the origins of shared intuitions of justice

Work In Progress – Do Not Quote December 18, 2006

Colin S. Diver Professor of Law, University of Pennsylvania. The authors are*

indebted to Michael Orchowski, Alex Shaw, Peter DeScioli, Lindsay Suttenberg, Martha K.

Presley, Shang Cao, and Susan Eisenberg for invaluable research assistance.

Assistant Professor of Psychology, University of Pennsylvania.**

Professor of Law, Professor of Biological Sciences, and FedEx Research Professor***

of Law, Vanderbilt University.

1

THE ORIGINS OF

SHARED INTUITIONS OF JUSTICE

Paul H. Robinson, Robert Kurzban, and Owen D. Jones* ** ***

Abstract

Contrary to the common wisdom among criminal law scholars, theempirical evidence reveals that people's intuitions of justice are often specific,nuanced, and widely shared. Indeed, with regard to the core harms and evils towhich criminal law addresses itself – physical aggression, takings withoutconsent, and deception in transactions – the shared intuitions are stunninglyconsistent, across cultures as well as demographics. It is puzzling thatjudgments of moral blameworthiness, which seem so complex and subjective,reflect such a remarkable consensus. What could explain this striking result?

The authors theorize that one explanation may be an evolvedpredisposition toward these shared intuitions of justice, arising from theadvantages that they provided, including stability, predictability, and thefacilitation of beneficial exchange – the cornerstones to cooperative action andits accompanying survival benefits. Recent studies in animal behavior and brainscience are consistent with this hypothesis, suggesting that moral judgment-making not only has biological underpinnings, but also reflects the effects ofevolutionary processes on the distinctly human mind. Similarly, the childdevelopment literature reveals predictable stages in the development of moraljudgment within each individual, from infant through adult, that are universalacross all demographics and cultures.

The current evidence does not preclude alternative explanations entirely.Shared views of justice might arise, for example, through general sociallearning. However, a social learning explanation faces a variety of difficulties.It assumes individuals will adopt norms good for the group at the expense of self-interest. It assumes an undemonstrated human capacity to assess extremelycomplex issues, such as what will be an efficient norm. It predicts the existenceof only differences in views of justice, to reflect the wide differences amonggroups, their situations, and cultures producing different efficient norms and


2

different effectiveness in teaching the same. It is inconsistent with thedevelopmental data that show intuitions of justice appearing early, before sociallearning of such complexity is possible, and that show identical intuitions arisingat parallel times despite cultural differences that would affect social learning.And, finally, a general social learning explanation predicts views of justice asaccessible, reasoned knowledge, rather than the inaccessible, intuitiveknowledge that we know it commonly to be.

Whatever the correct explanation for the consensus puzzle, intuitions ofjustice seem to be an inherent part of being human and this, in turn, can haveimportant implications for criminal law and criminal justice policy.

____________


3

Table of Contents

I. The Puzzle: Shared Intuitions of Justice. . . . . . . . . . . . . . . . . . . . . . . . . . . 4II. An Evolutionary Explanation.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

A. The Expression of Genes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91. Genetic Influences versus Genetic Determinism. . . . . . . . . 92. The Genetic Code as a Recipe for Development. . . . . . . . 10

B. Evolution as a Process of Solving Survival Problems. . . . . . . . . . 12C. The Evolution of General Moral and Proto-Legal Sentiments. . . 13D. The Evolution of Intuitions of Justice. . . . . . . . . . . . . . . . . . . . . . 15

1. Human Sociality and the Predisposition to Acquire SharedIntuitions of Justice. . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2. Intuitions of Justice as Generating Benefits for Individuals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

III. Empirical Support for an Evolutionary Explanation. . . . . . . . . . . . . . . . 24A. Animal Studies Evidence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24B. Brain Science Evidence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29C. Child Development Evidence.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

1. Predictable Stages of Development. . . . . . . . . . . . . . . . . . 342. Making Subtle and Complex Judgments at an Early Age

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36a. Distinguishing Moral, Conventional, and Prudential

Rules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37b. Judging the Relative Seriousness of Wrongful

Conduct. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39c. Judging Blameworthiness with Factors Beyond

Offense Seriousness. . . . . . . . . . . . . . . . . . . . . 423. Cross-Cultural Developmental Studies.. . . . . . . . . . . . . . . 44

D. Summary of Evidence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46IV. Alternative Explanations: General Social Learning and Efficient Norms

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47A. Spontaneous Social Learning.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

1. Preferring Group Interests Over Individual Interests. . . . . 492. Complexity of Determining Efficient Norms as Beyond

Individual Capacity. . . . . . . . . . . . . . . . . . . . . . . . . . . 50B. Accumulated Social Learning. . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

1. Absence of Variation in Intuitions of Justice Among GroupsDespite Large Differences in Situation and Culture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

2. Inconsistency with Developmental Data. . . . . . . . . . . . . . 533. Intuitional Knowledge as Having Distinct Characteristics

from Learned Knowledge. . . . . . . . . . . . . . . . . . . . . . 544. Difficulties in Teaching Inarticulable Lessons. . . . . . . . . . 56

Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57


See Paul H. Robinson, Competing Conceptions of Modern Desert: Vengeful,1

Deontological, and Empirical [Part II.D&E] (forthcoming 2007).

Id.2

See id. at [Part I.C]; Paul H. Robinson & John M. Darley, The Utility of Desert, 913

NW . U. L REV . 453 (1997).

Michael Tonry, Obsolescence and Immanence in Penal Theory and Policy, 1054

COLUM . L. REV . 1233, 1263 (2005).

4

I. THE PUZZLE: SHARED INTUITIONS OF JUSTICE

The role of justice in assigning criminal liability and punishment has beena matter of long-standing debate. A standard argument against a desertdistributive principle -- that is, against distributing punishment according to anoffender's blameworthiness -- has been that such a concept of "desert" is simplytoo vague and the subject of too much disagreement to operationalize. There1

may be some truth to these complaints when applied to a philosophical notion ofdesert, which has been the traditional basis of the desert school.2

But more recently, a utilitarian-based theory of desert has urged relianceupon an empirical notion of desert, drawn from the community's shared intuitionsof justice rather than from the reasoned concepts of moral philosophy. The3

same objections about vagueness and lack of agreement have been lodged againstthis empirical notion of desert. People's intuitions of justice, it is claimed, aresimply too vague to be relied upon and, in any case, there is too muchdisagreement about what constitutes deserved punishment to be able to constructa workable criminal justice system. It is the common wisdom that littleagreement exists among people's intuitions of justice.

[E]ven assuming retribution in distribution is appropriate, there is aclassic epistemological problem. How do we know how much censure,or "deserved punishment," a particular wrongdoer absolutely deserves?God may know, but as countless sentencing exercises have shown,peoples' intuitions about individual cases vary widely.4

There is . . . reason to doubt that anything like a consensus exists on theseriousness of criminal conduct. While there may be some agreement onrelative levels of harm, there appears to be great variation in perceptionsof the absolute magnitude of harm represented by various criminal acts,


John Monahan, The Case for Prediction in the Modified Desert Model of Criminal5

Sentencing, 5 INT 'L J. L. & PSYCHIATRY 103, 105 (1982). Other authors have made similar

arguments:

[Desert theorists argue that] we can work out one single, linear ordering of

crimes, from least to most "serious." Yet that scarcely seems a credible assumption.

Try, for instance, to rank the following crimes in order of their "seriousness": attempted

residential burglary, trading stock on inside information, negligent vehicular homicide,

bribing a mine-safety inspector, possessing an ounce of cocaine, and burning a cross on

the lawn of black newcomers to a previously all-white neighborhood. To view this

motley assortment along a single dimension of "seriousness" would seem no less

difficult than to perceive the inner logic behind the apocryphal Chinese encyclopedist

of Jorge Luis Borges's imagination.

David Dolinko, Three Mistakes of Retributivism , 39 UCLA L. REV . 1623, 1638-39 (1992).

Similarly, van den Haag argues:

[Desert theorists moreover appear] to believe that the comparative seriousness of crimes

can be determined in all cases. Not so. Comparative seriousness can be determined

only for some crimes, and it does not fully determine the comparative punishment

deserved. If rape is a crime and murder is a crime, rape-murder must be more serious

than either. Does rape-murder deserve the sum of the punishments meted out for rape

and for murder? More? Less? Even when crimes are nearly homogeneous, assigning

seriousness is arbitrary: Is rape more serious than assault with a deadly weapon? Is

burglary more serious than fraud when fraud does more harm? What about mishandling

toxic waste? Ordinal determinations of seriousness become altogether arbitrary when

the seriousness of heterogeneous crimes must be compared.

Earnest van den Haag, Punishment: Desert and Crime Control, 85 M ICH . L. REV . 1250, 1254

(1987).

5

and in either the relative or absolute level of culpability represented byvarious criminal actors.5

But in this respect the common wisdom is simply wrong. Part of theconfusion here arises from the failure to distinguish two distinct judgments:setting the endpoint of the punishment continuum and, once that endpoint hasbeen set, ordinally ranking all cases along that punishment continuum. The firstof these judgments is not really a distributive judgment, at least in the sense thatit is not part of the adjudication process. Rather, every society must decide whatpunishment it will allow for its most egregious case, be it the death penalty or lifeimprisonment or 15 years. Once that endpoint is set, the challenge for theadjudication system is to determine who should be punished and how muchpunishment each should be given. That process of distributing punishmentrequires only an ordinal ranking of offenders according to their blameworthiness.The end result is a specific amount of punishment, but that amount of punishmentis not selected because there is some magical connection between that offenseand that amount of punishment. Rather, it is selected because, given the large


For general discussion of these matters, see Robinson, Competing Conceptions, supra6

note 1 at [Part I.D.1].

For a general discussion of these matters, see Paul H. Robinson & Robert Kurzban,7

Concordance & Conflict in Intuitions of Justice, M INN . L. REV [Parts I-III] (forthcoming 2007).

Id. at Parts I & II.8

Id. at Part I.9

Id. at Part II.A.10

Id. at Part II.B.11

Id.12

Id. at 140.13

Id. at 141-143 (particularly Table 12). See also id. at 135-48 (discussing variability14

deriving from differences in views regarding how particular acts should be controlled or

punished). People from different cultures might share the intuition that an act is wrong, and even

acts’ relative seriousness, but differ in how punishment should be imposed, whether by the state,

family, or some other source. This discussion highlights the importance of assessing intuitions

(continued...)

6

number of cases with distinguishable blameworthiness and the limited range ofpunishment, that specific amount of punishment is required to set the offenderin his appropriate ordinal rank according to his relative blameworthiness. If thepunishment continuum endpoint is changed, the appropriate punishment for eachoffender changes accordingly.6

Empirical studies show broadly-shared intuitions that serious wrongdoingshould be punished and broadly shared intuitions about the relativeblameworthiness of different transgressions. The evidence comes from a widevariety of empirical studies. In some studies subjects were asked to put offenses7

or offense scenarios into one of a set of predetermined categories; in another kindof study, subjects were asked to rank order offenses or offense scenarios; in athird kind of study, subjects were asked to assign numerical values to each of anumber of offenses or offense scenarios.8

The results in all of these studies are consistent. First, subjects displayeda good deal of nuance in the judgments they made. Small changes in facts9

produce large and predictable changes in punishment. Second, subjects indicatedan interest in imposing punishment for serious wrongdoing. Finally, subjects10

indicated a good deal of agreement about the relative amount of punishment thatdifferent offenders deserved. Indeed, the ordinal ranking of different cases11

generally is consistent across demographics, including cultural differencesexamined in cross-cultural studies that replicated domestic studies. Typical of12

the conclusions in these studies, Newman reports that, “it is apparent that therewas considerable agreement as to the amount of punishment appropriate to eachact” and that looking at relative rankings indicates “general agreement in ranks13

across all countries.”14


(...continued)14

regarding seriousness as distinct from preferred punishments meted out by the state. While the

former might be correlated strongly with the latter in some contexts, it will be less so in others.

Id. at Part III.15

Id.16

CHARLES M.M. DE WEERT & NOUD A.W.H. VAN KRUYSBERGEN , ASSIM ILATION:17

CENTRAL AND PERIPHERAL EFFECTS, PERCEPTION 26, 1217-1224 (1997). The images are made

up of black and white splotches. The more white splotches, the brighter the image appears.

Baruch Fischhoff et al., Travel Risks in a Time of Terror: Judgments and Choices,18

24 RISK ANALYSIS 1301 (2004).

"These results unveil significant diversity in the journal quality perceptions among19

groups of economists despite the fact that our sample focused on AEA members. To test the

robustness of this claim, using Kendall’s W we examined the correlation in journal quality

perceptions between any two randomly selected economists in our sample. We found Kendall’s

W for the top ten journals in our rankings to be 0.396, which demonstrates a relatively low level

of agreement among economists. Once we extended this exercise to the top 20 journals in our

rankings, Kendall’s W dropped to only 0.095." Kostas Axarloglou & Vasilis Theoharakis,

Diversity in Economics: An Analysis of Journal Quality Perceptions, 1 J. EUROPEAN ECON . ASSN

1402, 1421 (2003).

7

The striking extent of the agreement on intuitions of justice is illustratedin a recent study that asked subjects to rank-order 24 crime scenario descriptionsaccording to the amount of punishment deserved. Most researchers would15

consider this a quite demanding task, perhaps asking for more concentration andeffort than most subjects are willing or able to provide. Yet the researchersfound that the subjects had little difficulty performing the task. The task was alsoquite complex, requiring subjects to compare the deserved punishment for eachscenario to that for each of the other 23 scenarios. Nonetheless, subjectsdisplayed an astounding level of agreement in the ordinal ranking of thescenarios.

A statistical measure of concordance is found in Kendall's W coefficientof concordance, in which 1.0 indicates perfect agreement and 0.0 indicates noagreement. In the study just described, the Kendall's W was .95 (with p <.001). This is a striking level of agreement. One might expect to get this high16

a Kendall's W if subjects were asked to judge the relative brightness of differentgroupings of spots, for example. When asked to perform more subjective or17

complex comparisons, such as asking travel magazine readers to rank theattractiveness of eight different travel destinations, one gets a Kendall's W of.52. When asking economists to rank the top 20 economics journals according18

to quality, one gets a Kendall's W. of .095.19

How can this be? How can it be that when asking people to perform acomparative task as complex and subjective as assessing the relativeblameworthiness of 24 different offenders, their level of concurrence matches


This extends an existing body of work that addresses the intersections of law,20

behavioral biology, and evolutionary psychology. See, e.g., Margaret Gruter, The Origins of

Legal Behavior, 2 J. SOC. & BIOLOGICAL STRUCTURES 43 (1979); THE SENSE OF JUSTICE:

BIOLOGICAL FOUNDATIONS OF LAW (Roger D. Masters & Margaret Gruter eds., 1992); FRANS

B. M. DE WAAL, GOOD NATURED: THE ORIGINS OF RIGHT AND WRONG IN HUMANS AND OTHER

ANIM ALS (1996); Michael T. McGuire, Moralistic Aggression, Processing Mechanisms, and the

Brain, in GRUTER , supra at 20. Recent work includes Owen D. Jones & Timothy H. Goldsmith,

Law & Behavioral Biology, 105 COLUM . L. REV . 405 (2005); Owen D. Jones, Proprioception,

(continued...)

8

that of comparative tasks much more simple and objective? That is the puzzlethat this Article attempts to explore.

One possible explanation is the effect of evolutionary processes, whichcan generate cognitive and behavioral predispositions that help individuals solvecommonly encountered problems in ancestral environments. Part II explainshow these processes might underlie the phenomenon of shared intuitions ofjustice. Part III examines evidence from animal behavior studies, brain science,and child development studies that are consistent with the evolutionaryexplanation. Part IV considers an alternative explanation: that shared intuitionsof justice might arise through general social learning if particular intuitions ofjustice provide efficient norms for group functioning, which seems likely. Butthe social learning explanation has a variety of difficulties. To note just one, thewide diversity in the life experiences of individuals, and in the social context andstructure of groups, one would expect a social learning process to produce farmore than the observed variations in intuitions of justice. On present evidence,we conclude, the evolutionary explanation of shared intuitions of justice seemsmore plausible than a universal social learning explanation.

II. AN EVOLUTIONARY EXPLANATION

Part I paints a striking picture of human intuitions of justice: Acrossdemographics, even across cultures, humans share nuanced intuitions about whatconstitutes serious wrongdoing, that serious wrongdoing should be punished, andabout the relative blameworthiness of offenders. How can matters of suchnuance and complexity show this degree of agreement despite dramaticdifferences in individuals’ social and cultural context and upbringing? How canit be that persons of dramatically different religious, economic, and educationalbackgrounds and of different races, ages, and genders share the same intuitionsof justice? We suggest that one explanation for this homogeneity of humanintuitions of justice derives from that which all humans share by virtue of beinghuman: their unique evolutionary history and resulting human nature.20


(...continued)20

Non-Law, and Biolegal History, 53 FLA . L. REV. 831 (2001); Owen D. Jones, Time-Shifted

Rationality and the Law of Law’s Leverage: Behavioral Economics Meets Behavioral Biology,

95 NW . U. L. REV . 1141 (2001); Owen D. Jones, Evolutionary Psychology and the Law, in THE

HANDBOOK OF EVOLUTIONARY PSYCHOLOGY 953 (David Buss ed. 2005); Dennis Krebs, The

Evolution of Morality, in THE HANDBOOK OF EVOLUTIONARY PSYCHOLOGY, id. Jones maintains

a bibliography of works in this area that can be accessed through the webpage of the Society for

Evolutionary Analysis in Law (SEAL) at <http://www.sealsite.org>.

