R1.1. Empirical issues related to the biology of sex differences

Of all the major findings that I described in the target article, a few empirical inferences were parsed by the commentaries. The first finding is greater facial expression–processing abilities in females, with the exception of anger, of which males are predicted to be more sensitive. Consistent with the SRFB, commentators LoBue & DeLoache show that females are better at detecting social, but not nonsocial, stimuli as compared with males. However, LoBue & DeLoache also present some data that suggest that both males and females detect threatening emotions (e.g., anger and fear) more efficiently than non-threatening emotions (e.g., happiness and sadness), leading the researchers to suggest that males and females may not differ in the ability to detect threat. I recently conducted a preliminary analysis that may shed light on the commentators' findings.

Using a large, representative sample of young adults (n=808), I found that women were just as likely as men to perceive threatening (i.e., anger, fear, disgust) versus non-threatening (i.e., joy, sadness, surprise) emotions from ambiguous facial stimuli. However, when the emotional interpretations were re-coded as either signaling dominance (i.e., joy, anger, disgust) or submissiveness (i.e., sadness, fear, surprise), males were more likely to perceive the former (Vigil, submitted a). These findings suggest that males and females differ in how they interpret dominant and submissive cues in others. The SRFB explains these phenomena as reflecting a natural history characterized by the interchange of higher levels of capacity cues among males and trustworthiness cues among females. Likewise, the SRFB predicts that most sex differences in (social) perceptual and expressive biases should become exaggerated upon adolescence (not at the ages of children that LoBue & DeLoache investigated, for example). Historically, adolescence is when males and females may have required, and thus developed, specialized social skills for regulating different types of relationships. This age-dependent constraint may also explain why Izard, Finlon, & Grossman (Izard et al.) did not find sex differences in emotional expression in their samples of preschoolers (though see the comments further regarding racial differences in emotional expressivity). Clearly, we need more research to examine the universality of age and sex differences in the expression of emotion across people from varying cultural backgrounds.

Other empirical concerns were for the predictions that males are more sensitive to take risks (displays of capacity) and that females are more sensitive to display altruism (e.g., kindness). To address these hypotheses, Basso & Oullier highlight findings from two experimental tasks (the “Dictator Game” and the “Ultimatum Game”) which involve the exchange of economic credits. The researchers found support for the prediction that males are less risk averse than females, but contend that females show no greater willingness to engage in true altruism (costly actions) than do males. The latter findings can be understood with a clearer description of what the SRFB actually predicts.

First, unconditional self-sacrifices are rarely adaptive in most contexts. It would therefore make no sense for either sex to endure an actual cost to personal fitness, without the possibility of a reciprocated reward. Rather, the SRFB predicts that females are only more sensitive to demonstrate the appearance of, or willingness to engage in, altruism, not the actual and unconditional provisioning of resources, as the experimental tasks demanded. I would therefore predict that, while not actually engaging in costly actions (e.g., expending one's credits), females will report greater compassion and felt guilt for their actions (e.g., not giving more credits to their experimental partners). Males, in contrast are predicted to demonstrate higher levels of capacity cues such as felt pride for successfully “cheating” their partners.

Further, the experimental tasks that Basso & Oullier describe are not as analogous to the functional display of emotions (especially trust cues) and for the overall negotiation of relationships as the commentators imply. This is because actual relationships are formed and maintained over time. In the experimental situations, participants are constrained by a limited number of options, aware of the eventual termination of the task, and not dependent on their experimental partners in any meaningful way (e.g., for sustained self-enhancement or protection). Another drawback of using economic games to simulate social interactions is that the games rely on the exchange of arbitrary (e.g., unearned) credits, again limiting the inherent interest of the tasks themselves. An extremely high level of creativity will surely be needed to design experimental protocols that can even remotely simulate the actual importance of non-kin relationships, as well as the scope of dynamic behaviors that humans use to regulate their relationships.

R1.2. Theoretical issues related to the biology of sex differences

The two major contentions that some commentators have with my theoretical models concern the utility of using an evolutionary approach to understanding human sex differences, as well as the specific types of sexual selection principles that would have favored their expression. The first set of issues is whether contemporary sex differences in social psychology even exist at all, and if so, whether they are solely the result of evolutionary forces. Alternative models to the evolutionary approach often highlight the importance of culture, proximate learning mechanisms, and motivational forces such as “norm formation” and “gender stereotyping.” Several of the commentators (e.g., Fugate, Gouzoules, & Barrett [Fugate et al.]; Vermeulen) took this approach, and Zayas, Tabak, Günaydýn, & Robertson [Zayas et al.] described several reasons why sex-typical emotionality can be better explained by models that emphasize individual learning processes rather than naturally selected behaviors. Specifically, Zayas et al. contend that: (a) patrilocality is not favored across all human cultures, (b) patrilocality does not result in unique social selection constraints for males compared with females, and (c) human males and females do not express emotionality differently and, if they do, that the differences are not the result of biological dispositions. The authors conclude by describing a social modeling/learning explanation of gender development which appears to be more complementary than contradictory to the SRFB and the overall thesis that sex differences in social behaviors are rooted in evolutionary design.

