Although we commend Niedenthal et al. for their effort to bring together the ethological functions of smiling and the neurocognitive mechanisms involved in its perception, we remain sceptical about their success, for the model they propose rests on a classification of smiles that is inconsistent with primate ethology and evolutionary theory.
Our reading of the ethological literature leads us to question the three types of smiles that Niedenthal et al. individuate (i.e., enjoyment, affiliative, and dominance smiles). In particular, the case made for the dominance smile category is based on anecdotal evidence and post hoc rationalization of disparate research. A better coverage of the abundant literature on the relationship between smiling and social status would have made a stronger argument in favor of, or against, a “dominance” smile category. Although dominant individuals have also been observed smiling during interactions with subordinates (LaFrance & Hecht Reference LaFrance, Hecht, Philippot, Feldman and Coats1999), smiling has often been associated with submission (Hall Reference Hall, Hogan and Carter2002; Schmid-Mast & Hall Reference Schmid-Mast and Hall2004). It is surprising to see in this target article a category of dominance smiles while the psychological and ethological literature could equally well support arguments in favor of a submission smile category (e.g., Mehu & Dunbar Reference Mehu and Dunbar2008), which would also be distinct from the affiliative smile that Niedenthal et al. define. Finally, we doubt that smiling functions to establish dominance. The ethological literature rather suggests that smiling mitigates the deleterious effect dominance may have on the maintenance of egalitarian relationships (Flack & de Waal Reference Flack and de Waal2007; Preuschoft & van Hooff Reference Preuschoft, van Hooff, Segerstrale and Molnàr1997).
A major asset of the target article is to propose a classification of smiles based on their function, rather than one based on their morphological features (sect. 2.2). However, from an evolutionary viewpoint, this functional typology is problematic because it mixes proximate and ultimate explanations of behavior (Tinbergen Reference Tinbergen1963). For example, the enjoyment smile is a category based on one of the proximate factors involved in smile production (i.e., positive emotion), whereas the “affiliative” and “dominance” smile categories reflect the ultimate function of smiles. As these criteria pertain to two different levels of explanation, they are not exclusive of each other. Indeed, research has shown that Duchenne smiles (also called enjoyment smiles) could be reliable signals of cooperative (and therefore affiliative) intentions; the honesty of the signal would be ensured by the emotional component that is difficult to strategically manipulate (Brown et al. Reference Brown, Palameta and Moore2003, Mehu et al. Reference Mehu, Grammer and Dunbar2007). The alternative view that enjoyment smiles are merely cues, that is, symptoms of positive emotional experience without communicative function (Ekman Reference Ekman1997), could not be invoked to justify the distinction between enjoyment smiles and affiliative smiles for the reason that cues are not functionally related to a receiver's response (Hasson Reference Hasson1994) and are, therefore, difficult to integrate within the sender-receiver process that Niedenthal et al. describe.
The confusion of proximal and ultimate factors leads to misunderstanding the crucial issue about the nature of smiles as communicative (smiles as social signals) versus denotative (smiles as symptoms of emotions) facial expressions. This is also illustrated in the case of dominance smiles in which Niedenthal et al. conflate the bodily expression of a positive feeling of superiority and the communicative intentions directed at social challengers. This conflation bears direct consequences on the embodied simulation processes that Niedenthal et al. claim to be central for interpreting the meaning of smiles. For example, their model could with difficulty explain how the perceiver of a dominance smile experiences negative feelings that have no obvious correspondence with the feelings actually experienced by the smiler, whereas alternative accounts regarding the functional role of such smiles (e.g., defusing social conflict) escape this contradiction.
We believe that an appropriate classification of smiles should be based on EITHER proximate causes (smiles with different motivational basis) OR ultimate causes (smiles with different adaptive functions). In the latter case, the classification is more laborious because the function of smiling is poorly documented in the literature (a situation that may explain Niedenthal et al.'s difficulty in achieving a satisfactory functional classification of smiles). Alternatively, a classification based on morphological and dynamic aspects of smiles may be useful (e.g., Duchenne vs. non-Duchenne). Such a categorization would be more relevant to the SIMS model, since the processing of a smile's physical features is central to embodied simulation (see sect. 3.2 of the target article).
