Though the goal of variationist sociolinguistics is explaining variation in language, the enterprise has traditionally levelled variation within groups and speakers in order to explain the variation between them. By doing so, we fail to account for the bulk of the linguistic variation occurring within any single interaction, much of which is affective in nature. And yet, affect has remained under-theorised in variationist sociolinguistics, though there are some recent proposals that affect matters for variation (e.g., Eckert 2010; Podesva 2016), which have generally characterized affect in terms of valence (positive vs. negative) or arousal (high vs. low).
My current work examines affect in a way that provides insights into a thorough different type of social factor: 'visceral' feeling. This is a qualitatively different dimension of social meaning, and one which is important for our theoretical models of language variation because it allows us to get at how people are feeling in interactions, as well as how these affective meanings tie into the wider social landscape. To get at visceral affect I draw on various traditions in the affective sciences (Barrett and Russell 2015; Hazlett 2006; Hochschild 1979) and introduce novel methods to the field of linguistics (see Mendoza-Denton et al. 2017 for similar use). By having pairs of participants chat while playing a video game and being audio/video/psycho-physically recorded, I can examine how vowel space realizations change relative to shifts in speakers' visceral affective experiences as well as their interactional negotiation of cooperative play. This work is uncovering a new way of dealing with affect in linguistic variation, one which emphasizes the highly complex and embodied nature of the phenomenon.
My interests in the relationship between language and the body led to a study of the semiotic connections among linguistic prosody, the body, and forms of physical activity (Esposito and Gratton 2020). A quantitative study of the instructional styles of bodybuilding and yoga instructors on YouTube showed that bodybuilding instructors employ faster articulation rates and higher pitch than yoga instructors. Articulation rate and pitch become semiotically linked to notions of energy, and the differences in the instructors' styles are rooted in differences in levels of embodied energy that bodybuilding and yoga are assumed to require. Instructors employ linguistic features that reflect these ideologies of their activities, and in doing so, present themselves as embodied instantiations of their respective practices. The study showed that the treatment of the body as a physically-active doer of things provides an important source for the generation of iconic, energy-related meanings. Crucially, notions of energy, and variables with iconic meanings, can drive group-level patterns of linguistic variation.
While a plethora of sociolinguistic work has explored various social meanings tied to variation in the vowel space — both related to variation in particular vowel classes (e.g., Eckert 2010; Podesva 2011) or to regional vowel shifts (e.g., D'Onofrio et al. 2018; Pratt and D'Onofrio 2017) — very little is known about the indexical potential of holistic changes in vowel space peripherality. My work reconceptualizes the meaning potential of changes in vowel space size, an understudied feature, showing that the size of a speaker's vowel space indexes their changing affiliations with interlocutors and affective states (Gratton in prep). Using two sets of production data, which highlight both inter- and intra-speaker variation, I show that vowel space size, specifically that compressing the vowel space from all directions, enables speakers to mark common ground, which can be exploited interactionally to construct closeness.
Using the Living Room Corpus at Stanford, I found that the more speakers reported feeling close to their interlocutor, the more compact their vowel spaces were. However, speaker's also used more compact vowel spaces when they enjoyed the interaction less. These meanings are both tied to centralization of the vowel space, though in different ways. While this only showed interspeaker variation in vowel space size, I also examined intrapseaker variation as well.
An analysis of vowel space peripherality in my Enby Corpus found that these speakers used more compact vowel spaces when speaking with other community members, and more dispersed ones when speaking to their cisgender friends. Also, in a moment-to-moment examination of variation in vowel space size, this data shows that speakers use centralization to construct solidarity and community unity. Together, this work suggests that the dynamic variation of vowel space size is socially meaningful.
While the important relationship between the voice — notably vocal pitch — and gender identity has been firmly attested for both cisgender and transgender speakers (Gratton 2017; Podesva 2007; Zimman 2012), pitch has important uses beyond gender. A quantitative analysis of stance and alignment in my Enby Corpus showed that they used pitch variation in particular interactive ways. When either converging with their interlocutor or expressing positive affective stances, speakers used higher vocal pitch. However, they also used higher pitch when diverging from their interlocutor when discussing topics that they felt strongly and positively about (particularly related to transgender life and experiences).
Both experimental pragmatic and discourse analytic branches are interested in how distinct intonational forms are tied to particular meanings, though they heavily differ in their approaches (Goodhue et al. 2016; Ladd 1980; Ward and Hirschberg 1985; Ogden 2006). One form which has been of interest to both is contrastive intonation (L+H*), and whether the addition of a high final rise (L-H\%) is necessary for contrastive inference. I examined whether contrastive intonation alone would trigger inference — in interactional contexts, specifically evaluations — comparing perceptions of the contours L+H* L-L% and L+H* L-H\%. It also examined the impact of using rich and naturalistic stimuli in experimental settings.
While contrastive intonation in context rich evaluations does trigger pragmatic inferences of disagreement, the high final rise was not required for drawing such inferences. Thus, the generalizable form of contrastive intonation is simply the contrastive pitch accent (L+H*). However, the contextual richness of the intreactions impacted the perception of the high final rise. Some speakers, but not all, were perceived as "trying to be nice" when disagreeing with L+H* L-H\%. This work found that rich and naturalistic stimuli can be used to elicit listener perceptions in a principle way and are greatly beneficial to exploring the complexity of perception.
My master's thesis examined the linguistic practices of a individuals from a non-binary transgender community of practice (from my Enby Corpus), located in the Greater Toronto Area, Canada. It explored how members within this community variably utilize vocal pitch and (ING) in different contexts: out-group (with cis-gender friends) and in-group discussions. It also examined the relationship between the use of these features and group hierarchy.
Speakers employed lower mean F0 and a greater use of the [ɪn] variant in the out-group discussions, as compared to the in-group ones. Long-standing members made greater use of the [ɪŋ] variant than newer memebers, and they also used smaller F0 ranges and shorter F0 excursions (particularly when taking stances of seriousness and earnestness towards the somber reality of non-binary lives). Additionally, in the in-group discussions, speakers used larger F0 ranges and F0 excursions during supportive passages.