Levels of Phonological Abstraction and Knowledge of Socially Motivated Speech-Sound Variation: A Review, a Proposal, and a Commentary on the Papers by Clopper, Pierrehumbert, and Tamati, Drager, Foulkes, Mack, and Smith, Hall, and Munson

Effects of Language Background on the Development of Sociolinguistic Awareness: The Perception of Accent Variation in Monolingual and Multilingual 5- to 7-Year-Old Children

“…This approach argues that linguistic knowledge is based on detailed memory representations which simultaneously encode nonlinguistic (e.g., who was speaking, in what situation) and linguistic (e.g., segmental features, voice quality, pitch) information (Docherty & Foulkes, 2014;Munson, 2010). Any abstract representations are thought to arise from common factors amongst the many memories a speaker-listener has (Hawkins, 2003), such that knowledge of phonetic and social variation derives from the clustering of similar forms, with knowledge of socially structured variation emerging where linguistic and social differences coincide (Docherty & Foulkes, 2014).…”

“…For example, hearing tokens of BATH produced with the northern English [ae] may "prime" a listener not only to expect more tokens of BATH words produced with this vowel, Evans (Strand, 1999), listeners may be better able to map the incoming signal to their own, underlying abstract representations of these categories. In terms of child language development, we might hypothesize that as children build up a bank of detailed representations through experience with speech that is rich in language-specific and potentially talker-specific phonetic detail, there is a developmental progression from specific to more abstract knowledge of sound structure (Munson, 2010). Acquisition might therefore in part be driven by frequency of exposure, with those differences that are both phonetically large and experienced more frequently (e.g., male-female differences based on pitch range) emerging earlier than those which are more arbitrary (e.g., ethnicity, social class) and which might be experienced relatively rarely within the speech community (see Foulkes & Docherty, 2006).…”

“…Previous research has suggested there is a developmental progression from specific and concrete knowledge (phonetically rich, potentially talker-specific) to more abstract knowledge of sound structure (generalizations), with age and exposure playing an important role in the development of an understanding of the social significance of variation (cf. Munson, 2010, Beckman et al, 2007. Our results provide some tentative support for this view; despite being unfamiliar with the accents tested, our bilingual children were able to successfully categorize talkers, perhaps because they had developed representations in which particular speaker attributes were associated with more fully developed macrosociological categories.…”

Sociophonetic Variation in Speech Perception

“…However, for simplicity of both experimental design and statistical analysis, much of the work in perceptual sociophonetics treats social categories as discrete and absolute (e.g., male versus female, Australia versus New Zealand). Consistent with arguments made by Munson (2010), a next step is to find ways where more subtle social information relating to a speaker’s construction of their identity can be incorporated into work on speech perception, investigating questions such as: how is a listener’s perception of a speaker’s style created? How do the different styles and stances of the listener affect speech processing?…”

Abstract social categories facilitate access to socially skewed words

“…Indeed, a number of recent models of lexical representation and perception very explicitly assume that our experiences of words are simultaneously interpreted for both linguistic and social meaning–activating, for example, both phonological categories, and social categories. For example, Munson [31] provides a schematic of a rich hybrid exemplar theory, in which the incoming signal is simultaneously indexed to multiple levels of representation–both social and phonological. And Sumner et al’s “dual route approach to speech perception”, assumes that “learned acoustic patterns are mapped simultaneously to linguistic representations and to social representations” [32](p.1).…”