Final H

This label marks a final high boundary tone for an intonation phrase (see section 4.4 below for discussion of this level of the prosodic hierarchy). This boundary tone typically occurs finally in interrogatives, such as in example nara_quest, but is also often found at the end of declarative sentences such as example utterances kazumi and mayumi, and also on utterance-medial phrases such as nibanme and others in this guide.

<<nara_quest>> ``Is that really the one from Nara?''

<<mayumi>> ``Mayumi drank too.''

The H% mark should be placed at the right edge of the intonation phrase, aligned exactly with the word and break index marks. In cases where this location does not correspond to the actual maximum value in the F0 contour, labellers should use the early F0 event label (`` >'') to pinpoint the actual event. Before pauses, very often the maximum F0 value of this H% boundary tone will occur before the cessation of phonation (i.e. before the word boundary), due to mistrackings of F0 caused by the rapid decrease in amplitude. The use of the early F0 event label to mark the F0 maximum is shown in example utterances kazumi and mayumi.

When marking a H% boundary tone on the right edge of an intonation phrase, labellers should also not forget to label the L% boundary tone of the final accentual phrase (the H- and L% tones delimit each accentual phrase). The combined tone label L%H% (or L%HL%) is thus used for convenience. In example utterances nara_quest and mayumi, the rise to the H% from the previous L% is obvious. However, in the utterance kazumi, it may not be so apparent. In this utterance, the pitch rises to the H- phrasal tone, then rises a second time to the high boundary tone utterance finally. We know that there is a L% present since the F0 does not continuously rise after the phrasal H-, but is leveled out by the L%, then rises again to the H%.

In addition to utterance-final rises, utterance-medial H% boundary tones also occur, as seen in examples nibanme and pinku_mado. In pinku_mado, the accentual phrases /pi'Nku no/ `pink-GEN' and /ma'do o/ `window-ACC' both form their own intonation phrase (see section 4.4 below), marked by a H% boundary tone and following pause (see [Nagahara and Iwasaki 1994] for other examples of utterance-internal high boundary tones).

<<pinku_mado>> ``I will place this pink window right in the center of the triangle roof.''

One last issue that deserves mention regarding the H% boundary tone is the qualitative difference in height between pitch rises. In examples kazumi, nara_quest, mayumi, and pinku_mado we observe fairly high F0 excursions to the top of the phrase's pitch range. However, there are also cases which appear to be qualitatively different from these, in which the pitch rises only part of the way. Compare the utterance in nara_quest to the so-called ``insisting'' declarative shown in nara_insist (see also the boundary tone on /siNsitu no ma'do wa/ `bedroom-GEN window-TOP' in example utterance nibanme).

<<nara_insist>> ``That's really the one from Nara.'' (insisting)

Since the tone tier in a Japanese ToBI transcription describes only shapes of contours (phonological tonal events), the amplitude of those F0 movements is not documented. There is therefore no way to distinguish between a high-rise and a mid-rise, which are both labelled as H%. For labellers interested in this difference in boundary tone height, a site-specific tier could be added to mark the distinction. This is a good example of how a basic J_ToBI transcription can be expanded to incorporate the research topics of a particular site.