The developmental roots of fairness: infants’ reactions to equal and unequal distributions of resources

Abstract The problem of how to distribute available resources among members of a group is a central aspect of social life. Adults reactnegatively to inequitable distributions and several studies have reported negative reactions to inequity also in non-humanprimates and dogs. We report two experiments on infants’ reactions to equal and unequal distributions. In two experiments,infants’ looking times and manual choices provide, for the first time, converging evidence suggesting that infants aged 12 to18 months (mean age 16 months) attend to the outcomes of distributive actions to evaluate agents’ actions and to reason aboutagents’ dispositions. The results provide support for recent theoretical proposals on the developmental roots of social evaluationskills and a sense of fairness. Introduction The problems concerning fairness of resources distribu-tion are ubiquitous in everyday reasoning and are centraltopics for social sciences and theories of ethics (e.g. Kant,1785⁄1964; Mill, 1861⁄1998; Sen, 2008). How do indi-viduals acquire the ability to reason about these prob-lems? Classic developmental theories (Piaget, 1932;Kohlberg, 1981) emphasized the effect of peer interac-tion, verbal competence and mathematical skills on howchildren and adolescents perform and evaluate distribu-tive actions (e.g. Damon, 1975; Gunzburger, Wegner A Hook, 1978; Lane & Coon, 1972;Larsen & Kellogg, 1974; Lerner, 1974). Studies haverepeatedly found that children below 5 years of age aremainly guided by self-interest, whereas older childrentend to prefer egalitarian distributions (Arsenio & Gold,2006; Fehr, Berhardt & Rockenbach, 2008; Carson B Lane & Coon, 1972; McGillicuddy-DeLisi, Daly & Neal, 2006; Sigelman & Waitzman, 1991). Itis only in late childhood that a systematic preference forproportional distributions linked to relative merit orneed is reported (but see McCrink, Bloom & Santos,2008, for evidence that even 5-year-olds can reasonproportionally to evaluate donations).There are several problems for classic theories onmoral judgment and the empirical research stimulated bythem. Explicit verbal reasoning is likely to confoundmoral competence and language skills and to revealmostly post-hoc constructions that individuals generateafter an implicit and automatic evaluative process hasbeen completed (Haidt, 2001). Also, while proportionalreasoning is surely required when distributions must takeinto account relative effort, merit or need, its develop-ment does not explain the origins of the evaluativecomponent of the process. School-aged children’s verbalresponses are useful to chart the development of explicitjudgments, but they are of little use in investigating theorigin of the sense of fairness and testing whetherhumans possess spontaneous evaluation skills that areapplied to agents’ distributive actions.An alternative theoretical view, defended by the Britishempiricists, emphasized the role of spontaneous senti-ments in the generation and development of moraljudgments (Smith, 1759⁄1948; Hume, 1740⁄1978). Toapply this view to distributive justice scenarios, one needsto imagine that spontaneous emotional reactions causedby the distress of an actual or potential victim maystimulateanaversionforunjustdistributions.Hypothesesderived by moral sentimentalism have recently receivedconsiderable empirical support from behavioural (Haidt,2001; Rozin, Lowery, Imada & Haidt, 1999), physiolog-ical (Blair, Mitchell & Blair, 2005) and neuroimagingstudies (e.g. Greene, Sommerville, Nystrom, Darley C Hsu, Anen & Quartz, 2008). Given thatsome empathic reactions emerge very early in develop-ment (Hoffman, 1991), this view would predict an earlyemergence of aversion to inequity in children.A third theoretical view can be traced back to Rawls’theory of moral competence (Rawls, 1971). Works