First, human patrilocality is an example of a plastic or facultative phenotype. Phenotypic plasticity, or variability in the expression of traits, is an essential characteristic of ontogeny because it enables the individual to develop phenotypes that are specialized for different types of ecological conditions. As I described in the target article, patrilocality is the predominant social migratory system in traditional societies; however, as evidenced through anthropological records, this pattern does vary somewhat according to local, ecological conditions. These conditions appear to encompass historical ties between environment factors (e.g., regions where resources are scarce and groom labor is used as a bride service) and social customs (e.g., historically low levels of inter-group hostility). Under these special conditions, humans may benefit from alternative locality customs. Under more typical conditions (i.e., involving inter-group hostility), male-biased philopatry is associated with numerous biological incentives (see Geary Reference Geary2009), as elaborated by Madison and as I describe in further detail below (see sect. R1.3).

Second, Zayas et al. suggest that because certain types of social ecologies (e.g., consisting of acquaintances) are associated with fitness-reducing costs, nature would not have favored women to expose themselves to these conditions. However, according to evolutionary reasoning, it is because interacting with different types of affiliates is associated with both fitness benefits (e.g., genetic out-breeding) as well as costs (e.g., risk of rejection), that these behaviors have been designed by evolutionary trial and error to evidence plasticity. The existence of cost-benefit fitness trade-offs is a necessary condition for the evolution of phenotypic flexibility, and as such, the facultative expression of social philopatry, differential motivations to form distinct types of relationships, and the development of expressive behaviors that facilitate these goals. The importance of these processes is what motivated me to focus on how variability in social behaviors can be understood as a function of certain fitness-related cost-benefit trade-offs that covary with specific conditional factors (e.g., sex, age, personal experiences, social network dynamics, ecological factors such as climate) and situational stimuli (e.g., the presence of different audiences).

Third, Zayas et al. question the evidence that human males and females have evolved the proclivity to develop specialized social behaviors, based on the modest effect sizes that are sometimes found in this type of research. They support this concern by stating that “prolonged directional selection would have resulted in relatively large sex differences in emotional expressivity” and “this is clearly not the case.” However, it is unclear how the commentators estimate what should be the appropriate effect size for psychological processes. From my perspective, nature selects biological designs to be highly specialized for their own set of environmental contingencies, and because such specialization often involves phenotypic plasticity, it is implausible to assume a degree of evolutionary design from the weight of a statistical effect size. In terms of sex differences, as long as a mean sex difference in psycho-biological processes exists, even at a proximate level of causation (e.g., memory retrieval, as suggested by Fugate et al.), the presumption should be that some degree of psychobiological specialization has taken place.

Finally, Zayas et al. suggest that sex differences in emotionality are driven by contemporary, individualistic factors (e.g., personal development of social norms) rather than evolutionary pressures. I don't agree with this dichotomy and instead believe that societal experiences (e.g., exposure to gender norms) and the psychological (e.g., learning) mechanisms that process these experiences are ultimately constrained by, and thus a reflection of, biological structures and sensitivities. As a result, models that rely exclusively on social learning/modeling explanations of sex differences, in the absence of evolutionary specialization, are not able to account for (a) why males and females usually identify with, and model, same-sex individuals; (b) why mothers and fathers interact with their children differently (e.g., talking vs. doing activities) across generations; (c) why males and females form distinct peer networks; (d) developmental and even prenatal sex differences in the rudiments of social behaviors (e.g., eye-contact and touching); and (e) cross-cultural universality of masculine (e.g., physical violence) and feminine (e.g., crying) behaviors.

R1.3. Theoretical issues related to the sexual selection of sex differences

Within the evolutionary school of thought, Lozano makes a number of excellent points regarding the utility of examining both intra-sexual and inter-sexual selection forces to account for human sex differences. Lozano highlights several biological scenarios that may be related to sex-typical behavior patterns, including the possibility that masculine and feminine dispositions are the products of inter-sexual selection pressures (e.g., similar to face and body shape) rather than skills that are needed to manipulate same-sex relationships. Of course, intra-sexual selection pressures often operate in parallel with mate preferences, resulting in behaviors with pleiotropic functions. According to the SRFB, for instance, traits that signal capacity (e.g., physical attractiveness) and trust (e.g., kindness) are essential for attracting all types of (non-kin) relationships. These traits should thus be advantageous for regulating interactions with romantic and non-romantic peers.

