Musical Beats and Intonational Phrases in French and English Traditional Songs

Résumé En

Traditional French songs are characterized by a special dialect of the language. In this dialect, words or phrases are “feminine” or “masculine”. A feminine expression ends phonetically with a posttonic syllable. In that case, the syllable nucleus is always a schwa. A masculine expression ends phonetically with a tonic syllable. The nucleus of this syllable cannot be a schwa. Some words/phrases are necessarily masculine (for example "été" or "train"). Other words are freely delivered as masculine or feminine: for example, the word “cadette” may be feminine ([ka.ˈdԑ.tə]) or masculine ([ka.ˈdԑt]). Similarly, vie may be pronounced as [ˈvi.ə] or [ˈvi].It has been argued that, in traditional French songs, the stress-to-beat matching is strictly constrained for a “line ending” (i.e. an intonational phrase ending): if the phrase is masculine, its last syllable cannot match a weaker beat than the penultimate syllable; if the phrase is feminine, its lat syllable must match a weaker beat that the penultimate syllables. In other words, there is a strict stress-to-beat matching at the end of an intonational phrase. The only exception is be in case of a melisma: in that case, the last two syllables are on strong beats (see Dell & Halle 2009).Nevertheless, there are other exceptions to Dell & Halle generalization. In this paper, I will show that an intonational phrase with a feminine ending may have its last two syllables on strong beats, even without a melisma. I will propose a new generalization to predict correctly the data.In English a feminine ending is posttonic as well, and a masculine ending is tonic. The only difference with French is that a specific word or phrase cannot be freely feminine or masculine. For example, science or working are necessarily feminine, and reboot or orange are necessarily masculine.I will show that the generalization explaining the stress-to-beat matching concerning the end of an intonational phrase is in English exactly what it is in French. This reinforce the generalization, which, at first sight, might been seen as ad hoc, i.e. made especially to explain the French data.