How can we judge if conversation partners will be good cooperation partners in other tasks? A recent proposal is that low-level linguistic similarity such as subconsciously matching others' language use may be a signal of cooperativeness. To elucidate the mechanisms behind this relationship, we conducted two experiments, in which we found that conversation partners that matched the participants' syntax and pitch were perceived as more cooperative and were chosen more often as cooperation partners. Our findings further suggest that the sheer act of adapting to someone's linguistic production was not as crucial for choosing cooperation partners, even if it involved an initial cognitive investment. Rather, the decisive factor was sharing someone's linguistic preferences and thereby indicating group membership. This may have important implications for understanding cooperation partner choice and for understanding the (co-)evolution of language and cooperation, which both are traits that are particularly prominent in humans.

Words that are learned early were shown to be semantically more stable, and vice versa (Cassani et al. 2021, Cognitive Science). Semantic change, however, has multiple aspects. In this diachronic corpus study, we examine the relationship between the age of acquisition (AoA) of words and a set of different measures of semantic change: change in a word's polysemy; overall semantic displacement; and average extent of semantic fluctuation. All measures are based on diachronically layered sense distributions (Hu et al. 2019, ACL) derived from the Corpus of Historical American English. AoA is taken from Kuperman et al. (2012, Behav. Res. Meth.). Taking interactions with frequency into account, we show that semantic displacement and fluctuation are positively associated with AoA as expected. Early acquisition is associated with an increase in polysemy. This hints at the relevance of semantic (metaphorical) extension in the early acquisition of the lexicon.