Acoustic representations of different dimensions of the gendered self

When representing relations of gender and speech, there is a long tradition of conceiving gender as a binary construct (women vs. men). Although there is a growing interest in acoustic gender markers beyond the mere difference between female and male speech, the complex interrelations between acoustic markers and the different dimensions of the gendered self (e.g., gender identity, gender socialization) are not yet fully disentangled. Within-group variation due to gender-role self-concept further complicates the picture. We therefore acoustically analyzed semi-spontaneous speech of cis female ( n = 17), cis male ( n = 17), and nonbinary speakers ( n = 18) and collected data of multiple gender dimensions (gender identity, gender assigned at birth, gender-role self-concept, gender socialization). Our exploratory factor analysis revealed that for these groups, the measures of the acoustic analysis can be subsumed into three factors, which relate to source and filter information (SourceFilter), to high spectral energy distribution (HighSpectrum), and to voice quality parameters (VoiceQuality). For the factor SourceFilter, nonbinary speakers showed broader distributions than both cis groups, for the factor HighSpectrum a more flexible use and a departure from cis -normative speech patterns. Every gender dimension is acoustically represented in the voices and speech patterns of this sample but not all dimensions in the same way. Also, different dimensions of the gendered self are highly correlated, which emphasizes the difficulty of exact assignments of individual acoustic markers to individual gender dimensions. However, by combining several acoustic markers and different dimensions of the gendered self, the present study extends previous knowledge on the complex interrelations of voice and gender. This interaction is of broad relevance for sociophonetic research and serves as a foundation for future studies, e.g., on voice gender perception and processing of acoustic representations of gender. The results also provide further evidence for theories of more nuanced speaker voice representations beyond the gender binary.