However, let me directly address the crux of Lozano's hypothesis: that dominance may signal maturation and submissiveness may signal youth, and thus human mate preferences drove the evolution of masculinity and femininity. The reasons why sex differences in emotionality were probably not selected by mate preferences, irrespective of within-sex competition pressures, is because mate preferences cannot account for (a) implicit preferences for same-sex friendships, (b) social motivations to construct unique peer networks, and (c) sexually dimorphic social styles in early development (i.e., prior to puberty). Moreover, (d) sex-typical emotive gestures such as crying in females and aggression/threat promotion in males are not directly preferred in prospective mates, at least not to the same extent of well-established mating characteristics such as age, beauty, and resource acquisition. Finally, males' and females' social styles could not have been selected from mate choices, because (e) these are the very distinctions that often result in “miscommunication” between the sexes, a phenomenon that is more likely to deteriorate, rather than strengthen, pair bonding.

Lozano is therefore correct in stating that intra-sexually selected traits can also affect mate choices; it is evident (and predicted) that they sometimes do. However, I strongly believe that the majority of the dimorphisms (e.g., social motivations, emotional expressivity, and speech styles) that I reviewed in the target article are probably the result of intra-sexual selection pressures. In support of this hypothesis, I recently found that the previously mentioned pattern of males and females to perceive differential cues of dominance or submissiveness in facial stimuli is moderated by sex-typical relationship dynamics (Vigil, submitted a). Males with larger social spheres (i.e., numbers of friendships) were more likely to perceive dominant emotions (e.g., joy and anger) than males with smaller social spheres, and as compared to females in general. Regarding Lozano's related comment on the evolution of romantic love, it is unclear whether this sensation is expressed differently by males and females, and, if it is, whether such differences were selected by mate preferences and thus used to facilitate mating and/or parenting strategies.

I nonetheless agree with Lozano that the SRFB can only be enhanced with the integration of models that can incorporate the simultaneous operation of additional forms of selection pressures that are known to drive sex differences in social behaviors (e.g., differential parental investment). In this sense, I see male-biased philopatry as a supplementary adaptation to the basic human mating strategy of resource acquisition in males. If males can enhance mate value through resource inheritance and if patrilocality can enhance resource inheritance, then this form of philopatry may directly result from human mating constraints. Thus, we may be able to organize a tentative chain of selection pressures (e.g., parental investment→mate preferences→male-biased philopatry→sex-typed social styles) that can more fully account for the evolution of human sex differences in emotionality.

R2.1. Ultimate versus proximate levels of analyses

It is important for social scientists to remember that psychological phenomena can be adequately explained through both proximate and ultimate levels of analyses (Tinbergen Reference Tinbergen1963). Proximate explanations incorporate physiological, situational, and experiential mechanisms and are able to answer what-type questions (e.g., what learning experiences contribute to sex differences). This level of analysis is essential for measuring individual differences in phenotypic expression. Ultimate explanations instead incorporate micro- and macro-evolutionary forces and are able to answer why-type questions (e.g., why are human males different or similar to human females). This level of analysis is essential for measuring the functionality (and thus often assumed existence) of the psychological phenomena. Sound ultimate levels of explanation operate in parallel with sound proximate levels of explanation, and to view them as contradictory is erroneous. At the same time, both proximate and ultimate levels of analyses are required to model the form and function of psychological adaptations. I attempted to do this in the target article by describing how some social, psychological processes in humans can be understood in the context of evolutionary cost-benefit fitness trade-offs that cause individuals to respond to personal life experiences and situational factors (e.g., audience characteristics) through sex-typical and sex-general behavior patterns. In the following subsections, I describe how two proximate sources of causation – individual life experiences and accompanying learning mechanisms – fit into the broader socio-relational framework.

R2.2. Range of reactivity

Some of the commentators contended that evolutionary approaches to understanding sex differences are too constrained, that they don't incorporate learning experiences, and that, because sex differences are sometimes not found, their existence should be denied all together. These contentions can be resolved through a brief description of the concept of range of reactivity. Ranges of reactivity simply refer to the continua of possibilities (and constraints) that any given phenotype can be expressed. Some types of phenotypes such as eye color are not as plastic and thus have very narrow ranges of reactivity; these phenotypes do not benefit from conditional modifications and are thus designed to be less influenced by environmental or experiential factors. Other types of phenotypes such as social behaviors are highly plastic and are more modifiable by life experiences. These phenotypes have wider ranges of reactivity that support the ontogeny of ecological specialization. As I mentioned earlier, phenotypic plasticity is driven by cost-benefit fitness trade-offs and operates to modify developmental trajectories in ways that optimize personal attributes, within the constraints and opportunities of the local environment.

An example of this concept for understanding sex differences in social behaviors is illustrated in Figure R1. Males and females have evolved different ranges of reactivity or proclivities to develop masculine and feminine behaviors. The specific points along the continuum at which people express their unique combinations of masculine/feminine traits are influenced by individual (e.g., genetic) and experiential factors and by the proximate learning mechanisms that process life experiences. Figure R1 shows that males and females both have wide ranges of reactivity to develop prototypically masculine/feminine behaviors, and more narrow ranges of possibility to develop atypical behaviors. Although there is a great deal of variance within each sex, males and females are sensitive to develop specialized expressive styles for regulating different types of social ecologies. This concept of range of reactivity thus makes it possible to integrate most of the “alternative,” mostly proximate learning-based models that the commentaries have highlighted.

