Research on the pragmatic device like has primarily focused on like as a quotative. The overall distribution of like has received comparatively little attention. Particularly lacking are accounts which show how like may vary for a speaker in different situations or contexts.
This study addresses this gap by testing the effects of speaker gender, addressee gender and familiarity on the frequency of like. This study is based on 24 20-minute dyadic conversations, 12 of which occurred between friends, and 12 between strangers. The dyads were split evenly between female-female, female-male and male-male pairings. The recordings were transcribed and coded according to a system based on Terraschke (2008). The data was then analysed using mixed effects logistic regression.
The results for discourse marker like showed a significant interaction between familiarity and gender relation. When participants were friends with their addressee, they were more likely to use “quotative”, “subjective stance” and “hesitation” like if their addressee was of the same gender. In the unfamiliar data, both males and females were more likely to use quotative like and subjective stance like when speaking to a female addressee. “Discourse link” like was more likely to occur in the male-male dyads. “Cut-off” like was more likely to occur between friends, and females were more likely to use it when speaking to males.
Audience design (Bell, 1984), social identity theory (see Meyerhoff, 1996) and observations on the communicative preferences of men and women (Coates, 2004) are used to account for these results. It is argued that because quotative like and discourse link like are used to structure discourse, or have a textual function, they are better explained by the communicative preferences of men and women. Subjective stance like and hesitation like are better accounted for using audience design and social identity theory because of their interactional function.
