Do H*L and L*H accents have similar target positions?

Abstract

Targets of pitch accents have generally been defined on the basis of production data. Obviously, most of the research on tonal targets has focussed on peaks associated with H*L accents as these are very well identifiable, in contrast to the lows associated with L*. In the perception experiments reported here it is made clear that, in Dutch, the conventional target positions of L* and H* (being the F0-valley and the F0-peak respectively) do not coincide. The valley associated with L* which gives rise to the perception of a sentence accent occurs much earlier in a syllable than the peak associated with H*. No significant effect was found for the position of the accent: nuclear or prenuclear. Confirmation was found for the existence of the phonological contrast between H* H% and L*H H% in Dutch.