The social life of phonetics and phonology
Introduction
Variability is one of the defining characteristics of human speech. No two voices are identical, no two utterances the same. Variability in speech is not, however, wholly random or chaotic. Rather, it results from a number of specific sources and may form rule-governed patterns.
Much research in phonetics and phonology has been concerned with identifying these patterns and their sources, with a view to furthering understanding of central issues within these fields. Explaining systematic variation in speech may yield insights into the operation of the speech production and/or perception mechanisms, or it may help in distinguishing phonological knowledge from natural phonetic processes. It has been shown, for example, that phones may differ in their intrinsic properties: Close vowels have a higher f0 than open vowels (Shadle, 1985; Whalen & Levitt, 1995), voice onset time durations and stop release bursts vary as a function of both place of articulation and the quality of adjacent vowels (Liberman, Delattre, & Cooper, 1952; Westbury, 1983), articulatory gestures tend to be longer and ‘stronger’ when in stressed contexts or preceding major prosodic boundaries (Kelso, Bateson, Saltzman, & Kay, 1985; Shattuck-Hufnagel, & Turk, 1996), and faster speech tends to produce more articulatory undershoot (Lindblom, 1963; Perkell, Zandipour, Matthies, & Lane, 2002). Such sources of variation are largely explicable with reference to articulatory constraints and/or the natural laws of aerodynamics and acoustics operating within the vocal tract (Ohala, 1983; Shadle, 1997; Stevens, 1998). By contrast, other variable patterns appear to be language specific and thus represent part of the phonological knowledge acquired by speakers of the language. English, e.g., displays contextual allophony such that voiceless stops are aspirated in initial prevocalic position, but not when in onset clusters preceded by /s/. Similarly, language-specific morphophonological alternations yield variable patterns illustrated by act [−t] versus action [−ʃ−] and elect [−t] versus election [−ʃ−].
While these examples show that certain types of variability have been studied extensively in phonetics and phonology, other types have been the subject of rather less attention. ‘Sociophonetic’ variation is a case in point. This term is one that has come to be used extensively in the last few years, referring usually to variation in speech that correlates with social factors like speaker gender, age, or social class. It is still something of an exception, though, to find experimental phonetic or phonological studies that include such social dimensions in their methodology, or to find reference to sociophonetics in textbooks, university course descriptions, or conference announcements.
There are several reasons why sociophonetic factors have remained peripheral to phonetics and phonology. Above all, the dominance of particular theoretical models and methodological traditions has meant that social factors have been partitioned de facto from the ‘purely linguistic’. The emphasis in the generative tradition, e.g., has been on describing the linguistic knowledge of the “ideal speaker–listener, in a completely homogeneous speech community” (Chomsky, 1965, p. 3). Differences between speakers of a given language have thus been less of a concern.
In this article, our aim is to assess the potential for sociophonetic variation to inform theoretical modeling in phonetics and phonology. We begin with an exploration of what is meant by ‘sociophonetics’ and ‘sociophonetic variation’ (Section 2), since these terms have not been particularly well defined. We also offer several illustrations of sociophonetic variation, drawing especially on the findings of our own research on non-standard varieties of British English. In Section 3, we examine the relationship between the sociophonetic and linguistic properties of speech from the perspective of a child acquiring a first language. We present data from studies of child-directed speech (henceforth CDS) and from children's speech production in which identifying sociophonetic variation was a principal aim. The findings of these studies form the basis for a discussion of the relevance of sociophonetic issues for phonetic and phonological theory (Section 4). In particular, we question whether the marginalization of sociophonetic factors is sustainable in the construction of a comprehensive account of the long-term storage of information about sounds and sound structure, and how that information is accessed in the on-line processes of speech production and perception. We suggest that an exemplar-based model is the most promising candidate for offering a unified account of how sociophonetic as well as linguistic material might be learned and stored. We also explore the implications of the findings of our sociophonetic studies for exemplar models. Finally, in Section 5, we highlight a number of important issues for future research in sociophonetics.
Section snippets
The scope of sociophonetics and sociophonetic variation
The definition of ‘sociophonetics’ is somewhat vague. The term has a reasonably long history (tracing back at least to Deshaies-Lafontaine, 1974), but it has been neither widely nor consistently used. There is, for example, no reference to it in the Oxford English Dictionary (second edition, 1989), or in popular linguistics glossaries such as Matthews (1997) or Trask (1997). In Crystal's (2003) Dictionary of Linguistics and Phonetics, it makes its first appearance in the fifth edition. It has
A child's perspective on sociophonetic variation
Given the pervasiveness of sociophonetic variation it is perhaps surprising that it has not played a more prominent role in phonetic and phonological work. Sociophonetic variation clearly reflects acquired knowledge about speech sounds, and sociophonetic information is conveyed through precisely the same medium as propositional information. Nevertheless, a clear account is still lacking of how sociophonetic information is represented cognitively. In particular, it is not clear how it is stored
Towards an integrated model
In light of these observations on sociophonetic variation, we turn now to a discussion of how such variation might be accommodated within a theoretical framework. In doing so, a specific question is the extent to which theoretical models can account for the intertwining of indexical and other information within the speech signal.
Many if not all theoretical models of phonetic and phonological structure have mechanisms to cater for variation. For example, within generative phonology variability
Concluding comments
In 1 Introduction, 2 Sociophonetic variation, 3 A child's perspective on sociophonetic variation of this article, we defined and illustrated sociophonetic variation, highlighting both its pervasiveness and also the relatively minor role it has played in the development of phonetic and phonological theory. In Section 4, we outlined a number of ways in which exemplar models hold appeal for modeling sociophonetic variation. This discussion also underlines the fact that, for all the potential that
Acknowledgments
We would like to record our thanks to Ghada Khattab, Leendert Plug, and Gareth Walker for their thoughtful comments on a draft of this article. Issue editors Jen Hay and Stefanie Jannedy provided many helpful suggestions for improving the paper as did the reviewers who included Cynthia Clopper and Sarah Hawkins. We are also grateful to Denise Deshaies, John Esling, Gillian Evans, Mike MacMahon and Gillian Sankoff for their help with specific issues raised here.
References (157)
- et al.
Variation and change in optimality theory
Lingua
(1998) - et al.
Children's perception of speaker sex
Journal of Phonetics
(1982) Lexical properties of postlexical rules: Postlexical derived environment and the elsewhere condition
Lingua
(1991)- et al.
Some acoustic cues for the perceptual categorization of American English regional dialects
Journal of Phonetics
(2004) - et al.
Stop epenthesis in English
Journal of Phonetics
(1986) - et al.
Pitch accent realization in four varieties of British English
Journal of Phonetics
(2000) Roles and representations of systematic fine phonetic detail in speech understanding
Journal of Phonetics
(2003)- et al.
Coarticulation in recent speech production models
Journal of Phonetics
(1977) Variable domains and variable relevance: Interpreting phonetic exponents
Journal of Phonetics
(2003)- Allen, W. (forthcoming). Phonological variation and change in the speech of Tyneside adolescents. Unpublished PhD...