Late draft of a paper now published in Weakness
of Will and Practical Irrationality, ed. Christine Tappolet
and Sarah Stroud, OUP 2003 pp. 274-297.
Dante pourquoi dis-tu qu'il n'est pire misère
Qu'un souvenir heureux dans un jour de malheur?
Alfred de MussetPourquoi indeed?
R. de Sousa
Competing Frameworks of Rationality
Weakness of will has sometimes been construed as a moral failing; but at least since Davidson (1980) it has been viewed as a pathology of agency. As such it infringes strategic rationality, which aims at maximizing the likelihood of success of a given course of action. On an alternative, Socratic view, it is an epistemic failing, violating the sort of rationality that aims to maximize the propensity for arriving at truth afforded by a given method of settling on a belief. Strategic and epistemic rationality are distinct. Each provide a familiar framework within which questions of rationality and irrationality are discussed; but where both seem applicable there is no obvious way to decide which of the two is the correct framework of evaluation.
This would pose no special problem, if the two frameworks could never clash. We are used to finding paradoxes and antinomies within each framework of rationality. Jon Elster's Ulysses problem and the Prisoners Dilemma are notorious examples that arise within the framework of strategic rationality.<1> The liar paradox with its innumerable epigones is the classic example in the epistemic framework. These problems have given rise to a large literature. But less often considered is the question of what we should say when a conflict arises between the two standard frameworks. Ultimatum games and Newcomb's paradox are two much discussed classes of examples that involve both issues of strategy and of belief, but they do not involve competition between the two frameworks.<2> If a genuine conflict is possible, the question arises whether one framework can subsume the other, or whether there might be a third, more general point of view in terms of which the conflict might be arbitrated. If, as I shall argue, such a third framework exists, might it not, in turn, yield its own paradoxes and antinomies?
If one framework is to subsume the other, the claim of the strategic may seem immediately preponderant. Settling on a belief may be regarded as a type of action, and we can devise strategies for discovering truth. More pertinently, the notion of correctness or truth itself can fall under the aegis of a pragmatic criterion, as urged by a long tradition in philosophy, running from Protagoras to Richard Rorty through Blaise Pascal and William James, for whom the point of seeking truth, or more radically the very test of truth, lies in the goal of making people fare better. Similarly, when Freud contrasts the "reality principle" with the "pleasure principle", the former is merely a means to satisfying the latter in the long run: truth is ancillary to the search for satisfaction, and its claims must presumably remain subservient to the larger goal.
Conversely, however, as Socrates pointed out to Protagoras in the Theaetetus, an exclusive focus on pragmatic success cannot avoid the epistemic burden of correctly predicting the consequences of a course of action. Epistemic rationality, a champion might claim, subsumes the strategic, in that in all strategic reasoning the real work is done by principles of epistemic inference that happen to be applied to conditional predictions. That premises and results may allude to practice is a mere accident of subject matter, not a reflection of the existence of a different sort of rationality. In every case of practical inference, it must be true that some goal is sought, that some means are appropriate and available, and so forth. Moreover, even if we regard the principles of epistemic rationality as simply designed to foster the beliefs most likely to underwrite successful action, the epistemic need not concede anything to the strategic. Epistemic norms are like those of gastronomy: there would be none, were it not for the imperatives of nourishment; yet nourishment does not wholly determine the norms of gastronomy. Similarly, pragmatic considerations may be thought to underlie our interest in truth; yet practical concerns are insufficient to account for the norms of rational belief.
Each framework, then, can make a case for swallowing up the other. Yet unless just one of these claims is spurious, it seems the two may indeed conflict when a belief is both more likely to lead to good consequences and less likely to be true. In such cases our commitment to the truth and our allegiance to the good part company. Pascal urged us to believe in God even if the probability of his existence is as close to zero as you like (provided it is not actually zero). For if God exists, and our faith earns a Heaven of infinite worth, the expected gain remains infinite even when multiplied by that tiny probability. Conversely, if disbelief incurs eternal damnation, the expected loss is infinite, and outweighs all earthly gains. If Pascal's options were exhaustive, the reasoning would be strategically sound.
But would it be right? Certainly it is epistemically worthless. And one might claim that no transcendent point of view is required, for the epistemic rationality Pascal recommends flouting and the strategic rationality he advocates are not really focused on the same object.<3> Pascal, like James (1979), recommends strategies of self-improvement that explicitly require cultivating self-deception. But the intrinsic aim or point of belief is truth: that is not a normative principle, but a definitional one. So recommending self-deception is inherently irrational, regardless of the desirability of the goal it is intended to serve. By even speaking of belief—if only to subvert it—the Pascalian move implicitly commits itself to the norms of belief and so must recognize the irrationality of flouting them. But what is the force of this 'must'? Surely it is not the 'must' of logical necessity. For if it were, then the views of Pascal and James would be not just irrational, but impossible. If it is the 'must' of epistemic necessity, then the point has already been granted, and is irrelevant to my dilemma as I try to decide, simply, what to believe. Is the 'must' then to be taken as a moral 'must? The claim that it is might be grounded in the observation that we don't just classify self-deception as irrational, but commonly condemn it even when it may have solid strategic justification.<4> The propensity for each side to claim moral high ground is attested by the title of William Clifford's (1866) famous article on the "Ethics of belief," no less than by the tone of James's (1979) rejoinder. The dispute is easy to describe in terms of self-righteous rhetoric: Should one not care more about truth than advantage? says one side (and your practical rationality be damned) . Should one not care about real consequences and not abstract truth? says the other (and your epistemic scruples be damned).
To each of these accusations, I shall argue, the best reply is to ask, quite literally, Why should I care?. Taking that question literally involves recognizing that reasons to care are powerless to move us unless they are grounded in what we actually do care about. The normative claims made on us by either of the two standard frameworks of rationality—the epistemic and the strategic—will work on us only if we care.
Emotions and Norms
I have suggested that neither strategic nor epistemic rationality can arbitrate between Pascal and Clifford, because each is antecedently committed to one framework or the other. If adjudication between them is to be achieved, therefore, it must be in virtue of a third kind of rationality. The hypothesis I shall advance is that there is indeed such a third framework of rationality, governing emotions capable of delivering objective axiological verdicts. Such emotions constitute perceptions of value. Only an emotional attitude towards the two distinct sorts of appropriateness defined by the frameworks of strategic and epistemic rationality can arbitrate between them. In this sense, emotions are capable not only of functioning as arbiters of appropriateness, but more particularly of adjudicating appropriateness of kinds of appropriateness.
