Volume 55, Issue 5 - Special issue: The Sociophonetics of /s/, Guest Editors: Erez Levon, Marie Maegaard, and Nicolai Pharao

This paper provides an introduction to the papers in this special issue on the sociophonetics of /s/. We begin by reviewing some of the principal findings on variation in the production and perception of /s/, summarizing studies in sociolinguistics, experimental phonetics, and laboratory phonology. We go on to identify similarities in the meanings associated with /s/ variation cross-linguistically, and briefly describe how theories of sound symbolism may help us to account for these patterns. We conclude this introductory article with a summary of the contributions to the special issue and a discussion of how together these articles help us to better understand that origin and trajectory of socially meaningful sociophonetic variation.

Sibilant consonants are well-established as resources for the negotiation of gender and sexuality, but the origin of these links is less clearly agreed upon. Some researchers have pointed to sex differentiation in the vocal anatomy as a potential cause for gender differences in /s/, though a review of the literature indicates that learned articulatory patterns play a critical role. This article focuses on the spectral qualities of /s/ among 15 English-speaking transgender men and transmasculine individuals. Because their early socialization and physiological development is not normatively aligned with their self-defined gender identities, trans people are well-positioned to illuminate the relative contribution of physiology and identity to the gendered voice. Two analyses are presented, one of which focuses on inter-speaker variation among all 15 participants, and the other of which compares one bilingual speaker’s productions of /s/ in English and Spanish. Together, these analyses demonstrate that sex category does not determine the gender-linked acoustic characteristics of /s/. Instead, a more complex, multidimensional framework for gender that distinguishes between gender assignment, role, identity, and presentation is necessary to account for the full range of gendered phonetic styles that speakers can employ and hence to understand the process through which gendered voices arise.

This article reports on the development of gender-specific speech patterns in Mandarin children’s production of voiceless sibilant fricatives. Ninety-four children aged from 2 to 5 (gender-balanced) participated in a word repetition task, producing a series of words beginning with fricatives. Their speech was digitally recorded and acoustically analyzed. The results indicate that gender-related differentiation in /s/ and /ɕ/ occurs around age 4 and becomes more robust at age 5. The differentiation occurs due to the more anterior articulation location that girls employ in their speech in comparison with boys. In addition, 10 adults (gender-balanced) were recorded to probe the origin of this gendered speech phenomenon. Unlike children, adults only show robust gender-related differences in /ɕ/. Results comparing the speech of adults and children suggest that gendered variation in /s/ is potentially the result of a chain shift initiated by gender-linked variation of /ɕ/. These results are discussed with respect to their ramifications for our understanding of the organization and acquisition of multiple components of children’s phonological capacity.

This paper proposes an empirical method for the quantitative analysis of stance-taking in interaction. Building on recent conceptualizations of stance as the primary building-block of variation in language style, we describe how to implement an analysis of stance within a variationist framework via an examination of the particular speech activities within which stances are embedded combined with a consideration of the specific interactional goals these activities achieve. We illustrate our proposals with an investigation of variation in /s/-quality in the speech of cast members from two British reality television programs. Examining nearly 2000 tokens of /s/ in over 6 hours of recorded speech, we demonstrate how different acoustic realizations of /s/ in the sample correlate with the level of “threat” of a given speech activity, and we argue that this interactionally based analysis provides a better explanatory account of the patterns in our data than an analysis based on large social categories would. Through this paper, we therefore hope to contribute not only to the development of a more robust method for examining stance in quantitative sociolinguistic research, but also to help clarify the relationship between stances, speech activities and speaker identities more broadly.

