052177179X cambridge university press explaining social behavior more nuts and bolts for the social sciences apr 2007

In twenty-sixsuccinct chapters, Jon Elster provides an account of the nature ofexplanation in the social sciences; an analysis of the mental states –beliefs, desires, and emotions – that

EXPLAINING SOCIAL BEHAVIOR

This book is an expanded and revised edition of the author’s criticallyacclaimed volume Nuts and Bolts for the Social Sciences In twenty-sixsuccinct chapters, Jon Elster provides an account of the nature ofexplanation in the social sciences; an analysis of the mental states –beliefs, desires, and emotions – that are precursors to action; a systematiccomparison of rational-choice models of behavior with alternativeaccounts; a discussion of what the social sciences may learn from neuro-science and evolutionary biology; and a review of mechanisms of socialinteraction ranging from strategic behavior to collective decision making

He offers an overview of key explanatory mechanisms in the socialsciences, relying on hundreds of examples and drawing on a large variety

of sources – psychology, behavioral economics, biology, political science,historical writings, philosophy, and fiction In accessible and jargon-freelanguage, Elster aims at accuracy and clarity while eschewing formalmodels In a provocative conclusion, he defends the centrality of quali-tative social science in a two-front war against soft (literary) and hard(mathematical) forms of obscurantism

Jon Elster is Professor (Chaire de Rationalite´ et Sciences Sociales) at theColle`ge de France A Fellow of the American Academy of Arts andSciences and Corresponding Fellow of the British Academy, he is arecipient of fellowships from the John Simon Guggenheim Foundationand the Russell Sage Foundation, among many others Dr Elster hastaught at the University of Chicago and Columbia University and hasheld visiting professorships at many universities in the United States andEurope He is the author or editor of thirty-four books, most recentlyClosing the Books: Transitional Justice in Historical Perspective and Retri-bution and Restitution in the Transition to Democracy

Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo Cambridge University Press

The Edinburgh Building, Cambridge CB2 8RU, UK

First published in print format

Information on this title: www.cambridge.org/9780521771795

This publication is in copyright Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press.

Published in the United States of America by Cambridge University Press, New York

www.cambridge.org

hardback paperback paperback

eBook (EBL) eBook (EBL) hardback

For Jonathan and Joanna

IV  LESSONS FROM THE NATURAL

vii

P REFACE

This book began as a revision of a book I published in 1989, Nuts andBolts for the Social Sciences It ended up as a quite different and moreambitious kind of book It covers a much greater variety of topics, inconsiderably more detail, and in a different spirit Although ninechapters have the same headings as chapters in the earlier book, onlyChapter 9 and Chapter24remain substantially the same

Although comprehensive in scope, the book is not a treatise It is bothless and more than that It is an elementary, informal, and personalpresentation of ideas that have, I believe, considerable potential forilluminating social behavior I use plenty of examples, many of themanecdotal or literary, others drawn from more systematic studies Thevery occasional use of algebra does not go beyond high school level Atthe same time, the book has a methodological and philosophical slantnot usual in introductory-level presentations There is an effort to placethe social sciences within the sciences more generally – the natural sci-ences as well as the humanities There is also an effort to make the readerkeep constantly in mind how general principles of scientific explanationconstrain the construction of theories with explanatory pretensions.The style of the bibliographical notes to each chapter reflects the rise

of the Internet, in particular of Wikipedia, Google.com, and Scholar.Google.com Since readers can find most relevant references in a matter

of minutes, I have omitted sources for many of the statements andfindings in the text Instead I try to point readers to important source-books, to some modern classics, to books and articles that are the sources

of claims that might be harder to track down on the Internet, and toauthors from whom I have taken so much that not mentioning themwould justify a pun on my name (Elster in German means magpie).Although the main text contains few references to contemporaryscholars, I refer extensively to Aristotle, Seneca, Montaigne, La Roche-foucauld, Samuel Johnson, H C Andersen, Stendhal, Tocqueville,Proust, and other classical writers who remain literally inexhaustible

ix

sources of causal hypotheses We would be cutting ourselves off frommany insights if we ignored the mechanisms suggested by philosophy,fiction, plays, and poetry If we neglect twenty-five centuries of reflectionabout mind, action, and interaction in favor of the last one hundredyears or the last ten, we do so at our peril and our loss I cite theseauthors not so much to appeal to their authority as to make the case that

it is worth one’s while to read widely rather than narrowly In directopposition to what I perceive as the relentless professionalization of(especially American) social science, which discourages students fromlearning foreign languages and reading old books, the present volume is

an extended plea for a more comprehensive approach to the study ofsociety

In preparing the manuscript I received assistance and comments frommany people I should first thank my students at Columbia Universityfor their incisive questioning and comments in the course where I firstpresented the material that turned into this book Suggestions fromPablo Kalmanovitz were particularly useful In Collioure, AanundHylland and Ole-Jørgen Skog spent three days with me discussing a draft

of the whole book In Oslo, Hylland, Karl O Moene, and John Roemercontinued the discussion over a day and a half Their comments not onlysaved me from many (many!) errors but also suggested how I couldsupplement and consolidate the exposition I am grateful to Roemer

in particular for urging me to write a conclusion I received writtencomments on the whole manuscript from Diego Gambetta, Raj Saah,and an anonymous reviewer Gambetta’s comments were particularlydetailed and helpful I had useful conversations with Walter Mischelabout the ideas – largely originating with him – presented in Chapter10

I also received valuable written comments from George Ainslie on theideas – many of them raised by him – presented in Part I of the book.Bernard Manin commented constructively on Chapter25.Robyn Dawesoffered incisive comments on Chapter7and Chapter12 Finally, over theseveral last years I have presented drafts of chapters for this book to the

members of the ‘‘Monday group’’ that has met weekly in New York Cityeach fall and more occasionally in the spring since 1995: John Ferejohn,Raquel Fernandez, Russell Hardin, Stephen Holmes, Steven Lukes,Bernard Manin, Pasquale Pasquino, Adam Przeworski, and John Roemer.

