When considering the role of smiles in social interactions, it is important to consider situations involving incomplete information about our partner in the interaction. In many social interactions uncertainty typically exists about the extent of cooperation to be expected from the interaction, and there may also be uncertainty about the stakes of the interaction and the possible choices available to the partner. A self-interested individual will never trust her partner if she believes that the partner will act selfishly in a situation that calls for cooperation. Meanwhile, mutual trust among a group of individuals might be socially beneficial, a situation typical for public goods. Given that trust and cooperation are widely observed in humans and other animals, we have to ask which signals might be used to identify trustworthy partners. Smiling may be just such a signal.
The evolution of such signals is not easy to explain. If trustworthy individuals have observable characteristics that distinguish them from non-trustworthy individuals, these signals must have an inherent cost that make them impossible or difficult to fake. Because a non-trustworthy individual would profit from interacting with a trustworthy individual, he has a large incentive to fake the signal. As has been known since the work of Zahavi (Reference Zahavi1975) in biology and Spence (Reference Spence1973) in economics, a signal can remain reliable in the face of incentives for mimicry only if its inherent cost is larger than this potential gain to the mimic–but it must still be a cost worth bearing for a genuine bearer of the signaled trait. Since trustworthiness might depend on the situation at hand, it seems unlikely that a fixed physiological characteristic would be an appropriate signal, but it is likely that behaviors revealing internal emotional and psychological states could be effective. Smiles are likely to be a signal identifying possibly valuable interaction partners via their correlation with psychological states that reliably indicate trustworthiness. The information content of the signal might concern either personal characteristics of the partner (the degree of altruism or strong reciprocity) or unobservable characteristics of the situation (possible gains to be made or the choice options available to the partner). There is a literature on smiling as a signaling device that has traditionally focused mainly on the former, but both might be equally important and influential.
The idea that emotions can serve as commitment and signaling devices has been developed in economics since Frank (Reference Frank1988) and Hirshleifer (Reference Hirshleifer and John1987). In particular, negative emotions can trigger costly punishment behavior of non-cooperative actions (Fehr & Gächter Reference Fehr and Gächter2000). Smiling has been identified as a signal interpreted as representing cooperative tendencies (Krumhuber et al. Reference Krumhuber, Manstead, Cosker, Marshall, Rosin and Kappas2007). When participants in an economic trust game are presented with pictures showing their partner either smiling or with a neutral face, they are more likely to trust the smiling partner. Whether this is an adaptive strategy is a separate question. In particular, the reliability of the picture as a guide to the partner's behavior might depend on when the picture was taken. Photos taken independently from the decision situation seem to contain little information, while photos taken when the partner was in the process of the relevant decision seem to contain nonverbal cues that allow subjects to deduce the other players' strategy. Indeed, it is not smiles as such, but rather smiles that are perceived as “honest” that are conducive to trust. “Genuine” or Duchenne smiles are reportedly difficult to fake and most easy to produce when the person is in a positive and sharing mood (Mehu et al. Reference Mehu, Grammer and Dunbar2007). Indeed, lying is known to induce a significant level of cognitive load on the liar, which makes it difficult to react spontaneously, and which reduces subtle movement (Vrij et al. Reference Vrij, Semin and Bull2006). Honest smiling, which involves a contraction of the orbicularis oculi, further has the obvious cost of reducing the smiler's visual field. Doing so might have fitness costs for the smiler, and thus should be done only sparingly, which can explain why smiling convincingly is cognitively costly. Genuine smiles can therefore be taken as an honest signal of the partner being in a positive mood, one that makes deception of the interaction partner less likely. In a situation concerning trust, this might mean that the trustee has already made a trustworthy decision, or that he is aware that he has something interesting to offer to the partner.
We can relate this idea to the smile types identified by Niedenthal et al. “Enjoyment smiles” are smiles elicited by a general positive atmosphere, and which are seen as honest signals that the partner feels comfortable in the current situation. “Affiliative smiles” are smiles that are used as a signal to the partner. Niedenthal et al. claim that affiliative smiles will be initially considered as “true smiles” (section 6.2) and as “false” when: “the perceiver believes that the smile was motivated by a desire to hide, moderate, or justify something negative” (section 6.2, para. 1, emphasis ours). We believe that it will depend partly on the specific situation of the smiler whether the smile is perceived as false or not. Whether an affiliative smile has characteristics that lead it to be judged true or false is therefore influenced not only by the situational knowledge of the observer, but also by unobservable characteristics of the smiler. Therefore, the smile can be seen as a valuable signal of this unobservable information.
Indeed, we have observed in a recent experimental study (Centorrino et al. Reference Centorrino, Djemai, Hopfensitz, Milinski and Seabright2010) that honest smiles lead to higher levels of trust, and that a smile is rated as more convincing when the trustee has larger stakes at his disposition. Therefore, honest smiles can be seen as a costly signal that can only be produced when in a situation potentially beneficial to the observer of the smile, and can thus serve as a signal valuable to the observer and to the smiler alike.
