Two problems in theories of tone-melody matching

It is now accepted that musical traditions in many tone languages aim for some correspondence between the intrinsic melody of a sequence of spoken syllables and the musical melody to which the syllables are sung.  An important feature of this correspondence involves the PITCH DIRECTION between pairs of syllables in sequence: for example, a syllable sequence L-H is best sung to a rising musical melody.  However, many questions remain, of which I will discuss two:

(1) The pitch direction principle is easy to apply to level tones, but not to contour tones.  So far the only reasonably secure generalisation is that in Cantonese pop songs, the END PITCH of a contour tone is what determines the way the tone is matched to music (Chan 1986), but it seems clear that this does not apply in all languages.

(2) The status of sequences of two identical musical notes is ambiguous. Given a sequence of two different tones (e.g. H-L), a “level” (identical-note) sequence is neither a match nor a mismatch, but something intermediate (‘non-opposing’; Schellenberg 2009); however, given a sequence of two identical tones (e.g. H-H), an identical-note sequence is a match, and mismatch appears undefined.

These two problems interact when we consider syllable pairs involving identical contour tones. I will review various empirical findings on how syllable pairs involving contour tones are treated in actual song corpora and show that there appear to be numerous specific subregularities that are difficult to express in simple general terms such as Chan’s end pitch principle.