Investigating potential MERGED and DISTINCT speakers in an ongoing merger — Evidence from production and perception

This study investigates whether it is possible to use evidence from production and perception to identify different types of speakers in the study of the ongoing merger of /ʃ/ and /ç/ in Norwegian. Specifically, by investigating the acoustic and perceptual characteristics of the contrast between /ʃ/ and /ç/, we examine whether we can draw a principled line between speakers who have the merger (MERGED speakers) and speakers who do not have the merger (DISTINCT speakers). Our findings indicate that speakers and listeners in part rely on the same parameters when producing and perceiving the contrast between /ʃ/ and /ç/ in Norwegian, most notably spectral standard deviation and intensity. In both production and perception, speakers form continua, from having small or no acoustic differences between /ʃ/ and /ç/ to having large differences, and from having productions which cannot be distinguished by listeners to having productions which are reliably distinguished by listeners. While we find an overlap between these distributions, indicating that some speakers at the extremities are MERGED and DISTINCT speakers, these continua do not allow for a clear boundary between groups. Any grouping of speakers will therefore have to rely on an arbitrary boundary between MERGED and DISTINCT speakers.