3

Interpretation

Interpretation and explanation

In many writings on the humanities, the focus has been on interpretation rather than explanation. In the German tradition, a contrast was often drawn between the “spiritual sciences” (Geisteswissenschaften) and the natural sciences (Naturwissenschaften). In the former, we are told, the proper procedure is that of interpretation or “understanding” (Verstehen). For the latter the appropriate language is that of explanation (Erklären). Max Weber wrote, for instance, that natural science does not aim at “understanding” the behavior of cells.

We may then ask whether the social sciences rely on understanding or on explanation. I believe this question is wrongly put. In my view, to interpret is to explain. Interpretation is nothing but a special case of the hypothetico-deductive method (Chapter 1). Scholars in the humanities cannot, for instance, use “empathy” as a privileged shortcut to the interpretation of behavior, since one scholar's empathetic understanding may differ from that of another. To decide among conflicting interpretations they have to confront these interpretive hunches or hypotheses (for that is what they are) with experience. As I argued in Chapter 1, experience includes not only the facts we are trying to understand, but also novel facts that we might not otherwise have thought about investigating.¹

Interpretation is directed to human actions and to the product of human actions, such as works of art. In Chapter 16 I address the issue of interpretation of literary works, more specifically works in which we need to understand the actions of the characters as well as the choices of the author. In trying to understand other literary works, as well as the “wordless arts” of painting, sculpture, or instrumental music, this two-tier issue does not arise. Yet in these art forms too, the choices of the artist lend themselves, in principle, to much of the same analysis as that which I shall propose for authorial decisions. The artists make choices according to some criterion of “betterness” that neither they nor we may be able to formulate explicitly, but that is revealed in practice when they discard one draft, one sketch, or one recording in favor of another. Yet the relation between the criterion of betterness and human psychology is more complicated and less well understood in the wordless arts than in (classical) fiction. I shall not attempt to deal with them.

Rationality and intelligibility

The remainder of this chapter, therefore, will be directed to the interpretation of action. Interpreting an action requires us to explain it in terms of the antecedent beliefs and desires (motivations) of the agent. Moreover, we should explain these mental states themselves in a way that makes sense of them, by locating them within the full desire-belief complex. An isolated desire or belief that does not have the normal kind of solidarity with other mental states is just a brute fact that may allow us to explain behavior but not to understand it.

A paradigm mode of explaining action is to demonstrate that it was performed because it was rational (Chapter 13). To do so, it is not enough to show that it had good consequences for the agent: it must be understood as optimal from the agent's point of view. Trying to explain the choice by its beneficial consequences is a form of “rational-choice functionalism” – combining the two approaches I warned against in the Introduction – that sheds no light on the meaning of the behavior. It is a fact, for instance, that if people attach high value to future consequences of present behavior, that is, have a low rate of time discounting (Chapter 6), their lives go better.² It also seems that higher education shapes time preferences in that direction.³ These two premises do not, however, amount to a rational-choice explanation of why people decide to become educated. For an explanation to get off the ground one would have to show that people have the requisite beliefs about the impact of education on the ability to delay gratification, and that they are subjectively motivated to acquire that ability.⁴

If behavior is rational, it is ipso facto also intelligible (but see Chapter 16 for an exception). Irrational behavior can also, however, be intelligible. I shall distinguish among three varieties of intelligible but irrational behavior and contrast them with some cases of unintelligible behavior.

The first arises when the machinery of decision making (see Figure 13.1) is truncated in one way or another. By virtue of its peculiar urgency, a strong emotion may prevent the agent from “looking around” (i.e. gathering information) before acting. Rather than adopting a waiting strategy similar to that of the Roman general Fabius the Cunctator (hesitator), the agent rushes into action without taking the time to consider all the consequences. Another form of truncation arises in weakness of will, traditionally understood as acting against one's own better judgment (Chapter 6). The person who has decided to quit smoking yet accepts the offer of a cigarette acts on a reason, namely, a desire to smoke. For an action to be rational, however, it has to be optimal in light of the totality of reasons, not just one of them. I shall have occasion, however, to question this understanding of weakness of will.

