WHY WE CAN'T UNDERSTAND THOUGHT FROM THE OUTSIDE 8 page
A Treatise of Human Nature, book 2, part 3, sec. 3 ( L. A. Selby-Bigge, ed., Oxford University Press, 1888), p. 416. I'm afraid it's unavoidable to revisit the subject of prudence in a discussion of practical reason, overworked as it is.
quite common, to lack it is not contrary to reason, any more than to lack sexual desire is contrary to reason. The problem is to show how this misrepresents the facts.
The fundamental issue is about the order of explanation, for there is no point in denying that people have such secondorder desires: the question is whether they are sources of motivation or simply the manifestation in our motives of the recognition of certain rational requirements. A parallel point could be made about theoretical reason. It is clear that the belief in modus ponens, for example, is not a rationally ungrounded assumption underlying our acceptance of deductive arguments that depend on modus ponens: Rather, it is simply a recognition of the validity of that form of argument. 2.
The question is whether something similar can be said of the "desire" for prudential consistency in the treatment of desires and interests located at different times. I think it can be and that if one tries instead to regard prudence as simply a desire among others, a desire one happens to have, the question of its appropriateness inevitably reappears as a normative question, and the answer can only be given in terms of the principle itself. The normative can't be displaced by the psychological.
If I think, for example, "What if I didn't care about what would happen to me in the future?" the appropriate reaction is not like what it would be to the supposition that I might not care about movies. True, I'd be missing something if I didn't care about movies, but there are many forms of art and entertainment, and we don't have to consume them all. Note that even this is a judgment of the rational acceptability of such
See Barry Stroud, "Inference, Belief, and Understanding," Mind 88 ( 1979), p. 187: "For every proposition or set of propositions the belief or acceptance of which is involved in someone's believing one proposition on the basis of another there must be something else, not simply a further proposition accepted, that is responsible for the one beliefs being based on the other."
variation--of there being no reason to regret it. The supposition that I might not care about my own future cannot be regarded with similar tolerance: It is the supposition of a real failure--the paradigm of something to be regretted--and my recognition of that failure does not reflect merely the antecedent presence in me of a contingent second-order desire. Rather, it reflects a judgment about what is and what is not relevant to the justification of action against a certain factual background.
Relevance and consistency both get a foothold when we adopt the standpoint of decision, based on the total circumstances, including our own condition. This standpoint introduces a subtle but profound gap between desire and action, into which the free exercise of reason enters. It forces us to the idea of the difference between doing the right thing and doing the wrong thing (here, without any specifically ethical meaning as yet)--given our total situation, including our desires. Once I see myself as the subject of certain desires, as well as the occupant of an objective situation, I still have to decide what to do, and that will include deciding what justificatory weight to give to those desires.
This step back, this opening of a slight space between inclination and decision, is the condition that permits the operation of reason with respect to belief as well as with respect to action, and that poses the demand for generalizable justification. The two kinds of reasoning are in this way parallel. It is only when, instead of simply being pushed along by impressions, memories, impulses, desires, or whatever, one stops to ask "What should I do?" or "What should I believe?" that reasoning becomes possible--and, having become possible, becomes necessary. Having stopped the direct operation of impulse by interposing the possibility of decision, one can get one's beliefs and actions into motion again only by thinking about what, in light of the circumstances, one should do.
The controversial but crucial point, here as everywhere in
the discussion of this subject, is that the standpoint from which one assesses one's choices after this step back is not just first-personal. One is suddenly in the position of judging what one ought to do, against the background of all one's desires and beliefs, in a way that does not merely flow from those desires and beliefs but operates on them--by an assessment that should enable anyone else also to see what is the right thing for you to do against that background.
It is not enough to find some higher order desires that one happens to have, to settle the matter: such desires would have to be placed among the background conditions of decision along with everything else. Rather, even in the case of a purely self-interested choice, one is seeking the right answer. One is trying to decide what, given the inner and outer circumstances, one should do--and that means not just what I should do but what this person should do. The same answer should be given to that question by anyone to whom the data are presented, whether or not he is in your circumstances and shares your desires. That is what gives practical reason its generality.
