to construe it as stagnant, impoverished and radically defective.
Nor should the theory theorist take much comfort in the thought that it's necessary to impute these vices to folk psychology in order to explain its predictive virtues, if the thought is that only science-like empirical theories can be predictive. Given folk psychology's conspicuous lack of many other scientific virtues, besides predictivity, if the thought is that predictive utility flows from scientific virtue, the considerable predictive utility of folk psychology, far from being explained is rendered inexplicable. I intend to show the thought that only something having the logical character of a science-like empirical theory can be predictive is an unsupported dogma. Not only do the extratheoretic functions of folk psychology require us to construe the logic of its central pattern of explanation as unscientific, but such a construction is by no means inconsistent with the considerable predictive utility such explanations (and folk psychology generally) so obviously have. Against an uncharitable interpretation which flies in the face of the untheorylike appearance of folk psychology generally, and the unscientific look of its central pattern of explanation (of actions by aims) in particular, I propose an interpretation which both fits the `text' and is charitable.
The questions that confront us, then, are these. 1) If folk psychology is not a science-like theoretical network, what kind of conceptual network is it? And 2) how could a network of this sort -- i.e., of whatever sort folk psychology is -- be predictive? "If reason and action illustrate a different pattern of explanation [than science], that pattern needs to be identified."{5}
Folk Psychological Explanation
Aristotle gives the following example of the sort of explanation in question: `Thus: why did he come? To get the money -- wherewith to pay a debt -- that he might thereby do what is right.'{6}
We think this explains why he came; and had we known beforehand that he needed money and knew getting it required coming, we would have expected he'd come. As to the basic constituents of such explanations, that they centrally invoke motives or ends (e.g., wishes, desires, objectives) and cognition (e.g., believing, seeing, detecting), there is little disagreement. The logical character of such explanations, however, is disputed. As motives come in all `flavors' -- wants, intentions, hopes, ideals, etc. -- it may even seem doubtful whether our various explanations of behavior in terms of aims or purposes comprise a single form at all. Aristotle's well chosen example cuts through this knot of varying terms and reveals, I think, one form of explanation underlying the seeming variety of our ordinary everyday accounts of human behavior. The revealing feature of the example, in this regard, is its neutrality with regard to the various motivating attitudes, e.g., wanting, intending, needing, etc. It simply invokes a chain{7} of propositions as aims (leaving the `flavor' of his aiming unspecified) to explain his coming. We might also have said "He wanted to get money" or "He intended to get money, or "wished" or "hoped" or "desired" to get money. On their face, these seem stylistic variants of one and the same explanation. Whether the aiming was lustful (as "desires" might suggest) or dutiful (as "intends" might) or reluctant (as "needs" might) the explanatory pattern is the same.
We even mix up these various motivating attitudes in a single account, any which way we choose, it seems. Perhaps: He needed money -- since he intended to pay a debt -- because he wanted to do what was right. Or perhaps: He wanted the money -- since he needed to pay a debt -- because he intended to do what was right. Aim (in whatever sense it explains action) gets transmitted across these various attitudes every which way, regardless of `flavor'. The different attitudes may each provide different supplemental information about the agent, yet all seem to carry the same explanatory force with regard to the particular act to be explained. What matters is just 1) that they are motivating and 2) that towards which they motivate, i.e. the intentional or propositional content of the attitude. We use these various motivational terms more or less interchangeably, as context and connotation suggest.
The question then is what is the explanatory force of attributions of aim: what sort of explanations do motives or reasons provide? Do they give grounds -- if accepted as true -- for expecting the act they purport to explain to occur;{8} and if so, are the grounds inductive or deductive? Lastly, if attributions of purpose provide grounds for expecting the acts they explain to occur, does their provision of such grounds involve the subsumption of their explananda under empirical generalizations or contingent "covering laws", like scientific explanations?
Potential Predictivity
It is surprising, I think, that anyone has ever thought to deny or seriously minimize{9} the predictive utility of explanation by motives or aims. Suppose, for instance, that I go to the bank for a loan. The loan officer will inquire concerning my intentions (e.g., she might want references, an explanation of what the loan is for, look me in the eye, etc.). She may even attempt to bring it about that I want to repay by various threats (e.g., to my credit rating). Of course she will want to know some other things about me as well: that I understand what repayment requires (the terms of the loan), and that I have the ability to fulfill those terms (e.g. that I'm employed). A full reconstruction of the pattern of motive explanation will obviously have to include cognition and capacity as well as aim, but given my intent and ability to repay, and my understanding the terms of the loan, the loan officer has reason to expect I'll repay. Motives in conjunction with beliefs and abilities obviously support expectations: money, love, and even lives frequently depend on this.
