Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Integrating Simulation and Multifaceted Evaluation in Affective Language Processing: A Multiple-Drivers Account of Corrugator Activity
Poster Session F, Friday, October 2, 2:45 - 4:45 pm, Wangari Maathai
Marijn Struiksma1, Li Kloostra1; 1Utrecht University
Across several studies we found that corrugator supercilii activity during affective language comprehension reflects multiple interacting affective processes (’t Hart et. al, 2018; 2019; 2021; in press). Emotion-describing language in short narratives reliably elicits frowning responses that are strongly modulated by moral context. When characters are described as morally good, corrugator activity tracks affective valence of linguistic stimuli, and negative affective states (e.g., angry, frustrated) evoke greater corrugator activity than positive states (e.g., happy). This result could be ascribed to simulation-based situation model construction. Additionally, the corrugator may indicate evaluation of the affective event in context. In contrast, when characters are described as morally bad, valence-based differentiation is attenuated, absent, or reversed, indicating that simulation-based affective responses are counteracted by evaluation. Temporal manipulation, of word-by-word presentation, shows that affective simulation of narrative events (e.g. ‘Mark is frustrated’) operates primarily at the level of situation model construction, rather than during lexical access. Critically, corrugator responses to later affective events do not simply track event valence, but instead reflect an integration of simulated emotion of the character and moral evaluation. Attempts to strengthen evaluation via an explicit rating task produced unexpected effects, suggesting that evaluative processes are not unitary, but instead consist of multiple facets that can be selectively engaged, disrupted, or reweighted by task demands. Together, these findings consistently support a multiple-drivers framework in which corrugator activity reflects the integrated outcome of at least three interacting components. First, an affective simulation driver that supports automatic, language-driven simulation of described emotional states during situation model construction. Second, an immediate affective evaluation driver that rapidly appraises the situation using empathy, potentially opposing the valence of the simulated state. Third, a reflective evaluation driver that engages slower, cognitively mediated evaluations using the context, own (moral) values and concerns, and experiences, that can amplify, suppress, or reorganize earlier affective responses. Such reflective evaluation can (also) be task- or trait-mediated (e.g. fairness judgments or justice trait). Corrugator activity thus reflects the net result of these interacting drivers, rather than any single driver. This architecture has direct implications for embodied cognition accounts. Facial electromyography responses should not be interpreted as a straightforward readout of affective simulation. Instead, embodiment must be understood as contextually embedded and evaluatively constrained, with simulation operating within a dynamically structured evaluation system. The data therefore call for a revision of strong simulation-based interpretations of peripheral facial muscle activity. We accordingly propose a multiple-drivers model of affective processing, in which emotional responses to language emerge from the dynamic integration of simulated affect, moral and normative evaluation, and higher-order, task-dependent evaluation. Such a model better captures the complexity of affective language comprehension in narratives and re-frames peripheral facial muscle activity as indices of composite affective states rather than simulation dominated.
Topic Areas: Meaning: Discourse and Pragmatics,