The objection that has to be answered, here as elsewhere, is that this sense of unconditioned, nonrelative judgment is an illusion--that we cannot, merely by stepping back and taking ourselves as objects of contemplation, find a secure platform from which such judgment is possible. On this view whatever we do, after engaging in such an intellectual ritual, will still inevitably be a manifestation of our individual or social nature, not the deliverance of impersonal reason--for there is no such thing.
But I do not believe that such a conclusion can be established a priori, and there is little reason to believe it could be established empirically. The subjectivist would have to show that all purportedly rational judgments about what people have reason to do are really expressions of rationally unmotivated desires or dispositions of the person making the judgment--desires or dispositions to which normative assessment
has no application. The motivational explanation would have to have the effect of displacing the normative one--showing it to be superficial and deceptive. It would be necessary to make out the case about many actual judgments of this kind and to offer reasons to believe that something similar was true in all cases. Subjectivism involves a positive claim of empirical psychology.
Is it conceivable that such an argument could succeed? In a sense, it would have to be shown that all our supposed practical reasoning is, at the limit, a form of rationalization. But the defender of practical reason has a general response to all psychological claims of this type. Even when some of his actual reasonings are convincingly analyzed away as the expression of merely parochial or personal inclinations, it will in general be reasonable for him to add this new information to the body of his beliefs about himself and then step back once more and ask, "What, in light of all this, do I have reason to do?" It is logically conceivable that the subjectivist's strategy might succeed by exhaustion; the rationalist might become so discouraged at the prospect of being once again undermined in his rational pretensions that he would give up trying to answer the recurrent normative question. But it is far more likely that the question will always be there, continuing to appear significant and to demand an answer. To give up would be nothing but moral laziness.
More important, as a matter of substance I do not think the subjectivist's project can be plausibly carried out. It is not possible to give a debunking psychological explanation of prudential rationality, at any rate. For suppose it is said, plausibly enough, that the disposition to provide for the future has survival value and that its implantation in us is the product of natural selection. As with any other instinct, we still have to decide whether acting on it is a good idea. With some biologically natural dispositions, both motivational and intellectual, there are good reasons to resist or limit their influence. That
this does not seem the right reaction to prudential motives (except insofar as we limit them for moral reasons) shows that they cannot be regarded simply as desires that there is no reason to have. If they were, they wouldn't give us the kind of reasons for action that they clearly do. 3. It will never be reasonable for the rationalist to concede that prudence is just a type of consistency in action that he happens, groundlessly, to care about, and that he would have no reason to care about if he didn't already.
The null hypothesis--that in this unconditional sense there are no reasons--is acceptable only if from the point of view of detached self-observation it is superior to the alternatives; and as elsewhere, I believe it fails that test.
Bernard Williams is a prominent contemporary representative of the opposite view. In chapter 4 of Ethics and the Limits of Philosophy, 4. he argues that reflective practical reason, unlike reflective theoretical reason, always remains first-personal: One is always trying to answer the question "What shall (or should) I do?" and the answer must derive from something internal to what he calls one's "motivational set." Williams says that in theoretical reasoning, by contrast, while it is true that one is trying to decide what to believe, the question "What should I believe?" is in general replaceable by a substantive question which need make no first-person reference: a question like "Did Wagner ever meet Verdi?" or "Is strontium a metal?" This means that the pursuit of freedom through the rational, reflective assessment of the influences on one's be-
For a very persuasive argument that brute desires or preferences in themselves never provide reasons for action, see Warren Quinn, "Putting Rationality in Its Place," in his Morality and Action ( Cambridge University Press, 1993).
Harvard University Press, 1985.
liefs leads, in the theoretical case, to the employment of objective, non-first-personal standards. To decide what to believe, I have to decide, in light of the evidence available to me, and by standards that it would be valid for anyone to use in drawing a conclusion from that evidence, what is probably true.
But Williams holds that in deciding what to do, even if I try to free myself from the blind pressures of my desires and instincts by reflecting on those influences and evaluating their suitability as reasons for action, such reflection will never take me outside of the domain of first-personal thought. Even at my most reflective, it will still be a decision about what I should do and will have to be based on my reflective assessment of my motives and reasons. To believe that at some point I will reach a level of reflection where I can consider truly objective reasons, valid for anyone, that reveal what should be done by this person in these circumstances, is to deceive myself. In the practical domain, there is no such standpoint of assessment. 5.