I thank them all for their friendly and constructive objections

I dedicate the book to Jonathan and Joanna Cole – they willknow why

I cite Montaigne’s Essays from the translation by M Screech (London:Penguin, 1971); Proust from the new translation edited by C Pren-dergast (London: Penguin, 2003); Pascal’s Pense´es from the translation

by A J Krailsheimer (London: Penguin, 1995); La Rochefoucauld’sMaxims from the translation by L Tancock (London: Penguin, 1981);

La Bruye`re’s Characters from the translation by H van Laun (New York:Scribner, 1885); Stendhal’s On Love from the translation by G Sale,

S Sale, and J Stewart (London: Penguin, 1975); and Tocqueville’sDemocracy in America from the new translation by A Goldhammer(New York: Library of America, 2004) Other translations from Frenchare mine

  

This book is about explaining social behavior In thefirstpart, I spell out

my conception of explanation, and in the remaining four parts, I struct a toolbox of concepts and mechanisms that apply to particularcases Needless to say, it does not aspire to completeness Rather thantrying to spell out the gaps, which will be obvious, let me begin byenumerating a sample of the puzzles that, I submit, can be illuminated

con-by the approach I am taking In the Conclusion, I return to the samepuzzles with brief references to the explanations I have cited in earlierchapters

The examples and the explanations must be taken with two caveats.First, I do not claim that all the explananda are well-established facts In

an actual explanation, this is of course a crucial first step – it makes nosense to try to explain what does not exist For the purpose of building

a toolbox, however, one can be less rigorous Second, even for theexplananda whose existence is well documented I do not claim that theexplanations I cite are the correct ones I only claim that they satisfy aminimal condition for an explanation – that they logically imply theexplananda The puzzles and explanations are intended to show ‘‘if thiskind of thing happens, here is the kind of mechanism that might explainit’’ as well as ‘‘if this mechanism operates, here is the kind of thing it canproduce.’’ Given these caveats, here are the puzzles, arranged somewhatarbitrarily (since many puzzles could fit in several categories) according

to the four substantive parts of the book.1

Why do other gamblers believe than when red has come up five times in a row, black is more likely than red to come up next?

Why do preferences sometimes change through the sheer passage

of time?

Why do many people who seem to believe in the afterlife want it

to arrive as late as possible?

Why are people reluctant to acknowledge, to themselves and others, that they are envious?

Why are people reluctant to acknowledge, to themselves and others, that they are ignorant?

Why, among sixteenth-century converts to Calvinism, did the belief that people were predestined either to heaven or to hell induce greater peace of mind than the belief that one could achieve salvation through good works?

Why is it (sometimes) true that ‘‘Who has offended, cannot forgive’’?

Why is shame more important than guilt in some cultures?

Why did the French victory in the 1998 soccer World Cup generate so much joy in the country, and why did the fact that the French team did not qualify beyond the opening rounds in 2002 cause so much despondency?

Why do women often feel shame after being raped?

Why do humiliating rituals of initiation produce greater rather than lesser loyalty to the group into which one is initiated?

Why are people unwilling to break self-imposed rules even when

it makes little sense to follow them?

Why is the pattern of revenge ‘‘Two eyes for an eye’’ instead of

‘‘An eye for an eye’’?

Why is the long-term yield on stocks much larger than that on bonds (i.e., why does not the value of stocks rise to equalize the yields)?

Why do suicide rates go down when dangerous medications are sold in blister packs rather than bottles?

Why did none of thirty-eight bystanders call the police when Kitty Genovese was beaten to death?

Why did some individuals hide or rescue Jews under the Nazi regimes?

Why did President Chirac call early elections in 1997, only to lose his majority in parliament?

Why are some divorcing parents willing to share child custody even when their preferred solution is sole custody, which they are likely to get were they to litigate?

Why are poor people less likely to emigrate?

Why do some people save in Christmas accounts that pay no interest and do not allow for withdrawal before Christmas?

Why do people pursue projects, such as building the Concorde airplane, that have negative expected value?

Why, in ‘‘transitional justice’’ (when agents of an autocratic regime are put on trial after the transition to democracy), are those tried immediately after the transition sentenced more severely than those who are tried later?

Why, in Shakespeare’s play, does Hamlet delay taking revenge until the last act?

IIILessons from the Natural Sciences

Why are parents much more likely to kill adopted children and stepchildren than to kill their biological children?

Why is sibling incest so rare, given the temptations and opportunities?

Why do people invest their money in projects undertaken by other agents even when the latter are free to keep all the profits for themselves?

Why do people take revenge at some material cost to them and with no material benefits?

Why do people jump to conclusions beyond what is warranted by the evidence?

Why does an individual vote in elections when his or her vote is virtually certain to have no effect on the outcome?

Why are economically successful individuals in modern Western societies usually slimmer than the average person?

Why do people refrain from transactions that could make everybody better off, as when they abstain from asking a person

in the front of a bus queue whether he is willing to sell his place?

Why did President Nixon try to present himself to the Soviets as being prone to irrational behavior?

Why do military commanders sometimes burn their bridges (or their ships)?

Why do people often attach great importance to intrinsically insignificant matters of etiquette?

Why do passengers tip taxi drivers and customers tip waiters even when visiting a foreign city to which they do not expect to return?

Why do firms invest in large inventories even when they do not anticipate any interruption of production?