A second variety arises in the short-circuiting of the machinery of decision that occurs when belief formation is biased by the agent's desires. Wishful thinking, for instance, is irrational, but fully intelligible. A subtler form of motivated belief formation arises when the agent stops gathering information when the evidence gathered so far supports the belief he would like to be true.⁵ These forms of motivated belief formation are, in their way, optimizing processes: they maximize the pleasure the agent derives from his beliefs about the world rather than the pleasure he can expect from his encounters with the world.

A third variety is what we might call a wire-crossing in the machinery of decision. We can easily understand why the mind might engage in cognitive dissonance reduction (of which wishful thinking is one variety), but why should it also pursue dissonance production? The idea, cited in Chapter 2, that we believe easily what we fear is an example. Why would fear of a bad outcome make us see it as more likely than is warranted by our evidence? If the belief is supported neither by the evidence nor by our desires, why adopt it? Clearly, nothing is being optimized. In one sense such behavior is harder to understand than actions arising from truncation and short-circuiting, since there is nothing in it for the agent, no partial or short-term goal that it satisfies. Nevertheless it is intelligible (as I understand that idea) because it arises from the belief–desire system of the agent.

Actions that elude interpretation include those caused by compulsions and obsessions, phobic behavior, self-mutilations, anorexia, and the like. To be sure, such behavior has the effect, which explains why it is performed, of relieving the anxiety the agent feels if she does not perform it. Yet washing one's hands fifty times a day or walking up fifty flights of stairs to avoid taking the elevator is not like using a tranquilizer. Taking Valium may be as rational and intelligible as taking aspirin, but compulsive and phobic behavior is unintelligible because it is not part of an interconnected system of beliefs and desires. Or to take an example from John Rawls, we would find it hard to understand the behavior of someone who devoted his time to counting blades of grass unless it was linked to some other goal, such as winning a bet.

Wishful thinking is intelligible, as is counterwishful thinking. The belief of a disturbed individual that the dentist in the building next door is directing X rays at him to destroy his mind is not. By contrast, paranoid beliefs in politics are intelligible because they are rooted in the desires of the agent. A strongly anti-Semitic person is motivated to entertain absurd beliefs about the omnipotent and evil nature of the Jews (see Chapter 7). It is not that she wants Jews to have these features, but she is motivated to believe they do because the belief can rationalize her urge to destroy them. Even contradictory beliefs may be intelligible. An anti-Semite may on different occasions characterize the Jews as “vermin” and assert their omnipotence.⁶ The very same people who say that “Jews are always trying to push in where they are not wanted” also believe that “Jews are clannish, always sticking together.” One and the same Muslim may assert that the Israeli intelligence service Mossad was behind the attacks on the World Trade Center on September 11, 2001, and take pride in the event.

Understanding civil wars

Let me give two extended examples to flesh out the ideas of intelligible beliefs and desires, both taken from studies of civil wars past and present. I shall then draw on the same studies and some others to address the basic hermeneutic question of how we can impute or establish motivations and beliefs.

Consider first the belief in predestination, which was a main issue dividing Calvinists and Catholics in the wars of religion. At its origin was the intense religious anxiety experienced by many believers in pre-Reformation times, due to uncertainty about their salvation. How could one be sure – could one ever be sure – that one had done enough to achieve it? Looking back to his earlier years, Calvin wrote in 1539 that even when he had satisfied the demands of the church to confess his sins and efface God's memory of them by doing good works and penance, “I was far removed from certainty and tranquility of conscience. For each time that I delved into myself or lifted my heart up to You, I was struck by such an extreme horror that neither purgations nor disculpations could cure me.”

What relieved him from anxiety was the shift from a conception of God as immanent in the world, an oppressive and threatening presence, to a conception of God as absolutely transcendent. Crucially, this idea was linked to the doctrine of double predestination: since God had chosen from eternity who would be saved and who would be damned, there was nothing one could do for one's salvation and hence no reason to worry that one had not done enough. The key interpretive issue concerns the link between this belief in predestination and the relief from anxiety. A priori, this effect of the doctrine seems unintelligible. Calvin taught that the elect were a small minority, ranging (in different statements) from one in a hundred to one in five. What could generate more anxiety than the belief that one was very likely to be among the damned and that there was nothing one could do to escape an eternity of burning in hell? Would not conversion from Catholicism to Calvinism be to go, literally, from the frying pan into the fire?

