Age and sex effects on sound change: One size does not fit all

Abstract

This study examines whether speaker age and sex predict production of Advanced Tongue Root (ATR) vowel harmony by urban Twi (Niger-Congo, Kwa) speakers in Ghana. Traditional descriptions of the ATR harmony system state that [-ATR] /a/ is phonetically realized as [ + ATR] [æ] before [ + ATR] vowels /i, u, o/ and palatal segments. However, recent research indicates that this description is no longer accurate for all Twi speakers: in contrast to suburban speakers, most urban speakers raise and front [æ] to [e], impinging on phonemic /e/, before /i, u/ and palatal segments (Kpogo, 2021). This pattern suggests a possible sound change (i.e., vowel merger) if suburban speakers are taken to represent a conservative variety of Twi. To explore the source of this innovation among urban speakers, variation in production across speaker age and sex was investigated. Preliminary acoustic data from a picture-naming task suggests that both factors are predictive of the innovative production pattern. Adolescents and younger adults, but not older adults, raise and front [æ] to [e], implicating the younger generation as leading this linguistic innovation. Additional findings suggest that men, but not women, are leading the innovation, a finding inconsistent with the overwhelming generalization that female speakers lead linguistic innovations.