For further introduction to core principles of evolutionary biology, see Jones &21

Goldsmith, Law and Behavioral Biology, supra note 20 (including sources there cited); and Owen

D. Jones, Evolutionary Analysis in Law: An Introduction and Application to Child Abuse, 75 N.

C. L. REV . 1117, 1127–57 (1997). Useful textbooks include TIMOTHY H. GOLDSM ITH & W ILLIAM

F. ZIMM ERMAN , BIOLOGY, EVOLUTION , AND HUMAN NATURE (2001); JOHN ALCOCK , ANIM AL

BEHAVIOR (8 ed. 2005); and JOHN CARTW RIGHT, EVOLUTION AND HUMAN BEHAVIOR (2000).th

9

In this Part, we first briefly offer a primer for the role of genes in thedevelopment of an individual’s traits, then describe how evolutionary processes,most notably natural selection, operate to preserve both anatomical andbehavioral traits that helped to solve challenges regularly encountered byancestors over evolutionary time. We then explain how the process of solvingthose challenges can lead to a variety of common psychological preferences inhumans, including some general moral preferences. And we subsequentlydiscuss how, more specifically, those evolutionary processes can result in sharedintuitions of justice in the context of punishments for wrongdoing.

A. The Expression of Genes

It is relatively uncontroversial that evolutionary processes, genes, andenvironments all influence human predispositions and resultant behaviors. Butprecisely how those influences interact, on which predispositions and behaviors,and with what consequences and implications, represent important questionsunder continuing investigation. Moreover, even what is known is oftenmisunderstood. Consequently, we start with a brief primer on the generalprinciples by which evolutionary processes, genes, and environments can affectbehaviors, and how they cannot.21

1. Genetic Influences versus Genetic Determinism. Geneticexplanations do not imply that a trait is “hard wired” or “inflexible.” All humantraits exist because of the interaction between an individual’s set of genes – asubset of all possible human genes – and the unique combination of


See, e.g., MARY JANE WEST-EBERHARD , DEVELOPMENTAL PLASTICITY AND22

EVOLUTION 65 (2003).

Id.23

See generally STEVEN PINKER, THE BLANK SLATE: THE MODERN DENIAL OF HUMAN24

NATURE 76, 113 (2002); JEFFREY L. ELMAN, ET AL, RETHINKING INNATENESS: A CONNECTIONIST

PERSPECTIVE ON DEVELOPM ENT (NEURAL NETW ORKS AND CONNECTIONIST MODELING) (1977);

MATT R IDLEY , NATURE V IA NURTURE: GENES, EXPERIENCE, AND WHAT MAKES US HUMAN

(2003); Jones & Goldsmith, Law & Behavioral Biology, supra note 20 at 485-88.

This still-common view was given powerful voice by John Locke. See, e.g., JOHN25

LOCKE, AN ESSAY CONCERNING HUMAN UNDERSTANDING 121 (John Maynard Hutchins ed.

1952).

PINKER, BLANK SLATE, supra note 24. 26

See RICHARD DAW KINS, THE BLIND WATCHM AKER 52 (1996) ("The genes, as we27

shall see, are more like a recipe than like a blueprint . . . .").

For an overview of recent developments in the field of evolutionary development, see28

SEAN B. CARROLL, ENDLESS FORM S MOST BEAUTIFUL: THE NEW SCIENCE OF EVO DEVO AND THE

MAKING OF THE ANIM AL K INGDOM (2005)

10

environmental conditions a person encounters while developing. The effect of22

any given gene depends on many factors, such as the presence or absence ofother genes, the body’s internal environment (e.g. cells), the individual’s externalenvironment, and so forth. Gene expression, therefore, is extremely complex,23

and is always influenced by the (broadly construed) environment the geneencounters.

It is clear that changing genes or changing environments changes traits,whether physical or psychological. And there is consequently widespreadconsensus that environmental determinism and genetic determinism are bothequally incorrect. Human minds are neither a “blank slate” to be written upon24

by experience, nor completely the product of genetics. They are, rather, the25

products of complex interactions between genes and their developmentalenvironments, all under the relentless influence of evolutionary pressures thathave favored some patterns of interactions over others.26

2. The Genetic Code as a Recipe for Development. Although“blueprint” is a tempting metaphor for the genetic code, a better metaphor is“recipe.” Without the presence of the appropriate environment (a cook and27

kitchen), genes (recipes) have no important causal consequences. In the realmof biological development, this process, the dynamic interplay between genesand environment, is exceedingly complex. Biochemical interactions at the28

cellular level and the building of connections in the brain are vastly intricate andstill little understood. However, this complexity does not imply that noprediction can be made about an individual's development, nor does it imply that


John Tooby & Leda Cosmides, The Psychological Foundations of Culture, in THE29

ADAPTED M IND at 83-84 (Jerome. Barkow et al., eds., 1992). For this reason, attempting to

dichotomize important human traits as only culturally or genetically influenced is misguided.

John Tooby et al., The Second Law of Thermodynamics is the First Law of30

Psychology, 129 PSYCHOL. BULL. 858, 863-864 (2003).

HENRY GRAY , GRAY’S ANATOM Y (1858) (Susan Standring ed., 2004). 31

See NOAM CHOMSKY , KNOW LEDGE OF LANGUAGE: ITS ORIGIN , NATURE, AND USE32

3 (1986) (discussing “Language Acquisition Device”).

11

development is so complex as to be essentially random. Rather, evenpsychological traits that are “learned,” such as language, are brought about by apredictable interaction between the individual's genes and the environment inwhich the individual develops.29

The genes shared by humans cause each person's structures to be built ina predictable and systematic way. Consider the human visual system. Everyhuman visual system receives different stimuli because every baby is born intoa unique perceptual world with different objects, people, animals, and so forth.However, the mature visual system for all normally developing adults isessentially the same. When a trait such as this emerges across a wide range ofenvironments, it is said to be “reliably developing.” Human genes ensure thateven these complex elements of the body are “reincarnated” each generation.30

Because this developmental process for human psychological traits is complex,it is easier to see such trait “reincarnation” in physical traits. Consider Gray’sAnatomy. A book about "the human body" is possible only because all humans31

develop in the same reliable, systematic fashion, despite the obviouscomplexities of that process. And, important to our purposes here, evenpsychological traits, such as the capacity to learn a language, follow a similarpattern.

Of course, development is not necessarily uniform across all individuals.In the context of language learning, the developing child has a special capacityto acquire the ability to understand and produce language that the child hearsspoken. But all normal humans have this special language-acquisition32

mechanism. Other species do not. A cat exposed to the same linguisticenvironment learns to understand few words, if any. We will argue below that,just as they do with language, people have a specific ability that allows them toacquire intuitions of justice. The intuitions that result from this acquisitionmechanism are not “innate” in a naïve sense of the term, such as “present atbirth.” Instead, we argue that there exists a predisposition by which peopleacquire these intuitions over the course of development both as a child and anadult.


Gary F. Marcus & Simon E. Fisher, FOXP2 in Focus: What Can Genes Tell Us About33

Speech and Language?, 7 TRENDS IN COG . SCI. 257 (2003).

STEVEN PINKER, THE LANGUAGE INSTINCT: HOW THE M IND CREATES LANGUAGE 10434

(1994).

This generations estimate is commonly used. See, e.g., Colm O’hUigin, et al., The35

Implications of Intergenic Polymorphism for Major Histocompatibility Complex Evolution, 156

GENETICS 867, 873 (2000). It corresponds to a period of roughly five million years, the estimated

time since human and chimpanzee ancestors diverged. Feng-Chi Chen & Wen-Hsiung Li,

Genomic Divergences Between Humans and Other Hominoids and the Effective Population Size

of the Common Ancestor of Humans and Chimpanzees, 68 AM . J. HUMAN GENETICS 444, 444

(2001).

Obviously, there is more to this issue than can be covered here. Processes that lead36

to evolutionary change include mutation, genetic drift, gene flow, and natural selection (including

sexual selection and kin selection). For a brief description, see Jones & Goldsmith, Law and

Behavioral Biology, supra note 20, at 438-29. Natural selection is commonly considered the one

process that can generate both increases in complexity and a close fit between the features of an

organism and its environment. Natural selection refers to the result of three conditions: a)

(continued...)

12

Which language a child acquires depends on the linguistic environment,but the genes that allow language acquisition play as important a role as thepresence of linguistic sounds. All normally developing humans in Topeka33

acquire the syntax and vocabulary of English, while those in Tokyo learnJapanese. In this instance, the environmental variation leads to differences in thedeveloped adult. In short, the effect of experience varies from one trait toanother, with some traits being relatively constant across environments (vision)and others being different across environments (language).

The reason for the differential effect of environment can be traced backto the logic of evolved adaptations. Evolution settled on a way to construct thevisual system that does not depend on what is in one’s visual world because, byvirtue of the universal principles of optics, the same solution to vision applies inevery setting. This is not the case for language, however; the best language tolearn depends on the language others around you are speaking. For this reason,evolution left specific elements of language acquisition to vary across groups.34

B. Evolution as a Process of Solving Survival Problems

Most of the significant genes that distinguish us as “human” are ones thatwere selected over the course of 250,000 generations because their effects led toreproductive success. More specifically, current human genes are in large35

measure the ones that caused individuals to survive and reproduce better, onaverage, than did alternative genes. The origin of all complex, functional humantraits, whether physical or psychological, is thus evolution by natural selection.36


(...continued)36

variation; b) heritability; and c) differential reproduction of individuals as a function of their

heritable variations. The process was famously identified and described in CHARLES DARW IN ,

THE ORIGIN OF SPECIES BY MEANS OF NATURAL SELECTION (1859) (Penguin Classics ed. 1985).

John Tooby & Leda Cosmides, The Past Explains The Present: Emotional37

Adaptations and the Structure of Ancestral Environments, 11 ETHOL. & SOCIOBIO . 375, 398

(1990).

13

The success of one gene over another is generally the result of its contributionto solving problems that humans faced, such as finding nutritious food, choosinggood mates, raising a family, avoiding predation, and so forth. Genes thatcontributed to success in confronting these problems were selected (meaningthey tended to appear in increasingly large percentages in successivegenerations). Thus, the traits these genes produced can be understood as havingfunctions – means by which these problems were solved.

Traits are typically specific in function. That is, to guide the organismtoward positive reproductive outcomes, specific traits are needed to solvedifferent problems. This is evident in physical organ systems, for example, asthe heart is good at pumping blood but bad at filtering it for toxins while thereverse is true of the liver. And the human visual system is good at using lightto learn about what is out in the world, but it is not well equipped to acquirelanguage. Specificity of function is the hallmark of natural selection, becausemechanisms such as the heart and the visual system tend to work best when theyhelp to solve a narrow task, in the same way that tools work best when they aredesigned for one particular job. The same holds for many psychological traits.In sum, much of human psychology consists of evolved, reliably developingtraits that have functions associated with the adaptive problems faced by ourancestors. Modern human minds contain psychological adaptations thatsuccessfully solved our ancestors' adaptive problems.37

C. The Evolution of General Moral and Proto-Legal Sentiments

No one yet knows the full suite of psychological adaptations that humanminds share. And given the complexity of the interactions between genes andculture, it is often unclear what can be said with confidence about how the humanbrain works. Yet in order to bridge from the general notion of evolvedpsychological adaptations to our specific suggestion that shared intuitions ofjustice reflect evolved adaptations, we need to lay a brief foundation for the ideathat preferences – and indeed loosely “moral” sentiments – can evolve.Fortunately, there is ample reason, at both theoretical and empirical levels, to


See generally Goldsmith & Zimmerman, Biology, supra note 21. 38

See generally FRANS B. M. DE WAAL, GOOD NATURED: THE ORIGINS OF RIGHT AND39

WRONG IN HUMANS AND OTHER ANIM ALS (1996); ROBERT WRIGHT, THE MORAL ANIM AL

(1994); INVESTIGATING THE BIOLOGICAL FOUNDATIONS OF HUMAN MORALITY (JAM ES P. HURD

ED ., 1996); THE HANDBOOK OF EVOLUTIONARY PSYCHOLOGY (David Buss ed. 2005); and

particularly Krebs, The Evolution of Morality, supra note 20 at 747.

See, e.g., OSTRACISM : A SOCIAL AND BIOLOGICAL PHENOMENON (Margaret Gruter40

& Roger D. Masters eds., 1986).

See generally EDVARD WESTERM ARCK , THE H ISTORY OF HUMAN MARRIAGE (1891);41

Irene Bevc & Irwin Silverman, Early Proximity and Intimacy BETWEEN SIBLINGS AND INCESTUOUS

BEHAVIOR: A TEST OF THE WESTERMARCK THEORY, 14 ETHOL. & SOCIOBIO . 171 (1993).

14

believe that human minds share a wide variety of preferences that range from thehedonic to the moral.

Like most evolved human capacities, the psychological adaptations thatsteer humans through the complex social world have functions that led toreproductive success in the past. The function and concurrent reproductiveadvantages conferred by some human social adaptations are relatively obvious.Few could doubt that our species-typical preferences to sleep at night (onaverage) rather than during the day, to prefer to provision our own children overthe children of other people, or to seek pleasure in sexual activity with others ofour own rather than another species, reflect complex cognitive processes that inturn reflect evolutionary adaptations as well as the cultural overlays that mayadjust the contexts and patterns in which we do so. The evolutionary logic isclear. And other species behave similarly, given the physical and behavioralniches they have evolved to exploit.38

The power of natural selection is sufficiently great, and the competitionfor genetic space within the genome sufficiently intense, that many researchersand theorists believe that a wide set of human psychological and behavioral traitsare, in part, the results of adaptation through natural selection. These include notonly traits relevant to mating and parenting and kinship, but also traits relatingto the particular challenges of group living, which include aggression,competition, cooperation, deception, and moral sentiments enabling evaluationof good and bad behaviors generally. Previous work has suggested that these39

preferences resulted in proto-legal systems of social primate interaction.40

Consider the sentiments leading to judgments and, later, laws surroundingincest. Because of the detrimental effects of inbreeding, evolution appears tohave selected for genes that cause organisms to develop behavioral systems thatlead them away from mating with close genetic relatives. In humans, this41

manifests itself as a shared sense that committing incest is wrong.


See, e.g., Debra Lieberman et al., Does Morality Have a Biological Basis? An42

Empirical Test of the Factors Governing Moral Sentiments Relating to Incest, Proceedings of the

Royal Society of London Series B, 270 BIOLOGICAL SCIENCES 819 (2003). See also Jonathan

Haidt & Matthew Hersh, Sexual Morality: The Cultures and Reasons of Liberals and

Conservatives, 31 J. APPLIED SOC. PSYCHOL. 191 (2001); Jonathan Haidt, The Emotional Dog

and Its Rational Tail: A Social Intuitionist Approach to Moral Judgment, 108 PSYCHOL. REV .

814 (2001). This insistence applies even to cases in which, for example, the consummated sex

act is between adult siblings, fully consensual, one time, and incapable of leading to pregnancy.

Neither a conscious motive to avoid deleterious inbreeding nor an ability to detect the43

degree of genetic relatedness is necessary. Even the inclination in many animals to leave their

natal group upon reaching sexual maturity can serve this function. See, e.g., Tim H. Clutton-

Brock, Female Transfer and Inbreeding Avoidance in Social Mammals, 337, NATURE, 70, 70

(1989). In humans as in many species, for example, natural selection appears to have favored a

predisposition against mating with those who were co-resident opposite-sex siblings during

childhood – which has historically been a reliable proxy for close genetic relatedness.

See WESTERM ARCK , supra note 41, at 189.44

15

The intuition has been shown through experimentation not to derive fromlogical rationales. Experimental subjects insist on the wrongfulness of incest,even if they are unable to articulate principled reasons for why incestuousrelationships are wrong. Incest-avoiding mechanisms serve their evolutionary42

function without awareness of this function on the part of the individualorganism. Such an intuition is highly adaptive, meaning that people who had43

this intuition, as opposed to those not having it, enjoyed greater long-termreproductive success, leaving more offspring who themselves successfullyreproduced. In this case, the means by which this was accomplished was44

directing the individual away from inbreeding and its concurrent genetic costs.Incest provides a clear example of how natural selection can favor the evolutionof intuitions about wrongdoing whose particular functions are relevant toreproductive success.

D. The Evolution of Intuitions of Justice

Given the influence of genes, environments, and evolutionary processeson human problem-solving through evolved behavioral predispositions, includingmoral ones, it is a small but significant logical step to suggest that sharedintuitions of justice, specific to punishments for transgressions of varyingseriousness, also reflect our common evolutionary history. Below we argue thathuman sociality has laid the foundation for an evolved predisposition to acquireshared intuitions of justice and that such intuitions benefit the individuals thatbear them. Further, we argue there is reason to believe that evolution has in


Recently, Marc Hauser suggested how evolution could have led to the human capacity45

for morality. Though similar in locating the origins of justice in the theory of evolution by natural

selection, and similarly using language as an analogy to motivate the analysis, our treatment

differs in its focus on the advantages conferred to individuals in having particular intuitions

regarding punishment of those who commit wrongful acts. Hauser links the intuition that

wrongdoing be punished to intuitions about fairness. MARC D. HAUSER , MORAL M INDS: HOW

NATURE DESIGNED OUR UNIVERSAL SENSE OF RIGHT AND WRONG 108-109 (2006) (“We can

safely assume that these intuitions evolved prior to or during our life as hunter-gatherers . . . In

such small-scale societies, fairness was most likely an effective proxy for judging punishable

acts.”).

Peter J. Richerson & Robert Boyd, The Evolution of Human Ultra-Sociality, in46

INDOCTRINABILITY , IDEOLOGY, AND WARFARE: EVOLUTIONARY PERSPECTIVES 71, 71 (Irenaus

Eibl-Eibisfeldt & Frank Salter eds., 1998).

