Introduction

All languages use vowels and consonants in the representation of their words,1 and a large number, referred to as ‘tone languages’, also employ tone for this purpose. In this chapter, the question of how pitch variation is used in language is continued from chapter 2 with a discussion of lexically contrastive tone patterns. Some background is provided in section 3.2, while in section 3.3 the notion Tone Bearing Unit is introduced and the autosegmental representations it implies. While rejecting the idea that there is a meaningful class of ‘accent languages’ (in the way that a class of ‘tone languages’ can be defined), we define accent as a location for a (lexical or intonational) tone contrast, following its use in the same sense in section 2.3 (section 3.5). Observe that the notion is distinct from pitch accent, which is a tone or a complex of tones occurring in the location designated as ‘accented’. When ‘accent’ is compared with the notion ‘word melody’, both can be seen as devices that explain distributional restrictions of tones. In a final section, the combined use of intonational and lexical tone is briefly illustrated.

Tone languages

The earliest distinction within the group of tone languages is between those that just have level tones, which require the syllable to reach a certain pitch height, and those that also have contour tones, which require the syllable to be said with a pitch movement. Pike (1948) termed these ‘register tone languages’ and ‘contour tone languages’, respectively.