Notes for February 4th and 11th Meetings:

Part 1: Well I g-

Some transcribers argued for an ip boundary between WELL and I, but they assigned different break indexes:

• 3 = sense of disjuncture as expected for an i.p. and clear tonal fall.
• 2 = weaker sense of disjuncture than expected, but clear tonal fall.

Transcribers disagreed about whether WELL and I are both accented.

Transcribers also disagreed about whether the speaker restarts the IP following the dysfluency. Some transcribers argued that the local prominence on WENT indicates restarting a new i.p., but others argued that the local prominence could easily be a continuation of the same i.p., perhaps with a L+H* pitch accent to account for the rise in pitch.

Part 2: went out one morning and I saw all around

As often happens, transcribers disagreed about whether the rises at WENT and AROUND sound like (and look like) H* or L+H*. In the case of WENT, if it begins a new i.p., ToBI Guidelines instruct you to use a H* pitch accent. On the other hand, you could argue that the pitch rises within the syllable enough to warrant a L+H* (and you might make a note in the misc tier pointing out that you are going against guidelines for that reason).

Transcribers similarly disagreed about the extent to which the phrase accents at MORNING (!H vs L-) and AROUND (H- vs !H-) fall.

In addition, not everyone agreed that there was the expected sense of disjuncture for the ip (i.e., 3- instead of 3). Indeed, one transcriber observed that perhaps there is no ip boundary between AROUND and IT given that there was a temporary garden path effect in which IT was parsed as the object of the preposition AROUND. However, the problem with this interpretation is that without some i.p. boundary between the !H* pitch accent on AROUND and the H* pitch accent on BACK, there is no accounting for the relatively low pitch of IT WAS LIKE. An i.p. boundary at AROUND allows for pitch range resetting in the next intermediate phrase. (One transcriber entertained a L- boundary at the end of LIKE, but wasn't completely comfortable with it, due to the lengthening on AROUND).

One transcriber thinks that the reason the i.p. boundaries don't sound like full i.p. boundaries to some ears is that the speaker is speaking very quickly (this transcriber used "3" instead of "3-").

Part 3: it was like on the back bumper

Some transcribers heard downstep from BACK to BUMPER, but others did not. Those who did not transcribe downstep did mark a delayed F0 peak in BUMPER.

Part 4: of the Honda too

Everyone agreed that TOO is in its own ip, and most transcribers heard a sequence of L+H* accents. If you are following ToBI Guidelines strictly, then TOO should be labeled as H* (though see note in Part 1 for WENT). There was disagreement about the phrase accent on Honda: one transcriber clearly heard this as level with the pitch accent. As for the phrase accent on TOO, there were different judgments about the degree of fall (!H- vs. L-).