Smile asymmetries and reputation as reliable indicators of likelihood to cooperate: An evolutionary analysis

Cooperating with individuals whose altruism is not motivated by genuine prosocial emotions could have been costly in ancestral division of labour partnerships. How do humans ‘know’ whether or not an individual has the prosocial emotions committing future cooperation? Frank (1988) has hypothesized two pathways for altruist-detection: (a) facial expressions of emotions signalling character; and (b) gossip regarding the target individual’s reputation. Detecting non-verbal cues signalling commitment to cooperate may be one way to avoid the costs of exploitation. Spontaneous smiles while cooperating may be reliable index cues because of the physiological constraints controlling the neural pathways mediating involuntary emotional expressions. Specifically, it is hypothesized that individuals whose help is mediated by a genuine sympathy will express involuntary smiles (which are observably different from posed smiles). To investigate this idea, 38 participants played dictator games (i.e. a unilateral resource allocation task) against cartoon faces with a benevolent emotional expression (i.e. concern furrows and smile). The faces were presented with information regarding reputation (e.g. descriptions of an altruistic character vs. a non-altruistic character). Half of the sample played against icons with symmetrical smiles (representing a spontaneous smile) while the other half played against asymmetrically smiling icons (representing a posed smile). Icons described as having altruistic motives received more resources than icons described as self-interested helpers. Faces with symmetrical smiles received more resources than faces with asymmetrical smiles. These results suggest that reputation and smile asymmetry influence the likelihood of cooperation and thus may be reliable cues to altruism. These cues may allow for altruists to garner more resources in division of labour situations. S.P. Shohov, ed., Advances in Psychology Research, vol. 11, 59-78. © Nova Science Publishers 2002. Huntington, New York. Printed in the United States of America. ISBN: 159033-186-9. All rights reserved. William Michael Brown and Chris Moore 60 How can genetic self-promotion design costly organism-level adaptations for delivering benefits to others? Simply put, how did altruism between non-kin evolve in humans? One adaptive problem for designing psychological adaptations mediating altruism is the costs associated with free-rider exploitation in division of labour partnerships. Imagine that you engage in a joint venture where you and your partner are task specialists reaping the benefits from your division of labour. Benefits are attained because neither of you would have been able to succeed at the goal without the other. However, if you chose your partner based on some arbitrary criteria (e.g. he or she told you “I am honest, you can trust me.”) you could be opening yourself up to exploitation. What compounds the costs of exploitation in a division of labour situation is that it is difficult to monitor cheating. Indeed in some conditions you may not even recognize cheating because you are not specialized in the task being performed. One hypothesis for the maintenance of altruism in division of labour situations is that altruists encode subtle nonverbal cues signalling willingness to cooperate in the future (Frank, 1988). These signals are hypothesized to manipulate the sensory systems of other altruists, facilitating the formation of a coalition with the sender. If altruists can detect one another and form social support networks, this may allow for the selection of genes predisposing altruism. In mathematical simulations and modelling it appears that alliance formation among altruists cognitively equipped to detect one another may evolve (Frank, 1988; Peck, 1995; Wilson & Dugatkin, 1997; Cooper & Wallace, 1998; de Vos, Smaniotto, & Elsas, 2001). Surprisingly there is little evidence testing the basic assumption that humans can detect altruists (Brown & Moore, 2000). Furthermore even if altruists are detectable, how is this accomplished (e.g. what are the nonverbal and paralinguistic signals involved, and what are the costs maintaining signal reliability)? Despite the theoretical importance of reliable signals indicating altruistic character, to date no studies have isolated the signals or even demonstrated costs maintaining signal reliability. This chapter has two goals. One is to integrate the literature on human altruism detection (i.e. Frank, 1988; Wilson & Dugatkin, 1997; Brown & Moore, 2000) with the field on animal signalling (Zahavi, 1987; Grafen, 1990; Adams & Mesterton-Gibbons, 1995; Guilford & Dawkins, 1995; Maynard Smith & Harper, 1988; 1995; Maynard Smith, 1991; Vehrencamp, 2000). The second goal is to report an experimental investigation of two potential signals of altruism (i.e. nonverbal and reputation cues). Both nonverbal and reputation cues to altruism have been hypothesized to maintain human cooperation via indirect reciprocity (Alexander, 1987; Hirshleifer, 1987; Frank, 1988) and multilevel selection processes (Wilson & Dugatkin, 1997). However, there is little experimental evidence supporting the hypothesis that reputation cues (e.g. being known as having sincere motives for helping others) and/or nonverbal expression of prosocial emotions lead to any tangible benefits. Section II of this chapter presents findings consistent with the hypothesis that nonverbal and reputation cues accrue tangible benefits in evolutionary games.