Phonetically untrained British English-speaking listeners were asked to provide and rank up to four unprompted and unguided verbal descriptions of the voice qualities heard in controlled audio samples in English and German spoken by one talker per language. These were then pooled and ranked by frequency, and popularly-chosen descriptive labels were extracted from the list. After having listened to the same audio samples as were heard by the first participant group, a second group of listeners was asked to select and rank up to five labels from the popular choices list. The terms chosen by lay listeners, and the intersection of these descriptions with the technical terminology used by phoneticians and forensic speech scientists, are considered from the perspective of the elicitation and reliability of earwitness testimony in the forensic domain.
|York Papers in Linguistics Series 2
|Accepted/In press - 14 Dec 2012