Much needs to be done to make this plausible, but my central concern here will not be to build a defense of the thesis that some emotions constitute perceptions of value.<5> Rather, it will be to explore the landscape that might be revealed if we could assume such a defense had been successful. And in that spirit my central claim is a strongly naturalizing one. It is that in the ultimate analysis the facts of emotion are all that can be appealed to by way of justification of the normative claims of rationality.
Emotional irrationality: close but no cigar
The easiest way to grasp the nature of a framework of rationality is to look at examples of cases that fail of it. I shall therefore offer a sampling of the ways in which our attitudes and emotions might be deplored as systematically paradoxical or irrational. (One needs to add: systematically, for we are not concerned with the fact that a rational being can be irrational on this or that particular occasion.) But rather than plunge straight into that list, I first want to offer, for the sake of contrast, some examples that illustrate less interesting forms of disorders. These are cases that may involve emotions, but can be accounted for without going outside the bounds of the standard norms.
My first two cases are inspired by Jon Elster.
(i) Emotions are typically possessed both of motivating power and of intrinsic positive or negative valence. (Elster, 1999 p. 329) Sometimes, states with intrinsic positive valence induce us to act in such a way as to land us in states of negative valence. In a state of euphoria, induced by coffee or other benign drugs, I may find everyone I meet delightful and judge them to be wholly trustworthy. This pleasant condition, however, is one I may live to regret. What's more, I may know it at the time, but repress the thought as mean.
(ii) Conversely, I may prefer the emotion with negative valence, for the sake of some other gain. Elster cites Robert Solomon's nice case of the woman who "continues to patronize a shop which she knows has cheated her [because] her small losses are more than compensated for by the self-righteous satisfaction of her continuing indignation."(quoted from Solomon 1993 by Elster 1999 p. 310)
(iii) If my irascible disposition induces my lover's disaffection, and I am moved to anger by that very disaffection, my anger is clearly counterproductive.
(iv) Rolling stone gathers no moss. Here is an anthropological fact.<6> Ask a European to explicate this proverb, and she will be most likely to interpret the moss as representing culture and accomplishments, and the proverb as recommending that one settle down. For most Americans, on the contrary, the proverb is in praise of roving: being a rolling stone is a good thing, avoiding parasitic growths and encumbrances. For Europeans, it seems, positive emotional valence attaches to the thought of moss; for Americans, the emotional valence of moss is negative.
Cases (i)-(iv) are essentially prudential cases. They illustrate straightforward economic choices based on calculations of gains and losses. In the first two, I am merely being squeezed between desires for outcomes that happen to be incompatible in practice. In the third case, there is nothing paradoxical about my emotion considered in itself. It's just bad luck that my emotional propensities haven't been fine-tuned well enough to avoid such counterproductive displays. As for the fourth, perhaps it merely exploits a disagreement about values. These examples may present a certain paradoxical flavour, but they involve no antinomies, nor do they signal any unavoidable irrationality.
The next example introduces a crucial element of additional complexity, in that it involves second-order attitudes adjudicating potential conflicts between first-order and second-order attitudes. But it is not yet a case of the kind I am looking for.
(v) Deterrence. As Jonathan Schell pointed out, the logic of deterrence
commits us in certain circumstances to do what we must never do in any circumstances.... Deterrence theory is indeed a marvel of circularity and contradiction. To obtain the benefit of the policy, we must threaten to perform an insane action. But the benefit we seek is precisely not to perform that action. We thus seek to avoid performing an act by threatening to perform it... (Schell 1984, p.64)
Each side must reason thus: I must assume you are rational in your willingness to be deterred. Otherwise there would be no point in my threat, since only a rational calculation can lead an agent to be deterred. But inasmuch as I am willing to be deterred, I must also assume you are irrational, indeed "insane" in your willingness to carry out your threat. For since retaliation guarantees an agent's own destruction, retaliating can never be rational. The equilibrium of deterrence is a symmetrical one, with both sides figuring as X in the deterring role and as Y in the role of the one deterred. X must believe Y to be rational as deterred but irrational as deterring. For X must believe he has convinced Y that he, X himself, is insane enough to carry out his threat, yet sane in his ability to be deterred by Y's threat of the same insane behaviour. Robert Frank (1988) presents a consequence of this situation, adverting to the case where, in a personal confrontation, it is rational to give free rein to the most irrational emotions in order to convince my opponent that I am capable of ignoring my own interests for the sake of harming him. Otherwise, he might think me ready to compromise when rational calculation shows that pursuing the case will cost me more than dropping it.
This last example is, I believe, a genuinely paradoxical result. Moreover, in Frank's version and perhaps in Schell's original version too, it clearly involves emotions. For deterrence will work only if it arouses fear, and it will arouse fear only if my opponent is persuaded that my emotions will overcome my concern for my own interests. Yet the paradox arises purely from the structure of a certain game-theoretical situation, and so it is not clear that it involves emotions essentially, in any sense in which emotion might not be simply reducible to preference.
In the cases I am looking for, the paradox or antinomy is even more intimately tied to the nature of the emotion itself. The following cases might do, in that they are not obviously describable in terms of the standard modes of rationality. They share two features: first, in one way or another they involve temporality; second (and partly as a consequence) they represent essentially contestable evaluations and can themselves be adjudicated only in terms of their holistic ties to one another. They will also help to make plain the powerful and yet curiously exposed and arbitrary position of axiological evaluations.
Questionable attitudes and emotional antinomies: a sampling
(1) Remembering happiness in sorrow. My epigraph from Alfred de Musset alludes to Dante, Inferno V:121-122: "Nessun maggior dolore / che ricordarsi del tempo felice / ne la miseria".<7> Musset continues: "Quel chagrin t’a dicté cette parole amère, / Cette offense au malheur?"<8> (Musset 1908, 79). Both attitudes are intelligible and neither seems to rest on a persuasive argument that could topple the other. The two judgments constitute an emotional antinomy, which can't be translated out of the domain of emotion into some more familiar strategic or epistemic terrain. It clearly isn't the case that Dante and Musset have some strategic disagreement about how best to pursue some goal in view. Nor are they disagreeing about truth or about the validity of inference rules.