Figure R1. Epigenesis of evolved proclivities within a range of reaction. Males and females are sensitive to develop unique behaviors styles. Individual differences (e.g., genetics [not shown]) and learning experiences moderate the degree to which the behaviors are expressed.

R2.3. Proximate learning mechanisms

For example, Swain describes interesting research showing that same-sex parent-infant dyads evidence more behavioral and arousal synchrony in their daily interactions as compared to opposite-sex dyads. These findings again highlight the utility of using intra-sexual models for understanding social behaviors, but also show how proximate learning experiences, such as classical conditioning, can strengthen sex-typical behavior patterns. As mentioned earlier (see Fugate et al.,Zayas et al.), parents also reinforce sex-typicality through social modeling (observation and mimicking) mechanisms. Likewise, Basso & Oullier show how group demands (expected rules in organizational settings) can lead to acceptance and rejection, and thus how operant conditioning can alter sex differences in social behaviors. Similar arguments are made by Fischer and Wiefel & Schepker, who contend that different types of relational demands, such as history of trust, play a pivotal role in the expression of emotion.

The SRFB hypothesizes that sex differences in emotionality are largely based on the differential sensitivity to advertise trust cues. As noted by Wiefel & Schepker, the emphasis on trust links the SRFB to other models of social psychology such as “Attachment Theory” (Bowlby Reference Bowlby1969). I believe the SRFB extends traditional applications of attachment models by showing why early childhood experiences (i.e., behavioral responses of others) are associated with the development of specialized and functional interaction styles. Infant-caregiver experiences probably form the basis of self-conceptualizations of reciprocity potential, as well as the basis for social expectations. For example, individuals who experience distrustful relationships in their life develop increased perceptual and expressive proclivities to detect and express anger (Vigil et al., submitted). Anger behaviors are in turn effective at provoking distancing responses from peers (e.g., Vigil Reference Vigil2008). Thus, this research shows that the types of learning that occur within relationships can and do alter the development of expressive styles (e.g., hostility) that primarily operate to regulate individuals' unique social conditions.

In an interesting caveat to the literature, Izard et al. found that minority children showed reduced sex differences in the expression of emotion. This finding is analogous to similar findings in adults showing reduced sex differences among African Americans as compared to European Americans (Vrana & Rollock Reference Vrana and Rollock2002) and Asian Americans (Vigil, in preparation). The ethnic discrepancies are probably the result of variation in social, structural (e.g., in terms of social spheres), and relational demands (e.g., relative earning capacity and perceptions of peer trust) among males compared with females, for people from different cultural backgrounds.

In a related commentary, Fischer describes how females express more antagonistic aggression in intimate situations with less traditional and more egalitarian relationship partners. I agree with Fischer that the findings can be explained according to the expected outcomes of the anger behaviors. Specifically, the SRFB predicts that individuals should express more risky forms of aggression, such as antagonistic anger, when their relational partners are perceived to have lower capacity (to retaliate) than themselves (target article, Fig. 1). Lastly, I would like to clarify Fischer's insinuation that females express higher levels of aggression than males. These findings are typical for research relying on self-report measures, which usually include a multitude of non-risky behaviors (e.g., arguing with peers). In ethnographic studies that measure acts of violence that involve a greater risk of death (e.g., homicide), males are far more likely to express these behaviors than are females (e.g., Archer Reference Archer2009).

R3.1. Human psychology as an exploitive system

It is interesting that some of the commentators, such as Goldstein Ferber, tend to view my models as too individualistic and not focusing on mutual goal attainment, whereas other commentators, such as Buss, instead imply that I could have emphasized individual fitness gains even more strongly. Buss distinguishes three types of resource acquisition strategies (i.e., personal efforts, cooperative efforts, and exploitive efforts), whereas I consider all three strategies as operating off the same exploitive, and hence personal-fitness-enhancing, motivations. I agree with Buss that anger and related capacity displays (e.g., signals of prowess) operate to exploit the reciprocity potential (e.g., material resources, fertility) of others. However, I also believe that trustworthiness displays (e.g., expressed kindness and vulnerability) are equally exploitive. By advertising trustworthiness cues (e.g., via crying) to other people, individuals are able to exploit the motivation of others to advertise their own reciprocity potential such as via sympathetic responses. Relationship formation may thus ultimately function as the context within which individuals can readily interchange reciprocal displays of capacity and trustworthiness with others in the form of expedient and continuous investment cues, respectively.