The answer is probably to be found along the lines first sketched by Max Weber. Given their belief in predestination, the Calvinists could not hold that rational and systematic effort would bring them salvation, but they could and did hold that it would give them the subjective certainty of salvation. Calvin himself wrote that “the vocation of the elect is like a demonstration and testimony of their election.” And it seems in fact that conversion to Calvinism effectively eliminated uncertainty about salvation. I return to this form of “magical thinking” in Chapter 7. Here I merely want to emphasize how the twin mechanisms of wishful thinking and magical thinking lend intelligibility to the belief in predestination.

Consider next the intelligibility of motivations. Why have young Palestinians been willing to give their life in suicide missions? Their main motivation – to obtain or defend a national homeland – is not difficult to understand.⁷ It is a cause that may be as compelling as was the defense of democracy in the struggle against Hitler. What may seem puzzling is the strength of the motivation. Some additional causal factors are needed to make it intelligible. I shall discuss half a dozen of these and conclude in favor of one of them.

Prior to September 11, 2001, there was a widespread belief that the typical suicide bomber in the Middle East was a single young unemployed man, perhaps sexually starved, for whom a religious movement could fill a vacuum that would otherwise be occupied by family and work. Then overnight, after the attack on the World Trade Center, experts on terrorism decided that they had to “rewrite the book.” Even before then, however, the frequent if fluctuating deployment of female suicide bombers should have led scholars to question this stereotype. In the second Intifada, the use of female suicide bombers, some of them mothers or highly educated people, was even more striking.

The often cited factors of poverty and illiteracy also seem to have limited causal efficacy, at least as features of the individual suicide attackers. Among Palestinian suicide bombers, income and education tended in fact to be higher than in the general population. Explanations in terms of poverty are also unsatisfying because it is not clear how poverty would generate the required motivation. In one common view, the gains from blowing oneself up have to be weighed against the cost of blowing oneself up – one's life. If life is not highly valued, the cost is less. According to this approach, a life in misery and poverty is worth so little to the individual that the costs of suicide become negligible. I am skeptical about this argument, since I think that poor people find their lives as worth living as anyone else. That people adjust their aspirations to their circumstances so that they maintain a more or less constant level of satisfaction (“the hedonic treadmill”) is a pretty well-established psychological finding.⁸

A more plausible factor than absolute deprivation is relative deprivation, that is, the gap between expectations and reality experienced by the many educated Palestinians who lacked the prospect of decent employment. Downward social mobility could have the same effect. Yet the most relevant features seem to be permanent feelings of inferiority and resentment. The first of these emotions is based on comparison between oneself and others, the second on interaction between oneself and others. Generally speaking, interaction-based emotions are more powerful than comparison-based ones. Many writers on the Palestinian suicide bombers emphasize the intense resentment caused by the daily humiliations that occur in the interaction with the Israeli forces. Beyond the degrading checks and controls to which the Palestinians are subject, there is also their awareness that many Israelis think all Arabs “lazy, cowardly, and cruel,” as a Jerusalem taxi driver said to me some thirty years ago.

If this account is correct, the strong resentment of those who currently occupy the desired homeland enables us to understand the willingness to die of the Palestinian suicide attackers. The desire to fight the Israelis derives its strength from being embedded in a larger motivational complex. There is, however, an alternative view. Palestinian suicide attackers have usually been kept on a short leash by their handlers, who are ready to provide additional pressures in case the primary motivation should fail when the time of action approaches. One would-be suicide attacker in Iraq, who was captured and disarmed because he was visibly nervous, said that for three days before his mission he had been locked up in a room with a mullah who had talked about paradise and fed him “a special soup that made him strong.” The mental state that actually triggers the act of detonating the bomb may therefore be ephemeral and something of an artifact rather than a stable feature of the person. While terms such as “brainwashed” or “hypnotized” may be too strong, there is evidence that some of the attackers were in a trancelike state in the minutes before they died. When, as in such cases, an intention is isolated from the overall desire-belief system of the person, no interpretation is possible. The behavior of the handlers and more generally the organizers of the mission can, of course, be the object of interpretation.