(2) Does it matter, now, that now won't matter when I'm dead? Thomas Nagel has advanced an argument about the incoherence of the thought that life is absurd because nothing we do now will matter in a million years. It goes like this: if it matters now that our lives will seem insignificant in a million years, then what is a million years away can matter. But if what is a million years away can matter, why can our present lives not matter in a million years? Conversely, if being a million years away makes an event meaningless, then not mattering in a million years can't matter now, and therefore cannot make our present lives absurd. (Nagel, 1979, p. 11)
This argument can form the basis of an anti-Epicurean argument concerning death. Epicurus's well-known argument to show that it is irrational to fear death seems irrefutable: When I am, death isn't; when death is, I am not there to care. What could be plainer? Rather than attempting to refute it, one may well wonder why such an irrefutable argument leaves so many people unmoved. Nagel's argument can be extended to attack it more directly. It would go something like this:
If Epicurus can console me now with the thought that I will feel nothing at future time d, then something matters to me now about a future time. So why shouldn't I be distressed now by that very same thought? Epicurus can't bring up his old argument to show that such an attitude would be irrational, for against the present objection his old argument merely begs the question.
Nagel's original argument is open to an objection from the asymmetry of time. Perhaps the future matters but the past doesn't. So it might be reasonable to mind what the distant future will think or fail to think of us, while it might also be reasonable for the denizens of that distant future to think we don't matter. The present argument about death, however, doesn't suffer from that defect, since both considerations face the same way, towards my future death. Nor is it liable to appraisal in terms of standard norms. The Epicurean argument comes as close as anything to an argument that it is epistemically sound; but, as we've just seen, it begs the question. And from the strategic point of view, to be sure, given the fear of death's negative valence, we might be better off without it. But that doesn't suffice to make it inappropriate. What would suffice? We need a standard capable of adjudicating the claims of emotional attitudes, as well as arbitrating between the types of appropriateness dictated by the epistemic and the strategic standards.
(3) Dessert Last, and related principles. Some people leave the best till last. Foolish, say others, for by the time the last is due you may be dead, or the best have spoiled. My three-year old daughter has offered this version of "Dessert Last". Teachers at day care take shifts, and Teresa, her "favourite teacher" comes either in the morning shift or for the afternoon. She sees her van in the parking lot as we arrive at day care. Often this has caused her to utter cries of delight, but today she is terribly disappointed.
-- Oh, I'm so disappointed, Teresa is here.
-- But aren't you glad that she's here?
-- No, if she's here this morning, she won't be there in the afternoon.
(4) Violations of the Philebus principle:<9> In the Philebus, Plato defends the thesis that there are "false pleasures". From his discussion we may distill the following principle:
[PP] A pleasure of anticipation should be proportional in intensity to the anticipated pleasure which it represents.
Taking great pleasure in the anticipation of something from which one expects none when it comes seems (to Plato and to me, at least) intuitively irrational. The principle does not say anything about what to believe about the future pleasure. It is not, therefore, captured by or reducible to an epistemic principle. Nor, for two reasons, does it seem to be reducible to strategic canons. First, one might think it great good fortune to be able to squeeze some pleasure in advance out of what isn't going to provide any when it actually occurs. That is surely the attitude enjoined by strategic considerations (such as maximize pleasant moments). Second, the principle doesn't say anything about choices, or desires, but only the pleasure or emotion itself. One might offer a justification of the PP on epistemic or on pragmatic grounds. Epistemically, one might note that the sheer quality of a pleasure of anticipation might afford information about the quality of the future pleasure. Strategically, one might appeal to the likelihood that the motivational aspect of disproportionate pleasures of anticipation might disrupt the course of practical planning. But neither of these considerations suffices to account for the PP, which in itself is about the quality and intensity of experience in pleasures of anticipation, not about their informational or motivating functions.
(5) Aspectual mismatch. Aristotle associates happiness with activity rather than with being in a certain state or having completed some achievement. Now a desire, it seems, can envisage a future event under one or another of different temporal aspects, in the sense of that word that corresponds to the grammatical sense of "aspect". Thus desire can focus on a future event under the description of the achievement of a certain goal, or on the activity that leads to its achievement. And in certain cases it seems plausible to claim that the focus is irrational or even mistaken, because the objectively desirable aspect of the event is its enduring quality, not its achievement. My stock examples here are tourism, art and sex: if the hurried tourist cares only about having completed a visit to some beautiful site, her attention is likely to be diverted from the experience of its contemplation; if the impatient lover or concert-goer is focused on the approach of a symphony's last chord, or on orgasm, they may be missing out on the more valuable experience of the music or the sexual caress.
It is intriguing to wonder about the relation of the requirement of aspectual match to the next phenomenon, which might seem to lend comfort to the hurried love's preoccupation with endings.
(6) The Peak-End Rule: Daniel Kahneman has offered evidence to show that one's assessment of a series of episodes is governed not by the total or the average of its hedonic intensity and duration, but largely by the measure of representative moments at the peak and at the end. The paradoxical consequence follows that by judiciously adjusting the pattern of intensity, you can make the subject rate as less aversive a sequence of pain episodes by actually making the pain last longer.
Consider the series 2-5-8 and 2-5-8-4, where the numbers refer to reports of pain provided on a 10-point scale every 5 minutes (10 was worst, 1 was least painful). Although it may seem obvious that the addition of 5 extra minutes of pain can only increase total discomfort, the mean subjective ratings were in fact lower for the longer sequence. The sequence 2-5-8 was rated 64 for total unpleasantness, but 2-5-8-4 was rated only 53. (Kahneman 2000, 696-7) In this case, then, people rated more pain as significantly less unpleasant.
This finding reveals a surprising discounting of the value of duration in the past. I return to it in a moment, after some remarks concerning discounting of the future.
(7) Future discounting and weakness of will. In his fascinating book Picoeconomics, (Ainslie 1992) gives a long list of pathological or irrational behaviours which can, he claims, be explained in terms of our propensity to discount the future at a hyperbolic rate.<10> All involve the propensity to preference reversal which is results from the hyperbolic pattern of discounting. Prospects at unequal temporal distances, like buildings unequally far away, will reverse the order of their apparent sizes as one approaches. Thus when walking towards a small building which stands in front of a large one, you first see clearly that the more distant one is larger. But as you get nearer, while the actual distance from both diminishes in a linear way, the ratio between the distance to the small building and the distance to the large building changes, and as it does so the smaller building comes to occlude the taller. So it goes too, with the future prospects that we value: a preference for a lucid morning tomorrow over a bibulous evening tonight may be clear by the early light of this morning, but as the evening approaches the "temptation" of the now looming bottle occludes the picture of to-morrow's hangover.