R3.2. Fundamental mechanisms of exploitation

One of the most important commentaries is from Todorov, who has been constructing a neurocognitive model of affective processing that is remarkably similar to my own. What is impressive about this convergence is that Todorov and I derived our conclusions from two very different analytical strategies. Todorov derived his models from a bottom-up approach, using empirical findings to build a conceptual model, whereas my models were constructed from a purely top-down or theory-driven analysis. As predicted from my models, and as was found by Todorov, people evaluate others along two dimensions, what Todorov refers to as valence/trustworthiness and dominance/power impressions.

The SRFB extends Todorov's findings in several ways that include: (a) conceptualizing the natural essence of these social properties as fundamental components of reciprocity potential; (b) extending the utility of the dimensional models to explain variation in expressive behaviors including displayed affect; and (c) describing some cost-benefit fitness trade-offs that support situation-based and condition-based variation in expressive behaviors. Todorov and I agree that social, perceptual processing of capacity and trust cues in others precipitates affective responses in the individual, and that affective responses ultimately function to induce affiliation versus avoidance from others. I simply extend this argument to model the fundamental dimensions of expressive behaviors as behavioral advertisements of these same social properties. According to the SRFB, all forms of expressive behaviors (i.e., behaviors that are both observable and modified by the social context) are dynamic advertisements of capacity and/or trustworthiness cues, which ultimately function to control how other people respond to the individual.

I recently found support for this hypothesis as it relates to affective processing by examining how perceptions of other people's capacity and trustworthiness trait levels are associated with interpersonal dispositions and discrete emotive reactions toward the people (Vigil, submitted b). Specifically, I showed that trustworthiness impressions are parsimonious predictors of the motivational desire either to affiliate with (i.e., “form a friendship”) or to avoid (i.e., “stay away from”) social objects. However, simultaneous impressions of capacity trait levels are necessary (and sufficient) for predicting discrete affiliative (e.g., sympathy vs. admiration) and avoidant (e.g., fear vs. disgust) emotions, as predicted in the target article's Figure 1. The types of emotional reactions that individuals express in turn affect whether other people respond to the individual with either affiliative or avoidant dispositions of their own (see Vigil Reference Vigil2008). Collectively, these findings suggest that both the perceptual processing and expression of human affect can be understood along several broad dimensions of social relevance. Affect behaviors operate by advertising the essential properties of reciprocity potential (i.e., capacity and trustworthiness cues) in order to regulate social fitness by selectively promoting affiliation versus avoidance across the individual's relationships.

In this regard, I disagree with Todorov's suggestion that broad conceptual dimensions of affective processing are insufficient for predicting specific emotional reactions in vivo. In fact, in the target article, I attempted to outline several overlapping dimensions that can be simultaneously applied for just this purpose. Again, some of these dimensions can be conceptualized as (a) the perception and (b) the expression of capacity/trustworthiness cues, (c) the motivation to promote affiliation versus avoidance (target article, Fig. 1), (d) the signaler's sex (Fig. R1), (e) recent life experiences (target article, Fig. 2), and (f) characteristics of the signaler's audience (target article, Fig. 3), among several other probable dimensions (e.g., climatic ecology and health status).

The predictive validity of these hypotheses will ultimately rest on the universality of what I presented as basic behavioral responses. For example, Goldstein Ferber questions whether people from different cultures (and whether different species) respond to expressions of vulnerability in trusted affiliates with increased social support. I believe that this, as with most of the broad response patterns that I described in the target article (e.g., distancing reactions toward angry peers), are universal to humans. I would also suggest that submissive displays (i.e., trustworthiness cues) are far rarer in nature than is the demonstration of dominant displays (i.e., capacity cues). Specifically, submissive behaviors should covary with the social complexity of each species. Species that form continuous relationships (e.g., certain primates, wolves, dolphins, elephants, and lions) should be most likely to signal trustworthiness gestures (e.g., pain behaviors, high-pitched utterances, non-threatening eye contact), as these mechanisms are predicted to be functional for regulating longer-term relationships.

R3.3. Variation in affective responses

Several of the commentators are concerned with the ability to predict certain social reactions (e.g., indecisiveness about others) and emotive gestures (e.g., different types of laughing and crying behaviors) that were not fully addressed in the target article. Lozano and Goldstein Ferber, for example, find my approach/withdrawal heuristic to be too constrained to integrate what were described as more dynamic social reactionary strategies, including wait and see, freezing, and simply revealing oneself (e.g., to potential predators). They also contend that individuals must monitor and implicitly process multiple cost-benefit fitness trade-offs that are involved with interacting with different people. I agree that interacting with any sort of environmental stimuli, and especially other people, which are the least predictable stimuli humans can encounter, present multiple and simultaneous costs and benefits, as described earlier. However, I suggest that a dichotomous (affiliative/avoidant) heuristic can sufficiently account for variant response behaviors as well as simultaneous appraisal processes. This is possible as long as humans have the heuristical algorithms for processing the net outcome of either affiliating with, or avoiding, others. From my perspective, wait-and-see strategies, such as experiences of curiosity, may operate as low-intensity or low-valence approach dispositions; for instance, motivating the future appraisal of others. Freezing behaviors, in contrast, are obviously more beneficial for evading dangerous stimuli, by using a behavioral strategy that is specialized differently than other forms of avoidant reactions (e.g., displays of fear or violence).

