The question of how the delicate pitch variations that humans can produce are employed in language has been one of the most fascinating topics in phonological and phonetic research at least since Joshua Steele's Essay towards establishing the Melody and Measure of Speech (Steele 1775), but has developed a particularly fruitful momentum in the past two decades. This book is an account of my current understanding of this issue.

Lexical pitch variations and intonational pitch variations are phonologically represented as tones, like H(igh) and L(ow), which form a string of elements running parallel to the string of vowels and consonants. Like vowels and consonants, tones may delete, assimilate, or change their value in particular contexts. They are organized temporally with reference to prosodic constituents, such as the mora, the phonological phrase, and the intonational phrase. Studying the phonology of tone and intonation can sharpen one's understanding of phonetics and phonology in a relatively brief time. The greater variation in the realization of tones, together with their relative sparsity compared with the denser occurrence of vowels and consonants, encourages a comprehensive view of the trajectory from underlying representation to phonetic surface form. As a result, the difference between phonology and phonetics as well as that between underlying phonology and surface phonology can more readily be appreciated.

The theory of intonational structure presented in this book owes a great deal to the work of Janet Pierrehumbert, whose 1980 thesis on American English intonation in effect provided the theoretical framework it has adopted, which work itself was intellectually indebted to Gösta Bruce's 1976 thesis on Stockholm Swedish.