A hermeneutic dilemma

It is well and good to claim that behavior must be explained in terms of the antecedent mental states – desires and beliefs – that cause them, but how do we establish these prior causes? On pains of circularity, we cannot use the behavior itself as evidence. We must look at other evidence, such as statements by the agent about his motivation, the consistency of his non-verbal behavior with these statements, the motives imputed to him by others, and the consistency of their non-verbal behavior with these imputations. Yet how can we exclude the possibility that these verbal and non-verbal forms of behavior were purposefully chosen to make an audience believe, falsely, that a particular motivation was at work? Professions and allegations of motivations can themselves be motivated. The question is central in collective decision making. As I argue in Chapter 24, all methods for consolidating individual preferences into a social decision create incentives for the participants in the process to misrepresent their preferences.

Consider, as an example, the motives of leaders and followers in civil wars. The parties have professed, or their opponents have imputed to them, one of three motives: religion, power, and money. Those who profess religious motives are often accused of using them as a disguise for their real motives, be they political or pecuniary. During the French wars of religion (1562–98), the warring parties constantly accused each other of using religion as a pretext for their political or even pecuniary aims. There were some bases for these charges. Henri de Navarre (later Henri IV) converted six times in his life, and the last conversion, in 1593, was widely suspected of opportunism. His father, Antoine de Bourbon, had already made it clear that his faith was for sale to the highest bidder. He accompanied the queen regent to mass, and his Protestant wife to communion. On his deathbed, he sought consolation from both religions. A leading reformer, Cardinal de Châtillon, married after his conversion but retained both his title as cardinal and the revenue from his bishopric. Another prelate, Antoine Carraciolo, bishop of Troyes, also wanted to combine a Protestant ministry with the income from his bishopric. A leading Catholic, Henri Duc de Guise, was perfectly willing to seek an alliance with the Calvinists against King Henri III.

In the contemporary world, too, religion is sometimes used as a pretext for politics, and politics as a pretext for money. The goals of the Chechnyan insurgents and of some Palestinian organizations, notably the Fatah, were originally exclusively political. When they took on a religious mantle, it was largely to attract a larger following. In Palestine, the rivalry with the unquestionably religious Hamas made this a necessity for organizational survival. In the Philippines, the terrorist group Abu Sayyaf has used the demand for an independent Islamic state as a pretext for kidnapping for huge ransoms. In Colombia, it remains uncertain whether the Revolutionary Armed Forces of Colombia (FARC) retains its original motivation to fight social injustice or whether it has by now degenerated into a mafia. In all these cases, as in the French wars of religion, the imputation of motives is often fraught with uncertainty. It may be hard to know, notably, whether the motives of leaders and of followers are completely concordant.

There are many reasons why people might want to misrepresent their motivations and those of their opponents. For one thing, each society has a normative hierarchy of motivations (Chapter 9) that induces a desire to present oneself as animated by a noble motivation rather than by a baser one, and to impute a low-ranked motivation to the opponent. In the French wars of religion as in the English Civil War, each side presented itself as religiously motivated and the other as merely hungry for power. For another, if one can make others attach credence to profession of a particular motivation, it may be easier to achieve one's aims. Because the image of a terrorist can be more daunting than that of a common criminal, mercenary kidnappers may increase the chances of concessions by waving the banners of a cause. In Colombia, many kidnappings are committed by common criminals who try to provoke fear among the families of victims by claiming to belong to a guerrilla group. Kidnappings are scarier if the terrorists are thought to be willing to take drastic measures if something goes wrong, and less willing to bargain over deadlines or haggle about money. If they cannot obtain what they demand, they can at least “make a statement” by killing their victims.

