Artistic Recitation of Metered Speech

9

A similar story (but with some significant differences) may be told about the end of the very first line of the play (excerpt 3). First, the line is followed by a 0.056 msec pause; this is less than negligible as an acoustic cue for line-ending (but, as we shall see, it has a different function). Secondly, the line ending is indicated by an unusually long, classical "terminal contour". Third, the last syllable "-tent" is, as expected, considerably lengthened. It is, indeed, the longest syllable in this line (0.490 msec), even though, in English, sound sequences in polysyllables are usually shorter than comparable sequences in monosyllables (compare, for instance, tailvs. tailor;I have elsewhere discussed this issue at some length; Tsur, 1998: 156-157). The only (monosyllabic) word that approximates its duration is now(0.486 msec). The length of this line-initial word is explained by rhetoric, not rhythmic, reasons. To appreciate the duration of this syllable (of a tri- syllabic), one might observe that the sequence wint-(in "winter") is slightly over half as long (0.291). Fourth, the word-final oral stop [t] is excessively overarticulated (overarticulating, by the same token, the word boundary and the line boundary as well).

nowisthewinterofour discontent Figure 4Wave plot and pitch contour of "Now is the winter of our discontent"

My perception-oriented theory of metre assumes that certain rhythmic problems can be solved by the overarticulation of certain syllables. All the gurus who instructed me in empirical research told me they were not aware of any possibility for the machine to indicate overarticulation. It seems to me now that the machine canshow overarticulation when, e.g., certain identifiable features of carefularticulation are slightly or greatly exaggerated, such as duration; but there are some additional, quite interesting, features. Language in everyday conversation is usually underarticulated. Especially in English, certain articulatory features of word boundaries are almost always suppressed, and words run into one another. Consider the pairs of back-to-back [s]s in figure 5. The word-final [s] in thisis run into the word-initial [s] in sun.This is the normal way of speaking.⁷By contrast, between the word-final [s] of gloriousand the word-initial [s] of summera minute 0.059 msec pause is inserted. The listener doesn't perceive it as a pause, but as an articulatory gesture intended to separate the back-to-back [s]s, a kind of

7

We still hear here an elusive separation between the two [s]s. The spectrograms give illuminating information about this, but I can't disscuss it here at any length.