Temporal Attitudes
My sample cases all involve comparisons or changes across time, or ways of envisaging temporal events. In addition they all illustrate a certain arbitrariness of our attitudes, particularly where they are subject to change, or where they affect our temporal perspectives. This is not surprising when we recall that the temporal domain is one on which classical utilitarian calculation is generally silent. Recent moral philosophy affords a number of discussions of the morality of increasing the number of sentient beings in order to increase the total amount of hedons in the world;<11> but contains very little about the weight that should be assigned to the duration of experiences. Apart from taking all too much for granted the preferability of long life over short, it seems we tend to ignore that issue altogether.<12>
On the question of rationality through time, self-evident principles are hard to come by. Any reasonably stable emotional dispositions with regard to such evaluation may therefore have a prima-facie claim to being deemed criterial, if only by default. From the biological point of view, one might expect that evaluations of the future and evaluations of the past might obey rather different rules (It seems reasonable, for example, to prefer having suffered some unpleasant episode to being about to suffer it). One might also expect that a number of principles will govern rational attitudes to temporal characteristics of experience. When such principles turn out to be violated in practice—as in the case of the "peak-end" phenomenon which contradicts the plausible idea that a shorter period of suffering will always be ranked as more desirable than a longer one—we might infer that human beings are systematically irrational. Alternatively we might think we have misinterpreted the principles actually at work, and assume that once we identify them correctly no general systematic irrationality will be found. Those are, crudely put, the two sides involved in the "Rationality Debate" pitting Kahneman and Tversky et al. against Gigerenzer et al. (Kahneman, Slovic & Tversky, 1982; Gigerenzer, Todd & ABC Research Group, 1999).
Part of what is at issue in the rationality debate is whether the question itself is an empirical one. From a methodological point of view, therefore, the issue is very difficult to settle, since either side can dispose of many stratagems of conceptual gerrimandering which will neutralize the empirical facts adduced by the other (Stein, 1996). My aim in this paper is to exploit this stand-off, in both a negative and a positive sense. The negative stage consists in pointing out that a stand-off is to be expected, because there is no higher and impartial court before which principles of temporal rationality can be assessed. Inevitably, if we try to justify such standards, we will travel in a circle. In the positive stage, the contention is that, in the final analysis, the facts of emotional attitudes are all that can be appealed to, both by way of justification of the normative claims of rationality, and by way of adjudication between such normative claims when they conflict. I believe that both the negative and the positive stages of this project are essentially part of Hume's legacy, in a way that I now briefly digress to explain.
Hume's Legacy
In relation to epistemic rationality, the view that the normative is rooted in the facts of entrenched practice is far from novel. Still, it is hard to swallow. Surely, we want to say, there is a difference between the factual question of what we do and the normative question of what we ought to do. Yet the justification of basic normative principles seems doomed to circularity. Hume showed this for the justification of induction. Nelson Goodman (1983) and Susan Haack (1993) have extended this respectively to our choice of projectible predicates and to deduction: there too, the question of justification can't be answered without begging the question.
The portion of the Humean argument which I shall attempt to extend to axiology goes something like this.
Take seriously for a moment the choice between induction and anti-induction. It might be suggested that induction doesn't need to be grounded in any deeper principle, because it is self-justifying. But anti-induction is self-justifying too: since inferring that the future will be different from the past has not worked in the past, anti-induction predicts that anti-induction will work next time, just as induction predicts that induction will work next time. So there is no difference there.
It might seem, however, that an asymmetry can be made out at the second-order level, in the following way. When we do expect that anti-induction will work—my gas tank has been non-empty for several hundred miles now, so it's bound to run out soon—we do so on inductive grounds. By contrast, it's not the case that when we are confident that induction will work, our confidence rests on anti-inductive grounds. So is there not here, at least, an irreducible asymmetry? Actually that appearance is illusory. It is akin to the illusion that there is an asymmetry between the predicates blue-green and Goodman's (1983) alternative grue-bleen: while the first pair is projectible and the second is not, that is not because the first enjoys any logical advantage. The only advantage that projectible predicates have is that they are "entrenched", that is, actually projected. (Goodman 1983 p. 94). Similarly our impression that it induction is asymmetrically related to anti-induction may be due to the entrenchment of our intuitions about what constitute reliable inferences from the past to the future. If those very intuitions are what is being called into question, therefore, they cannot be appealed to in their own defense without begging the question.
Hume's essential message is Don't ask why we do it, ask what we do. The asymmetry just discussed supports the answer: What we do is make inductions. That doesn't in itself support the claim that making inductions is what we should be doing: that is precisely Hume's "problem" of induction. But Hume's answer to the problem is that our practices are all the justification we will ever get. When extending the problem to the "new riddle of induction", this is a point which Nelson Goodman makes very explicit: "Principles of deductive inference are justified by their conformity to accepted deductive practice" (Goodman, 1983, p. 63). We may invoke increasingly deep and general principles, but in the end the same holds for the ultimate principles of rationality, no less than for our knowledge of the ultimate causes acting in the physical universe: "The most perfect philosophy of the natural kind only staves off our ignorance a little longer.'' (Hume 1975, 19)
Applying this to the principles of rationality may cast light on the current "rationality debate". To see how, consider this passage from Richard Thaler:
[(Rubinstein, 1982, p. 97)] explicitly distinguishes [the question of what will happen ... if both parties behave rationally] from two others, namely '(i) the positive question—what is the agreement reached in practice; (ii) the normative question—what is the just agreement'. (Thaler, 1992, p. 21)
By "the normative question" Rubinstein and Thaler apparently mean the moral question. Let us leave that aside for the moment. In a sufficiently strong sense, the first question is already a normative one, since "what would happen if the parties behave rationally" presupposes that there is an answer to what the parties should do (according to the canons of rationality). But while on a particular occasion it may seem obvious that this question is to be answered differently than "the positive question", it is far from clear that the questions can be kept apart all the way down.
Let me try to make this plausible for the some of the cases sketched above. As we shall see, some can be accounted for without appealing to the sort of paradoxical emotions I have in mind; but others, as far as I can see, cannot. (Those I leave out from further discussion are left as a challenge to the reader.)
