Attitudes and emotions are among the longest studied constructs in psychology, but there is still a lot of disagreement about the meaning of these concepts, their objective existence, and their function (e.g., Bohner & Dickel Reference Bohner and Dickel2011; Ekman Reference Ekman2016; Volz & Hertwig Reference Volz and Hertwig2016). Gervais & Fessler's (G&F's) target article provides a thoughtful review of these constructs and clarifies them by examining their evolutionary significance for human sociality, recognizing the intricate links and feedbacks between attitudes and emotions, and distinguishing the folk affect concepts from basic affect systems.
The sentiment construct in the Attitude–Scenario–Emotion (ASE) model might not be as illuminating as the other ideas in the target article. It seems to be yet another verbally described psychological construct that might be difficult to empirically distinguish from related experiences (other recent conceptualizations of sentiments are similarly vague; e.g., see Pang & Lee Reference Pang and Lee2008; Stets Reference Stets and Delamater2003). In the ASE model, this construct was formulated specifically to describe the concept of contempt, which at times may be more similar to different emotions and at other times to attitudes. The number of different sentiments and their expressions might be very large, and their characteristics might overlap with each other and with their constituent parts. It therefore seems difficult to measure empirically and either prove or disprove that a person is experiencing a particular hypothesized sentiment. The potential of the sentiment construct to usefully describe and, in particular, predict human judgments and experiences could be limited.
Instead of adding a new (or revived) vague construct to organize other unclear constructs, more clarity and predictive power might be achieved by attempting to build a simple, parsimonious, quantitative model that would more precisely describe and predict what a person might experience in a particular social situation. A potentially good starting point might be the constructionist models of attitudes sensu Tourangeau and Rasinski (Reference Tourangeau and Rasinski1988) and Schwarz and Bless (Reference Schwarz, Bless, Martin and Tesser1992) (note that this tradition is largely disconnected from the constructionist models of emotions discussed in the target article). For instance, Tourangeau et al. (Reference Tourangeau, Rips and Rasinski2000) proposed a belief-sampling model of attitudes in which a reported attitude is formed by sampling potentially relevant considerations from memory, which may include feelings, evaluations, images, or any other material that seems relevant in the current situational context. These models predict that the resulting attitudes will be affected by the context in which they are formed; but at the same time they can predict intra-individual consistency across time because of similar samples (Schwarz Reference Schwarz2007). The models can also produce quite precise predictions of differences in attitudes constructed in different contexts.
A similar, but broader approach might be used to quantitatively model not only attitudes, but also syndromes of attitudes and emotions, such as contempt and other sentiments in the ASE model. People might sample different emotions, attitudes, and other relevant considerations from memory and from the current context to form an appraisal of a social situation. Different considerations might receive different sampling weights, determined in part by the evolutionarily developed sensitivity for an other's relational value as a potential cooperation partner, a mate, or a source of useful information, and in part by individual circumstances and affordances of the particular situation. The content of the samples and the sampling weights can, at least in principle, be explicitly or implicitly measured, or inferred from evolutionary analysis and anthropological findings. The samples might be quantitatively described as, for instance, an individual's frequency distribution over different emotions (e.g., anger, disgust, mirth) and evaluative judgments (e.g., devaluation and diminution; examples taken from G&F's Figure 1 in the target article). These distributions might be thought of as quantitative representations of the sentiments in the ASE model, and might be used as building blocks of simple models (e.g., akin to Galesic et al. Reference Galesic, Olsson and Rieskamp2012) to predict the resulting motives, behaviors, and physical expressions.
To the extent that the same considerations receive similar weights in different cultures (e.g., because of their evolutionary significance), experiences in similar social contexts will translate across diverse populations. At the same time, to the extent that the social contexts differ between populations (e.g., in how easy it is to dissolve an unsatisfying relationship, or openly express hostility), the samples and the resulting experiences will be different. In other words, we might not need to assume an existence of specialized, evolutionarily developed functional networks of attitudes and emotions in order to explain similarities and differences across populations. Instead, sentiments such as contempt and respect might be conceived and perhaps more precisely modeled as experiences constructed from more basic emotions and relatively simple evaluative judgments sampled in a particular social context.