Why, in a group of students, would each think that others have understood an obscure text better than he has?

Why are votes in many political assemblies taken by roll call?

Why is logrolling more frequent in ordinary legislatures than in constituent assemblies?

Suggested explanations for these phenomena will be provided atvarious places in the book and briefly summarized in the Conclusion.Here I only want to make a general remark about two types of expla-nation that are not likely to be useful As readers will see in the very firstchapter, with several reminders along the road, one of the aims of thebook is to inculcate skepticism toward two common lines of reasoning.First, with very few exceptions the social sciences cannot rely on func-tional explanation, which accounts for actions or behavioral patterns byciting their consequences rather than their causes Do norms of tippingexist because it is more efficient to have customers monitor waiters than

to have the owner do it? I do not think so Second, I now believe thatrational-choice theory has less explanatory power than I used to think

Do real people act on the calculations that make up many pages ofmathematical appendixes in leading journals? I do not think so

On three counts at least, rational-choice theory is nevertheless avaluable part of the toolbox If understood in a qualitative commonsense

way, it is capable of explaining much everyday behavior Even when itdoes not explain much, it can have immense conceptual value Gametheory, in particular, has illuminated the structure of social interaction inways that go far beyond the insights achieved in earlier centuries Finally,human beings want to be rational The desire to have sufficient reasonsfor one’s behavior, and not simply be the plaything of psychic forcesacting ‘‘behind one’s back,’’ provides a permanent counterforce to themany irrationality-generating mechanisms that I survey in this book.Even though I am critical of many rational-choice explanations,

I believe the concept of choice is fundamental In the book I considerseveral alternatives to choice-based explanation and conclude thatalthough they may sometimes usefully supplement that approach, theycannot replace it The fact that people act under different constraints, forinstance, can often explain a great deal of variation in behavior Also, insome cases one may argue that selection of agents rather than choice byagents is responsible for the behavior we observe By and large, however,

I believe that the subjective factor of choice has greater explanatorypower than the objective factors of constraints and selection This isobviously an intuition that cannot be proved in any rigorous sense, and

in any case social scientists ought to have room for all the factors in theirtoolbox

I argue that all explanation is causal To explain a phenomenon (anexplanandum) is to cite an earlier phenomenon (the explanans) thatcaused it When advocating causal explanation, I do not intend toexclude the possibility of intentional explanation of behavior Intentionscan serve as causes A particular variety of intentional explanation isrational-choice explanation, which will be extensively discussed in laterchapters Many intentional explanations, however, rest on the assump-tion that agents are, in one way or another, irrational In itself, irra-tionality is just a negative or residual idea, everything that is not rational.For the idea to have any explanatory purchase, we need to appeal tospecific forms of irrationality with specific implications for behavior InChapter 12, for instance, I enumerate and illustrate eleven mechanismsthat can generate irrational behavior.

Sometimes, scientists explain phenomena by their consequences ratherthan by their causes They might say, for instance, that blood feuds areexplained by the fact that they keep populations down at sustainablelevels This might seem a metaphysical impossibility: how can the

existence or occurrence of something at one point in time be explained

by something that has not yet come into existence? As we shall see, theproblem can be restated so as to make explanation by consequences ameaningful concept In the biological sciences, evolutionary explanationoffers an example In the social sciences, however, successful instances ofsuch explanation are few and far between The blood-feud example isdefinitely not one of them

The natural sciences, especially physics and chemistry, offer tions by law; laws are general propositions that allow us to infer the truth

explana-of one statement at one time from the truth explana-of another statement at someearlier time Thus when we know the positions and the velocity of theplanets at one time, the laws of planetary motion enable us to deduce andpredict their positions at any later time This kind of explanation isdeterministic: given the antecedents, only one consequent is possible Thesocial sciences offer few if any law-like explanations of this kind Therelation between explanans and explanandum is not one-one or many-one, but one-many or many-many Many social scientists try to modelthis relation by using statistical methods Statistical explanations areincomplete by themselves, however, since they ultimately have to rely onintuitions about plausible causal mechanisms

Chapter 1Explanation

  

Explanation: General

The main task of the social sciences is to explain social phenomena It isnot the only task, but it is the most important one, to which others aresubordinated or on which they depend The basic type of explanandum

is an event To explain it is to give an account of why it happened, byciting an earlier event as its cause Thus we may explain Ronald Reagan’svictory in the 1980 presidential elections by Jimmy Carter’s failedattempt to rescue the Americans held hostage in Iran.1

Or we mightexplain the outbreak of World War II by citing any number of earlierevents, from the Munich agreement to the signing of the VersaillesTreaty Even though in both cases the fine structure of the causalexplanation will obviously be more complex, they do embody the basicevent-event pattern of explanation In a tradition originating with DavidHume, it is often referred to as the ‘‘billiard-ball’’ model of causalexplanation One event, ball A hitting ball B, is the cause of – and thusexplains – another event, namely, ball B’s beginning to move

Those who are familiar with the typical kind of explanation in thesocial sciences may not recognize this pattern, or not see it as privileged

In one way or another, social scientists tend to put more emphasis onfacts, or states of affairs, than on events The sentence ‘‘At 9 a.m the roadwas slippery’’ states a fact The sentence ‘‘At 9 a.m the car went off theroad’’ states an event As this example suggests, one might offer a fact-event explanation to account for a car accident.2

Conversely, one mightpropose an event-fact explanation to account for a given state of affairs, aswhen asserting that the attack on the World Trade Center in 2001

1

To anticipate a distinction discussed later, note that, Carter did not fail to attempt but attempted and failed A nonaction such as a failure to attempt cannot have causal efficacy, except in the indirect sense that if others perceive or infer that the agent fails to act, they may take actions that they otherwise would not have.