Clearly attributions of motive or aim (in conjunction with beliefs and abilities) are predictive. What remains to be seen is how.
Logical Considerations
One might suppose -- since a loan officer is never absolutely certain a creditor will repay -- that the reasoning that underwrites her (probable) expectation is inductive. Suppose it were inductive. Do we allow that seeking something might fail to result in our doing what we know it requires, despite our ability to perform the required acts, in some instances? Suppose payment on my loan is overdue. The bank has sent out a series of less and less gently worded reminders. Finally, a bill collector calls and questions my intentions. I could plead forgetfulness or disability; but I couldn't say, "I guess I'm just the improbable one who intends to repay but doesn't. It's not that I can't, repay," I say, "I just don't."
And the Bill Collector says, "You mean you won't".
"Who knows what the future holds?", I say. "I probably will -- I intend to -- but maybe I won't."
In fact, if intending to repay (knowing how and being able) doesn't always lead to repayment, why can't I say -- just straight out -- "I intend to but won't" On the inductive construal these words would not be (as I take it they are), self-refuting.{10} I might even have grounds -- not necessarily good grounds -- for supposing that I do just happen to be that unusual individual who doesn't repay despite their full intention to (and knowing how and being able). Perhaps I concluded this from my horoscope.
So, the underlying pattern of inference must -- if it is to support our practice of assigning blame and extenuation on the basis of agent's intentions -- be deductive.
We might further inquire as to whether it is nomological, i.e., whether the general principle required to validate the inference from intent, belief, and ability to action is a disconfirmable generalization. For much the same reasons as I adduced above -- against the inductivist proposal -- I believe, we must reject this suggestion also. Suppose it were nomological. Now when the collection agent calls I can maintain the innocence of my motives on the grounds that I'm a disconfirming instance: "those darn folk psychological laws", I might say, "are so unreliable." Again, might I not say -- having concluded from my horoscope (or whatever) -- "I intend to repay but won't" (though I know how and I'm able); and again this is surely unacceptable.
I conclude that in order to support the evaluative practices it does, the principle which validates our inferences from intention, belief, and ability to action must be held true a priori. We hold it true a priori in order to use attributions of intention (in connection with belief and ability) to hold people responsible for their actions.
Davidson rightly observes that "justifying and explaining an action, so often go hand in hand" and further, that "the justifying role of a reason ... depends on the explanatory role". But Davidson is wrong to think "the contrary does not hold";{11} for the evaluative force of attributions of purpose (along with beliefs and abilities) constrains the logical form of the explanatory accounts which invoke these concepts. It constrains them to be deductive, and not nomological.
A Priori Principles and Prediction
Here one might wonder how anything of this sort could possibly be predictive. But, "the point is not that nothing but action (accompanied by appropriate beliefs [and abilities]) reveals a desire, but that nothing can definitely override it."{12} Many things besides my eventual repayment can evidence my intent to repay prior to my actual repayment. The point is just that my subsequent failure to repay (unless excused by disability or ignorance of what's required) always suffices to override any previous estimate of my intent. But it does not follow from this -- as the objection here seems to suppose -- that I can never have justified belief that an agent intends something (and knows what's required, and is able to do what's required) prior to their actually doing the thing required. All that follows from the a priori status of the validating principle of our inference is that one can never know for certain whether someone's intentions, beliefs, and abilities are such that they'll act in some specific way. It does not follow that we can never know or have justified beliefs about anyone's intentions and beliefs and abilities prior to their actually performing the act their beliefs, intentions, and abilities explain. The thought that nothing of this sort could possibly be predictive seems to tacitly be requiring knowledge of initial conditions to be incorrigible. If this incorrigibility requirement is mistaken -- as it almost certainly is -- there seems no reason an inference validated a priori, as I have maintained that inferences from motive, belief, and ability to action are, could not predict.