The problem of self-serving bias in statements about the intentions of social agents is serious, but not insurmountable. A simple way around it might be to consider the objective interests of the agent and assume that in the absence of strong evidence to the contrary they coincide with her subjective motivation, regardless of what she says about the latter. Alternatively, one might identify the actual consequences of her action and assume that in the absence of strong evidence to the contrary they are what she intended to bring about. (Either idea might apply to the choice of higher education discussed earlier.) The fact that there exist these two procedures for shifting the burden of proof suggests, however, that neither is acceptable. Both objective interests and actual consequences can suggest useful hypotheses about subjective motives, but neither has a presumption in its favor.⁹

Historians and social scientists have developed other ways of handling the problem that, especially when combined, can yield reasonably certain conclusions. One technique is to go beyond statements made before an audience and to look for those less likely to be motivated by a desire for misrepresentation. Letters, diaries, reported conversations, drafts, and the like, can be invaluable sources. We know, from letters they wrote to their wives, that some delegates to the French Assemblée Constituante in 1789 voted against bicameralism and royal veto because they thought their lives might be in danger if they voted otherwise. In the assembly, they justified their votes by the public interest. During the Terror, however, the fear that private letters might be opened made them more cautious. Although contemporaries and scholars have found the motivations of Philip II of Spain impenetrable, his foremost historian writes that “one important exception exists: the dispatches of the dozen foreign ambassadors who resided at the court of Spain [and] dedicated their time and energy to removing the veil of ‘secrecy and dissimulation’ with which the king sought to conceal his decisions and plans from others.” Hume found the report by one clergyman historian credible because “it was contrary to the interests of his order to preserve the memory” of the events he wrote about. Lack of interest in the outcome may also enhance credibility. Commenting on an edict by the Emperor Galerius, Gibbon wrote that “It is not usually in the language of edicts and manifestos, that we should search for the real character or the secret motives of princes, but as these were the words of a dying emperor, his situation, perhaps, may be admitted as pledge of his sincerity.” In nineteenth-century England, deathbed statements were exempt from the usual rules about hearsay evidence. The first draft of a document may say more about the beliefs and motives of the author than a later published version. It is instructive, for example, to compare the drafts of Marx's The Civil War in France or of his letter to Vera Sassoulitch with the official versions.

There may also be a sharp contrast between what actors may say in public and what they say behind closed doors. Although the published debates of the French Assemblée Constituante in 1789–91 are endlessly fascinating, two factors conspire to make them less than reliable as evidence about mental states. On the one hand, the public setting constrained the delegates to use public-interest arguments only; naked group interest was inadmissible. On the other hand, their vanity was stimulated by speaking before a thousand fellow delegates and a thousand auditors in the galleries. In both respects, the American Federal Convention was more conducive to sincerity. Because the number of delegates was small (55, compared to 1,200 in Paris) and the proceedings were shrouded in secrecy, interest-based bargaining could and did occur. At the same time, as Madison wrote many years later, “Had the members committed themselves publicly at first, they would have afterwards supposed consistency required them to maintain their ground, whereas by secret discussion no man felt himself obliged to retain his opinions any longer than he was satisfied of their propriety and truth, and was open to the force of argument.” Nor did the fear of future revelations chill the debates, as the secrecy was supposed to extend indefinitely and was in fact broken only by the publication of Madison's notes many decades later. Strategic reasons for misrepresentations are blunted if sincerity carries no cost.

In Chapter 9 I discuss how American decision makers used historical analogies to argue for or against various options in the Vietnam War. If we ask whether they relied on these analogies when making up their minds or merely used them to justify decisions reached on other grounds, a comparison between what they said in public and in private is instructive. A detailed analysis of the analogies, most of them from the crucial years of 1965 and 1966, shows some striking differences (see Table 3.1).

Table 3.1

The comparison suggests that many of the public analogies with Cuba, Munich and Berlin were for external consumption only, while the comparisons with Korea and Malaya had greater impact on internal deliberations. (The analogy with Munich may have been downplayed in private because it was taken for granted.) The absence of public references to the perhaps most obvious comparison case, the disastrous French war in Indochina, is striking. The private references seem mainly to have been objections to George Ball's use of that analogy to argue for withdrawal from Vietnam. I also discuss the use of analogies by America decision makers in Chapter 9.

