Complete and incomplete neutralisation in Fuzhou tone sandhi

Abstract

This is a study of incomplete neutralisation using Fuzhou tone sandhi as a test case. In Fuzhou Min, Tone 44 and Tone 232 undergo putative neutralisation into Tone 53 preceding a set of low tones, while T44 and T53 are neutralised into T44 preceding T53. Data from 10 Fuzhou speakers show the former neutralisation to be acoustically incomplete, with the sandhi tone of underlying Tone 44 having a higher pitch onset, while the latter appears to be acoustically complete. This result expands the typology of incomplete neutralisation by showing that two distinct phonological objects can incompletely neutralise (A / B → C) into a third object, as previous studies predominantly feature neutralisation of one object into another: A / B → B, e.g. Beijing tone sandhi T2/T3 → T2 / __ T3. It is argued that a hybrid model with both abstract categories and stored exemplars can best account for the data (Pierrehumbert, 2002(Pierrehumbert, & 2006. Index Terms: tone sandhi, incomplete neutralization, Fuzhou Min, hybrid model http://dx.

F0 and voice quality of coarticulated Mandarin tones

“…Carryover effects have been found largely assimilatory, whereas anticipatory effects tend to be dissimilatory. Such is the case for Mandarin, Cantonese, Yoruba (Laniran & Clements, 2003), Thai, Fuzhou Min (Li, 2015), Tianjin Chinese, Taiwanese, and Itunyoso Trique (DiCanio, 2012;DiCanio & Nam, 2013). Still other studies found that carryover effects are not predominantly assimilatory, and anticipatory effects are not limited to dissimilation, as shown in Malaysian Hokkien (Chang & Hsieh, 2012), Vietnamese (Brunelle, 2009), Nanjing Chinese (Chen et al, 2018), and Mizo (Sarmah et al, 2015).…”

Incomplete neutralization and the blueprint model of production

“…Incomplete neutralization is not limited to the process of final devoicing as the following processes have also been documented as being incomplete: tapping/flapping in American English (Fox & Terbeek, 1977;Herd et al, 2010;Braver, 2014), intrusive stops in English (Fourakis & Port, 1986), schwa epenthesis in English (Davidson, 2006), coda aspiration in Andalusian Spanish (Gerfen, 2002;Bishop, 2007), coda liquids in Puerto Rican Spanish (Simonet et al, 2008;Beaton, 2016), French schwa deletion (Fougeron & Steriade, 1997), Russian voicing assimilation (Burton & Robblee, 1997), epenthesis in Lebanese Arabic (Gouskova & Hall, 2009), tone sandhi in Cantonese (Yu, 2007), Mandarin (Peng, 2000) and Fuzhou (Li, 2016), Japanese monomoraic lengthening (Braver & Kawahara, 2016), and word-final lenition in Chilean Spanish (Bolyanatz, 2020). Durational cues appear to be the majority of what is presented as being incomplete, but there is not enough evidence to suggest it is limited to the temporal domain as in some cases it was exclusively nondurational cues which were found to be incomplete (Burton & Robblee, 1997;Yu, 2007).…”