Previous research has shown that listeners’ identification of English anterior sibilant fricatives changes depending on whether they are primed to believe that the talker is a woman or a man. This article explored how stable this effect is across two types of priming. Listeners identified a nine-step shack-sack continuum created by combining an /s/-/ʃ/continuum with a natural production of a VC that had been acoustically manipulated to be gender-neutral. In the explicit-priming condition, the talker’s sex was primed by providing a picture of a man or a woman. In the implicit-priming condition, the talker’s sex was primed by having listeners judge the grammaticality of acoustically gender-neutral sentences that implied the talker was either a woman or a man. The effect of sex priming was strongest in the explicit condition, and it was in the expected direction: more sack judgments were elicited for trials where a man’s face was used as a prime. The influence of sex priming on fricative identification was weaker in the implicit condition, but in the expected direction. Together, these data show that gender normalization effects in fricative perception occur most strongly when the talker’s gender is suggested very explicitly by showing a picture of the talker. The small size of the effect questions an interpretation of Strand and Johnson’s effect as evidence that social variables influence speech perception pervasively.

This paper examines the indexical value of /s/-fronting in White Afrikaans and in White South African English (WSAfE). Prior research on this feature has shown that fronted articulations of /s/ in WSAfE serve as a regional and social indicator of the wealthy northern suburbs of Johannesburg, and anecdotal evidence suggests that the feature carries a similar meaning in White Afrikaans. This study therefore aims to examine whether the variable carries similar meanings across the two languages. Data are based on the evaluative reactions toward different experimental stimuli that were presented to 214 Afrikaans-English bilinguals in South Africa during a modified matched-guise task. The results indicate that /s/-fronting in a man’s voice is perceived in similar terms in White Afrikaans and WSAfE though it carries somewhat different meanings across the two languages when it occurs in a woman’s voice, a difference related in turn to different approaches to gender across the two speech communities. The results of this research, and the indexical value of /s/-fronting in the two languages, are therefore only understandable in terms of certain sociohistorical and sociological differences between the two speech communities. The article ends with some discussion relating to the possible source of the relevant similarities and differences, i.e., parallel innovation or sociophonetic transfer.

Previous studies have shown that the same sociolinguistic feature can be used as a sociolinguistic resource with different indexical potentials in different linguistic as well as social contexts. In this paper we present the results of a perceptual study of indexical meanings of fronted and palatalized variants of /t/ in combination with fronted /s/ in different registers of Copenhagen Danish. The data consist of responses to male speakers’ use of the two variants of /t/ in two different registers that we label “modern Copenhagen speech” and “street language.” Results show that the palatalized /t/ affects the indexicality of fronted /s/ with respect to perceived femininity and gayness when it occurs in “modern Copenhagen,” where fronted /s/ has previously been shown to index these traits. However, the variation has a different and less significant effect when occurring in “street language.” Furthermore, the effect is only found in “modern Copenhagen” if the speaker has not previously been heard to produce a fronted /s/. We discuss the implications for theories of the relation between features and clusters of features in listeners’ perceptions.

Earlier research has concluded that there is a strong symbolic relationship between Helsinki as a place and non-standard /s/ pronunciation. This phenomenon is likewise in continuous evidence in the Finnish media and social media. The notion of “Helsinki s” has become a folk linguistic fact although it lacks a clear linguistic correlate or even status as a linguistic fact. The only sibilant of the Finnish language is officially a voiceless alveolar, while the “Helsinki s” is most often discussed as “hissing”, “sharp” or “fronted”. However, according to recent research based on listening tasks, any /s/ may be designated and discussed as a “Helsinki s” if the speaker is regarded as a Helsinki speaker according to other criteria. The present article seeks to trace a development path to illustrate the layered semiotic processes that has led Finnish society to construct “Helsinki s”. In exploring the indexicalization process of this folk notion, the article focuses in particular on the origins and the circumstances under which /s/ pronunciation started to gain social recognition in Finnish society. The ideological link between the socially meaningful /s/ pronunciation and Helsinki is traced back to the early industrialization and urbanization era and (media) discourses of Helsinki in the late 19th century. The present-day associations and the 150-year path of “Helsinki s” are discussed.