The aim of the research reported here is to determine how people actually produce and perceive intonation and to differentiate clearly between the linguistic and emotional aspects of intonation and stress. The author presents evidence that, contrary to general assumptions, intonation is a central rather than peripheral linguistic feature; that it is an innate rather than acquired characteristic; that it is based on universal constants of human psychology, and as such is an invariant among specific language groups. At each stage he examines data regarding the articulatory, acoustic, perceptual, phonetic, and syntactic dimensions of intonation for American English as well as the available data for both related and unrelated languages, including Russian, Finnish, Japanese, and Swedish.
The descriptive structure of the analysis is based on the breath-group, a feature whose universal aspects are manifested in the cries of newborn infants. The unmarked or normal breath-group produces acoustic patterns that are characteristic of declarative sentences in English and numerous other languages. The marked breath-group characterizes the pattern of the “yes-no” question when “interrogative” words or particles are syntactically deleted. Evidence from many languages suggests that similar rules may be universally present.
The author points out that the breath-group is suprasegmental. A generative syntactic analysis of English indicates that it can span and delimit any constituent of the derived phase marker though it most often spans the primary constituents like the sentence, since within physiological limits it requires less effort to use longer breath-groups can, in certain instances, “disambiguate” an utterance that would otherwise be ambiguous by phonetically manifesting the derived constituent structure. Many of the examples that have hitherto been cited in linguistic studies as evidence for the “emotional” or “attitude-conveying” aspects of intonation appear actually to be examples of this phenomenon.
The linguistic feature prominence is introduced in the analysis in relation to linguistic stress and emphasis. The data presented indicate that the perception of intonation may involve the listener's knowledge of the archetypal or primary articulatory correlates of the breath-group and prominence. The author emphasizes that in the perception of language the listener makes use of a feedback mechanism of the “analysis-by-synthesis” type that is based on a knowledge of the grammar of the language as well as attention to the purely motor aspects of speech production.