16

particular contributed to intuitions that physical harm, taking of property, andcheating in exchanges are matters for particular attention and condemnation. 45

1. Human Sociality and the Predisposition to Acquire SharedIntuitions of Justice. The predisposition to acquire human intuitions of justicelikely arose from the deeply and irretrievably social nature of the human species.Humans have been described as “ultrasocial” by anthropologists because of themany ways in which humans interact with one another, particularly in the contextof cooperation between individuals and within groups. Group cooperation is46

rare in the biological world and captures the attention of biologists becauseevolution tends to favor genes that lead to selfish behavior. Humans, however,behave altruistically in many contexts: family relations, friendships, and within-group cooperation. How does such extensive cooperation come to pass? Theevolution of cooperation in humans and other species is the subject of a large


That literature suggests that several causal processes are simultaneously at work. The47

first, known as mutualism, reflects the fact that two can sometimes gain more by acting together

than the sum of their actions alone. The second, known as kin selection, reflects the extent to

which cooperatively furthering the reproductive success of kin increases the prevalence of any

heritable predisposition to do so, since many kin will also carry that predisposition. The third,

known as reciprocal altruism, reflects the fact that heritable predispositions to cooperate over time

can yield mutual benefits, even between unrelated cooperators. On the extent to which these

processes may be supplemented by “strong reciprocity,” see Herb Gintis, Strong Reciprocity and

Human Sociality, 206 J. THEORETICAL BIOL. 169 (2001).

There is debate surrounding why humans evolved to be the extremely social animals that

they are. For purposes of the current discussion, we remain agnostic on this debate and simply

assume that humans are social creatures, have been for many generations, and have many traits

that are adapted for social life. See, e.g., Robert Trivers, The Evolution of Reciprocal Altruism ,

46 Q. REV . BIO. 35, 35-39 (1971) (discussing the relationship between morality and cooperation);

Leda Cosmides, The Logic of Social Exchange: Has Natural Selection Shaped How Humans

Reason?, 31 COGNITION 187, 187-97 (1989); Tooby & Cosmides, Psychological Foundations,

supra note 29 at 19, 20. On the probable origins of social exchange in humans, and the inverse

relationship between social hierarchy and social exchange, see CHRISTOPHER BOEHM ,

H IERARCHY IN THE FOREST: THE EVOLUTION OF EGALITARIAN BEHAVIOR 197 (1999).

For recent empirical work, suggesting that those with a sense of fairness are likely48

to do better in cooperative interactions, see Ernst Fehr & K.M. Schmidt, A Theory of Fairness,

Competition, and Cooperation, 114 Q. J. OF ECON . 817 (1999).

17

theoretical and empirical literature in evolutionary biology, and the general47

outline is relevant to our argument.It has been shown both theoretically and empirically that cooperation can

evolve through several independent but overlapping processes. The one mostrelevant for our immediate purpose concerns the mutually beneficial effects ofreciprocity: if you share with me today in exchange for my sharing with youyesterday, we are both better off than if neither of us share. In social animals,reciprocity can involve such things as alerting other group members when foodhas been discovered, sharing food over time, and supporting a comrade in actionagainst others.

But underlying this rosy picture is a darker shadow. For while it isevident that reciprocators can outperform loners, a cheater (or defector) couldtheoretically outperform both if he were able to regularly take benefits withoutrepaying them. Consequently, an evolutionary arms race ensues in socialanimals between various predispositions toward cooperation and exploitation.In the end, the most successful cooperators are not those who always cooperate,but rather those who cooperate selectively with other cooperators, thusdiscriminating (passively or aggressively) against those who are not reliablepartners in cooperative endeavors. Put another way, effective cooperation48


For a detailed list of these requirements, see Cosmides & Tooby, Psychological49

Foundations, supra note 29.

That is, to the extent randomly arising genetic mutations may have inclined50

individuals toward these behavioral features, natural selection appears to have preserved those

mutations and increased their proportional prevalence within subsequent generations.

Social exchange is the simultaneous or sequential delivery of benefits between two51

individuals for mutual profit.

See generally ROBERT WRIGHT, NONZERO: THE LOGIC OF HUMAN DESTINY (2000).52

Exchange allows non-zero-sum interactions, in which each participant is better off exchanging

than alone. Id., at 26.

18

requires rewarding good behavior and punishing (or at least failing to reward)bad behavior.

For this to work, though, one must be able to evaluate the individual costsand benefits of various kinds of things (such as different objects, acts, andinformation) and similarly evaluate others’ valuation of those objects. One mustin turn be able to discern unfairness, such as an inequitable distribution ofresources or a violation of reciprocity. One must also be able to recognizeindividuals and remember recent interactions with them, so as to discriminate infuture in favor of some and against others. These are precisely the features we49

tend to observe in the most highly social animals.50

Those features are also, we argue, precisely the features that likelyunderlie the functional advantage of the specific learning capacity in humans foracquiring shared intuitions of justice. Humans have a universal and uniquelynuanced propensity for engaging in social exchange. Indeed, gains from social51

exchange form the basis of the modern economy and infiltrate nearly everyaspect of our lives, both in formal markets and in personal relationships. The52

psychology that underpins exchange requires deep intuitions and complexcomputational capacities to operate.

In particular, one critical capacity for successful social exchange is theintuition that one should punish individuals who injure others or cheat in anexchange. If one is engaged in transactions with the same person over time, thenallowing another individual to injure or to cheat with no punishment is aninvitation to exploitation without end. Therefore, to be most successful in socialexchange, one must have the capacity not only to detect but also to punish suchpersons.

This implies that there might have been selection in humans for thecognitive mechanisms designed to detect inequities and, similarly, for thecognitive mechanisms that yield intuitions that motivate the punishment ofpeople who violate the most ancient and fundamentally necessary principles of


For recent discussion, see Herb Gintis, Strong Reciprocity and Human Sociality, 20653

J. THEORETICAL BIOL. 169 (2001).

For suggestive nonhuman evidence, see infra notes 76 - 109. 54

See infra notes 76 - 109. 55

For early thinking on this subject, see Trivers, Reciprocal Altruism , supra note 47.56

See also MATT RIDLEY , THE ORIGINS OF V IRTUE: HUMAN INSTINCTS AND THE EVOLUTION OF

COOPERATION (1996) (tracing, among other things, the extent to which moralistic aggression can

police fairness in social exchanges, and thereby increase the prevalence of reciprocities by

increasing the cost of cheating).

19

social exchange. In other words, the evolutionary history of social exchange53

has likely led to a reliably developing psychological system that is able tocompute when someone has injured or cheated, as well as to a motivation topunish them.54

2. Intuitions of Justice as Generating Benefits for Individuals. Theargument sketched so far may explain why people have intuitions that lead themto believe that people who have harmed or cheated them ought to be punished.However, this does not explain why people have intuitions that punishmentshould be imposed on wrongdoers in transactions in which they have no part.This question -- why people care about third party transactions at all -- is animportant mystery. Most non-human organisms attend very little to transactionsthat occur between unrelated others (though there are some importantexceptions ). Why do humans care so much about these interactions? Our55

analysis follows along these lines:Once individuals possess intuitions about being wronged, detecting when

others have been wronged is possible using the same set of intuitions. This doesnot, itself, explain why individuals should care that someone else has beenwronged; it merely explains the origin of the ability to detect wrongful actstoward others.

What may explain the interest in wrongdoing to others is that individualsthemselves need to make decisions about the people with whom they interact,engage in social exchange, form groups and coalitions and, more generally,socialize. One observing a third party committing a wrong against anotherperson has important information about the perpetrator of the wrong, which maymake the observer less likely to choose that person for social interactions.

But also a puzzle is why people would have the intuition that wrongs toothers should be punished. The evolutionary advantage of this intuition can bedebated, but we advance some ideas that help explain the preference. The logic56

begins with the uncontroversial premise that humans evolved in relatively smallgroups that consisted of people with whom they had a large number of


See generally Boehm, Hierarchy, supra note 47.57

See Richerson & Boyd, supra note 46.58

Ozgur Gurek et al., The Competitive Advantage of Sanctioning Institutions, 31259

SCIENCE 108 (Apr. 2006).

Alexander made similar arguments regarding reputation and the effects on the group60

of having cooperative individuals. RICHARD D. ALEXANDER, THE BIOLOGY OF MORAL SYSTEM S

(1987). And Trivers made this argument explicitly. Robert L. Trivers, SOCIAL EVOLUTION 388

(1985) (“[A] sense of justice has evolved in human beings as the standard against which to

measure the behavior of other people, so as to guard against cheating in reciprocal

relationships.”). See also Krebs, supra note 20; The Evolution of Moral Behavior, in HANDBOOK

OF EVOLUTIONARY PSYCHOLOGY: IDEAS, ISSUES, AND APPLICATIONS 337, 342 (Charles Crawford

& Dennis L. Krebs eds. 1998); Robert Kurzban & Steven Neuberg, Stereotyping and

Discrimination: An Evolutionary Perspective, in EVOLUTIONARY PSYCHOLOGY HANDBOOK ,

supra note 20 at 653.

We have considered the benefits flowing to individuals, as is standard in biology. Some,

proposing “multi-level selection,” argue that considering benefits to the group yields an equal or

superior analysis. See, e.g., DAVID SLOAN W ILSON & ELLIOT SOBER, UNTO OTHERS : THE

EVOLUTION AND PSYCHOLOGY OF UNSELFISH BEHAVIOR (1998). In that view, groups that

effectively cooperated because of widely shared and specific intuitions of justice that benefitted

the group could have outperformed groups without those intuitions, leading over time to more

groups comprised of more people who share those intuitions. Game theoretic analysis suggests

that group-benefitting norms, and punishment through inflicting costs on those who do not follow

such norms, can be a powerful evolutionary force in the adoption, spread, and maintenance of

(continued...)

20

transactions over a sustained period of time. Consider two kinds of groups:57

first, one in which individuals whose acts of unprovoked violence, theft, orcheating in exchanges is left unpunished and, second, one in which these acts arepunished. In the second group, the individuals in the group benefit because theyare less likely to be subject to violence, theft, or cheating. Thus, each person, togain that benefit, has an incentive to take it upon himself to inflict punishmenton those who commit such offenses. Of course, such punishment is potentiallycostly, so the long-term benefits of punishing must outweigh the long-term costsif it is to be selected. And very recent experimental evidence in humans58

suggests that under such circumstances, given a choice, even prior transgressorschoose to be in a group that punishes over one that is punishment-free.59

However, personally punishing wrongdoers need not be the only way toincrease the chances of being in a group in which undesirable acts are punished.One need not inflict punishment one’s self. One can decrease the cost ofinflicting punishment on wrongdoers by supporting or defending those who doinflict the punishment. A third party punisher has a smaller risk of retaliation ifgroup members endorse the punishment meted out by the third party. So, anindividual benefits by having intuitions that support the punishment – by someindividual – of those who transgress.60


(...continued)60

these norms. See, e.g., Herbert Gintis et al, Explaining Altruistic Behavior In Humans, 24 EVO .

& HUM . BEHAV. 153, 154 (2003). See also Peter J. Richerson & Robert Boyd, The Evolution Of

Subjective Commitment To Groups: A Tribal Instincts Hypothesis, in EVOLUTION AND THE

CAPACITY FOR COMM ITM ENT 186-220 (Randolph M. Nesse ed., 2001).

There are multiple ways in which the punishment of wrongdoers can be implemented.61

One way is for the individual who has been harmed to punish. Another way is for individuals not

directly involved to inflict punishment. Among the Huron, for example, the obligation to slay

a murderer reportedly fell upon the kin of the murderer. BRUCE G. TRIGGER, THE CHILDREN OF

AATAENTSIC: A H ISTORY OF THE HURON PEOPLE TO 1660, 59, 60 (1976). And another way is for

institutions to inflict punishment on those who have committed wrongs. This comes the closest

to the way in which modern states impose punishment: through an official criminal justice

system that formally represents all citizens of the state.

This last element, contracts, is worth special consideration. Because of humans’62

abilities to represent abstract costs and benefits, the number of social exchanges that are possible

is much, much larger than in other organisms, who can exchange only narrowly delimited

commodities. If an organism cannot translate how beneficial or costly an object or act is in terms

of another object or act, she cannot make good decisions about what counts as a good exchange.

And as the range of possible exchanges increases, the advantage of the ability to enforce contracts

increases. See Leda Cosmides & John Tooby, Cognitive Adaptations for Social Exchange, in THE

ADAPTED M IND , EVOLUTIONARY PSYCHOLOGY AND THE GENERATION OF CULTURE 163, 177

(Jerome H. Barkow et al., eds. 1992) (discussing the computational requirements of exchange).

See JAM ES Q. W ILSON , THE MORAL SENSE 40 (1997).63

21

If it is true that (1) individuals prefer that wrongdoers be punished and (2)individuals may not want to bear the costs of inflicting punishment, it is notsurprising that means have been developed to satisfy these two preferences,though the means may differ across the centuries and cultures. For example, insome cultures (so-called “big man” societies ) one person has special61

responsibilities and duties in the context of such punishment. In the West,punishment is put into the hands of a system of criminal justice, includinglegislatures, police, judges, and so on. In the small groups in which humansevolved, the marginal benefit of having each individual support punishment forwrongdoings might have had an impact on making the group one in whichtransgressions were reduced, and thus protect an individual’s health, property,and ability to make contracts.62

In short, shared intuitions of justice contribute to the ability of anindividual or group to be able to punish, which in turn provides evolutionaryadvantage to all. In other words, there is an evolutionary advantage to anunderstanding of victimhood and to the concurrent condoning of the punishmentof one who has engaged in wrongful conduct. This is important because the63

imposition of punishment -- the imposition of costs against the wrongdoer -- isitself something that normally would be seen as a wrongful act. To produce


The problem alluded to above, supra note 62, is essentially identical for punishment64

because the means of punishment is often not in the same coin as the offense. If intuitions of

justice imply that a harm should be inflicted on a perpetrator, then computing what constitutes

a harm of appropriate magnitude is crucial. This is a richly debated topic in many modern areas

of law (e.g., drug offenses). However, we believe that there are strong intuitions about relatively

severity of offenses and suggest that the complexity of the psychological systems necessary to

make such comparisons should not be underestimated. The cycle of harm, retaliation, and

retaliation for retaliation can escalate if the original punishment is judged too great, given the

magnitude of the offense.

See Robert Boyd & Peter Richerson, Solving the Puzzle of Human Cooperation, in65

EVOLUTION AND CULTURE 105 (Stephen C. Levinson & Pierre Jaison eds 2005).

See generally ROBERT H. FRANK , PASSIONS W ITHIN REASON: THE STRATEGIC ROLE66

OF THE EM OTIONS (1988).

Dennis Krebs & Kathy Denton, Toward a More Pragmatic Approach to Morality:67

A Critical Evaluation of Kohlberg's Model, 112 PSYCHOL. REV . 629, 642 (2005).

22

conditions that avoid a never-ending spiral of wrongful acts, intuitions of justicemust specify that wrongful acts can be punished without eliciting the furtherintuition that the punishers should themselves be punished.

To be able to distinguish deserved punishment, which is permitted, fromundeserved punishment, which is not, the group obviously must share some senseof what constitutes wrongful conduct and the amount of punishment appropriatein any one case relative to that in other cases. If everyone in a group agrees on64

this -- what constitutes a wrongful act and its relative wrongfulness as comparedto other acts, hence the relative amount of punishment to be permitted withoutthe punisher themselves being subject to punishment -- stability and cooperation,and its consequent benefits to members of the group, can be achieved.65

Contributing to the institutional arrangement for punishment may not bethe only reason that intuitions of justice evolved. Because social exchange ismutually beneficial, participation in exchange is desirable. When there are manysocial interactants in a group, there is competition to be selected as a partner inexchange. We should therefore expect that natural selection would favorpredispositions that signal to others that one is a good candidate for participationin such exchanges. If one shares others’ moral intuitions -- that cheating in anexchange is condemnable, for example -- then one can expect selection to favorthose who signal having such an intuition. Krebs and Denton recently made66

this argument explicitly: “Persuading others that you are a fair, honest, generous,responsible and moral person who will make an attractive exchange partner mayinduce them to bestow benefits on you.”67

The emergence of intuitions of justice has a self-sustaining character.Once the intuitions exist in a group, actions that violate others’ intuitions invitecensure and punishment. Once the intuitions of justice exist, it is


Indeed, formal game theoretical analyses have shown that punishment can lead68

individuals to behave in accordance with an arbitrarily wide array of norms. Robert Boyd &

Peter J. Richerson, Punishment Allows the Evolution of Cooperation (Or Anything Else) In

Sizable Groups, 13 ETHOL. & SOCIOBIO 171 (1992).

Robert Kurzban & Mark R. Leary, Evolutionary Origins of Stigmatization: Social69

Exclusion, 127 PSYCHOL. BULL. 187, 192 (2001). See also Roy Baumeister & Mark R. Leary,

The Need to Belong: Desire for Interpersonal Attachments as a Fundamental Human Motivation,

117 PSYCHOL. BULL. 497, 521 (1995).

See, e.g., Joseph Henrich & Robert Boyd, The Evolution of Conformist Transmission70

and the Emergence of Between-Group Differences, 19 EVOL. & HUM . BEHAV. 215 (1998).

Our argument does not turn on absolutely exceptionless universality. Most features71

of most species appear in ranges reflecting variation in the genes and environments that give rise

to unique organisms. The breadth and distributions of features within those ranges is sensitive

to (among other things) the strength of selection pressures. Our argument is that there is good

reason to hypothesize that selection pressures in this context would yield very similar intuitions

of justice that are very wide-spread across the human species.