Explanatory Aims
The following considerations support the contention that the sense in which motives or aims are explanatory and evaluative is the `all out' or comprehensive sense in which I might want to take a ride on the fun wheel, say, despite not wanting (in the component or prima facie sense) to pay the price of the ticket. Indeed the probability of a component or prima facie motive leading to a particular act, as Davidson observes, "is vanishingly small."{13} Given any prima facie aim, you might still refrain from doing what you perceive the attainment of your prima facie aim requires, due to conflicting prima facie aims. So why did we expect him to come to get the money instead of, say, robbing a convenience store. Here -- e.g. if someone asks "Why not just rob a convenience store?" -- we supplement our characterization of his intent (to get money) by citing additional purposes or motives (e.g., not to commit armed robbery). It seems apparent, then, that the aims which explain and predict of action must be `all out' or total or comprehensive, not just component or prima facie aims.{14}
Of the various motive attributive terms some (e.g., `wants', `likes') seem more conducive to use in the prima facie sense just distinguished than others. On the other hand, `intend' (and perhaps also `seek') resist prima facie use, and generally seem to indicate `all out' (or the explanatory sort) of motivation.{15} Consider: can my intention to repay a loan conflict with my intention to pay for my child's medical treatment? If I took the loan knowing full well (but not confiding to the loan officer) that I was about to incur the large expense of Tiny Tim's orthopedic surgery, it should probably be said that I never really intended to repay, or at least that I never intended to repay on time, though I may intend to repay late all along. (But this intention and my intention to pay for Tiny Tim's surgery are not conflicting). Alternately, suppose I was about to repay, but at the last moment I used the money for Tiny Tim's emergency appendectomy. Here, we should say I did intend to repay on time until Tiny Tim's attack of appendicitis, but then, when I decided to pay for the appendectomy, my intention changed to repaying later. The conflicting intentions are again resolved (by this change) into two compatible intentions, or a single intent to pay for the appendectomy and then repay the loan. Or suppose -- a case of "weakness of will" -- I got the money to repay, but on the way to repay I was passing a tavern, went in, and spent the money. Here, I think, we would say that at some point between eyeing the tavern and spending the last of the money I ceased intending to repay on time (if not altogether).
Whether such conflicts can be resolved in all cases -- whether all must involve a change in intention -- may be debated, but seems, I think, likely. I might still intend to quit smoking, as I light up another cigarette, but presumably I don't intend to quit now (even as I light up) but -- as the saying goes -- "later".{16}
If the preceding discussion suffices to show that motive attributive terms have an `all out' or comprehensive sense, and that it is this sense we employ to explain and predict action, qualms about predictivity may arise again. How -- it might be wondered -- could we ever know that someone has a total goal or `all out' want such as to require them to do some specific thing prior to their actually doing it? The answer, I suggest, is that we often know well enough what someone -- particularly someone we know -- is `up to' or `after' in particular circumstances. The `totality' of `all out' aims is just their unlimited openness to further elaboration, as may be required. This may prevent us from ever knowing completely full well what anyone (including ourselves) is seeking, but does not prevent us from knowing well enough to know what to expect from them within a broad range of foreseeable circumstances. This no more undermines the predictive utility of the concept of `all out' aiming than the fact that scientific knowledge is never complete -- that scientific laws all seem, eventually, to get revised or replaced -- undermines the predictive utility of science.
This analogy between the openness of `all out' aim ascriptions, and the open endedness of scientific development reveals a difference in what might be termed the "epistemic aims" of folk psychology and science. If we think of folk psychology as a research program, its researches are aimed not at the discovery of more and more general laws of human behavior, but at the deeper and fuller understanding of the aims of particular individuals with whom we are concerned.{17} Thus, we may observe, that folk psychology not only has (evaluative and deliberative) functions that theories lack, but it performs the epistemic function of prediction, which it seems to have in common with scientific theories, in a different way. Theories produce knowledge in the form of increasingly shared (and publicly testable) laws. folk psychology produces wisdom in the form of subjective perceptual abilities, primarily, the ability to `read' the motives (along with the beliefs and abilities) of various individuals from their gross behavior (what they say and do) and its subtleties (e.g. gestures, tone of voice, etc.).
Alternately put, scientific explanations concentrate inductive risk in the "covering laws", and seek to minimize the inductive risk of the observation base, or the descriptions of the initial conditions, in the interest of theoretical advance. Motive explanations concentrate inductive risk in the descriptions of the initial conditions (i.e., a particular individuals motives, beliefs, and abilities) by removing all inductive risk from the validating principle (i.e. by holding it true a priori) in the interest of improving our estimates of particular individuals' aims. Here "the reliability of my prediction rests upon my understanding of the person ... it is because I understand him ... that I am able to say what he will do."{18} The refinements folk psychology undergoes in the course of being applied are refinements, of our individual empathic and perceptual abilities in relation to other particular individuals. "Progress" in folk psychology is not theoretic elaboration and refinement of a shared body of laws, but my refinement and qualification of my estimates of the aims of particular individuals. This may contribute somewhat to my general predictive and explanatory facility with folk psychology concepts, or my ability to estimate the motives of others (besides this particular individual), but hardly at all to our collective understanding or the development of a shared body of generalizations.