Social scientists may also remove the cost of sincerity by creating an artificial veil of ignorance. Suppose a scholar wants to study the relation between sexual orientation and some other variables of interest. It may be difficult to induce subjects to give true answers about whether they have ever had sexual experiences with members of the same sex, even if they are assured that answers will be anonymized. To get around this problem the researcher can instruct them to answer honestly in case they had any such experience and, if they never had, to flip a coin to decide whether to answer yes or no. If they comply – and they have no reason not to comply – and the sample is large enough, the data will be just as good as if everyone had answered truthfully.

Another technique to see whether the non-verbal behavior of the agents is consistent with their professed motivation is to ask: do they put their money where their mouth is? Whatever the system analysts Alain Enthoven and Daniel Ellsberg in the RAND corporation might say in public about the risks of nuclear war, their inner attitude was perhaps revealed by the fact that they decided not to sign up for RAND's lucrative retirement plan. They may have thought they would not be around long enough to benefit from it. When in 2003 the Bush administration cited its certainty that Sadaam Hussein had weapons of mass destruction as its main reason for invading Iraq, did it also deploy the requisite measures to protect American soldiers from this threat? In classical Athens, a litigator could not credibly claim to be poor if it was common knowledge that his speechwriter charged a high salary. As editor of the Soviet journal Our Achievements, Maxim Gorky unwittingly revealed “Our defects” in 1934, when he complained that paper shortage limited its print run. Some behavioral patterns may reveal the true motivation of kidnappers. In 1996 in Costa Rica, kidnappers (mainly, ex-contras from Nicaragua) demanded a $1 million ransom in addition to job guarantees for workers, a cut in food prices, a rise in the Costa Rican minimum wage, and the release of fellow rebels from prison. When they were offered $200,000, they were satisfied and did not insist on the political demands, a fact that persuaded the authorities that their Robin Hood/rebel stance was a ruse and that money had always been their goal. Or take the behavior of French aristocratic émigrés in London during the French Revolution. In this hotbed of rumor about the imminent restoration of the monarchy and competition to be more-royalist-than-thou, it was vital to convey one's willingness to serve the counterrevolution. Verbal assurances were not enough. Any person who rented an apartment for more than a month was badly regarded; it was better to rent by the week to leave no doubt that one was ready to be called back to France by the counterrevolution.

Historians and social scientists routinely use such behavioral indicators to judge the sincerity of public statements. Commenting on the decision of the Emperor Diocletian to burn all books on alchemy lest the Egyptians use them to enrich themselves and rebel against the empire, Gibbon remarks that “if Diocletian had been convinced of the reality of that valuable art, far from extinguishing the memory, he would have converted the operation of it to the benefit of the public revenue.” Toward the end of World War II, there was a marked degree of skepticism in occupied France about the prospects for German victory. It might not be safe to express this attitude, but it was reflected in behavior. The proportion of high school students who chose German as a foreign language (or whose parents chose it for them) doubled from 1939 to 1942 and fell rapidly thereafter. Many publishers who eagerly signed up for the right to translate German books chose not to use the option. In wartime, investors may be reluctant to state in public that they believe their country is losing, but the stock market will reveal their true beliefs.

Judges and jurors often proceed in the same way. Sometimes, they ask, “Did the accused have a motive for doing X?” hoping that an answer will help them decide whether she in fact did X. In this case, “having a motive” is an objective idea, namely, whether the accused would in some way benefit from doing X. In other cases, more relevant here, it is established that the accused did X and the question is “What was her motive for doing it?” To establish whether a killing was a crime of passion or a cold-blooded action judges and jurors do not mainly look at objective benefits, but try to establish the subjective state of mind of the accused. If the accused claims to have acted in a fit of anger or jealousy and later is shown to have bought the murder weapon ahead of time or to have taken her time over the killing,¹⁰ her credibility is weakened.

