Voicing contrast in Nepali infant-directed speech
thesisposted on 2022-03-29, 03:28 authored by Sujal Pokharel
Studies investigating voicing in oral stops using voice onset time (VOT) in IDS and ADS, present mixed results regarding the presence of hyper-articulation of VOT in IDS. However, there have been no studies of VOT in IDS in a language with a four-way contrasts of voicing in stops such as Nepali. Sixteen Nepali speaking mother-infant dyads were recruited. Four target pictures of minimal pair objects, each contrasting in onset consonant voicing (/ɡa.ɖa/ ‘bullock cart’, /ɡʱa.ʈi/‘neck’, /ka.ʈa/ ‘hairpin’, and /kʰa.na/ ‘food’) were selected as targets. Mothers were asked to play with their infant using the target pictures, thereby eliciting IDS. To elicit ADS, mothers were asked to interact with the adult experimenter. Acoustic analysis was then carried out for all word initial stops in the target words that occurred in sentence initial position or in isolation, in both IDS and ADS. Voicing cues were measured as lead time and lag time. In addition, the occurrence of devoicing (complete absence of lead time) in voiced consonants was recorded as a binary variable. Vowel duration was measured to control for speaking rate. The aim of the current study was to test two hypotheses. The first hypothesis was that there would be hyper-articulation of devoicing, lead time, and lag time in IDS compared to ADS.The second hypothesis was that the hyper-articulation might be a side-effect of speaking rate differences. The results showed the absence of hyper-articulation of devoicing, lead time, and lag time in IDS. Rather, the higher rate of devoicing and shorter lag time led to poorer voicing contrasts between the stop categories in IDS compared to ADS. Further, this difference between the categories could not be explained by speaking rate differences between the registers. The longer vowels in IDS suggest that mothers focus more on vowels while talking to infants, and this appears to have the effect of shortening the onset consonants. To our knowledge, this is the first time that the voicing contrast in IDS is being looked at this way. This then has wide ranging theoretical implications for better understanding the nature of IDS and its effects on language learning.