See Robinson & Kurzban, supra note 7, comparing Parts I-III with Part IV.72

See JEROM E KAGAN , THE NATURE OF THE CHILD 123, 189 (1984). See also Richard73

A. Shweder & Jonathan Haidt, The Future of Moral Psychology: Truth, Intuition and the

Pluralist Way, 4 PSYCHOL. SCI. 360, 363 (1993).

23

disadvantageous to publicly reject the principles of that system or to behave inways that conflict with others’ intuitions. Thus, persons who do not reliably68

behave consistent with those intuitions or do not signal their agreement in suchshared intuitions will be disadvantaged as against those who do.

To illustrate that “core” intuitions of justice should reliably develop in allhumans consider the fate of a mutation that causes individuals not to developsuch intuitions. Such individuals would be punished by those who they injured,stole from, or cheated, and would not be desired as social exchange partners oras members of groups. For a social creature, such a fate would have been areproductive disaster. The individual costs associated with having diverging69

intuitions from those shared by others are a powerful force for homogenizing theintuitions of justice. Thus, once such intuitions of justice are in place, there70

would be strong pressure favoring any variant in humans that acquires thoseintuitions, stabilizing them, and making them essentially universal in thespecies.71

One might wonder why some details of human intuitions of justice arewidely shared, while others are not. Intuitions of justice that were72

advantageous no matter what the context should be relatively consistent acrossindividuals, and should develop reliably regardless of one’s local environment.73

These “core” intuitions probably include the notions that unjustified physicalaggression, taking of another’s property, and cheating in exchanges are all wrong


See Jones, Proprioception, supra note 20 at 847-867 & Fig. 9 (arguing that these74

three notions, alongside several others, have evolutionary roots). As a consequence, and to the

extent that evolutionary processes influence the content of predispositions, emotions, intuitions

of justice, and morality, the major features of legal systems will tend to reflect that influence. See

id.; Jones & Goldsmith, supra note 20, at 466-475 & Table 1 (2005). For recent elaboration, see

Robin Bradley Kar, The Deep Structure of Law and Morality, 84 TEX . L. REV . 877 (2006).

24

and should be punished. In contrast, as conduct less directly prevented74

beneficial social interaction or was more dependent upon social, economic, orother context, the less likely it is to be included in the “core” intuitions thatreliably develop today in all environments. This is reflected in the cross-culturaland even within-group divergence on such things as drug use or prostitution, inwhich the core intuitions regarding harm, theft, and deception are not violated.

III. EMPIRICAL SUPPORT FOR AN EVOLUTIONARY EXPLANATION

In Part II we provided an evolutionary explanation for the little-acknowledged but stunningly consistent and widely shared intuitions of justice.On present evidence, one cannot know whether our specific evolutionaryexplanation is correct. Unlike bones, behavior does not fossilize, and one canonly test the accuracy of evolutionary accounts of behavior by triangulating frompresently available sources of information. That requires combining theories,offered in Part II, with observations and experiments, which we take up in thisPart III.

There is a good deal of empirical work that needs to be done, but in themeantime, we can look to three existing areas that shed some light on ourhypothesis. First in this Part, we consider animal studies that suggest evolved,rudimentary notions of fairness and blaming. Second, we consider brain sciencestudies that reveal identifiable physiological processes at work when humansmake moral judgments. Finally, we examine the field of developmentalpsychology, which shows a common path in the development of moral reasoningin the human child across demographics and cultures. We conclude that all threeof these provide evidence generally consistent with our evolutionary hypothesis.In contrast, we know of no data in direct conflict with it.

A. Animal Studies Evidence

If the evolutionary explanation for shared intuitions of justice is correct,one might expect to find in animals rudimentary forms of humans' core intuitionsabout what constitutes wrongdoing and that it should be punished, especially


For an alternative view, see J.R. Stevens et al., Psychological Constraints on the75

Evolution of Cooperation, 8 TRENDS IN COG . SCI. 60 (2005) (distinguishing punishment from

harassment).

Raoul A. Mulder & Naomi E. Langmore, Dominant Males Punish Helpers for76

Temporary Defection in Superb Fairy-wrens, 45 ANIM . BEHAV. 830 (1993).

Hudson K. Reeve, Queen Activation of Lazy Workers in Colonies of the Eusocial77

Naked Mole-Rat, 358 NATURE 147 (1992).

Marc Bekoff, Wild Justice, Cooperation, and Fair Play, in THE ORIGINS AND NATURE78

OF SOCIALITY 53 (Robert W. Sussman & Audrey R. Chapman eds., 2004).

Tom H. Clutton-Brock & Geoff A. Parker, Punishment in Animal Societies, 37379

NATURE 209, 211 (1995).

Clutton-Brock & Parker, supra note 79, at 212 (citing Joanne Reiter, et al., Northern80

Elephant Seal Development: The Transition from Weaning to Nutritional Independence, 3

BEHAV. ECOL & SOCIOBIO . 337 (1978); Joanne Reiter, et al, Female Competition and

Reproduction Success: Northern Elephant Seals, 29 ANIM . BEHAV. 670 (1981)).

Clutton-Brock & Parker, supra note 79, at 212 (citing Tom Clutton-Brock et al., The81

Logical Stag: Adaptive Aspects of Roaring and Fighting in Red Deer (cervus elaphus L.), 27

ANIM . BEHAV. 211 (1979); Tom Clutton-Brock et al., Passing the Buck: Resource Defence, Lek

Breeding and Mate Choice in Fallow Deer, 23 BEHAV. ECOL. SOCIOBIO . 281 (1988)).

25

among our close primate relatives. In fact, a number of socially cooperativespecies appear in some circumstances to exhibit these characteristics of“punishing” aggressors and cheaters. And a number of researchers now suggestthat such behaviors may reflect a rudimentary moral sense or intuition ofjustice. Individuals that deviate from various group or dyadic norms and75

expectations are sometimes avoided or aggressed against by the victims, thevictim’s relatives, or others. These phenomena are not limited to primates, oreven to mammals.

Here are a few of many examples. In one social species of wren,“helpers” assist by providing food when young are being raised. Helpersexperimentally removed from the group during that period are usually attackedand harassed upon their return, while helpers absent at other times of the year arenever attacked. Within the highly social naked mole rat communities, queens76

appear to focus attacks on lazy workers. Wolves apparently refuse to play with77

those who violate the social rule against injurious play-fighting, and the latterboth leave the group and die at higher rates than average.78

Researchers studying animal punishment patterns have found thatbehavior akin to theft is one particular target for retribution. For example,79

elephant seal pups caught trying to nurse from a female who is not their motherare not just shooed away, they are often bitten severely and sometimes killed.80

Young male deer attempting to sneak copulations with females being guarded byadult males are regularly attacked. 81


Sarah Brosnan, Nonhuman Species’ Reactions to Inequity and their Implications for82

Fairness (unpublished manuscript, on file with authors).

Sarah Brosnan & Frans de Waal, Monkeys Reject Unequal Pay, 425 NATURE 29783

(2003); Sarah Bosnan & Frans de Waal, Reply, 428 NATURE 140, 140 (2004).

Bosnan & de Waal, Reply, supra note 83, at 140.84

Brosnan, supra note 82; Sarah Brosnan et al., Tolerance for Inequity May Increase85

with Social Closeness in Chimpanzees, 272 PROC . R. SOC. B. 253 (2005).

Craig Packer, Reciprocal Altruism in Papio anubis, 265 NATURE 441 (1977).86

Julia A. Horrocks & Wayne Hunte, Maternal Rank and Offspring Rank in Vervet87

Monkeys, 31 ANIM . BEHAV. 772 (1983).

Frans de Waal & Lesleigh M. Luttrell, The Similarity Principle Underlying Social88

Bonding Among Female Rhesus Monkeys, 46 FOLIA PRIM ATOLIGICA 215 (1986).

26

With respect to an animal sense of fair treatment, a leading researcher hasconcluded that a wide variety of animals are capable of discerning when asituation is not equitable. For example, in a recent and widely-reported82

experiment with capuchin monkeys, different combinations of two adjacentmonkeys regularly alternated returning granite tokens for slices of cucumber.Once the experimenter began to provide one monkey in a dyad with a grape (amore highly valued food) in exchange for the same token that continued to yieldmere cucumbers for the other monkey, that other monkey often manifestedconsiderable distress. It sometimes jumped up and down, throwing the token orthe cucumber at the researcher, refusing to eat the cucumber, and the like. This83

lead the authors of the study to conclude that capuchins are capable of comparingtheir own reward to the reward others receive, and accepting or rejecting rewardsaccording to their relative, not absolute, value. Similarly, chimpanzees84

reportedly will often refuse to participate in an exchange once anotherchimpanzee is receiving a more valued reward for the same amount of effort.85

Thus, both capuchins and chimpanzees behave in ways suggesting both that theycan perceive unfairness and that it often agitates them.

Behavior suggesting an ability to perceive inequities appears to underliea great deal of social behavior in primates, in whom transgressive acts are mostsystematically punished. By way of background, many primates regularlyexhibit sophisticated cooperation, which ranges from simple reciprocal groomingand food sharing to complex tool-using and coalitional behavior. For example,in olive baboons there is a strong correlation between A’s prior support for afellow baboon B, in a conflict, and A’s successful recruitment of B to a newconflict in which A is later involved. In both vervet monkeys and macaques,86 87 88

individuals have been observed to support unrelated others, who have in the past


Jessica C. Flack & Frans de Waal, "Any Animal Whatever": Darwinian Building89

Blocks of Morality in Monkeys and Apes, in EVOLUTIONARY ORIGINS OF MORALITY (Leonard D.

Katz, ed., 2000).

Robert M. Seyfarth & Dorothy L. Cheney, Grooming, Alliances and Reciprocal90

Altruism in Vervet Monkeys, 308 NATURE 541 (1984).

Frans B.M. de Waal, Food Sharing and Reciprocal Obligations Among Chimpanzees,91

18 J. HUMAN EVOLUTION 433 (1989).

Alicia P. Melis et al., Chimpanzees Recruit the Best Collaborators, 311 SCI. 129792

(2006).

Clutton-Brock & Parker, supra note 79, at 211.93

Marc D. Hauser, Costs of Deception: Cheaters Are Punished in Rhesus Monkeys, 8994

PROC . NATL. ACAD . SCI. 12137, 12137 (1992); Marc D. Hauser & Peter Marler, Food-Associated

Calls in Rhesus Macaques (Macaca mulatta): II. Costs and Benefits of Call Production and

Suppression, 4 BEHAVIORAL ECOL. 206 (1993).

Frans B.M. de Waal, Food Sharing and Reciprocal Obligations Among Chimpanzees,95

18 J. HUMAN EVO . 433 (1989). See also GOOD NATURED , supra note 20, at 160 (concluding such

behavior “suggests a sense of justice and fairness”).

FRANS DE WAAL, CHIM PANZEE POLITICS 207 (1989). 96

Joan B. Silk, The Patterning of Intervention Among Male Bonnet Macaques:97

Reciprocity, Revenge, and Loyalty, 33 CURRENT ANTHRO 318 (1992); Flack & de Waal, supra

note 89 at 8.

27

supported them, by intervening in conflicts on their behalf. Vervet monkeys89

tend to preferentially groom individuals who have groomed them in the past.90

In chimpanzees, when one chimpanzee grooms another earlier in the day, thelatter is more likely to share food with him later. And chimpanzees91

experimentally engaged in tasks requiring collaboration with another chimpanzeequickly determine which among different potential collaborators is most effectiveand recruit that individual preferentially for subsequent collaborative tasks.92

On the other hand, failure to cooperate regularly yields sharpconsequences. Negative acts are the flip-side of the cooperation coin. Andauthors of a leading paper on the subject conclude that the intensity ofpunishments often increases with the increasing harm caused by transgressors.93

The worse the harm, the more severe the punishment. In rhesus macaques, forexample, those who discover food and are caught having failed to alert the groupto its discovery often become targets of significant aggression. In chimpanzee94

societies, those reluctant to share when they have food are more likely toencounter aggressive responses when they later approach those who have food.95

Chimpanzees will attack former allies who failed to assist them in conflicts withthird parties. And many primates, including chimpanzees, tend to intervene96

most often against those others who have most often intervened against them.97


See, e.g., Frans de Waal & Lesleigh M. Luttrell, Mechanisms of Social Reciprocity98

in Three Primate Species: Symmetrical Relationship Characteristics or Cognition, 9 ETHOL. &

SOCIOBIO . 101 (1998); Fillipo Aureli et al., Kin-oriented Redirection Among Japanese Macaques:

An Expression Of a Revenge System?, 44 ANIM AL BEHAV. 283 (1992).

DE WAAL, GOOD NATURED , supra note 20 at 157-58.99

de Waal, Food Sharing and Reciprocal Obligations Among Chimpanzees, supra note100

95 at 452.

de Waal, Food Sharing and Reciprocal Obligations Among Chimpanzees, supra note101

95 at 452.

GOOD NATURED , supra note 20, at 157-59. 102

FRANS DE WAAL, CHIM PANZEE POLITICS 207 (1989). See also Flack & de Waal, Any103

Animal, supra note 89, at 9 (“Monkeys and apes appear capable of holding received services in

mind, selectively repaying those individuals who performed the favours. They seem to hold

negative acts in mind as well, leading to retribution and revenge.”).

GOOD NATURED , supra note 20, at 159. See also Flack & de Waal, Any Animal,104

supra note 89, at 9.

Indeed, some dispute whether primates actually use negative sanctions to shape105

behavior of third parties or to punish deviation from social norms. Joan B. Silk, The Evolution

of Cooperation in Primate Groups at [??] in MORAL SENTIM ENTS AND MATERIAL INTERESTS: ON

THE FOUNDATIONS OF COOPERATION IN ECONOM IC LIFE (Herb Gintis et al. eds., 2005).

28

These patterns have prompted leading researchers to refer to such negativereciprocities as a “revenge system.” 98

Indeed, among chimpanzees (which, along with bonobos, are the closestrelatives of humans) retribution is sufficiently common that researchers considerretaliation “an integral part of [a] system of reciprocity.” Because sharing and99

other cooperative behavior exists in a “multi-faceted matrix of relationships,social pressures, delayed rewards, and mutual obligations,” the most successful100

individuals are those who can distinguish good colleagues from bad, and dealwith each accordingly. The leading chimpanzee scientist consequently describestheir community as “a ‘market’ of reward and punishment,” with “balance101

sheets” on social interactions, and reciprocity rules consisting of “‘one good102

turn deserves another’ and ‘an eye for an eye, a tooth for a tooth.’” “Not only103

are beneficial actions rewarded,” he concludes, “there seems to be a tendency toteach a lesson to those who act negatively.” Reward good deeds, punish bad.104

The data available, of which the foregoing is only a sample, do notthemselves prove that some core aspects of human intuitions of justice areevolved adaptations. Nor, even if they are adaptations, did they necessarily105

arise from the same roots as related intuitions in other species. However,tolerance of wrongdoing in highly cooperative species is evolutionarily unstable.Punishment of wrongdoing is likely a logically necessary companion ofcooperation in ultra-social species such as ours. And the intersecting vectors ofavailable animal evidence point toward that conclusion. Moreover, the close


GOOD NATURED , supra note 20, at 64.106

Flack & de Waal, Any Animal,, supra note 89. See also GOOD NATURED , supra note107

20, at 218 (“The fact that the human moral sense goes so far back in evolutionary history that

other species show signs of it plants morality firmly near the center of our . . . nature.”); MORAL

M INDS, supra note 45 at 108-109 (2006) (“We can safely assume that these intuitions evolved

prior to or during our life as hunter-gatherers . . . In such small-scale societies, fairness was most

likely an effective proxy for judging punishable acts.”).

Flack & de Waal, Any Animal, supra note 89, at 3.108

de Waal, GOOD NATURED , supra note 20 at 39. See also Brosnan, Nonhuman109

Species, supra note 82; Frans B.M. de Waal, The Chimpanzee’s Sense of Social Regularity and

its Relation to the Human Sense of Justice, 34 AM . BEHAV’L SCI. 335 (1991); Sarah Brosnan, A

Sense of Fairness in Monkeys, in THE ENCYCLOPEDIA OF ANIM AL BEHAVIOR (Marc Bekoff ed.,

2004); Bosnan & de Waal, Reply, supra note 83 at 140; RICHARD JOYCE, THE EVOLUTION OF

MORALITY (2006) (reaching same conclusion).

29

relationship of humans, chimpanzees, and other primates suggests that we oughtnot gratuitously, uneconomically, and unparsimoniously assume that there aredifferent causes in humans for the same behaviors in other animals. It is not106

surprising, then, that the leading animal researchers believe that evolution hassupplied the “building blocks” of human morality, and that the question is not107

whether, but rather to what degree, biology has influenced the development ofhuman moral systems. “Evolution” they believe, “has produced the requisites108

for morality: a tendency to develop social norms and enforce them.”109

B. Brain Science Evidence

Some people may balk at the notion that intuitions of justice could be theproduct of evolution even in an indirect way because views of justice areobviously matters of complex judgment making. The design of one's visualsystem might be the product of evolution, but how could evolution affect aperson's judgment concerning what punishment is deserved? Such judgments areobviously the product of mental reasoning processes not cell growth patterns.And, such an argument might continue, there is simply no mechanism by whichthe physiological nature of human beings could translate into complex judgment-making.

But, of course, specialized brain structures and brain processes determineall decision-making, no matter how complex. And neuroscience has begun toprovide details on what does what. As demonstrated below, we now know thatspecific brain regions perform certain kinds of reasoning tasks, and moraljudgment-making, in particular, has been shown to involve specific brain regionsand, indeed, those regions interact differently depending upon the exact natureof the moral decision being presented.


See generally PINKER, BLANK SLATE, supra note 24 (discussing such assumptions).110

See also STEVEN PINKER, HOW THE M IND WORKS (1999).

