Inferring Stress-Timed Rhythm from Written Songs: A Phonetic–Phonological Study of Elton John

Abstract

This study examines how English stress-timed rhythm can be inferred from written song lines, drawing on a corpus of twenty-five officially published Elton John lyrics processed in UAM CorpusTool. The approach is purely phonetic–phonological and text-based, with no reference to audio or melody. The question is whether simple orthographic cues suggest a natural stress pattern in neutral reading. Three cues are tracked. First, weak forms are treated as normally unstressed, and contractions in spelling are taken as further evidence of reduction. Second, three clear linking contexts are isolated: “and,” “of” and “to” followed by a vowel-initial word, sequences that typically invite coarticulation and compression of unstressed syllables between stronger beats. Third, main stress in neutral declaratives is modelled with a practical proxy: the final content word of the line, or the nearest preceding content item if the line ends on a function word. In UAM CorpusTool, each verse line receives four tags: clause type, list of function words, visible contractions and main-stress proxy. The analysis is qualitative and modest in scope. Recurrent weak-form zones, regular and/of/to links and consistent final-content-word prominence together outline a plausible phonetic–phonological profile compatible with stress-timed English, while the corpus tool keeps procedures transparent and replicable.