When considering the role of smiles in social interactions, it is important to consider situations involving incomplete information about our partner in the interaction. In many social interactions uncertainty typically exists about the extent of cooperation to be expected from the interaction, and there may also be uncertainty about the stakes of the interaction and the possible choices available to the partner. A self-interested individual will never trust her partner if she believes that the partner will act selfishly in a situation that calls for cooperation. Meanwhile, mutual trust among a group of individuals might be socially beneficial, a situation typical for public goods. Given that trust and cooperation are widely observed in humans and other animals, we have to ask which signals might be used to identify trustworthy partners. Smiling may be just such a signal.
The evolution of such signals is not easy to explain. If trustworthy individuals have observable characteristics that distinguish them from non-trustworthy individuals, these signals must have an inherent cost that make them impossible or difficult to fake. Because a non-trustworthy individual would profit from interacting with a trustworthy individual, he has a large incentive to fake the signal. As has been known since the work of Zahavi (Reference Zahavi1975) in biology and Spence (Reference Spence1973) in economics, a signal can remain reliable in the face of incentives for mimicry only if its inherent cost is larger than this potential gain to the mimic–but it must still be a cost worth bearing for a genuine bearer of the signaled trait. Since trustworthiness might depend on the situation at hand, it seems unlikely that a fixed physiological characteristic would be an appropriate signal, but it is likely that behaviors revealing internal emotional and psychological states could be effective. Smiles are likely to be a signal identifying possibly valuable interaction partners via their correlation with psychological states that reliably indicate trustworthiness. The information content of the signal might concern either personal characteristics of the partner (the degree of altruism or strong reciprocity) or unobservable characteristics of the situation (possible gains to be made or the choice options available to the partner). There is a literature on smiling as a signaling device that has traditionally focused mainly on the former, but both might be equally important and influential.
The idea that emotions can serve as commitment and signaling devices has been developed in economics since Frank (Reference Frank1988) and Hirshleifer (Reference Hirshleifer and John1987). In particular, negative emotions can trigger costly punishment behavior of non-cooperative actions (Fehr & Gächter Reference Fehr and Gächter2000). Smiling has been identified as a signal interpreted as representing cooperative tendencies (Krumhuber et al. Reference Krumhuber, Manstead, Cosker, Marshall, Rosin and Kappas2007). When participants in an economic trust game are presented with pictures showing their partner either smiling or with a neutral face, they are more likely to trust the smiling partner. Whether this is an adaptive strategy is a separate question. In particular, the reliability of the picture as a guide to the partner's behavior might depend on when the picture was taken. Photos taken independently from the decision situation seem to contain little information, while photos taken when the partner was in the process of the relevant decision seem to contain nonverbal cues that allow subjects to deduce the other players' strategy. Indeed, it is not smiles as such, but rather smiles that are perceived as “honest” that are conducive to trust. “Genuine” or Duchenne smiles are reportedly difficult to fake and most easy to produce when the person is in a positive and sharing mood (Mehu et al. Reference Mehu, Grammer and Dunbar2007). Indeed, lying is known to induce a significant level of cognitive load on the liar, which makes it difficult to react spontaneously, and which reduces subtle movement (Vrij et al. Reference Vrij, Semin and Bull2006). Honest smiling, which involves a contraction of the orbicularis oculi, further has the obvious cost of reducing the smiler's visual field. Doing so might have fitness costs for the smiler, and thus should be done only sparingly, which can explain why smiling convincingly is cognitively costly. Genuine smiles can therefore be taken as an honest signal of the partner being in a positive mood, one that makes deception of the interaction partner less likely. In a situation concerning trust, this might mean that the trustee has already made a trustworthy decision, or that he is aware that he has something interesting to offer to the partner.
We can relate this idea to the smile types identified by Niedenthal et al. “Enjoyment smiles” are smiles elicited by a general positive atmosphere, and which are seen as honest signals that the partner feels comfortable in the current situation. “Affiliative smiles” are smiles that are used as a signal to the partner. Niedenthal et al. claim that affiliative smiles will be initially considered as “true smiles” (section 6.2) and as “false” when: “the perceiver believes that the smile was motivated by a desire to hide, moderate, or justify something negative” (section 6.2, para. 1, emphasis ours). We believe that it will depend partly on the specific situation of the smiler whether the smile is perceived as false or not. Whether an affiliative smile has characteristics that lead it to be judged true or false is therefore influenced not only by the situational knowledge of the observer, but also by unobservable characteristics of the smiler. Therefore, the smile can be seen as a valuable signal of this unobservable information.
Indeed, we have observed in a recent experimental study (Centorrino et al. Reference Centorrino, Djemai, Hopfensitz, Milinski and Seabright2010) that honest smiles lead to higher levels of trust, and that a smile is rated as more convincing when the trustee has larger stakes at his disposition. Therefore, honest smiles can be seen as a costly signal that can only be produced when in a situation potentially beneficial to the observer of the smile, and can thus serve as a signal valuable to the observer and to the smiler alike.