It has to be admitted that phenomenologically, the subjectivist view is more plausible in ethics than in regard to theoretical reason. When I step back from my practical reasonings and ask whether I can endorse them as correct, it is possible to experience this as a move to a deeper region of myself rather than to a higher universal standpoint. Yet at the same time there seems to be no limit to the possibility of asking whether the first-personal reasoning I rely on in deciding what to do is also objectively acceptable. It always seems appropriate to ask, setting aside that the person in question is oneself, "What ought to happen? What is the right thing to do, in this case?"
That the question can take this form does not follow merely from the fact that it is always possible to step back from
Actually, there is a bit of obscurity in Williams's view on this point, since he may believe there is an objective answer, discoverable by anyone, to the question of what a particular person should do, given the contents of his "motivational set." See the essay Internal and External Reasons in his collection Moral Luck ( Cambridge University Press, 1981), pp. 103-5.
one's present intentions and motives and consider whether one wishes to change them. The fact that the question "What should I do?" is always open, or reopenable, is logically consistent with the answer's always being a first-personal answer. It might be, as Williams believes, that the highest freedom I can hope for is to ascend to higher order desires or values that are still irreducibly my own--values that determine what kind of person I as an individual wish to be--and that all apparently objective answers to the question are really just the first person masquerading as the third. But do values really disappear into thin air when we adopt the external point of view? Since we can reach a descriptive standpoint from which the first person has vanished and from which one regards oneself impersonally, the issue is whether at that point description outruns evaluation. If it does not, if evaluation of some sort keeps pace with it, then we will finally have to evaluate our conduct from a non-first-person standpoint.
Clearly, description can outrun some evaluations. If I don't like shrimp, there simply is no higher order evaluation to be made of this preference. All I can do is to observe that I have it; and no higher order value seems to be involved when it leads me to refrain from ordering a dish containing shrimp or to decline an offer of shrimp when the hors d'oeuvres are passed at a cocktail party. However external a view I may take of the preference, I am not called on either to defend it or to endorse it: I can just accept it. But there are other evaluations, by contrast, that seem at least potentially to be called into question by an external, descriptive view, and the issue is whether those questions always lead us finally to a first-person answer.
Suppose I reflect on my political preferences--my hope that candidate X will not win the next presidential election, for example. What external description of this preference, considered as a psychological state, is consistent with its stability? Can I regard my reasons for holding it simply as facts about myself, as my dislike of shrimp is a fact about myself? Or will
any purely descriptive observation of such facts give rise to a further evaluative question--one that cannot be answered simply by a reaffirmation that this is the kind of person I am?
Here, as elsewhere, I don't think we can hope for a decisive proof that we are asking objective questions and pursuing objective answers. The possibility that we are deceiving ourselves is genuine. But the only way to deal with that possibility is to think about it, and one must think about it by weighing the plausibility of the debunking explanation against the plausibility of the ethical reasoning at which it is aimed. The claim that, at the most objective level, the question of what we should do becomes meaningless has to compete head-to-head with specific claims about what in fact we should do, and their grounds. So in the end, the contest is between the credibility of substantive ethics and the credibility of an external psychological reduction of that activity.
There is a deep philosophical problem about the capacity to step back and evaluate either one's actions or one's beliefs; it is the problem of free will.
Suppose you became convinced that all your choices, decisions, and conclusions were determined by rationally arbitrary features of your psychological makeup or by external manipulation, and then tried to ask yourself what, in the light of this information, you should do or believe. There would really be no way to answer the question, because the arbitrary causal control of which you had become convinced would apply to whatever you said or decided. 6. You could not simultaneously believe this about yourself and try to make a free, rational choice. Not only that, but if the very belief in the causal system of control was itself a product of what you thought to be reasoning, then it too would lose its status as a belief freely
Recall the scrambled-brain hypothesis in chapter 4.
arrived at, and your attitude toward it would have to change. (Though even that is a rational argument, whose conclusion you are no longer in a position to draw!)
Doubt about your own rationality is unstable; it leaves you really with nothing to think. So although the hypothesis of nonrational control seems a contingent possibility, it is no more possible to entertain it with regard to yourself than it is to consider the possibility that you are not thinking. I have never known how to respond to this conundrum.
