부산대학교 도서관

Abstract Are the sound systems of languages ecologically adaptive like other aspects of human behavior? In previous substantive explorations of the climate–language nexus, the hypothesis that desiccation affects the tone systems of languages was not well supported. The lack of analysis of voice quality data from natural speech undermines the credibility of the following two key premises: the compromised voice quality caused by desiccated ambient air and constrained use of phonemic tone due to a desiccated larynx. Here, the full chain of causation, humidity → voice quality → number of tones, is for the first time strongly supported by direct experimental tests based on a large speech database (China’s Language Resources Protection Project). Voice quality data is sampled from a recording set that includes 997 language varieties in China. Each language is represented by about 1200 sound files, amounting to a total of 1,174,686 recordings. Tonally rich languages are distributed throughout China and vary in their number of tones and in the climatic conditions of their speakers. The results show that, first, the effect of humidity is large enough to influence the voice quality of common speakers in a naturalistic environment; secondly, poorer voice quality is more likely to be observed in speakers of non-tonal languages and languages with fewer tones. Objective measures of phonatory capabilities help to disentangle the humidity effect from the contribution of phylogenetic and areal relatedness to the tone system. The prediction of ecological adaptation of speech is first verified through voice quality analysis. Humidity is observed to be related to synchronic variation in tonality. Concurrently, the findings offer a potential trigger for diachronic changes in tone systems.