Researchers have argued that sound change may obtain when deviant percepts due to listeners' under-normalization for variation in speech become seeds for new perceptual and production norms (Ohala, 1993; Blevins, 2004). How deviant percepts accumulate in a systematic fashion to give rise to stable variation remains a vexing question. This study explores how variation in socio-cognitive processing may contribute to the emergence and propagation of sound change, showing that variations in several individual-difference cognitive dimensions, such as abilities to empathize, general "autistic traits'', and personality traits, significantly influence the way contextual information is integrated in speech perception. These findings suggest that the seeds for sound change may be distributed at the population level in such a non-random way that may ultimately facilitate the emergence of socially-differentiated sound change.