Extending the Humean moral to sui-generis emotional rationality
In the dessert last cases, one might find cultural as well as individual variation in the attitude that people take to ordering one's future pleasures. This blunts the paradoxical force of the competing attitudes. A similar cultural dependency is suggested by the contrast of interpretations of Rolling stone gathers no moss. In addition, the discrepancy noted between American and European interpretations of the proverb doesn't amount to an emotional antinomy, because the two evaluations are based on different mappings of the elements of the proverb. For the European, "moss" symbolizes culture and "rolling stones" are referred to persons who never focus on a stable project; for the American, "moss" symbolizes undesirable attachments and rolling stones are those who keep up the search for fresh experience.
The conflicting intuitions of Dante and Musset, on the other hand, might well pertain to the very same memory of happiness in times of sorrow. A dissenter might argue that the difference can be traced to the fact that each attends to a different aspect of the situation. Dante focuses on the present suffering afforded by the contrast of past happiness, whereas Musset attends to the past happiness itself. If each were to switch his focus of attention, he might concede the other's point. Perhaps: but then each might insist that the other's focus of attention was somehow perverse, inappropriate to the situation as a whole. And in the face of that persisting disagreement, what is called for is again some criterion capable of arbitrating not merely between two attitudes associated with different points of view, but also between the two standards of appropriateness favouring the contrasting points of view.
Much the same seems to be true of attitudes to death. We must, I argued, concede that the Epicurean consolation is question-begging. If so, to prefer it over the despair so well expressed in Philip Larkin's Aubade<13> is also question-begging in assuming that a strategic value placed on comfort should outweigh the value of an adequate emotional response. But what is an "adequate emotional response" to death? While there's death, there's hope, my friend Malcolm Deas is fond of saying. To me, that sometimes seems a very touchstone of rationality.
In my own previous discussions of the Philebus and Aspectual Matching principles, I have argued that both are irreducible to standard norms: both represent sui generis principles of rationality. In the context of the present argument, I would further stress that while one can find things to say in their favour, we are unlikely to find any knock-down proof of their validity. In fact, one will not find any stronger justification for them than an appeal to one's reasoned feelings about what kind of life is better. In the light of my remarks on Hume's legacy, that places them in just the same position as the standard norms themselves.
Is this claim excessively pusillanimous? To be sure, some have been more sanguine about the prospects for appraising emotional rationality. We can find stronger claims, explicitly or implicitly, both in Kahneman's discussion of "evaluation by moments" and in Ainslie's discussion of temporal discounting.
Kahneman elaborates as follows on the "heuristic of moments":
[A]s a good first approximation, the affective value of the representative moment is a simple average of the most extreme affect experienced during the episode (Peak) and of the affect experienced near its end. . . . The affective value of that representative moment, in turn, determines the global evaluation of the entire episode (Kahneman 2000, p. 697)
He concludes that this heuristic of evaluation by moments "leads to violations of logic, because the temporal dimension of experience is not directly included in the representations that are evaluated." (p. 707).
But where, actually, is the violation of logic? I see only a discrepancy between points of view: when you propose to inflict on me a continuous series of painful episodes, I might well prefer 2-5-8 to 2-5-8-4. That choice isn't contradicted by the fact that I may reverse my ranking of their desirability in retrospect. Since the temporal perspectives are as different as the future is from the past, there are no stable grounds even for accusing people of inconsistency when they insist that they would always prefer a shorter episode of pain to a longer one, and later choose the longer. But that may just show that they evaluate the past and the future differently: and why should they not?
Temporal reversals brought about by hyperbolic discounting often look very much like standard cases of weakness of will. Yet as described, such a reversal doesn't necessarily constitute or involve akrasia. The reason is this. On the usual view, whether an act (say, of drinking) involves akrasia depends on whether I continue, even when reaching for the bottle, to affirm sincerely that, all considered, I prefer not to drink now. But if I deny this and declare instead that I have changed my preference, and experience no concurrent conflict, I cannot automatically be convicted of akrasia. This observation may leave us uneasy, for we might like to leave room for the possibility that my sincerity in such a case simply indicates that I have successfully masked my akratic conflict with self-deception. The problem here is that what I have called "the usual view" assumes that an agent always has access to her own real preferences. But that assumption is surely gratuitous, for self-deception may indeed cover my tracks in the way just suggested. If we give up the assumption, then some cases of akrasia might be held to be only apparently irrational, on the ground that the agent's articulated over-all preferences don't really reflect her real preferences. The point has been forcefully made by Alison McIntyre (1991), who points out that if akrasia is defined as a conflict between one's avowed best reasons and the reason on the basis of which one acts, akrasia is not necessarily irrational, for one may be mistaken in one's belief that one's avowed best reasons reflect one's deepest commitments.
McIntyre applies this to the well-known Huck Finn example first discussed by Jonathan Bennett (1974). McIntyre claims that Bennett sees Huck as irrational because he characterizes the latter's dilemma "as one in which general moral principles and reasons conflict with "unreasoned emotional pulls"." (p. 381, quoting Bennett p. 127). Instead she suggests that in at least some similar cases one is not irrational in being akratic, because there is no inconsistency between one's actual values and one's act, even though there might be an inconsistency between what one believes to be one's values and one's acts: "agents might not see what reasons they have for acting in a certain way. If one accepts this, then .... the practical judgments that she arrives at will express what she believes that she has most reason to do, but they might fail to express what she actually has reason to do or what it would be most rational for her to do." (p. 386)
McIntyre's thesis isn't merely that some objective reasons might exist to justify the akratic action. It is rather that the so-called akratic action might be better from the point of view of the authentic preference ranking of the agent herself. Rationality in action, she contends, is "evaluative consistency", and evaluative consistency might exist, for example, in view of the fact that if the agent had had more time to reflect, she would have changed her explicit opinion about what would be best to do. Thus she would have been saved from akrasia not by changing her behaviour but by changing her evaluation of it.
But what is the ground on which one might ascribe to an agent a system of values different from the one avowed by the agent herself? McIntyre's point evinces a certain kind of epistemic opportunism in the defense of an akratic agent: luckily for the agent, she has mistaken her own preferences. This is opportunism, because it wouldn't occur to us to make such an ascription, prying apart avowed from true preferences, unless one were antecedently convinced (as in the case of Huck Finn) that the agent had done the right thing. For what other motive, in the face of the agent's explicit avowal, could there be for thus rehabilitating the akratic act? Such a reassessment is itself not without cost, since it still entails ascribing some form of irrationality, namely the inability to bring one's beliefs about one's own preferences into line with those actual preferences themselves.
