Sources of tonal variation in Mandarin

Yi Xu
Journal of Chinese Linguistics monograph series #17: 1-31.

It has been widely recognized in recent years that tones in Mandarin and in other tone languages vary extensively in connected speech. The nature of many of these variations, however, remain unclear. To better understand these variations, it is critical to identify their sources. It is argued in this paper that there are two fundamentally different sources of tonal variation: linguistic/paralinguistic demands and articulatory constraints. Linguistic/paralinguistic demands specify the underlying pitch targets of lexical tones and pitch ranges over which tones are implemented. Articulatory constraints, on the other hand, determine how underlying pitch targets are actually realized. Among the linguistic/paralinguistic demands are the following:

Lexical tones that specify the relative height and shape of tonal targets;
Tone sandhi rules that change the tone categories in different tonal contexts; and
Higher level prosodic factors that specify the pitch ranges of tones:
Focus that specifies the pitch ranges of pre-focus, on-focus, and post-focus words,
Topic initiation that specifies the pitch range of the initial
words or phrase in an utterance; and
Other intonation patterns that may have even more complex pitch range specifications.
Articulatory constraints determine the following (and possiblly more):
The overall pitch range of a speaker;
The maximum rate of pitch change;
The maximum speed of changing the direction of pitch movement;
Vowel intrinsic pitch; and
F0 perturbation by consonants.
In Mandarin, during speech production, each tone is assigned to a syllable and hence is implemented coordinately with the syllable. The larynx produces the F0 contours by implementing tonal targets coordinately with their associated syllables over the pitch ranges specified by factors such as focus, topic initiation and other intonation patterns. Due to the limit of maximum rate of pitch change, a tonal target is often only approximated rather than fully reached, and the best approximation is achieved only in the later portion of a syllable.

Phenomena such as carryover assimilation, downstep, peak delay, probably result from the interaction between linguistic/paralinguistic demands and articulatory constraints. Phenomena such as declination probably result from a combination of more than one interactions.

Finally, the exact source of anticipatory dissimilation is still a mystery. Some speculations will be discussed.

See other publications


Home