Acoustic convergence : exploring the influence of ambient noise on speech production.
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
Increasing work across disciplines has recognized various ways in which people can be influenced by aspects of their surroundings. In fact, people have even been found at times to take on certain properties of the environment or objects within it. This process of convergence, where people align in particular ways with characteristics of their surroundings, has also been noted in a variety of linguistic-research and is the focus of the present work. While acoustic-convergence between speech partners is well attested at this point, little is known about how speech production may be influenced by ambient noise in the speech environment. Aiming to address this gap in the literature, the present work outlines the execution and analyses of three speech-in-noise experiments designed to test whether or not productions become systematically more (and/or less) like ambient noise in the dimensions of pitch, intensity, and speech rate/tempo. Furthermore, this work also investigates potential drives for any such convergent/divergent behaviours, as rooted in primarily automatic, social, and experiential motivations.
Two forms of noise are explored in the present work. Specifically, background music appears in each of the three studies, and background speech is introduced during Experiment 3 in a way that allows for direct comparison of convergence effects across noise types. A number of experience- and socially-based variables are tested for potential contributions when convergent/divergent behaviours are shaped.
This dissertation provides evidence that speech production is affected in reliable ways by background music, such that speech converges and diverges acoustically with aspects of that noise. A self-reported estimate of music listened to per day and a speaker’s identified gender are recognized as the most powerful predictors across experiments. As argued by Babel (2009), neither a wholly automatic nor socially-based theory of convergence can sufficiently account for the present data, and it appears clear that both play a role in shaping these changes to speech production. The effects observed in this work closely resemble convergence between speakers in previous sociolinguistic research, where personal attitudes and alliances are reflected through either convergence or divergence – moreover, certain forms of personal experience are recognized to further mediate these effects. The present work therefore provides valuable results for future work investigating theories of speech perception/production, and suggests that we as speakers are constantly updating our production patterns to become more (or less) like our surroundings.