Notes for the week:

These notes, as well as the posted TextGrid, were generated from the discussion among the group in attendance, including Kim Silverman. I have summarized the contents and have posted one "consensus" transcription in the TextGrid. Possible alternatives are discussed within the notes below, and transcribed in the alt* tier. Some possibilities were ruled out during discussion so they are not included in the TextGrid. (-LM)

General Comment:
The speech is stylized and very deliberate. Some commented that it sounded like the narration for a broadcast or school video. It turns out that the speaker is a professional voice talent who has narrated documentaries. His style is consistent and representative of a genre that we hear often in news broadcasts, documentaries, advertisements, automated telephone systems and customer service centers, etc.

Sequence: "During the..." The question was whether there might be an i.p. boundary after "During". There is no word lengthening to suggest an i.p., and the contour can be accounted for with a L+!H* on "late" (down-stepped with respect to the peak on "During"), so there is no need (or much evidence) for a L- on "During". Also, it would be unlikely (though not impossible) that an i.p. boundary would be located in the middle of a noun phrase after an article.

Since the accent in the first word falls on the first syllable, a L+H* pitch accent on "During" is unnecessary, and following the ToBI conventions it would be perfectly reasonable to label with just a H*, unless you had strong feelings that this really was a L+H*. One labeler chose the L+H* pitch accent. Notice the delayed peak on the second syllable of "During". In case anyone is wondering, this is not a L*+H pitch accent on "During" (compare the examples of L*+H in the ToBI Guide). Kim S. says: "If we do not posit an intermediate phrase boundary after "During" then this is a perfectly fine environment for any accent to be aligned later than its associated syllable. The factors making it late are:

He adds as a side note: "If we don't model these factors (and other similar ones) carefully and quantitatively in speech synthesis, then the resultant speech tends to sound like it has a foreign accent, often described as sounding Swedish. cause the accents will not align with the segmental structure the way native speakers/listeners expect)."

Sequence: "...late..." Whether or not there is an intermediate phrase boundary on "late" is the question. Two people though so, especially given the short region of silence following that word; others did not. If there is an intermediate phrase boundary at "late", then the pitch accent on "eighteen" would not be down-stepped with respect to the accent on "late". If there is not an intermediate phrase boundary on "late", then there is a successive series of down-steps across 'late eighteen hundreds". The transcription with the intermediate phrase boundary is shown in the alt tier. In the tones tier is the transcription without the i.p. Note that the break index at "late" in this case could either be 1 or 2: 1 if you think that the silence is due to the deliberate speech style; 2, if you feel that there is more of a disjuncture here than between other word boundaries. I put a 2 in the tones tier and listed alternatives in the alt tier (LM).

Sequence: "...hundreds..." Some transcribed !H* on this word, and some L+!H*. A mere !H* would not be quite so prominent. A comment: "I think a characteristic of this "voice talent documentary" style is that he uses lots of L+H* accents whereas in normal day-to-day conversation H* is more common.

Sequence: "...the British..." Probably the most straightforward part of the contour.

Sequence: "...conquered..." The second syllable is much higher than the first, though this is not an interpolation to the subsequent accent on "Yoruba". If this is a delayed H* peak, then it is aligned unusually late (KS). A solution is to posit a H* on the first syllable of "conquered", with the delayed peak in the second syllable, and a H- phrase accent on the second syllable. If the peak is followed by a high phrase accent, it is hard to tell where in the second syllable the peak really is (early vs. late). The peak can be early in the syllable and the pitch stays level for a H- accent (LM).

Sequence: "...Yoruba homeland." Some thought that this was a sequence of 2 L+H* pitch accents. *Three of us decided it was a hat pattern (beginning with L+H* instead of H*) and that the pitch fall in the second syllable of "Yoruba" is due to the exaggerated pitch range of this vocal style (so the drop often seen in the classic hat pattern between two H* pitch accents is more exaggerated here).

The hat (originated by Hans 't Hart for Dutch, but quite common in English) starts with L+H*, then has one or more H*'s before the L-L%. It usually marks a whole single multi-word semantic unit, which is what I think "Yoruba homeland" is. (KS)

*The alt tier: is placed right below the breaks tier. This tier, along with a set of conventions for use, is currently under construction. The development of this tier grew out of the 5th ToBI workshop for Spontaneous Speech (Simmons College, Boston, MA, August, 2005). If you have questions about the alt tier or the conventions for using it, feel free to e-mail me at maynell@ling.ohio-state.edu.