This is all explicable -- and to be expected -- if the logic of folk psychological explanations (of actions by aims) is I have presented it. By holding the validating principle true a priori we guarantee that the refinements and qualifications our predictive failures occasion will be refinements in our descriptions of the motivational states of individuals. The modifications and qualifications which predictive failures occasion in folk psychology are invariably regressive from the point of view of theory construction. The progress here, in folk psychology, is towards less generality of `laws' and even, it seems of usage. Theoretic progress in science consolidates dialects: folk psychology seems rather to produce a profusion of idiolects among its various practitioners. But the appropriate conclusion to draw from this, I submit, is not that folk psychology is a stagnant and infertile scientific theory, but that it's not a science-like empirical theory at all all: neither would-be science nor has-been science. It's a cognitive implement of a different sort, to serve evaluative and deliberative purposes, which informs our expectations in a different way than scientific theories.
Email me: your comments.
Back to: Homepage; Curriculum Vitae; Virtual Office
Notes
1. Ludwig Wittgenstein, Philosophical Investigations, translated by G. E. M. Anscombe, 1958, Basil Blackwell, Sect.2.^
2. Here I invoke Aristotle's distinction between thought and discourse that is evaluative and directed toward accurate perception or description of the particulars of cases (practical reason), and thought and discourse that is primarily nonevaluative (explanatory or "contemplative") and directed toward the formulation and discovery of general laws (theoretical reason). (See especially Nichomachean Ethics, Book 6.)^
3. Paul Churchland, "Eliminative Materialism and the Propositional Attitudes" in Journal of Philosophy, 1981, vol. lxxviii, no. 2, p. 68.
4. As it has been called in some of the recent literature: e.g., John Bishop, "Is a Unified Theory of the Mind-Brain Possible?: Review of Patricia Churchland Neurophysiology: Toward a Unified Theory of the Mind/Brain" in Biology and Philosophy, 1988, vol.3, no.3.^
5. Donald Davidson, "Actions, Reasons, and Causes" in Essays on Actions and Events, 1963, Oxford University Press, p.261-275.^
6. Posterior Analytics, translated by G. R. G. Mure in The Basic Works of Aristotle, ed. Richard McKeon, 1941, Random House, 85b30: the case of final causes" (85b36).^
7. Cf., G. E. M. Anscombe, Intention, 1966, Cornell University Press, Sect.26.^
8. Hempel's requirement of potential predictivity: "Any adequate answer to the question why a given event occurred will have to provide information which, if accepted as true, would afford good grounds for believing that the event did occur." Carl G. Hempel, "Aspects of Scientific Explanation" in Aspects of Scientific Explanation, 1969, Free Press, pp. 470-471.^
9. See, e.g., Donald Davidson, "Hempel on Explaining Action" in Essays on Actions and Events, 1976, Oxford University Press, pp.264-265.^
10. See Anscombe, Intention, Sect.52, for an argument to the contrary.^
11. "Actions, Reasons, and Causes," p. 8.^
12. Davidson, "Hempel on Explaining Action, p. 267.^
13. Davidson, "Hempel on Explaining Action, p. 265.^
14. "One must measure through a single standard for one pursues the greater good. Hence one can form a single image out of many images." (Aristotle, On the Soul, translated by J. A. Smith, in The Basic Works of Aristotle, 434a5-9.) The single image is the object of Aristotelian wish (or rational appetite), and wish provides the major (i.e. the motive) premise of the practical syllogism. Similarly Hempel maintains that what explains the act, in what he terms "rational explanation" (which gives the agent's reasons for acting) must be a "total objective" ("Aspects of Scientific Explanation," p. 465).^
15. The "prima facie" versus "all out" terminology is Davidson's. Anscombe makes a similar distinction (Intention, Sect.36, p.67) between a mere "prick of desire" and the sort of wants that explain actions.^
16. This is obviously related to the earlier point about whether someone can intend 'all out' to p, and continue to do so all along, and still not p despite knowing how and being able. Anscombe offers the counter example of Peter's betrayal of Christ (John 18:15-27), of which she says "St. Peter could do what he intended not to, without changing his mind, and yet do it intentionally." (Intention, Sect.52, p.94)^
17. This is the second mark (the first being its evaluative character) Aristotle cites to distinguish practical from theoretical cognition. "That practical wisdom is not scientific knowledge is evident; for it is ... concerned with the ultimate particular fact." (Nichomachean Ethics, 1141a23)^
18. A. I. Melden, Free Action, 1961, Humanities Press, p. 208.^