Marlene R. Cohen & William T. Newsome, What Electrical Microstimulation has111

Revealed about the Neural Basis of Cognition, 14 CURRENT OP. IN NEUROBIO . 169 (2004)

(providing overview of studies).

Alvaro Pascual-Leone, et al., Transcranial Magnetic Stimulation in Cognitive112

Neuroscience – Virtual Lesion, Chronometry, and Functional Connectivity, 10 CURRENT OP. IN

NEUROBIO . 232 (2000).

See ANTONIO R. DAM ASIO , DESCARTES’ ERROR: EM OTION , REASON , AND THE113

HUMAN BRAIN (1994); Hanna Damasio, et al., The Return of Phineas Gage: Clues About the

Brain from the Skull of a Famous Patient, 264 SCI. 1102 (1994).

30

Contrary to prior assumptions, we know today that the brain is110

anatomically and functionally specialized. Different parts of the brain acting invarying combinations perform different information-processing tasks that in turninfluence a person’s behavior. Scientists have investigated what part of thehuman brain does what by observing what happens when: different brainlocations are experimentally stimulated; different regions of the brain areremoved or damaged; or a person within a brain scanning device isexperimentally engaged in decision-making.

A wealth of experimental evidence has shown that artificial stimulationof different areas of the brain affect perception, cognition, and behavior indifferent ways. With invasive electrical stimulation, for example (most studiesof which have for obvious reasons involved non-human primates), researchershave manipulated the activity of small groups of neurons, ultimately establishingcausal links between the activities of those cells and specific aspects ofperception or cognition. And with non-invasive stimulation studies on humans111

(such as transcranial magnetic stimulation) researchers have discoveredpredictable interference with different cognitive tasks when the activity of aspecific portion of the brain is artificially disrupted.112

With respect to discoveries arising from cases of brain damage, considerthe well-studied case of Phineas Gage, a reliable and personable railroad workerof normal social disposition, who had a metal tamping rod pass cleanly throughhis skull and brain, after a spark ignited the dynamite he was tamping into itshole. After the rod removed Gage’s medial prefrontal cortex, he retained all hismental faculties but became bizarrely anti-social. Similarly and more recently,113

a variety of studies have identified ways in which relatively localized anddiscrete brain damage and lesions in particular brain areas can result in severeimpingement on the acquisition of social knowledge, on social behavior, and on


Joshua Greene, Cognitive Neuroscience and the Structure of the Moral Mind114

(forthcoming in INNATENESS AND THE STRUCTURE OF THE M IND , Vol. I (Peter Carruthers et al.,

eds.) [at m/s 6] [hereinafter Cognitive Neuroscience].

DESCARTES’ ERROR, supra note 113 at 79 (describing defect in the ability to reason115

about choices and to make socially and ethically appropriate ones based on that reasoning).

Jorge Moll, et al., Morals and the Human Brain: A Working Model, 14116

NEUROREPORT 299, 300 (2003) (collecting citations). Other implicated areas include “certain

related subcortical neuclei, particularly the amygdala, ventromedial hypothalamus, dorsomedial

thalamus, and the head of the caudate nucleus or anterior limb of internal capsule.” Id.

Useful introductions include: M ICHAEL I. POSNER & MARCUS E. RAICHLE, IM AGES117

OF M IND (1997); SCOTT A. HUETTEL ET AL., FUNCTIONAL MAGNETIC RESONANCE IM AGING

(2004).

See, e.g., Jorge Moll et al., The Neural Correlates of Moral Sensitivity: A Functional118

Magnetic Resonance Imaging Investigation of Basic and Moral Emotions, 22 J. OF NEUROSCI.

2730 (2002); An fMRI Investigation, infra note 119; The Neural Basis, infra note 119 (citing

studies); An fMRI Study, infra note 119.

See generally Greene, Cognitive Neuroscience, supra note 114; Qian Luo, et al., The119

Neural Basis of Implicit Moral Attitude – An IAT Study Using Event-Related fMRI, 30

NEUROIMAGE 1449 (2005) [hereinafter, The Neural Basis] (providing useful bibliography); Jorge

Moll, et al., Morals and the Human Brain: A Working Model, 14 NEUROREPORT 299 (2003).

Other important works include Joshua Greene, et al., The Neural Bases of Cognitive Conflict and

Control in Moral Judgment, 44 NEURON 389 (2004); Joshua Greene & Jonathan Haidt, How (and

Where) Does Moral Judgment Work?, 6 TRENDS COGN. SCI. 517 (2002); Joshua Greene et al.,

An fMRI Investigation of Emotional Engagement in Moral Judgment, 283 SCI. 2105 (2001)

[hereinafter An fMRI Investigation]; Hauke R. Heekeren, et al., An fMRI Study of Simple Ethical

Decision-making, 14 COGNITIVE NEUROSCI & NEUROPHYS. 1215 (2003) [hereinafter, An fMRI

Study]; Oliver R. Goodenough, Mapping Cortical Areas Associated with Legal Reasoning and

Moral Intuition, 41 JURIM ETRICS J. 429 (2001); Oliver R. Goodenough & Kristin Prehn, A

Neuroscientific Approach to Normative Judgment in Law and Science, 359 PHIL. TRANS. R. SOC.

LOND. B 1709 (2004).

31

moral reasoning. For example, damage to the ventromedial prefrontal cortex,114

particularly of a young child, renders that person essentially morallyincompetent. And a condition known as “acquired sociopathy” often results115

from damage to the orbitomedial or polar frontal cortex, the anterior temporallobe, or the superomedial frontal lobe.116

Even more significant are the results of brain imaging studies. Functionalmagnetic resonance imaging (fMRI), which tracks both locations and amplitudesof activity in the brain, has shown that specific parts of the brain -- the orbital117

and medial sectors of the prefrontal cortex, as well as the superior temporalsulcus, among others -- are involved in moral reasoning, and also show how118

the subareas work together in predictable ways during the process of making amoral decision. 119

Not only does this work reveal the underappreciated extent to whichprocesses of evolution and development have yielded highly specialized brain


Joshua Greene, The Secret Joke of Kant’s Soul, in THE NEUROSCIENCE OF120

MORALITY: EMOTION, DISEASE, AND DEVELOPMENT (Walter Sinnott-Armstrong, ed. forthcoming)

at m/s 33.

Greene, Cognitive Neuroscience, supra note 114. Specifically, the more personal121

dilemma generated comparatively greater activity in the posterior cingulate cortex, the medial

prefrontal cortex, and the superior temporal sulcus.

See, e.g., Greene, Cognitive Neuroscience, supra note 114, at 6; Borg, infra note 125,122

at 816; Green & Haidt, supra note 119.

See sources cited supra note 119.123

Hauke R. Heekeren, et al., Influence of Bodily Harm on Neural Correlates of124

Semantic and Moral Decision-Making, 24 NEUROIM AGE 887 (2005).

32

operations, evolutionary premises explicitly underlie the hypotheses of some ofthe most significant studies of moral decisionmaking. For example, JoshuaGreen and colleagues started from the premise that elements of basic emotions(such as fear, jealousy, and anger) are evolved short-cuts to behaviors that wereadaptive for their bearers, on average, in deep ancestral environments. That is,compared to more cumbersome analysis and deliberation, emotions are generally“reliable, quick, and efficient” ways of solving commonly encounteredcircumstances.120

These researchers therefore predicted that the more closely a moralscenario aligns with regular features of ancestral environments, the more likelyit is to invoke evolved emotional reactions, rather than dispassionately cognitiveand deliberative ones. And the results were consistent with those predictions.Specifically, more physically personal dilemmas, such as whether to push oneperson into harm’s way, thereby deflecting that harm from several people, yieldsgreater activity in the brain areas associated with emotions, while a less personalbut analogous dilemma, such as whether to throw a switch that would deflect anoncoming harm onto one person instead of several, yields greater activity in adifferent brain area, associated with less passionate analysis and calculation.121

Of course, such evidence does not suggest that there is a single, evolved,discretely boundaried “morality module” in the brain. The situation is more122

complex. Some of the regions associated with moral decision-making are alsoassociated with other brain activities. But what is clear is that an identifiablesubset of the brain’s regions, despite anatomical dissociability, are regularly andintimately networked in assessing moral dilemmas. And the interaction of theseregions changes predictably as the nature of the moral dilemma changes.123

In another study, researchers found that people make faster decisions, andwith less activity in the deliberating temporal poles, when moral scenariosinvolve bodily harm than when they don’t. Another study showed124

significantly greater activity in the right amygdala, left medial orbitofrontal


The Neural Basis, supra note 119; Jana Schaich Borg, et al., Consequences, Action,125

and Intention as Factors in Moral Judgments: An fMRI Investigation, 18 J. COGN. NEUROSCI. 803

(2006).

Borg, supra note 125126

Borg, supra note 125, at 815-6.127

Jonathan Haidt & Fredrik Bjorklund, Social Intuitionists Answer Six Questions about128

Moral Psychology, in THE COGNITIVE SCIENCE OF MORALITY: INTUITION AND D IVERSITY ,

(Walter Sinnott-Armstrong, ed. forthcoming) at m/s 17.

See, e.g., Greene, Cognitive Neuroscience, supra note 114, at 11. 129

Marc Hauser, et al., Reviving Rawls’ Linguistic Analogy: Operative Principles and130

the Causal Structure of Moral Actions, in THE EVOLUTION OF MORALITY: INTUITION AND

D IVERSITY (Walter Sinnott-Armstrong, ed. forthcoming) [at m/s 1].

Greene, Cognitive Neuroscience, supra note 114, at 19. 131

33

cortex, and medial frontal gyrus when subjects are shown interpersonal violencethan when shown vandalism or violence to objects. Compared to scenarios125

with unintentional harm, scenarios involving intentional harm yield greateractivity in areas associated with emotion (such as the orbito-frontal cortex andthe temporal pole), with correspondingly less activity in areas associated withdeliberative cognition (such as the angular gyrus and superior frontal gyrus).126

Moral scenarios in which action and inaction result in the same amount of harmyield greater activity in areas associated with deliberative cognition (such as thedorsolateral prefrontal cortex) and correspondingly less activity in areasassociated with emotion (including the orbitofrontal cortex and temporal pole),than do analogous nonmoral scenarios. This suggests that “[i]ndividuals willutilize varying combinations of cognitive and emotive facilities to address moralchallenges, but, overall, certain types of moral scenarios are likely to beprocessed in characteristic areas.” And removing emotive components of127

decision-making does not render people hyperethical, it instead renders them“unable to feel the rightness and wrongness of simple decisions andjudgments.”128

This area of research is relatively new. But the foregoing, in conjunctionwith other studies, has led a variety of researchers to conclude that our moralintuitions are built on evolved and widely shared building blocks. Marc129

Hauser, for example, concludes that the evidence points to “all humans [being]endowed with a moral faculty.” Joshua Greene argues that the evidence130

suggests that "the form of human moral thought is importantly shaped by theinnate structure of the human mind and that some basic, pro-social tendenciesprobably provide human morality with innate content.” And both argue that,131

while moral decision-making is networked, rather than unitary, it is networkedin a way distinctly analogous to the brain’s innate grammar (which is the basicgrammar design one is born with, and which underlies all human languages


Greene, Cognitive Neuroscience supra note 114, at 19. Hauser, supra note 94, at m/s132

1. On innate grammar, see STEPHEN PINKER, THE LANGUAGE INSTINCT, supra note 34.

For example, Haidt and Bjorklund argue that “Moral beliefs and motivations come133

from a small set of intuitions that evolution has prepared the human brain to develop....” Jonathan

Haidt & Fredrik Bjorklund, Social Intuitionists Answer Six Questions about Moral Psychology,

in THE COGNITIVE SCIENCE OF MORALITY: INTUITION AND D IVERSITY , (Walter Sinnott-

Armstrong, ed. forthcoming), at m/s 2; and Greene & Haidt, supra note 119, at 517, conclude that

moral intuitions “are shaped by natural selection, as well as by cultural forces.”

See, e.g., LAW RENCE KOHLBERG , ESSAYS IN MORAL DEVELOPM ENT, VOL. II, The134

Psychology of Moral Development (1984); Lawrence Kohlberg, From is to Ought: How to

Commit the Naturalistic Fallacy and Get Away With It, in COGNITIVE DEVELOPM ENT AND

EPISTEM OLOGY (Theodore Mischel, ed., 1971). See also THE MEASUREM ENT OF MORAL

JUDGM ENT, Vols. 1-2 (Anne Colby & Lawrence Kohlberg eds., 1987).

LARRY P. NUCCI, EDUCATION IN THE MORAL DOM AIN 81 (2001).135

This view parallels Piaget’s views on development. See generally Lawrence136

Kohlberg, The Philosophy of Moral Development, in 1 ESSAYS ON MORAL DEV . (1981) (linking

cognitive and moral development).

34

despite regional cultural variation). That is to say, these authors and others132

conclude that the basic moral sentiments humans share are products ofevolutionary processes.133

C. Child Development Evidence

The well-established field of child development also provides muchcollateral support for the evolutionary theory offered in Part II. First, there issubstantial evidence that children across cultures go through a consistent set ofstages of development of their intuitions of justice. In the same way that babyteeth grow from gums and adult teeth replace baby teeth, intuitions aboutmorality and justice come online according to a predictable sequence. Second,the cognitive skills that underpin these intuitions are precocious, with childrenable to make important conceptual distinctions relevant to morality and justiceat young ages. Finally, the content of the intuitions of justice track the coreintuitions discussed above, with intuitions about injury, theft, and fairness beingamong the first principles of justice understood by young children.

1. Predictable Stages of Development. There is a great deal of dataabout the development of moral reasoning in children. The classic body ofresearch in the area is Kohlberg’s work, recently described as “the most widely134

researched description of moral development available.” Kohlberg’s view is135

that people go through a predictable sequence of phases of moral thinking. As136

Kagan suggests, this is as one would expect if the capacity for learning moral


Jerome Kagan, Introduction, in THE EM ERGENCE OF MORALITY IN YOUNG137

CHILDREN , X (Jerome Kagan & Sharon Lamb eds., 1987).

See generally BÄRBEL INHELDER & JEAN PIAGET, THE GROW TH OF LOGICAL138

THINKING FROM CHILDHOOD TO ADOLESCENCE (1958).

His view might be summarized as:139

Level

(age range)Stage Description of Moral Behavior

Pre-conventional

(5-10)

Obedience/

Punishment

Conform to norms because of potential for punishment

by authority figures.

Individualism/

Exchange

Do what is “good” (i.e., feels good) for the self. This

includes beneficial exchanges with others.

Conventional

(10-14)

Good/BadBehave so as to elicit approval from others.

Law and Order

Behave so as to discharge one’s duty as part of the

social order.

Post-Conventional

(14-adult)

Social Contract

Acting in a way that benefits others because of

rationally-based laws/norms in a society. Individual

rights and utilitarianism.

“Universal ethical

principle orientation”

“Right is defined by the decision of conscience in

accord with self-chosen ethical principles appealing to

logical comprehensiveness, universality, and

consistency.”

Lawrence Kohlberg, The Philosophy of Moral Development, in 1 ESSAYS ON MORAL DEV . 164

(1981).

35

reasoning were like the capacity for learning language. That is, reliable,137

consistent, cross-situational development is a hallmark of the existence of aspecific set of psychological mechanisms designed for this particular function bynatural selection.

Kohlberg believed that as children develop more sophisticated ways ofthinking (for example, from the concrete to the abstract ), they also develop138

more sophisticated moral conceptualizations. Kohlberg’s sequence is dividedinto three “levels” each with two “stages.” Kohlberg argued that people139

universally progressed from lower to higher stages of moral development. Hedid not argue that everyone reached the highest stage, “universal ethical principleorientation.” The early stages, the "pre-conventional" stages, commonly fromages 5 through 10, are essentially about self-interest. The intermediate“conventional” levels, commonly ages 10 through 14, in contrast, reflect anawareness of the benefits of having a positive reputation as a moral agent and of


Dennis L. Krebs & Kathy Denton, Toward a More Pragmatic Approach to Morality:140

A Critical Evaluation of Kohlberg’s Model, 112 PSYCHOL. REV . 629, 633 (2005).

The first evolutionary perspective of which we are aware on Kohlberg’s work141

appears in Alexander, supra note 60, at 131-39.

See Kagan, Introduction, supra note 137 at X. 142

For a fuller discussion of general social learning as an alterative explanation, see Part143

IV, infra.

36

fulfilling one’s duties in the context of social exchange. The highest level, the“post-conventional” stages, commonly from ages 14 onward, includes a genuineinterest in others’ welfare, a respect for others’ rights, and a recognition ofuniversal moral principles. More recent approaches to Kohlberg’s modelmaintain the idea that stages are reached in sequence, but instead of one stage“replacing” another, each stage is seen as supplementing the logic of previousstages. On this view, a person who has reached Stage 4 will still use Stage 3140

reasoning in circumstances where doing so is useful or desirable.For our purposes, it is not important whether Kohlberg's scheme is the

best way of describing the various stages of development. What is significant isthe essentially undisputed existence of a predictable developmental path for allhumans, however that path might be conceptualized and described. That childreneverywhere go through similar stages of moral reasoning about justice, atroughly the same ages, is consistent with our evolutionary explanation for theorigins of intuitions of justice. As suggested by Kagan, “temporal141

concordance implies a biologically based preparedness to judge acts as right orwrong, where preparedness is used with the same sense intended by linguistswho claim that two-year-old children are prepared to speak their language."142

Imagine the reverse case. If there were no specific developmental systemfor the acquisition of moral intuitions, if intuitions of justice were simply amatter of general social learning, then the developmental route of the acquisitionof intuitions of justice would depend exquisitely on the environment in which thechild developed. This would include acts the child witnessed, ideascommunicated through language, pedagogy from various sources, and so forth.Because all of these elements are likely to differ widely across cultures and evenacross family and peer groups within cultures, such a general learning systemwould yield very different paths and timing in the acquisition of intuitions ofjustice for different individuals.143

2. Making Subtle and Complex Judgments at an Early Age.Kohlberg's notion of a stage-like progression of abilities has remained thestandard account in developmental literature but, crucially, Kohlberg seems tohave vastly underestimated the sophistication of children. Additional research


See Lawrence Kohlberg, A Current Statement on Some Theoretical Issues, in144

LAW RENCE KOHLBERG: CONSENSUS AND CONTROVERSY 485, 491 (Sohan Modgil & Celia

Modgil, eds., 1986). See also John M. Darley & Thomas R. Shultz, Moral Rules: Their Content

and Acquisition, 41 ANN . REV . PSYCHOL. 525, 537 (1990).