The point can be clarified by considering the converse case: suppose you are always enkratic: you do always (and infuriatingly) just what your avowed preferences warrant. Armed with McIntyre's insight, I can impugn your every act. The very smoothness of that surface betrays you: no one could have such constant harmony between their values and their every choice! Your whole life is inauthentic! Why not? Well, because the issue of akrasia only arises if there is a disturbing factor, an apparent discrepancy. It arises, as in the case of Huck Finn, because there is an issue about the morality, that is about the overall value of his act. But that involves an axiological judgment, typically embodied in an emotional reaction.
A similar message of rehabilitation is conveyed by Luc Bovens (1999), elaborating on a charming "cure for akrasia" proposed by Roy Sorensen (1997). Sorensen's idea was this. First, send me $1000 in trust. Whenever you are about to commit an akratic act, let me know and I will refund your $1000. The act will then be rescued from akrasia, providing the value of the act plus $1000 is greater than the value of the alternative act. (If it is likely not to, then you had better invest more at the first stage.) The moral that Bovens draws is that we seem to have a second-order attitude which favours resistance to weak will, inasmuch as we tend to seek, whenever we feel conflicted, ways of struggling against "temptation". But why not instead take "temptations" as indications of our true natural preferences, and use them to transform our second-level preferences? Is this not, Bovens asks, just a sort of puritanical prejudice? In assessing what is and what isn't akratic, we need to assess the contrary values to which the agent is apparently committed by the act on the one hand and its repudiation on the other. What are at stake here are not just opposed desires, but attitudes. And while the criteria for contrariety of desires are already somewhat murky (cf. de Sousa 1974), the criteria for contrariety of attitudes or emotions are even more obscure. So much emerged from the antinomies that arise from the divergent attitudes about death, past happiness, postponement of satisfaction, and future-discounting.
These examples have been offered in the hope of making clear what exactly might be meant by the claim that emotion are sometimes paradoxical in a sui-generis sense. Just as the epistemic and strategic can move to swallow one another up, however, one might attempt to show that the axiological can also be swallowed up by one or the other of the standard norms. Fully to deploy the reasons for rejecting this suggestion would take us too far afield. But before turning to the somewhat pessimistic considerations that will form my conclusion, I want rather dogmatically to sketch the main reasons for thinking that the attempt to reduce the axiological either to strategic or to epistemic rationality must fail.
For anyone who sympathises with the view that emotions have a biological aspect and arise from functional capacities, there is an obvious temptation to think an emotion justified just in case it is practically useful. Frank's conception of deterrence cited above illustrates this nicely: the very irrationality of emotions of rage and indignation is the source of their utility as deterrents, and we might therefore say that precisely by virtue of their irrationality they are strategically rational. But the reason this is not persuasive as a reduction is that it is simply irrelevant to the question of their axiological appropriateness. In this respect appropriateness is just like truth: sometimes it may be pragmatically counterproductive, but that only establishes that it is strategically irrational, not that such strategic irrationality suffices to convict it of inappropriateness in its own terms.
Precisely because of this parallel with truth, the claim of the epistemic might seem stronger. If emotions are indeed (as I have assumed but not here attempted to argue) perceptions of value (Tappolet 2000), are they not simply true or false, as other perceptions are, according to whether or not they represent those values correctly? In one sense, this is unobjectionable. But it is also trivial, in the sense that while it can be agreed that emotions are true if they correctly apprehend evaluative facts, this says nothing about whether the "facts" in question are significantly different from other sorts of "facts". In my own ongoing attempt to make sense of emotional truth, I have argued that the sort of truth applicable to emotions—the sort of "facts" to which they relate—is indeed significantly different from the sort of truth typically ascribed to propositional or "factual" beliefs (de Sousa 2002). The main reason is that each emotion provides its own conditions of appropriateness, or "formal object", in terms of which its success or failure must be assessed. This is, in all cases, different from the criterion of truth as semantic satisfaction which constitutes the formal object, or condition of epistemic success, for ordinary propositional beliefs. In this more interesting sense, then, in which the notion of emotional truth is not trivial, it is also quite distinct from the notion of truth ordinarily dealt with in epistemology. The present essay is intended to exemplify more concretely some of the ways emotions can give rise to paradoxes and antinomies that are not reducible to those which traditional epistemology is accustomed to dealing with.
Paradoxical Emotions: Some biological sources of disharmony
In the past I've suggested three candidates for what I called "basic tragedies of life", which I defined as involving "a necessary condition of a fundamental good, where that condition itself conflicts directly with the enjoyment or the perpetuation of that good." (de Sousa 1987, 328-331) The three are death, solidarity, and biography. Death brings life, and therefore meaningful life, to an end, yet death is the very condition of the meaningfulness of life. Our inescapable solidarity with the social context in which we live is what makes it possible for each person to develop genuine individuality; yet it also makes possible and perhaps even necessary the existence of conflict between society and the individual. Finally, the values we espouse typically aspire to universality and objectivity; yet the psychological possibility of their acquisition depends on the specific details of our upbringing. The necessary force of the categorical imperative, if you subscribe to it, is but a contingent consequence of your biography.
In all three cases, my claim was that the status of the paradox was ontological, in the sense of arising not from some internal characteristic of the emotional disposition, nor from some remediable malfunction or maladaptation of the disposition to its context, but from some deeper necessary fact about the human situation.
That claim might well be overblown. I might now prefer to reconstrue this alleged metaphysical necessity in terms of constraints pertaining to our psychological dispositions and their biological origins. For example, the fact of death constitutes a "tragedy" in my sense only on the assumption that one has a certain attitude to it. Suppose I espouse a purely Epicurean attitude: then I effectively cease to care one way or another about death. ceases to give meaning to life, and it ceases to take it away. I might then claim that fear of death is a kind of perversion: a sentiment which is perhaps explicable in terms of certain mechanisms, but not one for which any supportable rationale can be constructed. (de Sousa 2003). Conceding that fear of death exists regardless, one might explain it in terms of its stimulating effect on organisms threatened with perishing before they had a chance to reproduce. Manifestly such an explanation would fall short of a justification; yet it might be sufficient to block an objectivist claim about the uniquely appropriate attitude to death.