Because gaze represents an important vector of communicative intentions, the overlook of communicative/denotative distinction also weakens the theory regarding the role of eye contact. According to Niedenthal et al., eye contact would be a sufficient, but not necessary, trigger of the embodied-simulation process: When eye contact occurs, it triggers simulation of the current expression leading to embodied simulation and a better understanding of smiles' meaning. On the other hand, when eye contact is obstructed, the simulation is based on conceptual knowledge and is more prone to error (sect. 5, para. 5 and sect. 6.2, para. 4). This implies that a simulation triggered by eye contact is more likely to result in accurate interpretations of smiles, making eye contact slightly more critical (if not necessary) than presented by the authors.
We are therefore surprised that Niedenthal et al. make little reference to the rich literature on eye contact and gaze direction on emotion recognition. Various studies on the effect of gaze direction on emotion recognition have shown that eye contact may indeed increase perceived emotion (relatively to averted gaze), but it may also decrease it, depending on the nature of the emotional expression (see Bindemann et al. [Reference Bindemann, Burton and Langton2008] for recent research and references therein). These data hardly fit with a general enhancing role of eye contact in emotion perception.
We propose that Niedenthal et al. consider an alternative account based on the functional role of eye contact according to which the effect of eye direction on emotion perception reflects the congruency of the sender's gaze direction with the relevance of the emotional signal for the receiver: Angry (and happy) faces are more relevant to the receiver when she is gazed at (i.e., direct gaze), while fearful faces with averted gaze may signal a relevant threat for the receiver (Sander et al. Reference Sander, Grandjean, Kaiser, Wehrle and Scherer2007). This view implies that the role of eye contact occurs at a later stage of facial expression recognition than that posited by Niedenthal et al. Instead of having a ubiquitous enhancing effect, eye contact would thus increase or decrease the perceived intensity of the emotional display, depending on its relevance for the observer. Therefore, the involvement of brain structures such as the amygdala need not reflect the uncertainty associated with smiles (as Niedenthal et al. presuppose; see sect. 6.1.1, para. 1) but, instead, the self-relevance of emotional expressions in their context of occurrence (N'Diaye Reference N'Diaye, Sander and Vuilleumier2009).
Although we commend Niedenthal et al. for their effort to bring together the ethological functions of smiling and the neurocognitive mechanisms involved in its perception, we remain sceptical about their success, for the model they propose rests on a classification of smiles that is inconsistent with primate ethology and evolutionary theory.
Our reading of the ethological literature leads us to question the three types of smiles that Niedenthal et al. individuate (i.e., enjoyment, affiliative, and dominance smiles). In particular, the case made for the dominance smile category is based on anecdotal evidence and post hoc rationalization of disparate research. A better coverage of the abundant literature on the relationship between smiling and social status would have made a stronger argument in favor of, or against, a “dominance” smile category. Although dominant individuals have also been observed smiling during interactions with subordinates (LaFrance & Hecht Reference LaFrance, Hecht, Philippot, Feldman and Coats1999), smiling has often been associated with submission (Hall Reference Hall, Hogan and Carter2002; Schmid-Mast & Hall Reference Schmid-Mast and Hall2004). It is surprising to see in this target article a category of dominance smiles while the psychological and ethological literature could equally well support arguments in favor of a submission smile category (e.g., Mehu & Dunbar Reference Mehu and Dunbar2008), which would also be distinct from the affiliative smile that Niedenthal et al. define. Finally, we doubt that smiling functions to establish dominance. The ethological literature rather suggests that smiling mitigates the deleterious effect dominance may have on the maintenance of egalitarian relationships (Flack & de Waal Reference Flack and de Waal2007; Preuschoft & van Hooff Reference Preuschoft, van Hooff, Segerstrale and Molnàr1997).