Darley & Shultz, Moral Rules, supra note 144, at 552. 145

We take “innate” ideas to be those ideas that develop reliably in each member of the146

species given the broad range of plausible environments in which the organism develops. That

is, an idea is innate if it does not rely on very general processes such as induction or social

transmission for its acquisition.

See Judith Smetana, Preschool Children’s Conceptions of Moral and Social Rules,147

52 CHILD DEV . 1333, 1333-34 (1981).

37

has shown that children are able to make much more sophisticated judgments andmuch finer distinctions than those predicted by Kohlberg’s proposed sequence.Indeed, although evidence now suggests that moral reasoning develops relativelyearly, it is likely that research still does not yet fully reveal the precociousnessof moral reasoning. Darley and Schultz, in their broad review, suggest that144

“children are capable of making moral judgments at a much earlier age thanpreviously thought. According to the recent literature, "moral capacity is welldeveloped, although by no means completely developed, in the third year oflife.” These precocious abilities of young children support the evolutionary145

argument in Part II. To the extent that intuition, knowledge, and conceptualdistinctions are made by very young children, especially universally, theprobability that each child acquires these by general learning processes becomesless likely, and a more innate developmental sequence becomes more likely.146

a. Distinguishing Moral, Conventional, and Prudential Rules. In oneearly and crucial experiment, Smetana tested very young children’s beliefs aboutjustice to determine if they make distinctions between violations of moral rules– acts that are wrong and deserve punishment (for example, one child hittinganother) – and violations of conventional rules – acts that deviate from aconvention (for example, failing to say grace before a snack). Smetana used147

pictures indicating the acts in order to demonstrate violations. To elicitresponses, a pictorial scale (different size frowns) was used to gauge seriousness.Smetana also used a verbal assessment of how harshly the offender should bepunished: not at all, a little, or a lot. One group of subjects consisted of childrenbetween two-and-a-half to roughly three-and-a-half years old. The other groupconsisted of children between roughly three-and-a-half and almost 5 years old.Moral offenses included both physical harm (hitting) and theft (taking someoneelse’s apple). Both groups indicated that the moral transgressions were moreserious and deserved more punishment than the conventional transgressions. The


Judith G. Smetana, Toddlers' Social Interactions Regarding Moral and Conventional148

Transgressions, 55 CHILD DEV . 1767 (1984).

Test-retest reliability was .66 for the 14 subjects so tested, a reasonable number by149

traditional standards in experimental developmental psychology.

See Judith G. Smetana, et al, Preschool Children's Judgments About Hypothetical150

and Actual Transgressions, 64 CHILD DEVELOPM ENT 202 (1993).

Judith Dunn, The Beginnings of Moral Understanding: Development in the Second151

Year, in THE EM ERGENCE OF MORALITY IN YOUNG CHILDREN 91, 93 (Jerome Kagan & Sharon

Lamb eds., 1987).

Charles C. Helwig et al., Morality: Its Structure, Functions and Vagaries, in THE152

EM ERGENCE OF MORALITY IN YOUNG CHILDREN 155, 170 (Jerome Kagan & Sharon Lamb eds.,

1987). See also Jonathan Haidt et al., Affect, Culture, and Morality, Or is it Wrong to Eat Your

Dog?, 65 J. PERSONALITY & SOC. PSYCHOL. 613, 621 (1993) (observing that children 10-12 years

old across the cultures investigated think that pushing another child off a swing should be

punished and that this behavior would be wrong in other countries as well).

38

children indicated that the moral offenses, compared with the violations ofconvention, would be wrong even if “there was no rule about it.” 148

These results suggest that children even at this very young age distinguishmoral wrongs from violations of convention and apply a different logic to them.Moreover, children seem to consider moral offenses serious; the majority ofsubjects in both groups rated the moral offenses a four on the four-point scale.When tested two to three weeks later, subjects’ responses were similar to theirearlier responses, suggesting a consistency in judgments.149

The importance of these results should not be underestimated. The factthat such young children believe immoral acts are wrong even in the absence ofrules, indicates a precocious, universalist view of morality. In sharp contrast,150

children believe that conventions apply to a particular group at a particular timeand can be changed. Dunn has concluded that the evidence suggests that “bythree and four years old, children respond differently to different kinds of rulebreaking or transgression” and that “[a]lthough the data do not showunambiguously a clear distinction between transgressions of social conventionand of moral standards, they demonstrate that three-year-olds are sensitive todifferent kinds of cultural breach.” Other scholars take a stronger view,151

suggesting that:

children at early ages . . . judge moral issues to be obligatory, applicableacross like situations, not contingent on specific social rules or authority-dictates, and not alterable on an arbitrary basis. They judge conventionsas contingent on social organization – such as rules, authority, andexisting arrangements.152


Marie Tisak & Elliot Turiel, Children's Conceptions of Moral and Prudential Rules,153

55 CHILD DEV . 1030 (1984).

Larry P. Nucci, EDUCATION IN THE MORAL DOMAIN [parenthetical or pin cite]. See154

also THE EM ERGENCE OF MORALITY IN YOUNG CHILDREN XI (Jerome Kagan & Sharon Lamb

eds. 1987).

Phillip David Zelazo et al., Intention, Act, and Outcome in Behavioral Prediction and155

Moral Judgment, 67 CHILD DEV . 2478 (1996).

Sharon A. Nelson-LeGall, Motive-Outcome Matching and Outcome Foreseeability:156

Effects on Attribution of Intentionality and Moral Judgments, 21 DEVELOPMENTAL PSYCHOL. 332

(1985).

Zelazo et al, supra note 155 at 2488. 157

Marie Tisak, et al., Preschool Children's Social Interactions Involving Moral and158

Prudential Transgressions: An Observational Study, 7 EARLY EDUCATION & DEV . 137, 139

(1996).

Tisak & Turiel, supra note 153 at 1036159

39

Somewhat older children similarly distinguish between moral rules and“prudential rules”– rules that are in place to protect people from harm (forexample, climbing on the back of a sofa). Tisak and Turiel gave children (ingroups of roughly age 7, 9, and 11) stories about acts that violated either moralrules (regarding theft, hitting) or prudential rules (regarding running andfalling). Subjects reported that the violation of the moral rule was more wrong153

and that it would be more wrong to change moral rules (that is, it’s moreacceptable for the prudential rule to be modified). The comparison betweenmoral rules and prudential rules indicates judgments that go beyondconsideration of consequences of rule-violating actions. Again, the results ofTisak and Turiel suggest the development of nuanced views at early ages.

b. Judging the Relative Seriousness of Wrongful Conduct. Studiessuggest that even young children intuitively appreciate the relative seriousnessof different kinds of wrongful conduct, and that the sophistication of thedistinctions they make increases with age. One of the first principles of moralityacquired by children is the idea that it is wrong to hurt others. By age three,154

this principle is understood, and by age four, the foreseeability of harm is taken155

into account. As Zalazo et al. put it, “substantive concepts of harm and156

welfare, however acquired, are guiding moral judgments by the fourth year oflife.”157

Findings also indicate that “children consider moral transgressionresulting in physical harm to be more wrong than moral transgressions resultingin property violations.” A study by Tisak and Turiel indicates that six year old158

children judge physical violence more serious than theft. This suggests that159

children have complex intuitions across different domains. Additional evidencecomes from studies in which children are asked to give examples of moral


Marie S. Tisak & Jeanne H. Block, Preschool Children's Evolving Conceptions of160

Badness: A Longitudinal Study, 1 EARLY EDUC. AND DEV . 300 (1990).

Richard E. Nisbett & Timothy D. Wilson, Telling More Than We Can Know: Verbal161

Reports on Mental Processes, 84 PSYCHOL. REV . 231 (1977).

See Nisbett & Wilson, supra note 161 ; W ILLIAM H IRNSTEIN , BRAIN FICTION: SELF-162

DECEPTION AND THE RIDDLE OF CONFABULATION (1967).

See Dale T. Miller & C. Douglas McCann, Children's Reactions to the Perpetrators163

and Victims of Injustices, 50 CHILD DEV . 861, 866 (1979).

NUCCI, Education in the Moral Domain, supra note 154.164

William Damon, Early Conceptions of Positive Justice as Related to the165

Development of Logical Operations, 46 CHILD DEV . 301, 302 (1975).

Larry P. Nucci, Because it is the Right Thing to Do, 45 HUMAN DEVELOPM ENT 125166

(2002). Nucci claims that at around age six, “children’s moral judgments become regulated by

conceptions of just reciprocity.” Id. at 128. While when fairness and reciprocity emerge can be

debated, it looks as though it is after intuitions regarding harm.

40

transgressions. Children give physical acts of harm as the most commonexamples. Acts of physical aggression are “prototypical” moral violations tochildren.160

Children are, of course, not always able to articulate the reason why it iswrong to hit, and might even use superficially self-serving language to expressthemselves; however, the inability to justify oneself is common among adults161

as well. Research shows that by age 7, ideas about injury are fairly well162

elaborated, and these ideas include a calculation as to the intention of the persondoing the injury when making judgments about blame and punishment.163

This conclusion – that the first moral concept to appear in children is theconcept that physical aggression is wrong – is relevant to our argument. It seemslikely to be more than mere coincidence that this is also the first step in ouraccount of the evolutionary origins of intuitive justice.

In contrast, very young children might not seem to have a firm grasp onnotions of fairness in exchange or social contracts. Young children arefrequently heard announcing that some outcome is “unfair.” However, this oftenmeans that the outcome is unfavorable to the speaker, rather than violating asocial contract or norm. Nonetheless, Damon has suggested that young164

children can show “consistent, patterned reasoning about [fair distribution ofproperty, ownership, and the like].”165

It has been argued that it is not until later in the development of a child,perhaps around nine years of age, that adult-like conceptions of fairnessemerge. Indeed, as Nucci put it, “[t]he great accomplishment of early-166

childhood moral development is the construction of moral action tied to


NUCCI, Education in the Moral Domain, supra note 154. 167

See Jay G. Hook, The Development of Equity and Logico-mathematical Thinking,168

49 CHILD DEV . 1035, 1041 (1978); Jay.G. Hook & Thomas. D. Cook, Equity Theory and the

Cognitive Ability of Children, 86 PSYCHOL. BULL. 429, 441 (1979).

David Elkind & Ruth F. Dabek, Personal Injury and Property Damage in the Moral169

Judgments of Children, 48 CHILD DEV . 518 (1977).

There is evidence that even three-year-olds take intentions into account when making170

judgments about actors. See Sharon A. Nelson, Factors Influencing Young Children’s Use of

Motives and Outcomes as Moral Criteria, 51 CHILD DEVELOPM ENT 823, 828-29 (1980).

Zalazo et al., supra note 155 at 2478-92.171

See Charles C. Helwig & Urszula Jasiobedzka, The Relation Between Law and172

Morality: Children's Reasoning About Socially Beneficial and Unjust Laws, 72 CHILD

(continued...)

41

structures of ‘just’ reciprocity.” Tangentially relevant is research showing that167

young children (5 or less) have been found to allocate rewards equally, whereasolder children tend to allocate rewards with a proportionality (equity) rule,168

suggesting a developmental trend in issues associated with exchange (effort forreward).

The integration of multiple factors relevant to estimation of seriousnessalso begins early, perhaps around age seven. In a study performed by Elkind andDabek, children were divided into groups based upon age (average age of thegroups was roughly five-and-a-half, seven-and-a-half, and nine). The children169

listened to stories about different crimes. The offenses described in the storiesvaried in terms of whether an offense was intentional or not and whether thedamage was to a person or to property. The children then assessed blame andwere asked about punishment. On average, the children viewed damage to aperson to be more serious, and they judged intentional damage to be more seriousthan non-intentional damage (though this varied by age). Children of all agesviewed intentional personal injury to be more serious than unintentional propertydamage. Interestingly, the youngest group of children (average age roughly 5)thought that unintentional personal injury was more serious than intentionalproperty damage, suggesting the importance of personal injury. This pattern170

was not found in the older children. These results are mirrored in later work thatsuggests that while very young children focus on either intention or harm, olderchildren (four and five) use both harm and intention when making decisionsabout punishment. Such conclusions strongly imply that children have171

sophisticated views on desert, and possess the ability to weigh multiple factorsby age 7.

Similar results were obtained by Helwig and Jasiobedzka, who hadchildren aged 6, 8, and 10 evaluate laws that were either socially beneficial (suchas traffic laws) or unjust (such as age discrimination). They found that all172


(...continued)172

DEVELOPM ENT 1382 (2001).

Id, at 1382.173

Id. at 1382.174

See Adrian F. Furnham & Steven H. Jones, Children’s Views Regarding Possessions175

and Their Theft, 16 J. MORAL EDUCATION 16, 18-30 (1987). See also Cecilia Wainryb,

Understanding Differences in Moral Judgments: The Role of Informational Assumptions, 62

CHILD DEV . 840, 847 (1991) (new information can considerably modify judgments of wrongness

among people aged 11 to 21); see John Darley et al, Intentions and Their Contexts in the Moral

Judgments of Children and Adults 49 CHILD DEVELOPM ENT, 66, 66 (1978) for evidence in

children as young as 6 that judgments regarding punishment vary depending on relevant

circumstances, such as provocation.

Darley & Shultz, Moral Rules, supra note 144, at 535.176

42

children, regardless of age, considered “the perceived justice of the law, its socialbeneficial purpose, and its potential for infringement on individual freedoms andrights.” Helwig and Jasiobedzka concluded that “children apply moral173

concepts of harm, rights, and justice to evaluate laws” and these findings imply174

that children develop textured views of legal issues by around age 6.c. Judging Blameworthiness with Factors Beyond Offense Seriousness.

Children's intuitions of justice are sophisticated enough to include more than justan assessment of the seriousness of the wrongful conduct. They will, forexample, take account of a person's culpable state of mind – whether theviolation was intentional or accidental – as noted above. They also take accountof various exculpating or mitigating circumstances. For example, thepunishment judged to be appropriate will change depending on the availabilityof information regarding mitigating circumstances (such as the offender had ahead injury, etc.), an effect seen in experiments with children as young as 7 yearsold. Darley and Schultz suggested that: “The results indicated a fairly175

sophisticated use of a variety of the moral concepts by children from 5 years ofage. Children revealed evidence of knowing that judgments of punishmentpresupposed judgments of moral responsibility and that moral responsibilityjudgments presuppose causal judgments. They also used information onintention and negligence to assign moral responsibility and information onrestitution to assign punishment.”176

More recently, research on children’s comprehension of provoked harmversus unprovoked harm has shown similar nuances. Studies by Smetana,Campione-Barr, and Yell involved children between the ages of 6 and 9 andinvolved the examination of pictures that showed provoked and unprovokedtransgressions involving either physical injury (hitting) or psychological harm


Judith G. Smetana et al., Children's Moral and Affective Judgments Regarding177

Provocation and Retaliation, 49 MERRILL-PALMER Q. 209, 213 (2003).

Id. at 209.178

Id. at 223.179

A vast research enterprise shows that children under four years of age cannot180

understand that others might have beliefs that are wrong or different from their own, which is one

possible reason for this developmental difference, having nothing to do with a change in moral

intuitions, but only more general abilities of understanding. See generally SIM ON BARON-COHEN ,

M INDBLINDNESS: AN ESSAY ON AUTISM AND THEORY OF M IND (1995); Alan M. Leslie, Pretense

and Representation: The Origins of Theory of Mind, 94 PSYCHOL. REV . 412 (1987); Alan M.

Leslie, Pretending and Believing: Issues in the Theory of ToMM , 50 COGNITION 211 (1994).

See Cecilia Wainryb & Sherrie Ford, Young Children's Evaluations of Acts Based181

on Beliefs Different from Their Own, 44 MERRILL-PALMER Q. 484 (1998).

Id.182

See, e.g., Larry P. Nucci & Elsa K. Weber, Social Interactions in the Home and the183

Development of Young Children's Concepts of the Personal, 66 CHILD DEV . 1438, 1445 (1995)

(“children differentiate personal issues from matters of moral or conventional regulation”);

Smetana, Toddlers’ Social Interactions Regarding Moral and Conventional Transgressions,

supra note 148 at 1774 (“toddlers initiate responses to moral transgressions [and are] more likely

(continued...)

43

(teasing). Smetana et al. concluded that “[c]hildren judged hypothetical moral177

transgressions to be more serious and more deserving of punishment, and theyreasoned more about concerns with others’ welfare for unprovoked than forprovoked transgressions and when retaliation involved hitting rather thanteasing.” The researchers also found that “all children judged that escalating178

the retaliatory response by hitting in response to being teased was . . . moreserious and more deserving of punishment than teasing in retaliation for eitherteasing or hitting,” suggesting that children’s understanding of morality and179

punishment includes the view that physical injury is most egregious.The sense of justice held by young children also incorporates a feature

present in adult intuitions of justice: the effect of a mistaken belief of anoffender. When they make judgments regarding blame, children aged 5 and 7(but not 3 ) take into account the fact that others might have incorrect beliefs.180 181

Their judgments are quite nuanced: if a person’s beliefs that differ from thechild’s relate to matters of fact – that is, beliefs concerning what is true (asopposed to what is morally right) – a mitigation often is permitted. However, ifthe belief that differs from the child’s concerns what is right and wrong (forexample, a teacher who thinks it is acceptable to discriminate against onedepending on gender) the person’s mistake does not exculpate. This implies182

that children have a sophisticated understanding of others’ beliefs and the rolethey play in the commission of moral offenses. Further research has shownsimilar results in older children.183


(...continued)183

to respond to moral than conventional transgressions with emotional reactions, physical

retaliation, and increasingly with age, statements regarding the harm or injury caused”).

