Jan 17
======

Bell/etal:2003
--------------

Why is the observation of longer durations with less reduction of
vowels and greater use of basic vowels by women and by older speakers
expected from the usual correlations of speaker status and levels of
formality in speech?

Why do the strong effects of predictability on both shortening and on
vowel reduction suggest the existence of separate sources for the 2
effects?

The authors argue that the interlabeler agreement rates of the ICSI
transcribers, which range between 72.4% and 76.9%, are quite
acceptable for this task. Is there a way to determine an objectively
acceptable degree of agreement for a specific task?

Would it be worthwhile to incorporate a temporal component into this
analysis? For example, examining whether the realization of words
changes over time during the discussion. As I understand it, the
participants are not familiar with each other at the start. If that's
the case, could increasing familiarity lead to fewer full-form words
being used>Are 6 minutes enough for familiarity to even be explored?
Could we argue that familiarity is at least partially accounted by
controlling for speech rate (if familiarity results in fewer full
forms, it might manifest as an increase in speech rate)?

The study focuses exclusively on high-frequency function words. How
might the patterns observed differ for lower-frequency function words
or content words?

The authors use conversational data from the Switchboard corpus, which
may not fully capture natural speech due to the constraints of the
telephone conversation setting. How might findings differ in more
spontaneous, face-to-face conversational contexts?

I was surprised that utterance-initial lengthening effects are only
evident for "and" despite the comparably high incidence of
utterance-initial "I" and the authors' earlier insistence that words
would have stronger, less lenited forms at the beginning of
utterances.  Why is the result for utterance-initial "and" the only
function word to follow this pattern and what potential explanation
could there be?

Also, I was hoping to discuss repetitions more because my intuition is
that the repeated word would be reduced more than the first instance,
but the authors actually argue for a lengthening effect of repetitions
unless I am misunderstanding.