2

The voter turnout example discussed later provides another illustration.

explains the pervasive state of fear of many Americans Finally, standardsocial-science explanations often have a fact-fact pattern To take anexample at random, it has been claimed that the level of education ofwomen explains per capita income in the developing world.

Let us consider the explanation of one particular fact, that 65 percent

of Americans favor, or say that they favor, the death penalty.3

In ciple, this issue can be restated in terms of events: How did theseAmericans come to favor the death penalty? What were the formativeevents – interactions with parents, peers, or teachers – that caused thisattitude to emerge? In practice, social scientists are usually not interested

prin-in this question Rather than tryprin-ing to explaprin-in a brute statistic of thiskind, they want to understand changes in attitudes over time or differences

in attitudes across populations The reason, perhaps, is that they do notthink the brute fact very informative If one asks whether 65 percent ismuch or little, the obvious retort is, ‘‘Compared to what?’’ Compared tothe attitudes of Americans around 1990, when about 80 percent favoredthe death penalty, it is a low number Compared to the attitudes in someEuropean countries, it is a high number

Longitudinal studies consider variations over time in the dependentvariable Cross-sectional studies consider variations across populations Ineither case, the explanandum is transformed Rather than trying toexplain the phenomenon ‘‘in and of itself,’’ we try to explain how itvaries in time or space The success of an explanation is measured, inpart, by how much of the ‘‘variance’’ (a technical measure of variation) itcan account for.4

Complete success would explain all observed variation

In a cross-national study we might find, for instance, that the percentage

of individuals favoring the death penalty was strictly proportional to thenumber of homicides per 100,000 inhabitants Although this findingwould provide no explanation of the absolute numbers, it would offer aperfect explanation of the difference among them.5

perfect success is never achieved, but the same point holds Explanations

of variance do not say anything about the explanandum ‘‘in and ofitself.’’

An example may be taken from the study of voting behavior As weshall see later (Chapter12), it is not clear why voters bother to vote at all

in national elections, when it is morally certain that a single vote willmake no difference Yet a substantial fraction of the electorate do turnout on voting day Why do they bother?

Instead of trying to solve this mystery, empirical social scientistsusually address a different question: Why does turnout vary acrosselections? One hypothesis is that voters are less likely to turn out ininclement weather, because rain or cold makes it more attractive tostay home If the data match this hypothesis, as indicated by line C inFigure 1.1, one might claim to have explained (at least part of ) thevariance in turnout Yet one would not have offered any explanation ofwhy the line C intersects the vertical axis at P rather than at Q or R It is

as if one took the first decimal as given and focused on explaining thesecond For predictive purposes, this might be all one needs Forexplanatory purposes, it is unsatisfactory The ‘‘brute event’’ that 45

R

percent or more of the electorate usually turn out to vote is an interestingone, which cries out for an explanation.

The ideal procedure, in an event-event perspective, would be thefollowing Consider two elections, A and B For each of them, identifythe events that cause a given percentage of voters to turn out Once wehave thus explained the turnout in election A and the turnout in election

B, the explanation of the difference (if any) follows automatically, as aby-product As a bonus, we might also be able to explain whetheridentical turnouts in A and B are accidental, that is, due to differencesthat exactly offset each other, or not In practice, this procedure might betoo demanding The data or the available theories might not allow us toexplain the phenomena ‘‘in and of themselves.’’ We should be aware,however, that if we do resort to explanations of variance, we are engaging

in a second-best explanatory practice

Sometimes, social scientists try to explain nonevents Why do manypeople fail to claim social benefits they are entitled to? Why did nobodycall the police in the Kitty Genovese case?6

Considering the first tion, the explanation might be that the individuals in question decidenot to claim their benefits, because of fear of stigma or concerns withself-image Since making a decision is an event, this would provide a fullysatisfactory account If it fails, social scientists would, once again, look atthe differences between those who are entitled to benefits and claim themand those who are and do not Suppose the only difference is that thelatter are unaware of their entitlement As an explanation, this is helpfulbut insufficient To go beyond it, we would want to explain why someentitled individuals are unaware of their entitlement To discover thatbecause they are illiterate, they are unable to read the letters informingthem about their rights would also be helpful but insufficient At somepoint in the explanatory regress, we must either come to a positive event,such as a conscious decision not to become literate or a conscious

ques-6

For more than half an hour on March 27, 1964, thirty-eight respectable, law-abiding citizens in Queens, New York, watched a killer stalk and stab a woman in three separate attacks in Kew Gardens Twice their chatter and the sudden glow of their bedroom lights interrupted him and frightened him off Each time he returned, sought her out, and stabbed her again Not one person telephoned the police during the assault; one witness called after the woman was dead.

decision by officials to withhold information, or turn to those who doseek the benefits to which they are entitled Once we have explained thebehavior of the latter, the explanation why others fail to seek their benefitwill emerge as a by-product.

