Author response

The current response discusses the insightful commentaries by Dale Hay (2009) and Karen Wynn (2009) on the proposal that human altruism has deep roots both in phylogeny and ontogeny (Warneken & Tomasello, 2009). In particular, I focus on (a) what observational and experimental methods can reveal about altruistic motivations in children, (b) Wynn’s idea that early altruism might confer a selective advantage to the infants themselves, and (c) how recent findings on young children’s social ontology will enable us to test the hypothesis that ontogeny proceeds from rather global to more differentiated altruistic behaviours.

The British Psychological Society Why would such behavioursarise so early,before infants' helping is of alargetangible benefit to others?W ynn suggests that perhaps by acting altruistically, the infants may signal their cooperative attitude to the membersoftheir social environment-including the likelihoodt hat these infants will be valuable and skillful cooperators in their adult future.Parents therefore gete xcited to see their infant willingly offering their drooling pacifier not because of the object at hand,but because of the things yet to come once their offspring earnasalary.Borrowing from Dawkins't hought experiment in which altruism can emerge when cooperative individuals wear greenbeards to recognize and preferentially interact with each other (Dawkins, 1976), one could then say that it pays to grow abeard early in life because children'sfirstgreen stubble signals to others that theyshould payattention and facilitate their growing cooperativeness.Importantly,this scenario can only worki ft he appearance of early altruism actually does foreshadow later skillfula ltruism -o therwise it would be af alse signal.It seems to me like the subjects to be tested in this case should be the adults:Does the adult'sresponsetocute helperscorrelate with the mature cooperativeness of the grown up child later in life?
But there is another reasonw hy altruistic motivations may emerge early in lifesupposedly before other factorss uch as normsa nd reciprocity begin to play am ajor role.Think of reciprocal altruism.Mutually beneficial interactions can only geto ff the ground if individuals startout cooperating.Specifically,aplayer in ap risoner dilemma game promotes mutual cooperation when the player uses what Axelrod calls a' nice' strategy (by never being the first one to defect), i.e. cooperating first and then just copying what the other player dido nt he previous move (tit-for-tat;A xelrod, 1984).Thus, maybe mother nature gives children aj ump startt owards altruistici nteractions in this way to make' nice' strategies morel ikely.I ti sa no penq uestion to what extent this early altruistict endency is complemented by safety-measures such as ac heater detector,b ut perhapsi ndividuals without such ad evice up and running early in ontogeny are not at adisadvantagebecause theygrow up in the protected environment of their family.As the number of interactions with non-family members increases over ontogeny,however,altruisticchildrenshould exercise morecaution in deciding whom to help or not.

Selectivity
In herc ommentary,W ynn endorses our proposalt oe mpirically test to what extent children direct their altruism selectivelytowards differentsocial partners.This is driven by our hypothesis that children develop from rather global to more selective altruism throughout ontogeny.H owever,s he refers to recent researchw hich suggests that infants might actually be more selective from early on than we hypothesized.
First of all, we are also very impressed by the discriminatoryabilities of these young infants who, fore xample, differentiate between and preferentially interact with agents who speakt he language that is more familiar to the child (Kinzler,D upoux, &S pelke, 2007) or an agentw ho provedt ob em oreh elpful towards others ( Hamlin, Wynn, & Bloom, 2007).Y et, even though these discriminatory abilities and the altruistic motivations are present early in ontogeny,i tm ight represent am ajor developmental step to bring them together.N amely,i tr emains to be tested to what extent these discriminatorya bilities actually guide rather complexb ehaviourss uch as instrumental helping.There are other examples in which discriminatory abilities are present in early ontogeny,but still not used forcertain purposesuntil later in ontogeny.One example is selective trust: despitethe ability to discriminate between small and largequantities in infancy and the emerging capacity to learnand use words in toddlerhood, it appearsto take until middle childhood that children will preferentially learnanovel label endorsed by the group majority over am inority (Corriveau, Fusaro, &H arris, 2009).A nother example more related to the currenttopic is that in anonymous sharing tasksinwhich 3to 8-year-old children can donate to in-group or out-group members, the identity of the recipient as an in-group or out-group member becomes increasingly more important with age, with amajor transition towards parochialism at around 7-8 years of age (Fehr, Bernhard, &R ockenbach, 2008).T hus, the abilityt od iscriminate betweend ifferent types of people is one component(and anecessaryprerequisite) forselectively directed altruism, but when this discrimination becomes( increasingly more) relevant fort he children fortheir altruistic and other prosocial behaviours is another question which, as Wynn agrees, needs furtherempirical investigation.
Secondly,itisimportant to emphasizethat acritical feature of our proposal is not to investigate selectivity per se,but selectivity along the dimensions that are known to be important forthe evolution of altruism.Developmental researchshould seek out those proximate mechanisms that are purportedt oe xplain altruisticb ehavioursw ithin evolutionaryt heory,s uch as kinship, potential reciprocation and group membership.Moreover,our hypothesis suggests that thereshould be major agedifferences in the kind of selectivity that childrene mploy.A sw ep roposed,a na ltruistic bias towards familiar individuals (asacue forkinship)might be one of the earliestmechanisms forselectivity in the domain of altruistic behavioursalready emerging in children at two years of age (Young, Fox, &Zahn-Waxler,1999).However,other selective mechanisms may emerge later either because theya re more complexc ognitively, or because theyr equire more social experience.The types of selective mechanisms falling in this categorym ight includecontingentreciprocity or the internalization of social norms.Thus,eventhough already young children may act selectively early on along one dimension of the social world (such as cues forkinship), thereare many additional dimensions that children can take into account -a nd many of which may be beyond the scope of young children.Given the novel researcho ni nfant'sc ategorization of social partnersa nd their motivations to act altruistically towardsthem, we should be optimistic that this question will finda nanswer in the near future.