The Catholic University of America
Language is the repository of culture, identity and behavior labels define social situations, and the identities that we enact in them—personally performed but culturally contextualized—motivate our behaviors. Throughout these social interactions, felt emotion serves as a barometer concerning the success or failure of our attempted identity affirmations. In other words, who we are affects how we feel; when we interact with other people, we use how we feel to tell us who we seem to be (in the moment) and whether we are being affirmed in who we are (fundamental nature). I was interested in investigating how we practically navigate through—and determining whether I could predict and model mechanisms within—the intricacies of these complex relationships.
In my dissertation, I utilized the tenets and quantitative rating metrics of affect control theory and affect control theory of self (formal computational theories of culture) to generate and test novel theoretical predictions concerning felt-emotion feedback. In a primary longitudinal dataset of respondents in an especially self-referential and other-contingent interaction space (selves in romantic relationship dyads), I gathered two significant EPA profile (quantitative ratings along three affective dimensions: evaluation, potency, and activity) measures of identity: persona (affective ratings of “myself as I really am”) and relational self (affective ratings of “myself in this relationship”), and two significant EPA profile measures of emotion: ideal emotion (“the emotion I should feel in this relationship”) and actual emotion (“the emotion I actually feel in this relationship”). I also computed characteristic emotions (theoretically predicted based on affirming Actor identity) and structural emotions (theoretically predicted based on attempting to affirm Actor and Object-Person identities) profiles for both personas and relational selves.
My two-wave survey sample of young adults (ages 18-22 years) in romantic relationships (N=93, 71% valid-case response rate) included relationship types of two distinct cultural frameworks: stable (defined, N=50) and mutable (undefined/ambiguous, a subset of relationships within hookup culture, N=43) romantic relationship dyads. Cultural framework (defined vs. undefined) served as a main predictor variable along with two hypothesis-derived computed predictor variables: identity discrepancy (squared Euclidian distance between persona and relational-self sentiment profiles), and emotion discrepancy (squared Euclidian distance between ideal emotion and actual emotion sentiment profiles). I posited that relationship dissolution might be partially explained by the emotional feedback information concerning respondents’ attempted identity affirmations: when people get signals from their experienced emotions that are counter to those that would affirm the identities they wish to enact, they will act to rectify the situation by exercising their agency (i.e., redefine/discontinue the relationship generating the discordant signals). I posited that dissolution would be more likely in undefined relationships (as found in the literature) and that this would be partially predicted by identity-fueled emotion discrepancy.
Results from a series of two-sample t-tests, MANOVAs, regressions, and structural equation models statistically distinguished a new set of cultural identity labels (cultural meanings of ambiguous relationship labels (e.g., “talking to,” “having a thing”) are not interchangeable), validated several theoretical predictions (e.g., computed emotions significantly predicted reported expected and experienced emotions), and demonstrated that emotion information in the present significantly predicts relational behavior in the future (mediation through identity discrepancy emotion discrepancy accounts for a 17% reduction in the direct effect of relationship type on W2 relationship dissolution likelihood).
Additionally, personas better predict emotion expectations than do relational selves, and a robust standard errors OLS regression suggests a nearly 1-to-1 relationship (.95 (.32), p =.004, R2 =.47) in emotion discrepancy change as predicted by identity discrepancy magnitude. This implies that individuals are not self-selecting into undefined relationships because of negative self-views (on the contrary, average persona ratings were highly positive and statistically indistinguishable (p =.3617) by relationship type). Rather, seeing oneself as evaluatively less in an undefined relationship (p <.0001) is a product of holding that subculturally contextualized relational identity (defined participant personas and relational selves are statistically indistinguishable). Respondents attempt to affirm personas while being relational selves; when affirming relational self fails to affirm persona, disconfirming feedback via emotion discrepancy is the net result.
In both cultural frameworks, predicted values for respondents’ structural emotions were higher than predicted values for characteristic emotions. This suggests an inherent relational optimism: as social creatures who crave connection with others, we expect things to be better together than when alone. This emotional optimism is validated in the experiences of defined, but not undefined, relationship participants: defined participants experience the emotional boost they expect, but undefined participants find their emotional hopes—tempered by cultural insight to be statistically significantly lower already—unrealized in their experiences. Thus, enacting the relational self in undefined relationships causes a gulf between the emotions one expects to feel (informed by persona) and the emotions one does feel (informed by relational self). Essentially, respondents ask themselves, “Am I myself when I’m with you?” When the emotion-supplied answer is “no” (more likely for undefined participants) agentic individuals act.
In different definitions of a situation, we enact different identities. As we do, we feel different emotions, and expect different behaviors from ourselves and others, depending upon which facet of ourselves we are being at any given time. We and our identities are also embedded in specific cultures and subcultures; these too guide expectations, feelings, and behaviors. Relationships, and the relational identities associated with them, add another layer of complexity—how one feels as the same identity will vary depending upon with whom that identity is interacting. However, harnessing quantitative measures of emotion allows one to model these processes well. This research showcases the negative effects of an emotion-signaled disruption in the sense of synchrony for identities in relationships, provides an explanatory pathway for likelihood of dyadic relationship dissolution, and empirically demonstrates that emotion information in the present can significantly predict relationship behavior in the future. Because the theoretical underpinnings of these tests are formalized (not necessarily confined to the romance realm in which they were tested), results may generalize across social institutions. If so, measuring emotion discrepancy in the present could provide a practical diagnostic metric signal of those at increased mental health risk.