Considering the Kitty Genovese case, there is no variation in behavior

to explain, since nobody called the police Accounts of the case indicatethat several of the observers decided not to call the police In terms ofproximate causes this provides a fully satisfactory account, although

we might want to know the reasons for their decision Was it becausethey feared ‘‘getting involved’’ or because each observer assumed thatsomeone else would call the police (‘‘Too many shepherds make a poorguard’’)? Some of the observers, however, apparently did not even thinkabout calling the police One man and his wife watched the episode forits entertainment value, while another man said he was tired and went

to bed To explain why they did not react more strongly one might citetheir shallow emotions, but that, too, would be to account for a negativeexplanandum by citing a negative explanans Once again, their behaviorcan only be explained as a by-product or residual If we have a satis-factory explanation of why some individuals thought about calling thepolice, even if in the end they decided not to, we shall have the onlyexplanation we are likely to get of why some did not even think about it

In the rest of this book I shall often relax this purist or rigoristapproach of what counts as a relevant explanandum and an appropriateexplanation The insistence on event-focused explanations is a bit like theprinciple of methodological individualism, which is another premise ofthe book In principle, explanations in the social sciences should referonly to individuals and their actions In practice, social scientists oftenrefer to supraindividual entities such as families, firms, or nations, either

as a harmless shorthand or as a second-best approach forced upon them bylack of data or of fine-grained theories These two justifications alsoapply to the use of facts as explananda or as explanantia, to explanations

of variance rather than of the phenomena ‘‘in and of themselves,’’ and tothe analysis of negative explananda (nonevents or nonfacts) The purpose

of the preceding discussion is not to hold social scientists to pointless orimpossible standards, but to argue that at the level of first principles the

event-based approach is intrinsically superior If scholars keep that fact inmind they may, at least sometimes, come up with better and morefruitful explanations.

Sometimes, we might want to explain an event (or rather a pattern ofevents) by its consequences rather than by its causes I do not have inmind explanation by intended consequences, since intentions exist prior

to the choices or actions they explain Rather, the idea is that events may

be explained by their actual consequences, typically, their beneficialconsequences for someone or something As a cause must precede itseffect, this idea might seem to be incompatible with causal explanation.Yet causal explanation can also take the form of explanation by con-sequences, if there is a loop from the consequences back to their causes Achild may initially cry simply because it feels pain, but if the crying alsogets it attention from the parents, it may start crying more than it wouldhave otherwise I argue in Chapters 16 and 17 that this kind of expla-nation is somewhat marginal in the study of human behavior In most ofthe book, I shall be concerned with the simple variety of causal expla-nation in which the explanans – which might include beliefs andintentions oriented toward the future – precedes the occurrence of theexplanandum.7

In addition to the fully respectable form of functional explanation thatrests on specific feedback mechanisms, there are more disreputable formsthat simply point to the production of consequences that are beneficial insome respect and then without further argument assume that thesesuffice to explain the behavior that causes them When the explanandum

is a token, such as a single action or event, this kind of explanation failsfor purely metaphysical reasons To take an example from biology, wecannot explain the occurrence of a neutral or harmful mutation byobserving that it was a necessary condition for a further, advantageousone When the explanandum is a type, such as a recurrent pattern ofbehavior, it may or may not be valid Yet as long as it is not supported by

7

For some purposes, it may be useful to distinguish among causal, intentional, and functional explanation Physics employs only causal explanation; biology additionally admits functional explanation; and the social sciences further admit intentional explanation At the most fundamental level, though, all explanation is causal.

a specific feedback mechanism, we should treat it as if it were invalid.Anthropologists have argued, for instance, that revenge behavior hasbeneficial consequences of various kinds, ranging from populationcontrol to decentralized norm enforcement (Chapter 22 offers manyother examples.) Assuming that these benefits are in fact produced, theymight still obtain by accident To show that they arise nonaccidentally,that is, that they sustain the revenge behavior that causes them, thedemonstration of a feedback mechanism is indispensable And evenwhen one is provided, the initial occurrence of the explanandum must bedue to something else.

The Structure of Explanations

Let me now turn to a more detailed account of explanation in the socialsciences (and, to some extent, more generally) The first step is easilyoverlooked: before we try to explain a fact or an event we have toestablish that the fact is a fact or that the event actually did take place AsMontaigne wrote, ‘‘I realize that if you ask people to account for ‘facts,’they usually spend more time finding reasons for them than finding outwhether they are true They skip over the facts but carefully deduceinferences They normally begin thus: ‘How does this come about?’ Butdoes it do so? That is what they ought to be asking.’’

Thus before trying to explain, say, why there are more suicides in onecountry than in another, we have to make sure that the latter does nottend, perhaps for religious reasons, to underreport suicides Before we try

to explain why Spain has a higher unemployment rate than France, wehave to make sure that the reported differences are not due to differentdefinitions of unemployment or to the presence of a large undergroundeconomy in Spain If we want to explain why youth unemployment ishigher in France than in the United Kingdom, we need to decidewhether the explanandum is the rate of unemployment among youngpeople who are actively searching for jobs or the rate among youngpeople overall, including students If we compare unemployment inEurope and the United States, we have to decide whether the expla-nandum is the unemployed in the literal sense, which includes the

incarcerated population, or in the technical sense, which only includesthose searching for work.8

Before we try to explain why revenge takes theform of ‘‘tit for tat’’ (I kill one of yours each time you or yours kill one ofmine), we should verify that this is actually what we observe rather than,say, ‘‘two tits for a tat’’ (I kill two of yours each time you or yours killone of mine) Much of science, including social science, tries to explainthings we all know, but science can also make a contribution by estab-lishing that some of the things we all think we know simply are not so Inthat case, social science may also try to explain why we think we knowthings that are not so, adding as it were a piece of knowledge to replacethe one that has been taken away.9

Suppose now that we have a well-established explanandum for whichthere is no well-established explanation – a puzzle The puzzle may be asurprising or counterintuitive fact, or simply an unexplained correlation.One small-scale example is ‘‘Why are more theology books stolen fromOxford libraries than books on other subjects?’’ Another small-scaleexample, which I shall explore in more detail shortly, is ‘‘Why do moreBroadway shows receive standing ovations today than twenty years ago?’’Ideally, explanatory puzzles should be addressed in the five-stepsequence spelled out in the following In practice, however, steps (1), (2),and (3) often occur in a different order We may play around withdifferent hypotheses until one of them emerges as the most promising,and then look around for a theory that would justify it If steps (4) and(5) are carried out properly, we may still have a high level of confidence

in the preferred hypothesis Yet for reasons I discuss toward the end of

choose among hypotheses

8

In either of the last two cases, some individuals may take up a career as criminals or students because they do not think they would get a job if they tried For some purposes, one might want to count these among the unemployed; for other purposes, not.