Smetana et al., Children’s Moral and Affective Judgments, supra note 177, at 210.184

Citations omitted from quotation.

See, e.g., Marsha D. Walton & Andrea J. Sedlak, Making Amends: A185

Grammar-Based Analysis of Children's Social Interaction, 28 MERRILL-PALMER Q. 389 (1982)

(suggesting sophisticated understanding of transgressions among 5-10 year olds in classroom

settings, including implicit knowledge of what others will think relevant to the way in which a

transgression is construed).

EDUCATION IN THE MORAL DOM AIN , supra note 154 at 95-96.186

Jenny Yau & Judith G. Smetana, Conceptions of Moral, Social-Conventional, and187

Personal Events Among Chinese Preschoolers in Hong Kong, 74 CHILD DEV . 647, 647 (2003).

Smetana, Preschool Children’s Conceptions, supra note 147.188

44

Summing up this research, Smetana et al. suggest that “by 4 years of age,middle-class, primarily European American children, as well as AfricanAmerican preschoolers from lower socioeconomic backgrounds, reliablyevaluate moral transgressions pertaining to physical or psychological harm andfairness as very serious, deserving of punishment, generalizably wrong, andwrong regardless of whether a rule or a teacher prohibits the act.” In short,184

even young children seem to have textured and specific views regarding deservedpunishment, and these views are not derived simply from the dictates ofauthority.185

3. Cross-Cultural Developmental Studies. Substantial researchdemonstrates that the developmental sequence for moral reasoning is not uniqueto the Western world. Nucci suggests that “there is considerable cross-culturalevidence that children and adults across a wide range of the world’s culturesconceptualize prototypical moral issues pertaining to fairness and others’ welfarein ways very similar to children and adults in Western contexts, and differentiatesuch issues from prototypical matters of convention.” In a recent review of the186

evidence, Yau and Smetana conclude that despite cultural differences, “Childrenas young as 3½ to 4 years of age have been found to treat moral transgressionsas very serious, generalizably wrong, and wrong independent of rules andauthority sanctions. In contrast, they treat conventional transgressions as lessserious, contextually relative, and contingent on rules and authorities.”187

This last point, that children believe moral rules to be universallyapplicable, bears expansion. As discussed above, children distinguish between188

fundamental issues of justice – such as the belief that wrongdoing should bepunished, property should not be forcibly taken, and so on – and issues of


Cf. Leigh Shaw & Cecilia Wainryb, The Outsider's Perspective: Young Adults'189

Judgments of Social Practices of Other Cultures, 17 BRIT. J. DEV . PSYCHOL. 451 (1999)

(reaching similar conclusions with a Chinese sample).

Richard A. Schweder et al., Culture and Moral Development, in THE EM ERGENCE190

OF MORALITY IN YOUNG CHILDREN 1-83 (Jerome Kagan & Sharon Lamb eds., 1987).

Id. at 40. 191

See generally Cecelia Wainryb, Understanding Differences in Moral Judgments:192

The Role of Informational Assumptions, 62 CHILD DEV . 840 (1991); Charles C. Helwig et al.,

Morality: Its Structure, Functions and Vagaries, in THE EM ERGENCE OF MORALITY IN YOUNG

CHILDREN 155-243 (Jerome Kagan & Sharon Lamb eds., 1987).

45

convention or collateral societal issues -- such as the objects over whichindividuals can be said to have a property right and the conventions that a societyestablished for which clothing can be worn by which type of people. Childrenacknowledge that the latter – conventions – could be other than as they are, andthat if another group of people had a different convention, it would not be“wrong.”189

The capacity of children to make this distinction is relevant to thefrequently-cited argument that cultures differ in their morality, found inSchweder et al.’s cross-cultural investigation of the relative rankings of wrongfulacts as judged by Hindu Brahman children. Of the 39 acts listed, the children190

judged the most serious wrong on the list to be this: “[t]he day after his father’sdeath, the eldest son had a haircut and ate chicken.” They judged as one of the191

least serious transgressions-- ranked 35th of the 39 -- the case of a husband whobeats a wife because she repeatedly goes alone to the movies. These resultsmight seem to conflict with the notion of shared intuitions of justice; Westernchildren would make no such ranking. However, the rankings must beunderstood in the context of the underlying beliefs held by the raters. Becausethe father’s soul is put in jeopardy by the son’s act, the son's act is judged as amost serious harm. Similarly, the raters’ understanding that the wife’s act is aserious breach of a clear contractual and social obligation helps explain why thehusband’s beating of her is not ranked as being more wrongful. In short, whileunderlying intuitions regarding principles of justice may be similar, differentsocial conventions create different perceptions about the wrongfulness of specificconduct. As Wainryb put it, “what appears to be moral variation may be due192

not to diversity in ethical concepts but rather to differences in informational


Cecelia Wainryb, Understanding Differences in Moral Judgments: The Role of193

Informational Assumptions, 62 CHILD DEV . 840, 849 (1991). See also Cecilia Wainryb, Values

and Truths: The Making and Judging of Moral Decisions, in RIGHTS AND WRONGS: HOW

CHILDREN EVALUATE THE WORLD , NEW D IRECTIONS FOR CHILD DEVELOPM ENT 33-46 (Marta

Laupa ed. 2000) (arguing that the same underlying moral values in adults and children (or adults

in different cultures) can lead to differing views on what is an immoral act because of different

information about relevant facts).

46

assumptions.” This is an important point in judging the locus of genuine193

variability in beliefs surrounding intuitions of justice.

D. Summary of Evidence

Our hypothesis that shared intuitions of justice derive in large measurefrom the relentless effects of evolutionary processes on human brains andconsequent sentiments and behavioral predispositions, connects at a deep levelwith modern developments in biology and psychology. The hypothesis alsoappears to explain why these intuitions appear to be so stunningly consistentacross the species, so subtle in their complexities, and so non-randomly focusedon the harms to which their attention is particularly keen. We have not, in thisPart, attempted a definitive proof of our hypothesis. However we have exploredin overview three different areas in which the data are seemingly consistent withthe data in humans, and also consistent with our hypothesis.

The animals studies suggest that humans are not alone in punishingviolations of norms of cooperation, reciprocity, and fairness. That so manysocial species exhibit these behaviors, and in such very specific ways, suggeststhat there is an evolutionary and adaptive root to these behavioralpredispositions. And this in turn suggests that similar human behavior may stemfrom the same root. The data are consistent with the idea that tolerance ofwrongdoing is an evolutionarily unstable and unsuccessful trait. Some form ofpunishment appears to exist as a prerequisite for certain kinds of cooperationsuch as social exchange. This requires an ability to perceive wrongful behavior,to remember transgressors, and to treat them differently than non-transgressors.And we not only see that in other social animals, we also see that specifically incontexts of inflictions of physical harm, the taking of resources, and theviolations of norms of reciprocity.

From the rapidly expanding field of brain science, we see howevolutionary processes and development have yielded very anatomically andfunctionally specialized human brains – that tend to function very similarly inindividuals across the species. Damage to specific regions can result in dramaticsocial and moral behavioral aberrations (suggesting that those areas of the brain


47

are sufficiently specialized to be necessary for social and moral activity).Imaging studies reveal widely shared patterns in neural activity, consistent withevolutionary hypotheses, when subjects are confronted with various moralscenarios. More personal and evolutionarily salient scenarios invoke evolvedneural shortcuts to a greater extent than alternative scenarios. And far fromresponding with generic neural activity, human brains are intriguingly similarand differentially focused when considering such things as bodily injury,violations of cooperation norms, and the like. This all suggests that humanbrains typically process certain kinds of transgressions in characteristic ways,which in turn raises the possibility that they have evolved that specific capacityto do so.

In the child development literature, the experimental evidencedemonstrates an understanding at a very early age that some acts are wrong anddeserve punishment. Indeed, the intuition appears to exist at the point at whichchildren are first able to communicate these types of ideas. The notion ofphysical injury as a wrong seems to be one of the first elements to emerge inchildren, followed soon after by notions of fairness and reciprocity. Theseintuitions become increasingly textured as children grow older and change fromsimplistic notions of strict equality to more sophisticated notions such as equity,which take into account a person's capacities, contributions, and needs. Moststrikingly, the development of these rather sophisticated intuitions is predictablein essentially all children, without regard to the wide variety of family, social,and cultural contexts that different children experience.

While no single study or field of research conclusively proves ourevolutionary hypothesis for the origins of shared intuitions of justice, thetriangulation of the theoretical foundations from biology and psychologygenerally, alongside behavioral data in humans and other species, recent studiesof human brain operations, and broad research into characteristically humandevelopment of moral psychology, present a strong case.

IV. ALTERNATIVE EXPLANATIONS:

GENERAL SOCIAL LEARNING AND EFFICIENT NORMS

Evolutionary explanations inevitably evoke the riposte that perhaps thephenomenon in question is simply due to “learning” or “culture.” Could theconsensus we have documented here be the result of social learning rather than


Of course, any explanation for the very complex features of intuitions of justice must194

include evolution by natural selection as part of their origin, including explanations based upon

social learning. The ability to learn is itself an evolved trait. See Tooby & Cosmides,

Psychological Foundations, supra note 29, at 119 (discussing relationship between evolution and

learning mechanisms). Thus, an argument that groups discover efficient norms regarding justice

(see below) requires an evolutionary explanation for how the psychological mechanisms that

perform this calculation – whether conscious or unconscious – came to exist. Similarly, an

argument that people have a predisposition for social learning must also have an evolutionary

explanation for this highly sophisticated, complex, functional capacity. In short, “learning”

and/or “culture” do not obviate evolutionary explanations; they necessitate them.

See supra section [III.C.1].195

The distinction we draw here is between an evolved “general” social learning ability,196

which may result in, among many other things, the acquisition of intuitions of justice and an

evolved mechanism “specific” to acquiring intuitions of justice. We take it as uncontroversial

that a psychological system that acquires many different kinds of information from the social

world – a general social learning system – is the product of evolution by natural selection. But

the narrow question is whether the system for acquiring intuitions of justice was designed by

natural selection for that relatively narrow function, as we argue is the best-reasoned conclusion,

or was designed to acquire information across a wide range of content domains. Evolutionary

psychologists refer to this distinction as between “domain specific” and “domain general” mental

systems. See, e.g., Tooby & Cosmides, Psychological Foundations, supra note 29, at 97 (arguing

(continued...)

48

evolution? Perhaps the norms adopted simply reflect human propensity to194

copy others’ norms, hence the observed consensus without the need for ourevolutionary explanation. Or, perhaps groups adopt the same norms because thenorms are efficient for all groups.

In one sense, the social learning claim -- that shared intuitions of justiceare the product of social learning from the surrounding culture -- is not really inconflict with our evolutionary argument in Part II, for we suggest there thatintuitions of justice are indeed "learned" in a sense. Babies do not come into theworld with intuitions of justice intact; rather, they acquire them over time in apredictable and largely sequential way through interaction with theirenvironment. In this sense, the evolutionary explanation envisions some195

involvement of the surrounding culture, just as the special human mechanism foracquiring language involves the surrounding culture.

But non-evolutionary explanations go a step further. Their claim is thatintuitions of justice result entirely from general social learning with no relianceupon an evolutionarily-developed special mechanism for acquiring intuitions ofjustice. In other words, the non-evolutionary explanations claim that peopleacquire views of justice in the same way that they learn most other knowledgeor come to most other opinions -- that views of justice are learned, reasoned out,or chosen through rational decision-making.196


(...continued)196

that framing psychological issues in terms of how “domain specific” cognitive mechanisms are

has been and will continue to be productive in psychological research.). For a discussion of tests

for domain-specificity, in the context of logical reasoning, see Laurence Fiddick et al., No

Interpretation Without Representation: The Role of Domain-specific Representations in the

Wason Selection Task 77 COGNITION 1, 2 (2002).

RICHARD DAW KINS, THE SELFISH GENE (1976) (arguing that genes, rather than whole197

organisms, are the entities on which natural selection acts and so, except under certain conditions,

selection will act to favor genes that cause their own replication rather as opposed to genes that

cause individuals or groups or ecologies to prosper).

Robyn Dawes, Social Dilemmas 31 Ann’l Rev. of Psych. 169 (1980).198

SAM UEL S. KOMORITA & CRAIG D. PARKS, SOCIAL D ILEM M AS (1996).199

49

A. Spontaneous Social Learning

One non-evolutionary explanation, which might be called "spontaneoussocial learning," could go something like this: The views of justice upon whichthere is so much agreement are those views that are most efficient for a group tohold, perhaps because they permit productive social cooperation, much like theadvantages we discuss in the evolutionary account we give in Part II. Thisexplanation, notice, relies on no special mechanism for the acquisition ofintuitions of justice. But this explanation for the shared views of justice facesseveral difficulties.

1. Preferring Group Interests Over Individual Interests. First, forthis explanation to work, it depends upon the individual choosing to adopt viewsof justice that are efficient for the group. But there is little reason to believe thatpeople are predisposed toward preferring norms that are efficient for the group,rather than norms that are efficient for their own self interest. To make thisexplanation plausible, one would have to show that people would prefer what isgood for the group over what is good for themselves, a preference that theavailable data does not support. Instead, evidence from both the laboratory197

and the real world support the view that when group interests are pitted againstself-interest, what has been referred to as “social dilemmas,” it more likely thatself-interest will prevail. Acts of littering, pollution, over-fishing, and limitless198

other examples are all cases in which people choose to accommodate their owninterest rather than to do what is good for the group. 199

In the lab, the question of self-interest versus group interest has beenaddressed by using “public goods” games or “commons dilemma” games. Inthese games, people are put into groups and given an allotment of money.Subjects can keep some or all of the money, or put some or all of it in a group


See, e.g., Robert Kurzban et al., Incremental Commitment and Reciprocity in a Real200

Time Public Goods Game, 27 Personality & Soc. Psych. Bull. 1662 (2001).

John Ledyard, Public Goods: A Survey of Experimental Research, in HANDBOOK201

OF EXPERIM ENTAL ECONOM ICS 111-194 (J. Kagel, & A. Roth eds. 1995).

Cognitive Adaptations for Social Exchange, supra note 62 at 183 (arguing and202

providing evidence for the view that humans do not perform well on tasks that require a general

ability to use the rules of formal logic).

For a brief discussion, see Colin F. Camerer, BEHAVIORAL GAM E THEORY:203

EXPERIM ENTS IN STRATEGIC INTERACTION 60-62 (“multiplying the stake by two or ten makes

little difference,” id. at 60).

Herbert Simon, A Behavioral Model of Rational Choice 69 Q. J. ECON . 99, 101204

(1955) (“actual human rationality-striving can at best be an extremely crude and simplified

approximation to the kind of global rationality that is implied . . . by game theoretical models”).

See also GERD G IGERENZER & REINHARD SELTON , BOUNDED RATIONALITY , THE ADAPTIVE

(continued...)

50

pot. Money in the group pot is multiplied by the experimenter at some rate andsubsequently divided among all the members of the group. Because money in thepot is multiplied and subsequently split among the group members, everyone inthe group is made best off if everyone contributes all their money to the grouppot, in effect making the “pie” everyone shares larger. However, each person isindividually better off by keeping all their own allotment of money forthemselves. With a few exceptions, when people play these games in the same200

group over time, people keep greater and greater amounts of their endowment,reaching nearly purely selfish behavior over time. In short, when individual201

selfishness is pitted against group interest in the lab, selfishness eventually wins.

2. Complexity of Determining Efficient Norms as Beyond IndividualCapacity. Even if people preferred a norm that is efficient for the group as awhole, though not necessarily good for them personally, another difficulty for thespontaneous social learning explanation arises from the fact that determiningwhat is an efficient norm is an exceedingly complex matter, probably wellbeyond the capacity of an individual. Studies of human cognitive skills confirmtheir limitations. Substantial experimental evidence suggests that when peopleare presented with tasks that require the use of a general ability to reason usingthe rules of formal logic, they typically do not perform well. Even in202

experiments in which the financial stakes are very high, in which one mightassume that people would have an incentive to reason as logically and carefullyas possible, behavior deviates substantially from wealth-maximizingrationality. As Nobel Laureate Herb Simon put it “[A]ctual human rationality-203

striving can at best be an extremely crude and simplified approximation to thekind of global rationality that is implied . . . by game theoretical models.”204


(...continued)204

TOOLBOX (2002) (arguing that there are important constraints on human decision making, making

them only “boundedly rational”).

J ARED D IAMOND , COLLAPSE: HOW SOCIETIES CHOOSE TO FAIL OR SUCCEED (2005).205

ADAM SM ITH , AN INQUIRY INTO THE NATURE AND CAUSES OF THE WEALTH OF206

NATIONS (1776); F. A. HAYEK THE FATAL CONCEIT: THE ERRORS OF SOCIALISM (1989).

COLLAPSE, supra note 205, at 177 (“we have to wonder why the kings and nobles207

failed to recognize and solve these seemingly obvious problems undermining their society . . .

Like most leaders throughout human history, the Maya kings and nobles did not heed long-term

problems, insofar as they perceived them.”)

