A wealth of research has identified several social categories (e.g., sex, socioeconomic status, age) that reliably characterize the leaders of sound change, but individuals within a given social category are far from uniform in their use of linguistic variables. Which other factors might be relevant for predicting who is most likely to acquire new variants when language variation leads to language change? Usage-based approaches assert that an individual’s patterns of exposure and language use are integral to their behavior during language processing and, moreover, that the mechanisms utilized when processing language stem from domain-general cognitive functions. Indeed, psycholinguistic research over the last two decades has established links between cognitive control or executive function and language processing in the laboratory. In this talk, I will argue that cognitive processing strategies, when interpreted under a dual mechanisms framework, can also be used to predict whether an individual will adopt sound changes underway in their community. I will then demonstrate how the hypotheses that arise from this approach are supported in naturalistic discourse, focusing on two sound changes in-progress with varying degrees of social awareness in the speech of Puerto Rican Philadelphians. Finally, I will discuss a model of sound change intermixing social categories and individual differences in cognitive processing to arrive at expected patterns of use.