The alleged tragedies of solidarity and of biography could be dismissed in a similar spirit. The very whiff of paradox that attends my attitude to society, or my attitude to the objectivity and to the origin of my moral values, should be sufficient to motivate me to stop worrying about either. Why not espouse wholeheartedly my embeddedness in society? (when communitarians urge me to do this, is not the frisson of repulsion this injunction arouses in me merely a manifestation of my own irrationality?) And as for the objectivity of my values, the attitudinally correct stance here is surely that of irony in (Rorty 1979)'s sense: These are my values, and I stand by them—but I might be wrong, and what's more I might be shown to be wrong, in a way that might carry conviction not merely with others but with myself as well.
These observations have the apparent effect of casting doubt on the suggestion that we have any access to objective standards of emoting and therefore of apprehending value through emotion. This skepticism might seem vindicated in the light of a number of biological factors that seem calculated to promote disharmony in our emotional life.
First, organ design, including the design of brain functions, is constrained by the nature of adaptation by natural selection, which of necessity is a kind of "tinkering".<14> Our emotions have been cobbled together at different times in response to different selective pressures. As a result, our most basic emotional capacities are very likely to be relatively independent modules, often driven by unrelated biological needs (Cosmides & Tooby, 2000; Ledoux 2000; Panksepp 1998). There is no reason to think that they will work harmoniously together, any more than we can hope that the need to flee an enemy will never interfere with a peaceful digestion.
Second, most biologists have tended to agree that group selection can be assigned only a minimal role in the explanation of our emotional dispositions. This entails that selfish motives will frequently conflict with group interests (defined either as the sum of the interests of other members of the group, or in some other way that might appeal to emergent group interests). That, in turn, entails that the dispositions fostered by biology are likely to conflict with those nurtured (historically if not genetically) by the constraints of group interaction.
Third, constraints present in the environment of evolutionary adaptation do not necessarily correspond to those that affect contemporary life. Vestigial emotions, analogues of our vestigial appendix, might conflict with more currently functional ones.
Fourth, some of our dispositions may have their origins in the peculiar features of "Baldwinian selection". This has recently given rise to a variety of interpretations.<15> The Baldwin effect was originally conceived as a way to explain certain Lamarckian appearances in strictly Darwinian terms. In its most general form, it consists in a feedback loop that begins with non-instinctive behaviour, bringing about a change in environment, which change, in turn, leads to new selective pressures that favour genetic dispositions for certain types of behaviour. When the behaviour at the end of this loop is of the same type as the behaviour at the beginning, we may have something that looks rather like an episode of Lamarckian evolution, since it will actually be the case that a particular (group of) organisms’ choosing a certain mode of behaviour can ultimately lead to a predisposition for that type of behaviour to be coded in the genes. Sexual selection, which involves a positive feedback loop in which the hypertrophy of a certain trait results from the (random or functional) mate's preference for the trait, can be seen as a special case of Baldwinian evolution. It favours the survival value of the trait, by favouring the predisposition to choose it on the part of the mate. In its full generality, the Baldwin effect merely records the fact that behaviour can influence genes by affecting the environment in which selection takes place. The indirect character of the influence will usually guarantee that the results will be entirely unpredictable, so that what the Baldwin effect actually does is take the evolution of behaving organisms on a novel course which may for a time look as if it has a direction. Notoriously, that "direction" may be maladaptive: sexual selection need not result in any trait that is in the ordinary sense adaptive, and that is generally true of other forms of Baldwinian evolution. For my purposes here, the interest of the Baldwin effect lies in the fact that at the level of species evolution it presents a particularly vivid example of the extreme contingency and unpredictability of biological change. Yet from the point of view of a given organism in specific circumstances, the constraints stemming from Baldwinian selection are just as rigid as any other strictly compulsory hereditary trait or any ineluctable environmental constraint The result is that Baldwinian evolution could have saddled us with disparate emotional propensities unlikely to be easily harmonized. These could easily exacerbate the arbitrariness and disparity between our different emotional predispositions.
Pessimistic Concluding Remarks
The puzzling cases I have adduced as examples of putative axiological irrationality share three features. First, they all concern emotions and attitudes, rather than simple preferences or desires. Second, they are specifically concerned with evaluating experience in and through time. Third, they do not appear to be susceptible of being accounted for in terms of a recognizably coherent authoritative point of view. On the contrary, as I have illustrated in the last section, our emotional capabilities constitute an anarchic, disparate, and potentially conflicted amalgam of dispositions. Despite the obvious drawbacks of this situation, I have contended that the search for equilibrium—or more modestly accommodation—between the components of this anarchic amalgam is the only game in town. The Humean perspective I have argued for implies that no privileged pragmatic or epistemic vantage point can claim to stand in judgment on the attitudes in question, any more than they are able to arbitrate between themselves. Despite the incoherence apparently built into our emotional capabilities, only they can constitute the ultimate arbiters of our axiological judgments. Emotions remain the last court of appeal in the judgment of the appropriateness of other emotions. At the same time, they are also the last court of appeal in the conflicts between different modes of rationality, strategic and epistemic. The paradoxes and antinomies that characterize our emotions must therefore infect the whole field of our "ultimate" values.
The view presented here has an important (though hardly novel) political
consequence. This is that an education into rationality—and equally,
and for similar reasons, an education into the capacity to lead a moral
life—must rest, as Aristotle well knew, on an education of the
emotions. Needless to say, this view of rationality and morality will condone
shortcuts, much in the way that Mill points out that one can hardly be
expected to calculate the pertinent entry of the Nautical Almanac at every
turn of the tiller. But despite the intrinsic incoherence of our fundamental
emotional attitudes, the best hope of emotional rationality lies in the
broadest possible assessment of our emotions, by our emotions themselves,
in the light of a comprehensive educated emotional range. The heart of
both rationality and morality lies in a holistic assessment of one's emotional
dispositions, constantly tested against one another. It therefore becomes
crucial that those dispositions be rooted in a comprehensive a set of experience
as possible. The word partiality incorporates an inspired pun; for
partiality,
the failure of impartiality, has its roots in the inevitably partial
(that is, piecemeal and incomplete) character of emotional experience.
However desirable the ideal of impartiality or comprehensiveness, the biological
considerations I have raised make it unlikely that such an ideal can ever
be realised. That is no reason not to pursue it.
ENDNOTES [hit the BACK button to return to text]
Ainslie, George. 1992. Picoeconomics: The Strategic Interaction of
Successive Motivational States Within the Person. Cambridge: Cambridge
University Press.
Baldwin, James Mark. 1896. "Consciousness and Evolution."
Psychological
Review 3:300-308.
Bennett, Jonathan. 1974. "The Conscience of Huckleberry Finn."
Philosophy
49:123-34.