According to the SRFB, discrete affective sensations (e.g., feelings of sadness) should covary with, and could thus be predicted by, discrete expressive displays (e.g., sadness behaviors) and the systematic reaction of others (e.g., approach from intimate affiliates and avoidance from unfamiliar affiliates). This thesis could be applied to the study of variant forms of expressive behaviors throughout the life span, including crying behaviors in infancy and adulthood. I agree with Wiefel & Schepker that babies utilize crying as a powerful tool for manipulating others, and that caregivers play a key role in shaping the development of affective processes in children, such as through the proximate learning mechanisms I have mentioned. Wiefel & Schepker also describe how infants use crying to solicit attention, and that the attention can be needed for various reasons, including hunger, fatigue, and overstimulation. It makes sense that infants primarily rely on trustworthiness rather than capacity cues to manipulate others, as submissive gestures can best accentuate an infant's actual vulnerability and because these behaviors are most effective for regulating intimate, co-dependent relationships.

Lyons also does an excellent job of outlining the many possible functions (exploitive benefits) of crying behaviors, but misinterprets an associated premise from the SRFB. Lyons implies that I suggest that vulnerability displays (e.g., crying) are only functional by displaying one's actual willingness to reciprocate with others. Rather, vulnerability displays such as crying, worrying, and perhaps pain sensations are also adaptive by signaling reduced threat, which may simply be effective at signaling a safe context within which other people can advertise their own reciprocity potential (e.g., via sympathizing behaviors). I believe that humans are systemically motivated to advertise capacity and trust cues continuously to other people, given every available opportunity, including conditional and situational openings. This would create an inherent fitness incentive for providing other people with the opportunity to demonstrate their own reciprocity potential. It is therefore possible that humans produce certain behaviors, such as playful aggression and crying, to disarm the threat interpretations of others and to signal the opportunity to reciprocate social demonstrations of reciprocity with others, in a safe relational context, irrespective of more specialized relationship behaviors.

On a related note, Lyons mentions that crying isn't always perceived as attractive, a fact that is predicted by the SRFB. According to the target article's Figure 3, for instance, crying and other displays of vulnerability should be expressed and most positively received by proximate affiliates (e.g., family and close friends). Instead, these behaviors should be attenuated and aversively responded to by distal affiliates (e.g., acquaintances). Again, these are the types of cost-benefit fitness trade-offs that would have selected for the facultative adjustment of behaviors such as crying that are effective at soliciting social support from certain types of affiliates, while simultaneously averting interactions with other affiliates. These trade-offs are part of the foundation of my social spheres hypothesis (target article, Fig. 2).

I agree with Lyons that smiling is also an affiliative gesture, and hypothesize that these behaviors should be especially attractive (e.g., in terms of increasing “positive” trait impressions) among more distal affiliates, rather than intimate relationship partners. Unlike crying, which serves as more of a relationship maintenance behavior, smiling and laughing are largely used to solicit potential relationship partners. Still, Fischer describes several forms and functioning of smiling behaviors, including serving as a signal of affiliation (Duchenne smile), appeasement (closed-mouth smile), dominance (pride smile), or experiencing negative self-conscious emotions (e.g., embarrassed smile). Some of these smiles signal capacity, namely the types of smiles that display the teeth; these smiles should covary with dominant emotions (e.g., joy and anger). Other smiles, especially those that conceal the teeth, should instead covary with submissive emotions (e.g., sympathy and shame). Thus, it would appear as though humans use the teeth (e.g., canines) to signal capacity, perhaps through demonstrations of bilateral symmetry and overall healthiness. By concealing the teeth, humans may instead produce heuristical demonstrations of modesty, which may ultimately reduce threat perceptions in others. Again, the SRFB predicts that submissive (i.e., trust) cues are just as powerful at exploiting the reciprocity potential of others as are displays of dominance (i.e., capacity).

The function of smiles differs somewhat by sex. This is supported by the work of Vazire, Naumann, Rentfrow, & Gosling (Vazire et al.), which shows that males and females evidence unique associations between smiling behaviors and felt emotional sensations. In females, smiling covaries with affiliative moods, such as pride, enthusiasm, and inspiration (referred to by the commentators as positive emotions). In males, however, smiling is more strongly and negatively associated with avoidant emotions, such as anger, fear, and shame (referred to as negative emotions). Because males have evolved the proclivity to advertise their capacity (e.g., prowess) and to conceal their vulnerability (e.g., shame, pain, crying, frustration, worry), it makes sense that they should advertise various forms of dominance behaviors (e.g., teeth-baring, threat stare, erect posture, lowering voice-pitch) in coordination with submissive emotional states, relative to females. Of course the opposite pattern – for females to display submissive cues (e.g., concealing teeth, head lowering, raising voice-pitch) when experiencing dominant emotional states – is predicted as well. In any event, Vazire et al.'s research shows how biological sex is an important dimension of affective processing in humans.

