Disembodied voices, heard without a human body present, have become a familiar presence in everyday life through digital voice assistants (DVAs) such as Amazon’s Alexa and Apple’s Siri. These voices are not neutral technological tools but fetishized artifacts that appear natural and emotionally responsive while concealing the social and ideological labor embedded in their design. This paper offers a critical sociophonetic analysis of two German-language DVAs to investigate how their vocal profiles index distinct forms of femininity. Drawing on theories of gender as performative and the concept of fetishization (Butler 1990; Marx et al., 1922: 35–47), I argue that DVA voice design operates as a site of ideological mediation, where social beliefs about gender and gendered voices are transformed into sonic form and naturalized through interaction. Acoustic analyses reveal that Alexa’s breathy voice indexes a traditional domestic femininity, whereas Siri’s lower pitch and creaky voice align with a professional, authoritative femininity, mirroring their respective roles as a home assistant and a mobile assistant. These vocal features are not incidental but reflect language ideologies about what voices should sound like, beliefs that link vocal traits to social personae and gender norms. Embedded in everyday routines, such voices become domesticated (Berker et al. 2006), rendering the ideological work of design inaudible. This study illustrates how gendered language ideologies are materially encoded in voice technologies and invites a rethinking of the synthetic voice itself, as an unstable intersection of embodiment, mediation, and market ideology.
