Linguistics

This article presents two studies that investigate the phonetics and phonology of the Karbi (Tibeto-Burman; Assam) lexical tone system: an acoustic-statistical study and a follow-up perception study. Although currently called a tone system, the two studies show how calling it such is not uncontroversial.
A previous account of the Karbi tone system proposes three tonal categories (low, mid, high) for open or sonorant-final syllables, whereas checked syllables (ending in /p,t,k/) may only be low or high; additionally, while low and high tones are realized as differences in pitch height (F0), the mid tone also
exhibits characteristic glottalization word-finally (Grüßner 1978). In fact, a preliminary acoustic investigation by Konnerth (2010) suggests that the pitch height of the mid tone is non-unique and especially overlaps with the high tone in the case of bare monosyllabic stems. It is proposed that the
word-final glottalization is the more important phonemic cue. The current acoustic-statistical study expands the scope of investigation by looking not only at bare monosyllabic stems but also the same stems featuring one of two suffixes: a low tone suffix -pò and a high tone suffix -jí. By obtaining expanded
data sets from both a male and a female speaker, we were able to include a preliminary statistical analysis of F0 contours, which confirms the pitch height overlap between mid and high tones for bare stems, -pò suffixed stems, and -jí suffixed stems in the case of the female speaker. The male speaker shows
an interestingly different pattern with sufficient differentiation between mid and high tones in all cases, which other native speakers easily replicate when contrasting minimal pairs. To investigate the phonemic basis of the mid vs. high tone contrastive production, we followed up with a perception study, for which seven native speakers engaged in an identification task of tonal minimal pair items, listening to the
stimuli employed in the original acoustic study from the male and the female speaker. Our results suggest that the pitch height differentiation between the mid and high tones as produced by the male speaker is not phonemic but may in fact indicate a covert contrast.