9

Just as science can help explain popular beliefs in nonfacts, it can help explain popular beliefs in false explanations For instance, most of those who suffer from arthritis believe arthritic pain is triggered by bad weather Studies suggest, however, that there is no such connection Perhaps we should drop the search for the causal link between bad weather and arthritic pain and instead try to explain why arthritics believe there is one Most likely they were once told there was a connection and subsequently paid more attention to instances that confirmed the belief than to those that did not.

1 Choose the theory – a set of interrelated causal propositions – that holds out the greatest promise of a successful explanation.

2 Specify a hypothesis that applies the theory to the puzzle, in the sense that the explanandum follows logically from the hypothesis.

3 Identify or imagine plausible accounts that might provide alternative explanations, also in the sense that the explanandum follows logically from each of them.

4 For each of these rival accounts, refute it by pointing to additional testable implications that are in fact not observed.

5 Strengthen the proposed hypothesis by showing that it has additional testable implications, preferably of ‘‘novel facts,’’ that are in fact observed.

These procedures define what is often called the hypothetico-deductivemethod In a given case, they might take the form shown in Figure1.2

I shall illustrate it by the puzzle of increasing frequency of standingovations on Broadway It is not based on systematic observations orcontrolled experiments, but on my casual impressions confirmed bynewspaper reports For the present purposes, however, the shaky status ofthe explanandum does not matter If there are in fact more standingovations on Broadway than there were twenty years ago, how could we

go about explaining it?

I shall consider an explanation in terms of the rising prices ofBroadway tickets One newspaper reports the playwright Arthur Miller

as saying, ‘‘I guess the audience just feels having paid $75 to sit down, it’stheir time to stand up I don’t mean to be a cynic but it probably allchanged when the price went up.’’ When people have to pay seventy-fivedollars or more for a seat, many cannot admit to themselves that the showwas poor or mediocre, and that they have wasted their money To confirm

to themselves that they had a good time, they applaud wildly

More formally, the explanation is sought in the hypothesis ‘‘Whenpeople have paid a great deal of money or effort to obtain a good,they tend (other things being equal) to value it more highly than whenthey paid less for it.’’10

Given the factual premise of rising prices, this

10

A similar idea is sometimes used to defend the high fees of psychotherapists: patients wouldn’t believe in the therapy unless they paid a lot for it But no therapists to my knowledge state that they donate 50 percent of their fee to Red Cross.

proposition passes the minimal test that any explanatory hypothesis mustsatisfy: If it is true, we can infer the explanandum But this is a trulyminimal test, which many propositions could pass.11

To strengthen ourbelief in this particular explanation, we must show that it is supportedfrom below, from above, and laterally

Explanandum Hypothesis

Implication 5 (observed)

An explanation is supported from below if we can deduce and verifyobservable facts from the hypothesis over and above the fact that thehypothesis is intended to explain It must have ‘‘excess explanatorypower.’’ In the case of the Broadway shows, we would expect fewerstanding ovations in shows whose prices for some reason have not gone

up.12

Also, we would expect fewer standing ovations if large numbers oftickets to a show are sold to firms and given by them to their employees.(This would count as a ‘‘novel fact.’’) Even if these tickets are expensive,the spectators have not paid for them out of their own pocket andhence do not need to tell themselves that they are getting their money’sworth

An explanation is supported from above if the explanatory hypothesiscan be deduced from a more general theory.13

In the present case, theexplanatory proposition is a specification of the theory of cognitivedissonance proposed by Leon Festinger The theory says that when aperson experiences an internal inconsistency or dissonance among herbeliefs and values, we can expect some kind of mental readjustment thatwill eliminate or reduce the dissonance Typically, the adjustment willchoose the path of least resistance A person who has spent seventy-fivedollars to see a show that turns out to be bad cannot easily make herselfbelieve that she paid less than that amount It is easier to persuade herselfthat the show was in fact quite good

Although not without problems, the theory of cognitive dissonance ispretty well supported Some of the support is from cases that are verydifferent from the one we are considering here, as when a person who hasjust bought a car avidly seeks out ads for that very brand of car, to bolsterhis conviction that he made a good decision Some of the support arisesfrom quite similar cases, as when the painful and humiliating initiation

12

We would not necessarily expect fewer people to rise to their feet in the cheaper sections They might feel foolish sitting when others are rising; also, they might have to get up to see the actors who would otherwise be blocked from view by those standing in front of them.

13

More accurately: if it is a specification of a more general theory The relation between a general theory and a specific explanatory hypothesis is rarely a deductive one For one thing, there may be some slack in the theory itself (see Chapter 2 ) For another, a given theory can usually be operationalized in many different ways.

rituals of college fraternities and sororities induce strong feelings ofloyalty I am not saying that people would consciously tell themselves,

‘‘Because I suffered so much to join this group, it must be a good group

to belong to.’’ The mechanism by which the suffering induces loyaltymust be an unconscious one

An explanation receives lateral support if we can think of and thenrefute alternative explanations that also pass the minimal test Perhapsthere are more standing ovations because today’s audiences, arriving inbusloads from New Jersey, are less sophisticated than the traditionalaudience of blase´ New York denizens Or perhaps it is because shows arebetter than they used to be For each of these alternatives, we must think

of and then disconfirm additional facts that would obtain if they werecorrect If standing ovations are more frequent because audiences aremore impressionable, we would expect them also to have been frequent

in out-of-town performances twenty years ago If shows are better thanthey used to be, we would expect this to be reflected in how well they arereviewed and how long they play before folding