Taken individually, each of these techniques may fail. A deputy might not be willing to admit to his wife that he was afraid for his life, or he might claim he was afraid in order to hide a less reputable motive (e.g. taking a bribe). In nineteenth-century India, deathbed statements were seen as unreliable since people sometimes used their dying moments to harm their enemies. In the émigré example, both true believers and disbelievers would be motivated to lease by the week, the former to facilitate their return to France when the day arrived and the latter to escape the criticism of being defeatist. There are limits, however, on people's ability to weave the tangled web of deceit without revealing their true motives. Hypocrisy, Somerset Maugham said, is a full-time profession. Even Tartuffe slipped in the end. To argue for the sincerity of Henri IV's religious beliefs, his biographer not only quotes the positive evidence of “numerous episodes where his religious spirit manifested itself without any advertising intention” but also argues, “Had there been any hypocrisy, it would have showed its horns on this or that pleasant occasion.” Regarding Oliver Cromwell, by contrast, Hume asserted that “notwithstanding his habits of profound dissimulation, he could not so carefully guard his expressions, but that sometimes his favourite notions would escape him.”

Along the same lines we may quote Montaigne:

Those who counter what I profess by calling my frankness, my simplicity and my naturalness of manner mere artifice and cunning-prudence rather than goodness, purposive rather than natural, good sense rather than good hap – give me more honour than they take from me. They certainly make my cunning too cunning. If anyone of those men would follow me closely about and spy on me, I would declare him the winner if he does not admit that there is no teaching in his sect which would counterfeit my natural way of proceeding and keep up an appearance of such equable liberty along such tortuous paths, nor of maintaining so uncompromising a freedom of action along paths so diverse, and concede that all their striving and cleverness could never bring them to act the same.

While the benefits of misrepresentation may be considerable, the costs can be prohibitive. To some extent, the instrumental profession of motives is self-limiting. Because any given motive is embedded in a vast network of other motives and beliefs, the number of adjustments to be made in sustaining hypocrisy can be crippling. A single false note may be enough for the whole construction to crumble. Many proverbs testify to the irreversibility of the breakdown of trust. Although the folk belief “Who tells one lie will tell a hundred” needs to be severely qualified (Chapter 12), the unqualified belief is in fact widely held and serves to some extent as a deterrent for lying. For this reason, among others, Descartes may have been right in saying that “the greatest subtlety of all is never to make use of subtlety.”

Bibliographical note

The debate on Erklären versus Verstehen is covered in Part III of M. Martin and L. McIntyre (eds.), Readings in the Philosophy of Social Science (Cambridge, MA: MIT Press, 1994). The chapter in that volume by Dagfinn Føllesdal, “Hermeneutics and the hypothetico-deductive method,” argues for a position close to my own. The quote from Weber is in his essay “The interpretive understanding of social action,” in M. Brodbeck (ed.), Readings in the Philosophy of the Social Sciences (London: Macmillan, 1969), p. 33. The comment on Gregor Mendel's statistical methods is from R. Abelson, Statistics as Principled Argument (Hillsdale, NJ: Lawrence Erlbaum, 1995), pp. 96–7. The incoherence of anti-Semitic attitudes is touched on in J. Telushkin, Jewish Humor (New York: Morrow, 1992), which is also my source for other remarks on and by Jews about their alleged characteristics. The comments on the contradictory beliefs of American slaveholders is from A. Taylor, The Internal Enemy (New York: Norton, 2013), p. 324. The remarks on educational choice are an implicit polemic against G. Becker and C. Mulligan, “The endogenous determination of time preferences,” Quarterly Journal of Economics 112 (1997), 729–58, further discussed in the Conclusion. The evidentiary value of deathbed confessions is discussed in J. F. Stephen, A History of English Criminal Law (London: Macmillan, 1883; Buffalo, NY: Hein, 1964), vol. I, pp. 447–9. H. Sass, “Affektdelikte,” Nervenarzt 54 (1983), 557–72, lists thirteen reasons why a claim to have committed a crime out of passion might lack credibility. An outstanding interpretive discussion of motivations in the French wars of religion is D. Crouzet, Les guerriers de Dieu (Paris: Champ Vallon, 1990). Interpretive analyses of the motivations and beliefs of suicide attackers are found in the essays by S. Holmes, L. Ricolfi, and J. Elster in D. Gambetta (ed.), Making Sense of Suicide Missions (Oxford University Press, 2005). The comment on the ambassadors to Spain is in G. Parker, The Imprudent King (New Haven, CT: Yale University Press, 2014), p. xvi. The enumeration of public and private analogies in statements about Vietnam is a condensed version of Tables 3.1 and 3.2 in Y. Khong, Analogies at War (Princeton University Press, 1992). The report that Enthoven and Ellsberg did not sign up for retirement benefits is from F. Kaplan, Wizards of Armageddon (Stanford University Press, 1983), p. 124. The comment on Our Achievements is from S. Fitzpatrick, Everyday Stalinism (University of Chicago Press, 1999), p. 68. The behavioral indicators of the beliefs of the French in a German victory are cited from P. Burrin, France à l'heure allemande (Paris: Seuil, 1995). The comment on Henri IV's religious belief is in J.-P. Babelon, Henri IV (Paris: Fayard, 1982), p. 554. Excessive skepticism about motives is discussed in G. Mackie, “Are all men liars?” in J. Elster (ed.), Deliberative Democracy (Cambridge University Press, 1998).

