On the Dual Relativity of Tone

Introduction A longstanding debate in tonal phonology concerns how to characterize the relative nature of tone. Two types of tonal relativity are observed. First, tones are relative to a speaker’s pitch range. Thus a canonical lexical High tone is higher in a given speaker’s pitch range than a canonical lexical Low tone in the same position. Second, tones are relative to one another in sequence. Thus a High tone is canonically higher than an adjacent Low tone. The necessity of defining both types of relativity in the grammar is exemplified by the fact that tones can drift down over the course of a phrase, such that a High tone near the end of the phrase may be lower than a Low tone near the beginning of the phrase. In spite of this drift, alternating High and Low tones canonically maintain their heights relative to one another. This suggests that tones are defined both with respect to the speaker’s pitch range, as well as with respect to one another in sequence. The question of how best to capture this dual relativity in a formal linguistic theory has posed a challenge for many years.