Along similar lines, Provine accurately describes the effectiveness and sexual dimorphism of other types of affiliative gestures such as laughing. According to Provine, laughter is an honest signal of reciprocation because it occurs implicitly, often without conscious awareness, and because it is hard to voluntarily produce and give the impression of sincerity. I agree with Provine that laughing is a behavioral mechanism that is used to show appeasement and hence trust cues (e.g., via high-pitched vocal utterances) to others. Indeed, people selectively laugh for (certain) other people and not about the humorous content itself. Li & Balliet provide support for this hypothesis by describing how people initiate humor to indicate affiliative intentions, and humor is in turn associated with, and effective at demonstrating, affiliative dispositions to other people. I agree with Li and Balliet that smiling, laughing, and humor operate by adjusting the display of both capacity and trustworthiness in ways that induce affiliation from others. I also agree that different types of humor may operate to serve specialized functions (e.g., maintenance of existing relationships vs. solicitation of novel relationships), which should covary with the structural properties of the humorous content itself (e.g., self-degradation vs. degradation of others).

Tickling and tear production are also elaborate affiliative gestures that are used to strengthen bonding with proximate (e.g., intimate) affiliates via the behavioral display of vulnerability. In the case of tickling, vulnerability is exaggerated by providing access to sensitive areas of the body (e.g., neck, abdomen) and becoming catatonic during intense laughter. In the case of tear production, vulnerability is exaggerated by occluding visual acuity with a bodily fluid. Here again, we see the natural organization of phenotypic forms, functions, and the reactions of other people. Given the power of these basic behavioral mechanisms for regulating social fitness, it is surprising that they are given much less scientific attention than more “cognitive” social, psychological processes.

R3.4. Social behaviors as exploitive defenses

Buss highlighted an interesting concept: behaviors that protect the self from being exploited by other people. From my perspective, this concept is captured by the entire set of responses that I referred to as avoidant behaviors. Individuals should produce these responses when they perceive a risk of being exploited by other people, either through direct interactions with a dangerous person or via indirect fitness-losses (e.g., reputational consequences and comparisons with higher-status people). I agree with Buss that defensive heuristics are sometimes manifested as hegemonic masculinity (e.g., physical prowess and less risk aversion) such as through exaggerated aggression by males. However, it should also be recognized that defensive mechanisms can also operate through trust cues, such as appeasement and vulnerability displays. Submissive gestures such as self-reported shame, guilt, and subservience (e.g., asking questions) may be effective for protecting oneself by lowering threat interpretations and inviting reciprocal displays of kindness or mercy from others, as described earlier.

R4.1. The hierarchical organization of social psychology

The psychological sciences are currently hindered by the lack of unity on the organizational primacy and supporting roles of human thoughts versus feelings versus behaviors. Do thoughts and feelings ultimately support the adaptive qualities of expressive behaviors, or do behaviors ultimately facilitate the fitness objectives of thoughts? Alternatively, both thoughts and behaviors may be codependent, evolving in parallel and reliant upon support from the other for fitness enhancement. The answers to these hypotheses are imperative for understanding the form and function of human emotionality.

I ascribe to the general view that only behaviors can impact personal fitness. This is because a thought or feeling in and of itself cannot result in self-sustainment (e.g., survival) or self-enhancement (e.g., reproduction) without an associated modification in one's own behavior or in the behavioral reactions of other people (James Reference James1884). It therefore makes sense that, across all animals, including humans, basic learning mechanisms and associated cognitive processes (e.g., attention, perception, sensational awareness, information processing, and rationalization) can affect fitness only by altering actual behaviors. From the basis of this perspective, emotions primarily serve social expressive functions. This position is further supported by studies that show that: (a) blind and perhaps cognitively impaired people are emotionally expressive; (b) children (e.g., infants) are sensitive to mimic and express emotions at earlier ages than they are generally believed to engage in operational learning; (c) normative emotional development unfolds through social interactions; and (d) emotive gestures (e.g., teeth or weapon baring) are more universal in nature than the sensations that we often refer to as “feelings.” As Lozano aptly states, “evolutionarily, it only matters what emotions do, not how they feel.”

Still, many of the commentaries took the contrary approach, instead emphasizing intra-individual cognitive processes (e.g., self-reflection, rationalization, and cultural norm appraisal) over the primacy of social expressive mechanisms. According to Izard et al., these cognitive-based approaches represent the focus of the majority of emotion researchers. The commentators justified this position by citing the fact that affect is sometimes experienced in the absence of an apparent social situation. Vermeulen elaborates on this theme by describing how congenitally blind children express smiles similar to sighted individuals, the implication being that emotions are not always socially relevant and may therefore serve intrapersonal functions (e.g., self-reflection). As I suggested in the target article, this inference is analogous to the reasoning that: because people sometimes talk to themselves, and because deaf children can learn to speak, human language evolved to communicate to the self. In the following section I describe some potential reasons why emotions may consume the human consciousness and feel like important, self-reflective processes.