Attitudes and emotions are among the longest studied constructs in psychology, but there is still a lot of disagreement about the meaning of these concepts, their objective existence, and their function (e.g., Bohner & Dickel Reference Bohner and Dickel2011; Ekman Reference Ekman2016; Volz & Hertwig Reference Volz and Hertwig2016). Gervais & Fessler's (G&F's) target article provides a thoughtful review of these constructs and clarifies them by examining their evolutionary significance for human sociality, recognizing the intricate links and feedbacks between attitudes and emotions, and distinguishing the folk affect concepts from basic affect systems.
The sentiment construct in the Attitude–Scenario–Emotion (ASE) model might not be as illuminating as the other ideas in the target article. It seems to be yet another verbally described psychological construct that might be difficult to empirically distinguish from related experiences (other recent conceptualizations of sentiments are similarly vague; e.g., see Pang & Lee Reference Pang and Lee2008; Stets Reference Stets and Delamater2003). In the ASE model, this construct was formulated specifically to describe the concept of contempt, which at times may be more similar to different emotions and at other times to attitudes. The number of different sentiments and their expressions might be very large, and their characteristics might overlap with each other and with their constituent parts. It therefore seems difficult to measure empirically and either prove or disprove that a person is experiencing a particular hypothesized sentiment. The potential of the sentiment construct to usefully describe and, in particular, predict human judgments and experiences could be limited.
Instead of adding a new (or revived) vague construct to organize other unclear constructs, more clarity and predictive power might be achieved by attempting to build a simple, parsimonious, quantitative model that would more precisely describe and predict what a person might experience in a particular social situation. A potentially good starting point might be the constructionist models of attitudes sensu Tourangeau and Rasinski (Reference Tourangeau and Rasinski1988) and Schwarz and Bless (Reference Schwarz, Bless, Martin and Tesser1992) (note that this tradition is largely disconnected from the constructionist models of emotions discussed in the target article). For instance, Tourangeau et al. (Reference Tourangeau, Rips and Rasinski2000) proposed a belief-sampling model of attitudes in which a reported attitude is formed by sampling potentially relevant considerations from memory, which may include feelings, evaluations, images, or any other material that seems relevant in the current situational context. These models predict that the resulting attitudes will be affected by the context in which they are formed; but at the same time they can predict intra-individual consistency across time because of similar samples (Schwarz Reference Schwarz2007). The models can also produce quite precise predictions of differences in attitudes constructed in different contexts.
A similar, but broader approach might be used to quantitatively model not only attitudes, but also syndromes of attitudes and emotions, such as contempt and other sentiments in the ASE model. People might sample different emotions, attitudes, and other relevant considerations from memory and from the current context to form an appraisal of a social situation. Different considerations might receive different sampling weights, determined in part by the evolutionarily developed sensitivity for an other's relational value as a potential cooperation partner, a mate, or a source of useful information, and in part by individual circumstances and affordances of the particular situation. The content of the samples and the sampling weights can, at least in principle, be explicitly or implicitly measured, or inferred from evolutionary analysis and anthropological findings. The samples might be quantitatively described as, for instance, an individual's frequency distribution over different emotions (e.g., anger, disgust, mirth) and evaluative judgments (e.g., devaluation and diminution; examples taken from G&F's Figure 1 in the target article). These distributions might be thought of as quantitative representations of the sentiments in the ASE model, and might be used as building blocks of simple models (e.g., akin to Galesic et al. Reference Galesic, Olsson and Rieskamp2012) to predict the resulting motives, behaviors, and physical expressions.
To the extent that the same considerations receive similar weights in different cultures (e.g., because of their evolutionary significance), experiences in similar social contexts will translate across diverse populations. At the same time, to the extent that the social contexts differ between populations (e.g., in how easy it is to dissolve an unsatisfying relationship, or openly express hostility), the samples and the resulting experiences will be different. In other words, we might not need to assume an existence of specialized, evolutionarily developed functional networks of attitudes and emotions in order to explain similarities and differences across populations. Instead, sentiments such as contempt and respect might be conceived and perhaps more precisely modeled as experiences constructed from more basic emotions and relatively simple evaluative judgments sampled in a particular social context.