Sociolinguistic effects of speaking in masking noise
A classic study by Mahl (1972) found that when participants spoke in loud masking noise, they used ‘more colloquial’ language variants, such as glottal stops in place of intervocalic /t/. This was interpreted as demonstrating that participants’ attention had been diverted from self-monitoring their speech by the masking noise, causing them to revert to less prestigious language variants rather than using ‘more formal’, Received Pronunciation forms to increase their intelligibility. However, this conclusion was based on speculation rather than analysis of participants’ intelligibility, and the sample size was very small—results were drawn from close analysis of one participant, and anecdotal reports of others.
Here, we attempt a partial replication of Mahl (1972), looking at phonetic variation in masking noise. Participants were recorded as they spoke in silence and over white noise played at 60, 70 and 80dB SPL through circumaural closed-back headphones. Recording speech in a range of masker intensities allows us to correlate noise intensity with the degree of phonetic change. The transcribed interviews were then analysed for th-fronting and intervocalic /t/ glottalisation. Results are discussed with reference to linguistic and psycholological theories of auditory self-monitoring and speech production in noise.