R4.2. The form and function of emotional experiences

The feelings or experience component of affective responses is usually the first concept that people think of when asked to define an emotion. However, according to the reasoning mentioned above, felt experiences are limited to a supplementary or facultative role in the evolution of emotionality; that is, they are only capable of enhancing fitness vis-à-vis modifications to specific behaviors. In the target article, I provided no justification for the inclusion of emotional experiences and may have implicitly de-emphasized the importance of felt sensations for daily functioning. If, as I proposed, the selective interchange of heuristical expressive cues is effective for regulating individual relationships (i.e., promoting attraction vs. aversion) and hence overall social fitness, then why do humans (need to) feel emotional experiences at all? This question can be further parsed by asking: Why are humans cognizant of emotional experiences; why are emotional experiences valence-based (e.g., felt along pleasant and aversive dimensions); and how does feeling an emotion enhance personal fitness?

The first question is difficult to analyze, but can be viewed along two opposing hypotheses. One hypothesis is that emotional awareness is simply a by-product of a broader adaptation to be consciously aware. Another hypothesis is that emotional awareness is instead specific and functional and hence an evolved adaptation in and of itself. I tend to lean towards the latter hypothesis for several reasons. First, humans are not aware of all bodily sensations (e.g., what it feels like to store iron in the liver), but only certain ones, suggesting a special design for the ability to acknowledge emotional sensations. Second, emotional experiences are not just consciously observed, but are also felt in seemingly important ways. Third, several cognitive psychologists have suggested that many of humans' comparatively unique mental faculties, such as intelligence, consciousness, and voluntary thought processing, were the products of, and ultimately serve, social manipulatory functions (e.g., Dunbar Reference Dunbar1998; Geary Reference Geary2005; Humphrey Reference Humphrey, Bateson and Hinde1976). If these complex cognitive abilities evolved to regulate social relationships, then it is certainly possible that the awareness and experience of felt emotions may be designed for related purposes.

An associated hypothesis is that visceral experiences of pleasantness and aversion may have evolved to calibrate or otherwise differentiate the impact of significant life events in ways that enhance the efficacy of interpersonal interactions. By experiencing varying degrees of felt sensations (e.g., feeling slightly down vs. extremely down) in coordination with different types of life experiences, individuals may be better able to solicit sufficient degrees of responses (e.g., provisioning) from others. Likewise, inter-subjectivity (dual awareness) of the feelings of others (e.g., knowing what is feels like to experience mild vs. severe pain) may enable individuals to better qualify their own responses toward others without overextending personal resources such as time. From this perspective, humans are not just aware of arbitrary cognitive sensations, but rather, that these sensations exist and become accessible for fitness-enhancing purposes, by facilitating the selective interchange of reciprocity potential with other people.

Finally, I propose the thesis that humans may experience the biological affects (e.g., emotions, moods, anxiety, pain) for the sole purpose of showing or talking about them to other people. For example, one hypothesis is that emotional experiences may operate to sustain the behavioral advertisement of the felt emotions; this would be functional for prolonging the ability to solicit beneficial behavioral responses from others. A complimentary hypothesis is that humans may experience emotions (e.g., pride and guilt) in order to better convince others that one's behavioral advertisements are genuine. That is, by feeling emotions (or contextualizing emotions, as Fugate et al. suggest), individuals may be more effective at communicating the sincerity of one's relative state of capacity and trustworthiness attributes to others. From this perspective, it therefore makes sense that emotions feel important; they may be designed to do just that. By convincing oneself of the relevance of an emotional representation, humans may be better able to demonstrate to others that one's abilities and intentions are sincere.

Empirically, it is very difficult to separate cognitive processes (e.g., rationalization) that may be involved in emotion processes from the expressive properties of a self-report. Taken further, this confound opens up the possibility that many forms of self-reported information, such as self-descriptions (e.g., self-esteem) and social opinions (e.g., political ideologies), could largely operate to convey specific social impressions (e.g., demonstrations of dominance or submissiveness) to others (e.g., Vigil, submitted c). It is therefore likely that many types of self-reported information may be more closely associated with behaviorism rather than outcomes of cognitive reasoning processes. At the very least, the fact that the content of some self-reported information cannot be easily separated from the social impressions that the information communicates should give researchers caution to consider the possibility that they may be measuring behavioral expressions in addition to, or rather than products of complex computations. For these purposes, my definition of a social expression – a behavior that is both observable and moderated by the social context – should be especially useful for distinguishing communicative versus non-communicative mental processes.