THE BIG QUESTION IN TODAY'S GROUP FOR WHATis1: Do we need to have a high target relatively soon after the low (as in tones_b or tones_c) to account for the sharp rise from WHAT to WHAT? Asked a different way: Should a contour for tones_a actually have a gradual rise from the low pitch accent through the high phrase accent to the high boundary tone?
If we look at the utterances in the ToBI Guidelines that have a tones_a transcription (i.e., manitowoc, made3-utterance4, names, money-utterance4), they look and sound quite similar to this one. There is an early L* target that is followed by a sharp rise to the H-. The H- fills the space between the pitch accent and boundary tone, and there is upstep to the H%. While it might seem that there is no obvious upstep to the H%, this might simply be because the speaker is already quite high in her pitch range at the end of the utterance.
So where does this leave us with the transcriptions in tones_b and tones_c? Well, both are still possible transcriptions, but perhaps weaker candidates than tones_a. In tones_b, it is possible for the associated H tone to occur outside out of the accented word, but that is not its canonical representation. In tones_c, the shape of the rise between WHAT and WHAT is consistent with a rise from a L* to a H*, and but as is also the case with tones_b, few in the group perceived WHAT and POINTING to be pitch accented. If they are, then they are indeed H* accents.
In terms of TOO2: The group disagreed as to whether there is an ip boundary between THAT and TOO. Those who perceived an ip boundary downstepped the phrase accent relative to the pitch accent on THAT and assigned TOO to its own ip. Those who perceived lengthening on THAT, but no disjuncture, assigned a 2p break index and downstepped the pitch accent on TOO relative to THAT.