In this procedure, the advocate for the original hypotheses also has to

be the devil’s advocate One has consistently to think against oneself – tomake matters as difficult for oneself as one can We should select thestrongest and most plausible alternative rival explanations, rather thanaccounts that can easily be refuted For similar reasons, when seeking todemonstrate the excess explanatory power of the hypothesis, we shouldtry to deduce and confirm implications that are novel, counterintuitive,and as different from the original explanandum as possible These twocriteria – refuting the most plausible alternatives and generating novelfacts – are decisive for the credibility of an explanation Support fromabove helps but can never be decisive In the long run it is the theory that

is supported by the successful explanations it generates, not the other wayaround Emilio Segre`, a Nobel Prize winner in physics, said that somewinners confer honor on the prize whereas others derive honor from it.The latter are, however, parasitic on the former Similarly, a theory isparasitic on the number of successful explanations it generates If it isable to confer support on a given explanation, it is only because it hasreceived support from earlier explanations

What Explanation Is Not

Statements that purport to explain an event must be distinguished fromseven other types of statement

First, causal explanations must be distinguished from true causalstatements To cite a cause is not enough: the causal mechanism must also

be provided, or at least suggested In everyday language, in good novels,

in good historical writings, and in many social scientific analyses, themechanism is not explicitly cited Instead, it is suggested by the way inwhich the cause is described Any given event can be described in manyways In (good) narrative explanations, it is tacitly presupposed that onlycausally relevant features of the event are used to identify it If told that aperson died as a result of having eaten rotten food, we assume that themechanism was food poisoning If told that he died as a result of eatingfood to which he was allergic, we assume that the mechanism was anallergic reaction Suppose now that he actually died because of foodpoisoning, but that he was also allergic to the food in question, lobster

To say that he died because he ate food to which he had an allergy would

be true, but misleading To say that he died because he ate lobster would

be true, but uninformative It would suggest no causal mechanism at alland be consistent with many, such as that he was killed by someone whohad taken an oath to kill the next lobster eater he observed

Second, causal explanations must be distinguished from statementsabout correlations Sometimes, we are in a position to say that an event of

a certain type is invariably or usually followed by an event of anotherkind This does not allow us to say that events of the first type causeevents of the second, because there is another possibility: the two might

be common effects of a third event In his Life of Johnson, Boswell reportsthat a certain Macaulay, although ‘‘with a prejudice against prejudice,’’affirmed that when a ship arrived at St Kilda in the Hebrides, ‘‘all theinhabitants are seized with a cold.’’ While some offered a causal expla-nation of this (alleged) fact, a correspondent of Boswell’s informed himthat ‘‘the situation of St Kilda renders a North-East Wind indispensablebefore a stranger can land The wind, not the stranger, occasions an

epidemick cold.’’ Or consider the finding that children in contestedcustody cases are more disturbed than children whose parents havereached a private custody agreement It could be that the custody disputeitself explains the difference, by causing pain and guilt in the children Itcould also be, however, that custody disputes are more likely to occurwhen the parents are bitterly hostile toward each other and that children

of two such parents tend to be disturbed To distinguish between the twointerpretations, we would have to measure suffering before and after thedivorce A third possibility is canvassed later

Here is a more complex example, my favorite example, in fact, of thiskind of ambiguity In Democracy in America, Alexis de Tocquevillediscussed the alleged causal connection between marrying for love andhaving an unhappy marriage He points out that this connection obtainsonly in societies in which such marriages are the exception and arrangedmarriages the rule Only stubborn people will go against the current, andtwo stubborn persons are not likely to have a very happy marriage.14

Inaddition, people who go against the current are treated badly by theirmore conformist peers, inducing bitterness and unhappiness Of thesearguments, the first rests on a noncausal correlation, due to a ‘‘thirdfactor,’’ between marrying for love and unhappiness The second points

to a true causal connection, but not the one that the critics of lovemarriages to whom Tocqueville addressed his argument had in mind.Marrying for love causes unhappiness only in a context where thispractice is exceptional Biologists often refer to such effects as ‘‘frequencydependent.’’15

In addition to the ‘‘third-factor’’ problem, correlation may leave usuncertain about the direction of causality Consider an old joke:

Psychologist: You should be kind to Johnny He comes from a broken

home.

Teacher: I’m not surprised Johnny could break any home.

Or as the comedian Sam Levinson said, ‘‘Insanity is hereditary Youcan get it from your children.’’ The implication is that a disturbed childmay cause the parents to divorce rather than that a divorce causesthe disturbance Similarly, a negative correlation between how much theparents know about what their adolescent children are doing and thechildren’s tendency to get into trouble need not show that parentalmonitoring works, but only that teenagers intent on getting into troubleare unlikely to keep their parents informed about what they are doing.Third, causal explanations must be distinguished from statementsabout necessitation To explain an event is to give an account of why ithappened as it happened That it might also have happened in someother way, and would have happened in some other way had it nothappened the way it did, does not provide an answer to the samequestion Consider a person who suffers from cancer of the pancreas,which is certain to kill her within a year When the pain becomesunendurable, she kills herself To explain why she died within a certainperiod, it is pointless to say that she had to die in that period because shehad cancer.16

If all we know about the case are the onset of cancer, thelimited life span of persons with that type of cancer, and the death ofthe person, it is plausible to infer that she died because of the cancer Wehave the earlier event and a causal mechanism sufficient to bring aboutthe later event But the mechanism is not necessary: it could be pre-empted by another (In the example the preempting cause is itself aneffect of the preempted cause, but this need not be the case; she mightalso die in a car accident.) To find out what actually happened, we need