A major asset of the target article is to propose a classification of smiles based on their function, rather than one based on their morphological features (sect. 2.2). However, from an evolutionary viewpoint, this functional typology is problematic because it mixes proximate and ultimate explanations of behavior (Tinbergen Reference Tinbergen1963). For example, the enjoyment smile is a category based on one of the proximate factors involved in smile production (i.e., positive emotion), whereas the “affiliative” and “dominance” smile categories reflect the ultimate function of smiles. As these criteria pertain to two different levels of explanation, they are not exclusive of each other. Indeed, research has shown that Duchenne smiles (also called enjoyment smiles) could be reliable signals of cooperative (and therefore affiliative) intentions; the honesty of the signal would be ensured by the emotional component that is difficult to strategically manipulate (Brown et al. Reference Brown, Palameta and Moore2003, Mehu et al. Reference Mehu, Grammer and Dunbar2007). The alternative view that enjoyment smiles are merely cues, that is, symptoms of positive emotional experience without communicative function (Ekman Reference Ekman1997), could not be invoked to justify the distinction between enjoyment smiles and affiliative smiles for the reason that cues are not functionally related to a receiver's response (Hasson Reference Hasson1994) and are, therefore, difficult to integrate within the sender-receiver process that Niedenthal et al. describe.
The confusion of proximal and ultimate factors leads to misunderstanding the crucial issue about the nature of smiles as communicative (smiles as social signals) versus denotative (smiles as symptoms of emotions) facial expressions. This is also illustrated in the case of dominance smiles in which Niedenthal et al. conflate the bodily expression of a positive feeling of superiority and the communicative intentions directed at social challengers. This conflation bears direct consequences on the embodied simulation processes that Niedenthal et al. claim to be central for interpreting the meaning of smiles. For example, their model could with difficulty explain how the perceiver of a dominance smile experiences negative feelings that have no obvious correspondence with the feelings actually experienced by the smiler, whereas alternative accounts regarding the functional role of such smiles (e.g., defusing social conflict) escape this contradiction.
We believe that an appropriate classification of smiles should be based on EITHER proximate causes (smiles with different motivational basis) OR ultimate causes (smiles with different adaptive functions). In the latter case, the classification is more laborious because the function of smiling is poorly documented in the literature (a situation that may explain Niedenthal et al.'s difficulty in achieving a satisfactory functional classification of smiles). Alternatively, a classification based on morphological and dynamic aspects of smiles may be useful (e.g., Duchenne vs. non-Duchenne). Such a categorization would be more relevant to the SIMS model, since the processing of a smile's physical features is central to embodied simulation (see sect. 3.2 of the target article).
Because gaze represents an important vector of communicative intentions, the overlook of communicative/denotative distinction also weakens the theory regarding the role of eye contact. According to Niedenthal et al., eye contact would be a sufficient, but not necessary, trigger of the embodied-simulation process: When eye contact occurs, it triggers simulation of the current expression leading to embodied simulation and a better understanding of smiles' meaning. On the other hand, when eye contact is obstructed, the simulation is based on conceptual knowledge and is more prone to error (sect. 5, para. 5 and sect. 6.2, para. 4). This implies that a simulation triggered by eye contact is more likely to result in accurate interpretations of smiles, making eye contact slightly more critical (if not necessary) than presented by the authors.
We are therefore surprised that Niedenthal et al. make little reference to the rich literature on eye contact and gaze direction on emotion recognition. Various studies on the effect of gaze direction on emotion recognition have shown that eye contact may indeed increase perceived emotion (relatively to averted gaze), but it may also decrease it, depending on the nature of the emotional expression (see Bindemann et al. [Reference Bindemann, Burton and Langton2008] for recent research and references therein). These data hardly fit with a general enhancing role of eye contact in emotion perception.
We propose that Niedenthal et al. consider an alternative account based on the functional role of eye contact according to which the effect of eye direction on emotion perception reflects the congruency of the sender's gaze direction with the relevance of the emotional signal for the receiver: Angry (and happy) faces are more relevant to the receiver when she is gazed at (i.e., direct gaze), while fearful faces with averted gaze may signal a relevant threat for the receiver (Sander et al. Reference Sander, Grandjean, Kaiser, Wehrle and Scherer2007). This view implies that the role of eye contact occurs at a later stage of facial expression recognition than that posited by Niedenthal et al. Instead of having a ubiquitous enhancing effect, eye contact would thus increase or decrease the perceived intensity of the emotional display, depending on its relevance for the observer. Therefore, the involvement of brain structures such as the amygdala need not reflect the uncertainty associated with smiles (as Niedenthal et al. presuppose; see sect. 6.1.1, para. 1) but, instead, the self-relevance of emotional expressions in their context of occurrence (N'Diaye Reference N'Diaye, Sander and Vuilleumier2009).