Bovens, Luc. 1999. "The Two Faces of Akratics Anonymous." Analysis
59: 230-36.
Clifford, William And. 1866. "The Ethics of Belief." In Lectures
and Essays. London: Macmillan.
Cosmides, Leda, and John Tooby. 2000. "Evolutionary Psychology and
the Emotions." In Handbook of Emotions, Michael Lewis and Jeannette
M. Haviland-Jones, 91-115. New York: Guilford Press.
Davidson, Donald. 1980. "How is Weakness of the Will Possible?" In
Essays
on Actions and Events, 21-43. Oxford: Oxford University Press.
Deacon, Terrence W. 1997. The Symbolic Species: The Coevolution
of Language and the Brain. New York: Norton.
Dennett, Daniel C. 1991. Consciousness Explained. Boston, Toronto,
London: Little, Brown.
de Sousa, Ronald. 1974. "The Good and the True." Mind 83: 534-51.
de Sousa, Ronald. 1987. The Rationality of Emotion. Cambridge,
MA: MIT Press.
de Sousa, Ronald. 1997. "What Can’t we Do with Economics?
Reflections on Ainslie’s Picoeconomics." Journal of Philosophical
Research 22: 197-209.
de Sousa, Ronald. 2000. "Deux Maximes de Rationalité Émotive."
In
Emotion Und Vernunft; Émotion et Rationalité, Sudia
Philosophical vol 59, edited by Emil Angehrn and Bernard Baertschi, 15-32.
Bern; Stuttgart; Wien: Paul Haupt.
de Sousa, Ronald. 2002. "Emotional Truth". Aristotelian Society
Proceedings Suppl. Vol. 76:247-263.
de Sousa, Ronald. 2003. "Perversion and Death". The Monist,
forthcoming,
January.
Elster, Jon. 1979. Ulysses and the Sirens: Studies in Rationality
and Irrationality. Cambridge: Cambridge University Press, and Paris:
Editions de la Maison des Sciences de l'Homme.
Elster, Jon. 1999. Alchemies of the Mind: Rationality and the Emotions.
Cambridge, New York, Melbourne: Cambridge University Press.
Frank, Robert H. 1988. Passion Within Reason. New York: Norton.
Gigerenzer, Gerd, Peter Todd, and ABC Research Group. 1999. Simple
Heuristics That Make Us Smart. New York: Oxford University Press.
Goodman, Nelson 1983. Fact, Fiction, and Forecast. 4th ed. Cambridge,
MA: Harvard.
Jacob, Francois. 1976. The Logic of Life: a History of Heredity.
Betty Spillman, trans. New York: Vintage.
Haack, Susan. 1993. "The Justification of Deduction." Reprinted from
Mind, 85 (1976) in A Philosophical Companion to First-Order Logic,
compiled and edited by R.I.G. Hughes. 1976, 76-84. Indianapolis: Hackett.
Hume, David. 1975 <1977>. Enquiry Concerning Human Understanding;
A Letter from a Gentleman to His Friend in Edinburgh. Edited by Eric
Steinberg. Indianapolis: Hackett.
James, William. 1979 <1896>. "The Will to Believe." Edited by Frederick
H. Burkhardt. In The Will to Believe and Other Essay in Popular Philosophy.
Cambridge, MA: Harvard University Press.
Kahneman, Daniel, Paul Slovic, and Amos Tversky, eds. 1982. Judgment
Under Uncertainty: Heuristics and Biases. D. Kahneman, P. Slovic, and
A. Tversky. Cambridge and New York: Cambridge University Press.
Kahneman, Daniel. 2000. "Evaluation by Moments: Past and Future." In
Choices,
Values, and Frames, edited by Daniel Kahneman and Amos Tversky, 693-708.
Cambridge New York: Cambridge University Press Russell Sage Foundation.
Ledoux, Joseph. 2000. "Emotional Circuits in the Brain." Annual
Review of Neuroscience 23:155-84.
Lockard, Joan and Paulhus, Delroy, eds. 1988. Self-Deception: An
Adaptive Mechanism?, Englewood Cliffs: Prentice Hall.
McIntyre, Alison. 1990. "Is Akratic Action Always Irrational?" In
Identity,
Character, and Morality, eds Amelie Rorty and Owen Flanagan, 379-400.
Cambridge, MA: MIT Press.
Murphy, Dominic, and Stephen Stich. 2000. "Darwin in the Madhouse:
Evolutionary Psychology and the Classification of Mental Disorders." In
Evolution
and the Human Mind: Modularity, Language and Meta-Cognition, edited
by Peter Carruthers and Andrew Chamberlain, 62-82. Cambridge and New York:
Cambridge University Press.
Musset, Alfred de. 1908. Les Chef-d’Oeuvres Lyriques d’Alfred
de Musset. Choix et notice de Auguste Dorchain. Paris: Perche.
Nagel, Thomas. 1979. Mortal Questions. Cambridge: Cambridge
University Press.
Nozick, Robert. 1986. "Newcomb's problem and two principles of choice".
In Essays in honor of Carl G. Hempel. Dordrecht: Reidel.
Panksepp, Jaak. 1998. Affective Neuroscience: The Foundations of
Human and Animal Emotions. New York; Oxford: Oxford University Press.
Parfit, Derek. 1984. Reasons and Persons. Oxford: Oxford University
Press, Clarendon.
Richards, Robert J. 1987. Darwin and the Emergence of Evolutionary
Theories of Mind and Behavior. Science and Its Conceptual Foundations.
Chicago: University of Chicago Press.
Rorty, Richard. 1979. Contingency, irony, and solidarity. Cambridge:
Cambridge University Press.
Rubinstein, Ariel. 1982. "Perfect Equilibrium in a Bargaining Model."
Econometrica
50:97-109.
Schell, Jonathan. 1984. The Abolition. New York: Knopf.
Singer, Peter, 1993. Practical Ethics. Second Edition. Cambridge:
Cambridge University Press.
Sorensen, Roy. 1997. "Advertisement for a Cure for Incontinence."
Mind
106:743-43.
Stein, Edward. 1996. Without Good Reason: The Rationality Debate
in Philosophy and Cognitive Science. Oxford: Clarendon Press.
Tappolet, Christine. 2000. Emotions et Valeurs. Paris: Presses
Universitaires de France.
Thaler, Richard H. 1992. The Winner’s Curse: Paradoxes
and Anomalies of Economic Life. New York: The Free Press.