This study investigates the perception-production dynamics of [æ] and [e] in an ongoing merger in Asante Twi’s Advanced Tongue Root (ATR) harmony system. Traditional descriptions state that [-ATR] /a/ is realized as [+ATR] [æ] before [+ATR] vowels /i, u, o/. Yet, recent acoustic evidence points to an ongoing sound change in different Twi-speaking communities: in contrast to traditional (rural/suburban) Twi speakers, urban speakers raise and merge [æ] with phonemic /e/ before /i, u/, a change that is more advanced in men and younger speakers (Kpogo, 2023). The present study aims to ascertain whether merger in production correlates with perceptual merger among the same participants in Kpogo (2023). Data from a forced-choice identification task suggests that the presence of a perceptual merger is associated with the participant’s locality. Traditional participants consistently identified [æ] and [e] correctly, suggesting separate mental representations for these vowels. Conversely, urban participants often perceived [æ] as [e], suggesting less distinct representations influenced by their own production. Concerning the perception-production relationship, evidence of a link was observed across localities but, crucially, not within each locality. This study contributes to theoretical discussions on the perception-production relationship during sound change, hinting at a co-evolving link between the two modalities.