Introduction: Tone and intonation from a typological perspective

Abstract

Traditionally, tone languages and intonation languages are considered to be typologically distinct. Tone languages use pitch changes to distinguish morphemes, while intonation languages use meaningful pitch changes at the sentence level, often anchoring them to positions of phrasal stress (Ladd, 1996). However, tone and intonation have come to be considered more and more as overlapping areas of research since Pierrehumbert’s (1980) influential Ph.D. dissertation persuasively argued that intonation melodies are best represented as sequences of tones, using the autosegmental formalism developed for the analysis of tone languages. Modern textbooks (Ladd, 1996; Yip, 2002; Gussenhoven, 2004) recognize this fundamental compositional similarity between tone and intonation. Recent research shows, moreover, that there is no clear-cut distinction between tone and intonation languages. Tone languages also make use of intonation, defined as meaningful alternations in pitch across the sentence (e.g., Xu, 1999 for focus in Chinese, Downing et al., 2004 for Chichewa and Rialland (this issue)). Similarly, recent research argues that intonational melodies show tonal behavior in some non-tonal languages (Fery and Kugler, 2008). For reasons like these, Jun (2005) argues for a more fine-grained feature set to classify languages according to how they use pitch changes, placing them along a typological continuum rather than in absolute classes like tone vs. intonation language. For instance, the tone language Mandarin Chinese has both lexical pitch (i.e., tone) and stress, whereas Cantonese only has lexical pitch (Jun, 2005:444). Mandarin thus shares the property of having stress with well-known intonation languages like German or English, even though it is also clearly a tone language (i.e., a language with phonemic tone) while English and German are not. Some of the articles in this www.elsevier.com/locate/lingua Available online at www.sciencedirect.com