Intonational variation and social meaning: Categorical and phonetic aspects

Abstract

The field of sociolinguistics has in recent years witnessed an increasing interest in the social meaning of variation, examining the ways in which speakers exercise linguistic choices to construct identities and personae (Eckert 2000, Zhang 2005). While studies adopting this approach, known as the third wave of variation, have typically investigated the meaning of particular categories of variant (e.g., -in vs. -ing), the phonetic details differentiating tokens belonging to the same category have been disregarded. I argue here that social meaning is encoded not only in the choice of variant, but also in the phonetic properties characterizing variants. Tapping into the full range of social meanings conveyed by phonological variation requires looking at both categorical and phonetic aspects of variation, a point I illustrate here by considering variation in declarative intonation. In addition to arguing for the social significance of phonetic variation, this paper seeks to bring intonational variation into the third wave, as studies concerned with the social meaning of variation have thus far concentrated on segmental variation. In order for linguistic forms to carry social meanings, they must be sufficiently salient when they are uttered. I draw a distinction between two kinds of salience: categorical salience and phonetic salience. Categorical salience, as summarized in (1), can be determined by the frequency with which a given category of variant occurs. Under this view, infrequent variants stand out more than frequent variants. Thus if the word ain't is uncommon in my speech, it stands out when I use it. Ideologies linked to a variant may also endow it with categorical salience. For example, the alveolar form of the variable (ing) constitutes a stereotype in the Labovian sense (Labov 1972) because people explicitly comment on the feature, with phrases like 'dropping your Gs.'

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 19 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.