The socio-relational framework presented by Vigil is a new challenge in understanding sex differences in emotion expression. Vigil's model accounts for the fact that there are general sex differences in specific emotional expressiveness: women cry more, laugh more, or show more depressive symptoms, whereas men display more contempt or antagonistic anger. These sex differences would reflect the extent to which men and women aim to maintain different social spheres and are motivated to display different social cues.

Although I appreciate the argument that social sphere plays an important role in this explanatory framework, the model lacks flexibility and sensitivity to the immediate social context (see also Deaux & Major Reference Deaux and Major1987; Eagly & Wood Reference Eagly and Wood1999). Men and women may indeed differ in the ways in which they emotionally adapt in order to enhance their social fitness. However, social fitness should not be exclusively described in evolutionary terms, but should be defined as being able to adjust to one's current social environment, and to form and maintain social relationships in order to achieve one's life goals. These goals may be different for men and women, partly because of their different biological heritage, but also because of their current social roles (e.g., Diekman & Eagly Reference Diekman, Eagly, Shah and Gardner2008; Eagly Reference Eagly1997; Eagly & Wood Reference Eagly and Wood1999). In many societies, these social roles still co-occur with the different biological capacities of men and women; however, it is also clear that in an increasing number of societies men and women have more egalitarian social roles, and therefore meet similar social demands and social restraints.

Social demands and restraints that individuals are facing can be described in terms of both long-term and short-term social-relational goals in a specific situation. These can be roughly distinguished along similar lines as in Vigil's model; namely, approach and avoidance, or to promote relationships (either intimate or more formal) or to exclude oneself from relationships (running away, avoiding, excluding others). We have suggested that emotions have social functions (Fischer & Manstead Reference Fischer, Manstead, Lewis, Haviland and Feldman Barrett2008) that are derived from these social-relational goals. In other words, discrete emotions may serve either affiliation or distancing goals. A goal can be reached through different emotion expressions, however, depending on the social context and one's relational history with a specific person. For example, in a dispute with one's partner about a broken promise, the most important goal is to change the other's behavior. This goal can be reached via submissive behavior (disappointment, crying), but also via antagonistic behavior (direct anger, contempt). On the basis of Vigil's framework, we would predict that in an intimate context, women show submissiveness in order to maintain their intimate social sphere, whereas men would be aggressive in order to secure a larger social network. There is evidence, however, that sex differences in emotional expressions diverge from this general pattern, depending on specific social contexts.

First of all, results from meta-analyses have demonstrated that women show more aggressive behavior in intimate relations than do men (e.g., Archer Reference Archer2004), but more importantly for the present argument, the type of anger expression varies with the type of social role. Women in egalitarian relations report more direct, or antagonistic anger compared to women in traditional relations (Fischer & Evers, under review); moreover, it was shown that women in traditional societies report less antagonistic anger in intimate settings than do women in societies with more egalitarian roles (Fischer et al. Reference Fischer, Rodriguez Mosquera, van Vianen and Manstead2004). In other words, women's anger expressions seem to depend on the direct social demands that they are facing. This is supported by experimental evidence suggesting that women's anger can be predicted by the expected negative consequences of their overt anger (Evers et al. Reference Evers, Fischer, Rodriguez Mosquera and Manstead2005).

Another example of the context specificity of sex differences in emotion expressions is the case of smiling. It is true that women generally tend to smile more than men, but results from a meta-analysis suggest that they do so especially in situations where they are expected to smile more; for example, in situations in which they are observed or evaluated (LaFrance & Hecht Reference LaFrance, Hecht and Fischer2000; La-France et al. Reference LaFrance, Hecht and Paluck2003). Sex differences in smiling therefore seem partly based on gender-specific display rules (see also Stoppard & Gruchy Reference Stoppard and Gruchy1993). This is also evident from the fact that in social contexts with an explicit affiliation goal – for example, in caretaker roles – the sex difference in smiling decreases. This supports the general argument that social goals and one's social role in an interaction determine which emotional behaviors are functional to meet social demands.

A second critical issue with respect to Vigil's explanatory framework relates to the suggestion that making dispositional inferences in terms of capacity and trustworthiness cues is merely based on the characteristics of the display. It is likely, however, that such inferences are also context dependent. Smiling, for example, can be seen as signal of affiliation, appeasement, dominance, or negative self-conscious emotions, depending on the context (e.g., LaFrance & Hecht Reference LaFrance, Hecht and Fischer2000; Shields Reference Shields2002). The same applies to other – less ambiguous – emotional displays, such as crying. Crying can be interpreted as a sign of dispositional powerlessness (someone who is emotional), but also as a sign of temporary powerlessness (someone who is very sad), which may lead to quite different reactions. Perceivers may thus interpret emotional displays in different ways, not only depending on the actual dynamics of the display, but also on the construal of the perceiver, who may take into account the social role and identity of the displayer, and the social context.

We may thus wonder what we actually infer from emotional displays and why. There is as much evidence that we infer individual dispositions, such as capacity and trustworthiness, as social motives (Fridlund Reference Fridlund1994), status characteristics (Tiedens Reference Tiedens2001), or emotivational states (Roseman et al. Reference Roseman, Wiest and Swartz1994). Moreover, why would those inferences be context independent? Given that emotions are elicited in order to help an individual cope with problems in his or her environment, expressive displays are directly intended to change our social relations. This implies that we would infer information not only about the other person, but also about our relationship with this other person. In other words, we do not only infer from an angry face that someone has resources, but also that this person is more powerful and ready to retaliate.