Date: Sun, 27 Jun 2004 14:40:51 -0400
From: maynell@ling.ohio-state.edu
Subject: Friday's questions

Hi Mary,

The following are some questions from Friday that I felt I didn't have 
complete answers to, so I promised the class that I would forward them to you 
so that you could address them further on Monday.

1)  Anouschka asked: what does a L* H-L% contour look like?

    I find these hard to produce, and didn't have a ready example.  My 
intuition says that the f0 contour would rise from the L* but due to the low 
boundary tone would not rise quite as high as for a L* H-H%.

2)  For the utterance <<atlanta>> in Practice 2 INTERMEDIATE the transcription 
for the last few words is

     Philadelphia     to    Atlanta     ?
           |     ] ]           |   ] ]
           H*    H-H%          L*  H-H%

     I do agree with this transcription, though some felt that the L* 
on "Atlanta" should really be H*, because the f0 contour on "Atlanta" is 
similar to the contour for "Philadelphia".   I didn't have a real good 
argument for why it should be L*, other than the fact that it is lower than 
the H* on "Philadelphia", and so the slope for the H- is a little steeper than 
for the H- on "Philadelphia".

3)  In the utterance <<good-aft>>, the transcription for the last two words is

     Information    Services.
           |                 ] ]
           H*                L-H%

Some felt that there should be a L* on "Services".  But then we compared it to 
the utterance from Prac 1 <<eileen-leaving>>, specifically the second 
production of it:

     Eileen is leaving.
         |             ] ]
        L+H*           L-H%

in which "leaving" sounds as prominent as "services", but it does not have a 
L* pitch accent.  I think these are both ambiguous cases, and that perhaps the 
fact that the L- does not span a longer region makes it sound more prominent.  
Also my gut feeling is that the H% provides a contrast to the L- target so it 
stands out more, but it still isn't a pitch accent.  A L- followed by a L% 
does not have this contrast.

4)  Lisa asked, "When do you trust the f0 contour and when don't you?"

    For example, sometimes an intonation phrase ends in a H% but the f0 
contour drops slightly at the end.  This tempts people to transcribe it as a 
L%.  My answer to the general question was that it is necessary to be aware 
that obstruents can cause perturbations in the f0 contour (as can creaky 
voice, etc.), and this is also true of utterance final-obstruents.  I also 
shared my strategy of trusting your ears more than your eyes, especially when 
there is a conflict between what you see and what you hear).   However, I do 
realize that just because this works for me, it won't necessarily work for 
everyone (different transcribers having slightly different strategies, and 
there is a real element of having to account for the f0 contour when you 
transcribe - you just have to know what is junk and what isn't).

5)  There was one utterance for which we changed the transcription, but I 
failed to write it down.  Can any of you in the class remember what it is and 
let me know, in case we revise the canonical transcriptions in the future?

I think that is everything.

Laurie