16

James Fitzjames Stephen writes that ‘‘the law is perfectly clear that, if by reason of [an] assault [a man] died in the spring of a disease which must have killed him, say, in the summer, the assault was the cause of his death.’’

more finely grained knowledge The quest never ends: right up to the lastsecond, some other cause could preempt the cancer.17

Statements about necessitation are sometimes called ‘‘structuralexplanations.’’ Tocqueville’s analysis of the French Revolution is anexample In his published book on the topic, he cites a number of eventsand trends from the fifteenth century to the 1780s and asserts that therevolution, against this background, was ‘‘inevitable.’’ By this he prob-ably meant (1) that any number of small or medium-sized events wouldhave been sufficient to trigger it and (2) that it was a virtual certainty thatsome triggering events would occur, although not necessarily the onesthat actually did happen or when they did happen He also seems toargue (3) that after 1750 or perhaps 1770 there was nothing anyone couldhave done to prevent the revolution Although Tocqueville left notes for

a second volume in which he intended to account for the revolution as itdid happen, one might argue that if he successfully established (1), (2),and (3), there was no need to take this further step The problem withthis line of reasoning is that in many interesting social-science questions(and in contrast to the cancer example), claims such as (1), (2), and (3) arevery hard to establish by methods untainted by hindsight.18

A strongerargument can be made when similar events happen independently ofeach other at the same time, suggesting that they were ‘‘in the air.’’ Thestudy of simultaneous discoveries in science provides an example.Fourth, causal explanation must be distinguished from storytelling Agenuine explanation accounts for what happened, as it happened To tell

a story is to account for what happened as it might have happened (andperhaps did happen) I have just argued that scientific explanations differfrom accounts of what had to happen I am now saying that they also

17

Causal preemption should be distinguished from causal overdetermination The latter is illustrated

by a person’s being hit simultaneously by two bullets, each of which would have been sufficient to kill her The former is illustrated by a person’s being killed by one bullet, preempting the operation

of another fired a few seconds later.

18

The American Revolution is perhaps a more plausible candidate for a structural explanation An acute neutral observer such as the French minister Choiseul observed as early as 1765 that the independence of the American colonies was inevitable For a detached French commentator such as Raymond Aron, the independence of Algeria was also a foregone conclusion well before it came about The French Revolution is more akin to the collapse of Communism – inevitable mainly in hindsight.

differ from accounts of what may have happened The point may seemtrivial, or strange Why would anyone want to come up with a purelyconjectural account of an event? Is there any place in science for spec-ulations of this sort? The answer is yes – but their place must not beconfused with that of explanation.

Storytelling can suggest new, parsimonious explanations Suppose thatsomeone asserts that self-sacrificing or helping behavior is conclusiveproof that not all action is self-interested, and that emotional behavior isconclusive proof that not all action is rational One might conclude thatthere are three irreducibly different forms of behavior: rational andselfish, rational and nonselfish, and irrational The drive for parsimonythat characterizes good science should lead us to question this view.Might it not be the case that when people help others it is because theyexpect reciprocation, and that when they become angry it is because thathelps them to get their way? By telling a story about how rational self-interest might generate altruistic and emotional behavior, one cantransform an issue from a philosophical one into one that is amenable

As long as the model provides predictions with a good fit with theobserved behavior, we are entitled (it is claimed) to assume that agentsact ‘‘as if ’’ they are rational This is the operationalist or instrumentalistview of explanation, which originated in physics and was later adopted

19

In this particular case, the just-so stories happen to be false, since people also help others in one-shot interactions and getting angry may cause others to refrain from interacting with them.

by Milton Friedman for the social sciences The reason, it is claimed, wecan assume that a good billiards player knows the law of physics and cancarry out complex calculations in his head is that this assumption enables

us to predict and explain his behavior with great accuracy To askwhether the assumption is true is to miss the point

This argument may be valid in some situations, in which the agentscan learn by trial and error over time It is valid, however, preciselybecause we can point to a mechanism that brings about nonintentionallythe same outcome that a superrational agent could have calculatedintentionally.20

In the absence of such a mechanism, we might stillaccept the instrumentalist view if the assumption enabled us to predictbehavior with very great accuracy The law of gravitation seemed mys-terious for a long time, as it seemed to be based on the unintelligible idea

of action at a distance Yet because it made possible predictions that wereaccurate to many decimal points, Newton’s theory was uncontroversiallyaccepted untill the advent of the theory of general relativity The mys-terious workings of quantum mechanics are also accepted, albeit notalways without qualms, because they allow for predictions with evenmore incredible accuracy

Rational-choice social science can rely on neither of these two ports There is no general nonintentional mechanism that can simulate ormimic rationality Reinforcement learning (Chapter 16) may do it insome cases, although in others it produces systematic deviations fromrationality Some kind of social analog to natural selection might do it inother cases, at least roughly, if the rate of change of the environment isless than the speed of adjustment (Chapter17) In one-shot situations or

sup-in rapidly changsup-ing environments, I do not know of any mechanism thatwould simulate rationality At the same time, the empirical support forrational-choice explanations of complex phenomena tends to be quiteweak This is of course a sweeping statement Rather than having toexplain what I mean by ‘‘weak,’’ let me simply point to the high level of

20

At this first occurrence in the book of the word ‘‘agent’’ it may be worthwhile to note that many scholars prefer ‘‘actor.’’ Perhaps economists think in terms of agents, sociologists in terms of actors Although it does not really matter which term we use, I prefer ‘‘agent’’ because it suggests agency;

‘‘actor,’’ by contrast, suggests an audience that may or may not be present.