However, a more specialized version of the problem can be raised about practical reason alone. The hypothesis that practical reason does not exist is not self-contradictory. In spite of everything I have said, one might intelligibly suppose, without having to abandon all one's reasoning, that decisions to act are all ultimately due to arbitrary desires and dispositions--perhaps higher order and partly unconscious--that lie beyond the possibility of rational assessment. Consider the hypothesis that this is true in particular whenever we take ourselves to be engaged in practical reasoning. If someone actually believed this, he could not ask, "In light of all that, what should I do?" To ask that, hoping for a genuinely rational evaluation of the alternatives, would contradict the supposition of nonrational determination, which is supposed to apply to all choices, including this one. So if one really accepted the hypothesis, one would have to abandon the practice of rational assessment, all things considered, as an illusion, and limit the practical employment of reason to an instrumental role.
But is that possible? I don't think so; rather, I think the illusion is on the other side, in trying to see oneself as nonrationally determined. What we have here is a face-off between two attitudes--not, as in the case of subjectivism about theoretical reason, between two theories about how things are. The opposition here is between a theory about how things are and a practice that would be impossible if this was how things
were. If we go on trying to make up our minds about what to do on the basis of the best reasons, we implicitly reject the hypothesis of an ultimately nonrational determination of what we do. (I leave open the possibility that there is a form of causal determination that is compatible with rationality; if so, we could simultaneously engage in practical and theoretical reasoning and believe that we were so determined--including being so determined to believe that we were.)
The unquenchable persistence of the conviction that it is up to me to decide, all things considered, what I should do, is what Kant called the fact of reason. 7. It reveals itself in decision, not in contemplation--in the permanent capacity we have to contemplate all the personal, contingent features of our motivational circumstances and ask, once again, "What should I do?"--and in our persistent attempts to answer the question, even if it is very difficult. The sense of freedom depends on the decision's not being merely from my point of view. It is not just a working out of the implications of my own perspective, but the demand that my actions conform to universally applicable standards that make them potentially part of a harmonious collective system. Thus I find within myself the universal standards that enable me to get outside of myself. (In Kant's example, I am directly aware of the fact of reason when the tyrant threatens to kill me unless I bear false witness against an innocent man: I know that I can refuse--whether or not I will be brave enough to do so--because I know that I ought to refuse.) 8.
There is a direct analogy here with the operation of theoretical reason, which employs universal principles of belief formation to bring my thoughts into harmony with a consis-
Critique of Practical Reason (trans. Lewis White Beck, Bobbs-Merrill, 1956), original in vol. 5 of the Prussian Academy edition of Kant's works, pp. 31, 42.
Critique of Practical Reason, pp. 30, 155-9.
tent system of objective beliefs in which others can also hold a share--more commonly known as the truth. Reason is an attempt to turn myself into a local representative of the truth, and in action of the right. Freedom requires holding oneself in one's hands and choosing a direction in thought or action for the highly contingent and particular individual that one is, from a point of view outside oneself, that one can nevertheless reach from inside oneself.
This picture is opposed to the Humean alternative which limits reason to thought and gives it no direct application to conduct. According to that view, we may transcend ourselves to develop a truer and more objective conception of how the world is, but this transcendence influences our conduct only instrumentally--by revealing how we may most effectively act on our motives, which remain entirely perspectival. Even where an objective view of the facts leads us to pursue practical harmony with others, the motives remain personal.
But I believe that alternative is untenable. Even a moral system like that of Hobbes, based on the rational construction of collective self-interest, affirms the rationality of the selfinterest on which it depends. And that puts it in competition with other conceptions of what is rational.
We cannot evade our freedom. Once we have developed the capacity to recognize our own desires and motives, we are faced with the choice of whether to act as they incline us to act, and in facing that choice we are inevitably faced with an evaluative question. Even if we refuse to think about it, that refusal can itself be evaluated. In this sense I believe Kant was right: The applicability to us of moral concepts is the consequence of our freedom--freedom that comes from the ability to see ourselves objectively, through the new choices which that ability forces on us. 9.
