Stress assignment in letter‐to‐sound rules for speech synthesis

This paper will describe a program that attempts to determine word stress from spelling. Four sources of constraints will be discussed: (1) syllable weight, (2) part of speech, (3) morphology, and (4) etymology. Stress assignment is a well‐established weak point for many speech synthesizers because stress dependencies cannot be determined locally. It is impossible to determine the stress of a word by looking through a five‐ or six‐character window, as many speech synthesizers do. Well‐known examples such as degrade/degradation and telegraph/telegraphy demonstrate that stress dependencies can span over two and three syllables. In addition, examples such as calculi and tortoni demonstrate that stress is sensitive to etymology, a nonlocal property of the entire word. Calculi, like most other latinate forms, obeys the main stress rule of English, which assigns antipenultimate stress to nouns ending in two light syllables. In contrast, tortoni, like most other Italian and Japanese loan words, receives penultim...