Tone in Connected Discourse

Connected discourse has a number of substantial impacts on tone. First, the ideal form of tone is routinely compromised due to articulatory sluggishness. Second, the underlying tonal target itself is sometimes changed. Third, the surface tonal form is further distorted by other linguistic functions that also change pitch either deliberately or inadvertently. This article discusses how the surface acoustic form of tone can be understood in terms of the interaction between tone and these relatively independent processes. In connected discourse lexical tones in a sentence have to be said in quick succession, frequently with no pauses in between. In addition to tone, a sentence also carries many other communicative functions that either use pitch as their primary or secondary carrier or change pitch inadvertently. As a result, the acoustic manifestation of tone, mainly in the form of fundamental frequency (F0) contours, rarely resembles its underlying form. A clear understanding of tone is therefore impossible without comprehending the effects of these factors on the acoustic form of tone. This article discusses the realization of tone in connected discourse in terms of the interaction between tone and other pitch-modifying factors. The discussion is mainly based on data from Mandarin, for which extensive experimental data have been collected in recent years.