¹ In the experimental sciences, “novel facts” can mean facts that are literally new, as when one exposes rats or human beings to conditions that do not occur naturally. In the humanities and non-experimental social sciences, “novel” must be taken in the epistemic sense of “previously unsuspected” rather than in the ontological sense of “previously non-existing.”

² Thus five year olds who were willing to wait longer to obtain a larger marshmallow had better SAT scores when observed later.

³ Thus when Mexican applicants to college were randomly accepted, those who were successful in the admission lottery were, when measured two years later, more patient on average.

⁴ In Chapter 13 I argue that the idea of being motivated to be motivated by long-term consequences is conceptually incoherent, but that is a separate point from the one I am making here.

⁵ In an age of greater statistical innocence, Gregor Mendel, the discoverer of the laws of inheritance, apparently practiced this method of “quitting when you're ahead” in his experiments.

⁶ Commenting on the attitude of Virginian slaveholders, the foremost historian on the topic writes that they were against arming and freeing slaves to fight against the British in the War of 1812, arguing “that blacks were too cowardly to fight, although they also dreaded their slaves as a formidable internal enemy: living with slavery required such contradictions of belief.”

⁷ By claiming that this is their main motivation, I am not denying that there may be others, such as the desire for posthumous glory or fame, the material benefits that will accrue to the family of the suicide attacker, revenge for the Israeli killing of a friend or relative, or social pressure to volunteer for a mission. As I note in the introduction to Part II, I am skeptical about the motivational power of religious benefits in the form of a privileged access to paradise.

⁸ The idea of a hedonic treadmill must be handled with some care. It must not be confused with the idea of sour grapes, that is, with the tendency of people to downgrade what they cannot have (Chapter 9). If paraplegics report being as happy after a disabling accident as before, it is surely not because they devalue the state of full mobility. Nor must it be confused with Seneca's claim that “I am not sure that [the poor] are not happier [than the rich], because they have fewer things to distract their minds,” or with a similar assertion by an eighteenth-century physician, Thomas Percival, “It is one of the circumstances which soften the lot of the poor, that they are exempt from the solicitude attendant on the disposal of property.”

⁹ Gibbon, whom I often cite in this book, has a constant and puzzling habit of explaining behavior by a disjunction of motivations. To cite two examples among very many, he cites “the esteem or partiality” of a father to explain the inheritance he left to his son, and asserts that one political group was “persuaded or compelled” to acknowledge the supremacy of another. In general it is impossible to tell whether he made a backward inference from behavior to several possible motivations or whether he had some, but insufficient evidence, about each of them.

¹⁰ In a British decision (R. v. McPherson) from 1957, Lord Goddard asked rhetorically, “How can it be said that the appellant was acting in a gust of passion when he fired not one shot but four shots, and each shot involved the breaking of the gun to reload and the taking out of cartridges four separate times?”