For an illuminating treatment of this subject, see Christine Korsgaard , The Sources of Normativity ( Cambridge University Press, 1996).
Even a 'subjective'-seeming solution to this problem--like the answer that there are no universal standards for determining what we should do, and that each person may follow his own inclinations--is itself an objective, universal claim and therefore a limiting case of a moral position. But that position obviously has competitors, and one or another of the moralities that require some kind of impartial consideration for everyone is much more plausible. Let me now sketch out in a series of rough steps the familiar kinds of substantive practical reasoning that lead to this conclusion and that resist a Humean reduction.
The first step on the path to ethics is the admission of generality in practical judgments. That is actually equivalent to the admission of the existence of reasons, for a reason is something one person can have only if others would also have it if they were in the same circumstances (internal as well as external). In taking an objective view of myself, the first question to answer is whether I have, in this generalizable sense, any reason to do anything, and a negative answer is nearly as implausible as a negative answer to the analogous question of whether I have any reason to believe anything. Neither of those questions--though they are, to begin with, about me--is essentially first-personal, since they are supposed not to depend for their answers on the fact that I am asking them.
It is perhaps less impossible to answer the question about practical reasons in the negative than the question about theoretical reasons. (And by a negative answer, remember, we mean the position that there are no reasons, not merely that I have no reason to believe, or do, anything rather than anything else--the skeptical position, which is also universal in its grounds and implications.) If one ceased to recognize theoretical reasons, having reached a reflective standpoint, it would make no sense to go on having beliefs, though one
might be unable to stop. But perhaps action wouldn't likewise become senseless if one denied the existence of practical reasons: One could still be moved by impulse and habit, without thinking that what one did was justified in any sense--even by one's inclinations--in a way that admitted generalization.
However, this seems a very implausible option. It implies, for example, that none of your desires and aversions, pleasures and sufferings, or your survival or death, give you any generalizable reason to do anything--that all we can do from an objective standpoint is to observe, and perhaps try to predict, what you will do. The application of this view to my own case is outlandish: I can't seriously believe that I have no reason to get out of the way of a truck that is bearing down on me in the street--that my motive is a purely psychological reaction not subject to rational endorsement. Clearly I have a reason, and clearly it is generalizable.
The second step on the path to familiar moral territory is the big one: the choice between agent-relative, essentially egoistic (but still general) reasons and some alternative that admits agent-neutral reasons 10. or in some other way acknowledges that each person has a noninstrumental reason to consider the interests of others. It is possible to understand this choice partly as a choice of the way in which one is going to value oneself and one's own interests. It has strong implications in that regard.
Morality is possible only for beings capable of seeing themselves as one individual among others more or less similar in general respects--capable, in other words, of seeing themselves as others see them. When we recognize that although we occupy only our own point of view and not that of anyone else, there is nothing cosmically unique about it, we are faced with a choice. This choice has to do with the relation
For this terminology, see The View from Nowhere ( Oxford University Press, 1986), pp. 152-3.
between the value we naturally accord to ourselves and our fates from our own point of view, and the attitude we take toward these same things when viewed from the impersonal standpoint that assigns to us no unique status apart from anyone else.
One alternative would be not to "transfer" to the impersonal standpoint in any form those values which concern us from the personal standpoint. That would mean that the impersonal standpoint would remain purely descriptive and our lives and what matters to us as we live them (including the lives of other people we care about) would not be regarded as mattering at all if considered apart from the fact that they are ours, or personally related to us. Each of us, then, would have a system of values centering on his own perspective and would recognize that others were in exactly the same situation.
The other alternative would be to assign to one's life and what goes on in it some form of impersonal as well as purely perspectival value, not dependent on its being one's own. This would then imply that everyone else was also the subject of impersonal value of a similar kind.
The agent-relative position that all of a person's reasons derive from his own interests, desires, and attachments means that I have no reason to care about what happens to other people unless what happens to them matters to me, either directly or instrumentally. This is compatible with the existence of strong derivative reasons for consideration of others--reasons for accepting systems of general rights, and so forth--but it does not include those reasons at the ground level. It also means, of course, that others have no reason to care about what happens to me--again, unless it matters to them in some way, emotionally or instrumentally. All the practical reasons that any of us have, on this theory, depend on what is valuable to us.