AbstractIt has been well established for some time (Lehiste, 1970, Nooteboom, 1972, Lindblom … Rapp, 1973, Klatt, 1976) that measurable sound duration is the result of the operation of several linguistic and phonetic factors. The domains of these factors vary in extension from discourse sections via single sentences and their syntactic or prosodic parts to the syllable and the individual sound. Related to a decreasing order of these domains the following hierarchy of temporal control factors may be set up for speech: (a) the global speech rate, which is determined by the individual speaker or by the context of situation and which affects large chunks of monologue and dialogue; (b) the semantic structuring of utterances, resulting in a speeding up of sentence sections with peripheral information content and in a slowing down for points of central interest; (c) the position of syllables and words at the end of syntactic and semantic units, whose boundaries may be additionally marked by pauses: compared with the occurrence inside such units final syllables are lengthened with increasing strength towards the end (this is a special case of (b)); (d) the position of a syllable in mono-, di- or polysyllabic words: a stressed syllable is, for example, shorter when there are further unstressed syllables following it in the word (this is a special case of (c)); (e) stress: stressed syllables are longer than unstressed ones; (f) phonological quantity: phonologically long segments have greater duration than phonologically short ones under otherwise equal conditions; (g) phonetic context; e.g. - vowel duration before voiced/voiceless (lenis/fortis), as in English “bead/beat”, German “leiden/leiten” - vowel duration before fricative/plosive (e.g. vowel lengthening before voiced fricatives, but not before voiced stops in French) (h) intrinsic segment duration, e.g. - high/low vowels - plosives/fricatives - voiced/voiceless (lenis/fortis) obstruents.