51

These limitations are vividly illustrated in Jared Diamond’s recentaccount of the failure of civilizations. In these cases, the human inability to205

foresee the consequences of the way that societies organize themselves and toregulate the use of resources, played crucial roles in their eventual downfall.Among the Mayans and the Easter Islanders, for example, the mismanagementof resources, due to the inability to foresee how cultural practices would affectlong term growth and stability, played central roles in the collapse of thosecivilizations. This differs little from modern times, in which economists andothers reach little consensus on policies that will improve economic growth,because of the complexity of the system in question and the limitation of humancognitive abilities to understand and predict complex social systems. Diamond’sanalysis in many ways continues analyses by economists such as Smith and asHayek, that neither the aggregate decisions of a democratic society nor the206

guiding hand of a dictator (for example, among the Mayans) is sufficient.207

In short, the spontaneous social learning explanation for the agreementon views of justice is inconsistent with what is known about the limits of humancalculation, including the sharp limits on the human cognitive capacity fordeduction, the common human tendency to rely on simplifying heuristics, andother limitations on cognition. In contrast, the evolutionary explanation forintuitions of justice that we offer in Part II involves a process that “tests”alternative intuitional systems (or any other features of an organism) through thesieve of natural selection. The feedback loop on genes allows only a narrowrange of developmental programs to be selected. The spontaneous sociallearning explanation has no such power.

B. Accumulated Social Learning

One might try to save a social learning explanation from this last flaw --that it is beyond the cognitive ability of a single individual, or a single group --by altering the explanation in one respect: One might argue that such efficient


This type of view has its historical roots in the writings of anthropologists who208

developed ideas surrounding “functionalism,” the notion that societal institutions keep the society

working. The founders of the most relevant areas of this discipline are Emile Durkheim, discussed

above, and Alfred Reginald Radcliffe-Brown. See, e.g., EM ILE DURKHEIM , D IVISION OF LABOR

IN SOCIETY (1893); ALFRED REGINALD RADCLIFFE-BROW N , STRUCTURE AND FUNCTION IN

PRIM ITIVE SOCIETY (1952).

In the more modern anthropological literature, this is referred to as “cultural group

selection.” See, e.g., Joseph Soltis et al., Can Group-functional Behaviors Evolve by Cultural

Group Selection? An Empirical Test 63 CURRENT ANTHRO . 473, 474 (1995) (extending

functionalist arguments with a formal analysis and ethnographic evidence). For a recent,

comprehensive treatment of this and related issues, see PETER J. RICHERSON & ROBERT BOYD,

NOT BY GENES ALONE: HOW CULTURE TRANSFORM ED HUMAN EVOLUTION (2004).

Learning models are more sophisticated than we allude to here, but beyond the scope209

of the present discussion. Probably two of the most important transmission models of this type

are based on 1) conformity to the most common norm and 2) imitation of those who are of high

prestige. For a discussion of the former, see ROBERT BOYD & PETER J. RICHERSON , CULTURE

AND THE EVOLUTIONARY PROCESS (1985) and Joseph Henrich & Robert Boyd, The Evolution of

Conformist Transmission and the Emergence of Between-group Differences 19 EVO . & HUMAN

BEHAV. 215 (1998). For a discussion of the latter, see Joseph Henrich & Francisco Gil-White,

The Evolution of Prestige: Freely Conferred Deference as a Mechanism for Enhancing the

Benefits of Cultural Transmission, 22 EVO . & HUMAN BEHAV. 165 (2001) (arguing that there are

benefits to deferring to people who are successful in an environment so that one can learn from

these individuals). Joseph Henrich, Cultural Group Selection, Coevolutionary Processes and

Large-scale Cooperation, 53 J. OF ECON . BEHAV. & ORG . 33 (2004) (presenting a formal model

explaining how cultural group selection can lead to the spread of norms for prosocial behavior,

including the punishment of norm violators); Robert Boyd & Peter J. Richerson, Group

Beneficial Norms Can Spread Rapidly in a Structured Population, 132 J. OF THEORET. BIO . 287

(2002).

52

norms are not learned spontaneously by each person or each group but rather arethe product of accumulated social learning across several generations. That is,groups may move toward efficient norms over many generations and mayperpetuate those norms because they are efficient and lead to success.

In a sense, this argument borrows the refinement mechanism ofevolutionary theory and and makes it available to the general social learningexplanation. In other words, the claim is that groups with efficient norms grow,do better, and spread their norms, and thus over time there comes to be a greatdeal of agreement because of the diffusion of the efficient norms. The208

explanation is “evolutionary” in that it involves change over time, with lesssuccessful norms losing out to more successful ones. However, it differs fromthe evolutionary explanation in Part II, of course, because here the selectionoccurs strictly through the operation of social learning, rather than through geneselection. But this revised social learning explanation, which might be called209

"accumulated social learning," still faces a number of serious difficulties.


See DONALD BROW N, HUMAN UNIVERSALS (1989). Brown discusses the relationship210

between evolutionary biology and human cultural universals, suggesting that “...human biology

is a key to understanding many human universals” and that “...evolutionary psychology is a key

to understanding many of the universals that are of greatest interest to anthropology.” Id. at 6

(emphasis in original). See also John Tooby & Leda Cosmides, On the Universality of Human

Nature and the Uniqueness of the Individual: the Role of Genetics and Adaptation, 58 J. OF

PERSONALITY 17, 23-24 (1990) (arguing that the process of evolution by natural selection leads

to a universal “design” for any given species, including humans, leading to many universal

physiological and psychological features).

53

1. Absence of Variation in Intuitions of Justice Among GroupsDespite Large Differences in Situation and Culture. There exist vastdifferences in ecology, history, demographics, social structure, and many othervariables from one group and culture to another. These differences are sostriking and of such a nature that it seems odd that all groups would find thesame norms to be the most efficient and, further, that each group is equallyeffective in teaching these norms to each successive generation. It would seemsurprising that dramatic differences in social structure and social resources wouldhave no effect on how, when, and what a generation would learn about justicefrom each previous generation. One would expect some groups to be better thanothers at approximating the efficient norm and some groups to be better thanothers at teaching that approximation to the succeeding generation.

Yet, as we document in Part I, there is a consensus on the core intuitionsof justice even across demographics and cultures. How could one explain thisconsensus if those views simply reflect the efficient norm for each group? Thisseems a difficult task. It would require finding a set of conditions that areuniversal to all human groups no matter their circumstances and culture andshowing that it is these universal conditions alone that set the efficient norms forthe group's views of justice. If any other factor, other than these universalconditions, had influence in shaping the core views of justice, then one wouldfind differences between groups according to the differences of these non-universal factors. Perhaps others will take up this challenge -- finding theseuniversal conditions and showing how these alone determine core views ofjustice -- but until such can be found, or even imagined, it would seem to makethe spontaneous social learning explanation implausible. In contrast, theevolutionary explanation suffers no such problem. Evolutionary explanations,dependent as they are upon evolutionarily-evolved mechanisms that all humansshare, fit naturally with human universals.210

2. Inconsistency with Developmental Data. Even more problematicfor the accumulated social learning explanation is the fact that it simply does not


See supra [III.C].211

See supra [III.C.2.a].212

See supra [III.C.2].213

54

fit the developmental data, which we summarize in Part III.C. For example, weknow that children’s views of morality emerge early, and are nuanced from ayoung age. Recall that even very young children – around three years of age –are precocious universal moralists, believing that certain offenses are wrong evenif there were no rule about the act, and distinguish moral rules from rules ofconvention. In particular, children believe that unprovoked physical harm iswrong and, further, very young children distinguish among offenses (harm versustheft) and take intent and extenuating circumstances into account in evaluatingblameworthiness for violations, suggesting nuanced views of justice.211

If views of justice were the result of social learning, one would expectthat young children would have to be taught distinctions such as those betweenprecautionary rules and moral rules. But it seems implausible that these212

abstractions could be taught to such young children, who seem to come to theseimportant distinctions precociously. The development of these intuitions seem213

to parallel more closely, say, the growth of teeth – they simply develop accordingto a pre-programmed timetable for all children, as opposed to, say, chess, whichmust be taught systematically and over a period of time with explicit instruction.Indeed, chess has a small number of rules, as compared to the intricacies ofmoral judgment, and yet the learning of morality, somewhat paradoxically, seemsto come much more easily and naturally to children than chess.

While it is not impossible that the ideas that children have about wrongsand punishment are transmitted to them by parents and peers while they areyoung, the fact that these intuitions come online across children in a reliablydeveloping sequence at roughly the same time throughout the world, withoutdifferences reflecting the different practical, social, and cultural conditions,makes these intuitions look as though they are part of a specific, developedlearning mechanism, rather than the result of general social learning which isnecessarily dependent upon the child’s environment, which will differsubstantially from place to place.

3. Intuitional Knowledge as Having Distinct Characteristics fromLearned Knowledge. Even if the social learning explanation were notinconsistent with the developmental data, it nonetheless would appearproblematic because it is inconsistent with what we know about the nature ofviews of justice: their intuitional nature. If views of justice were learned fromothers, one would expect such views to have the character of other learned


See Haidt, Emotional Dog, supra note 42, at 814.214

Philippa Foot, The Problem of Abortion and the Doctrine of Double Effect, 5215

OXFORD REVIEW 5 (1967).

Marc Hauser et al., Reviving Rawls' Linguistic Analogy, in MORAL PSYCHOLOGY216

AND BIOLOGY [currently 20-21] (W. Sinnott-Armstrong, ed. forthcoming). Even the 30 percent

who gave what the authors classed as a "sufficient justification" may have been making purely

ex post attempts at explanation rather than reporting reasoning they used in reaching their

(continued...)

55

knowledge. That is, a social learning explanation assumes that our judgmentsabout justice are like many other judgments that we make in our daily lives, suchas: how fast should I drive on this stretch of road, or how long should I cook thissandwich in the microwave, or how carefully should I take notes of this lecture?These judgments, we assume, are the product of reasoning, promoted by lifeexperience and education.

But social science evidence suggests that judgments about justice,especially for violations that might be called the core of criminal wrongdoing,are more the product of intuition than reasoning. Their intuitional nature means,among other things, that they are judgments quickly arrived at, even by peoplewith little education or life experience, that they frequently are held with strongfeelings of certainty, and that the reasons we hold such judgments are generallynot consciously accessible to us.

In Jon Haidt’s work on “moral dumbfounding,” for example, peoplereport strong intuitions about things that are morally wrong, such as consensual,non-reproductive incest, but are unable to provide a principled explanation fortheir judgments. Similarly, Hauser, Young, and Cushman looked at judgments214

of morally permissible actions using the "trolley problem," in which some215

number of people can be saved from being killed by a runaway trolley if someaction is taken (or not taken), resulting in the death of a smaller number ofpeople, often a single individual. People are asked what choice should be madein these difficult situations, and asked to explain their reasoning. While subjectscommonly have strong and clear views on the proper result, they also commonlyare unable to offer an explanation for their conclusions. For example, in onevariation of the trolley problem, the subject can avoid the death of five people onthe track by (a) pushing a bystander on the track whose body and backpack willjam the trolley's wheels, stopping it, or by (b) throwing a switch to divert thetrolley to a side track where it will kill one person. Despite the fact that theresults of the two actions are identical, 89 percent of the subjects considered thelatter action moral but only 11 percent judged the former to be moral. Moreinteresting for our purposes, 70 percent of the subjects could give no plausibleexplanation for their judgment. This mirrors Haidt’s results concerning incest.216


(...continued)216

conclusion. The authors used what they called an "extremely liberal criterion": "A sufficient

justification was one that correctly identified any factual difference between the two scenarios

and claimed the difference to be the basis of moral judgment." Id. at 19.

Haidt, Emotional Dog, supra note 42, at 814. For a discussion of the automaticity217

of such judgments, see id. at 819-820.

MORAL M INDS, supra note 45 at 125. For a recent review of relevant literature, see218

Cass R. Sunnstein, Moral Heuristics, 28 BEHAV. & BRAIN SCI. 531 (2005).

For a recent example, see Bjørn Sætrevik et al., The Utility of Implicit Learning in219

the Teaching of Rules, 16 LEARNING & INSTRUCTION 363 (2006) (showing that for teaching rules

about chemistry to students, “learning was much more effective when more explicit ways of

teaching were employed”).

56

Based on his review of the existing social science literature, Haidtconcludes that “moral judgments” derive from “quick automatic evaluations(intuitions).” Similarly, Hauser concludes that “much of our knowledge of217

morality is . . . intuitive, based on unconscious and inaccessible principles . . ..” In sum, there is at least some degree of consensus that many moral218

judgments are made by a deeply intuitive system. If such judgments were theproduct of a set of principles of morality learned from others, it would seem tobe a straightforward matter to derive the “wrongness” of acts from theseprinciples, just as mathematical inferences can be made from a set of axioms andsubsequently explained with reference to them. “Moral dumbfounding” andrelated effects in the psychological literature suggest that this is not how thesejudgments are made.

4. Difficulties in Teaching Inarticulable Lessons. Further underminingthe social learning explanation for shared views of justice is the fact that, to theextent that people are unable to articulate the principles that underlie theirintuitions of justice, it would seem difficult to transmit those principles verballyto others. In other words, the intuitional nature of judgments about justiceundermines the argument that they are socially learned from others throughlanguage. Of course, learning through non-linguistic means is possible, but thesocial transmission of information is vastly more difficult without naturallanguage. The psychological literature is replete with examples of this generalprinciple, but the general idea is well illustrated by two relevant phenomena.219

First, consider that social transmission is rare in non-humans, restrictedto narrow domains such as song-learning in birds. Without language, non-humananimals are confined to very modest social learning, so much so that each case


See Michael Tomasello, Culture and Cognitive Development 9 CURRENT220

D IRECTIONS IN COG . SCI. 37, 37 (2000) (discussing the human adaptation for cultural transmission

of information, contrasting it with non-humans’ capacities, in particular non-human primates).

See also Bennett G. Galef, Jr., The Question of Animal Culture, 3 HUMAN NATURE 157, 162

(arguing that non-human animals do not imitate one another in a way that closely parallels human

imitation, looking in detail at two of the most prominent cases of putative non-human cultural

transmission).

57

of social learning is considered an extraordinary case. Second, consider some220

vivid examples from everyday. The game of “charades” (or, more recently,“Pictionary”) is fun precisely because, deprived of natural language, conveyingsimple ideas, even individual words, is difficult. Similar examples can be seenin the difficulty in understanding what pre-linguistic infants are trying tocommunicate, and familiar to anyone who has tried to ask even rudimentaryquestions of someone with whom one does not share a common language. Sociallearning without language is not impossible, but is certainly difficult. Ifindividual words and simple questions already pose a challenge, consider howmuch more difficult abstract complex notions of deontic principles would be toconvey non-linguistically. If people do not have explicit access to theseprinciples, it is unlikely that they are being socially transmitted.

In addition to these additional failings of the accumulated social learningexplanation, recall that it also suffers the objection raised with regard to theoriginal "spontaneous social learning" explanation: that such an explanationdepends upon individuals putting group interests above their own individualinterests. As noted above, the available evidence suggests that such altruism hasserious limits. Until it can be shown that there are special circumstances andreasons that prompt such altruistic behavior, the accumulated social learningexplanation is on this ground as implausible as the spontaneous social learningexplanation.

CONCLUSION

On present evidence, we think the explanation for the "puzzle" of theexistence of shared intuitions of justice is more likely a specific evolved humanmechanism for acquiring these core intuitions than general social learningderived from some set of conditions and life experiences universal to all humansand all human groups. The latter cannot be ruled out on present evidence butseems implausible, while the former is consistent with all available data.

Ultimately, whatever conclusion one comes to on this question, it doesnot alter the existence of important implications that we think follow. Whicheveris the source of shared intuitions of justice, it is a source that is not easily altered.


See PAUL H. ROBINSON ET AL., INTUITIONS OF JUSTICE: THEIR NATURE AND221

IM PLICATIONS FOR CRIM INAL LAW AND JUSTICE POLICY (forthcoming 2007).

58

Even if the source is general social learning, it must be social learning arisingonly from an aspect of human life experience that is so fundamental as to beessentially universal to all persons without regard to circumstances or culture.In other words, it is not a source that is open to easy change or manipulation. Ifit were not so fixed, if it were easily manipulable, then the natural variations incircumstance and culture would have altered it and produced variations inintuitions of justice; there would not be the agreement that does exist.

While a full account of these implications is another project, consider221

the possibilities. For example, it may be unrealistic to expect the population toall "rise above" their desire to punish wrongdoers and to expect the governmentto "reeducate" people away from their interest in punishing wrongdoers, as isurged by some reformers. It seems unlikely that the shared intuition that seriouswrongdoing should be punished can be changed through social engineering, atleast not through methods short of the kind of coercive indoctrination that liberaldemocracies would find unacceptable.

For another example, a criminal justice system that regularly fails to dojustice or that regularly does injustice, as judged by shared intuitions of justice,will inevitably be widely seen as failing in a mission thought important by thecommunity, even foundational, unless the system's unjust operation can behidden, something that would be hard to do without breaching notions of pressfreedom and government transparency to which liberal democracies aspire.

For a final example, any realistic criminal justice system or program forits reform must acknowledge and engage with the community's shared intuitionsof justice. This does not mean that law can never deviate from those intuitionsor try to change them. There is nothing sacred or immutable about our currentintuitions of justice. But a criminal justice system must be realistic about howdifferent kinds of changes may require different – and sometimes high – financialand social costs. The greatest success in shaping the perceived wrongfulness ofparticular conduct may not be to fight people's intuitions of justice but rather totry to harness them, by providing information or arguments that strengthen (orweaken) the analogy between the target conduct and the core wrongdoing onwhich people have strong intuitions.

the origins of shared intuitions of justice

Documents