Abstract
Building on the importance of sociality in grounding abstract concepts, this study investigated how concept type shapes recognition memory for social stimuli. Participants engaged in text-based dialogues about abstract or concrete concepts with virtual agents, after which recognition of names and faces, as well as perceived closeness, was measured. The procedure included a preliminary, context-based validation of linguistic dimensions adapted from Italian into Romanian, given the absence of prior norming studies. Discussions of abstract concepts elicited more uncertainty expressions, questions, turn-taking, general statements, affective content, and overt beliefs, thus offering converging evidence for the adequacy of the translated stimuli and previously reported differences in socio-cognitive engagement. Differential recognition effects facilitated by abstract concept conversations were observed for faces, but not for names. The higher salience of faces could not be ascribed straightforwardly to cognitive load or faster processing, as this interpretation was incompatible with the reaction times observed, which did not differ across conditions in either the encoding or the recognition phase. This, coupled with the absence of corresponding effects for names, which are typically recognized at least as well as faces in matched recognition paradigms, is consistent with a greater involvement of social cognitive processes during abstract concept processing. Overall, the results indicate that abstract concept processing modulates recognition memory for social stimuli. The experimental design is intended to be simple and reproducible, enabling cumulative future research on the relationship between concept types and other cognitive processes.