This paper discusses symbolic violence in sociolinguistic research on multilingualism. It revisits an archived recording of a group discussion between four boys about their chances of having sex with a female researcher. The data is rife with symbolic violence. Most obviously, the conversation enacted a heterosexist form of symbolic violence. This was, however, not the only direction in which violence was exerted. As argued by (Bourdieu & Wacquant. 1992. "An invitation to reflexive sociology." Cambridge: Polity), symbolic violence involves two fundamental elements -- domination and complicity. In the case at hand, the boys' sexist banter conformed to dominant expectations about their linguistic behavior, imbued in the research event. This is symbolic complicity of the kind that the Bourdieusian notion foresees. Yet another subordination to the dominant vision occurred when the researchers captured the conversation on tape, but decided to exempt it from publication. Here, we argue that giving deepened attention to sociolinguists' own run-ins with symbolic violence during research is valuable, because it provides an opportunity to reflexively consider the social conditions of the research practices, in relation to the data produced and analyzed. Ultimately, this reflexive exercise may help sociolinguists sharpen their tools for understanding the give and take of dominance and complicity unfolding in their data.