Format of the transcription list file:

<filename>,<ascii_transcription>

where

<filename> - name of file: <utt_num><mode><utt_type><spkr_id>. See
		vahadb.doc for explanation.

<ascii_transcription> - transcription of the utterance. See transcription 
			conventions document.

The ascii_transcription field has been prepared with initial and final
space characters.  This was done to simplify word searches: all
occurrences of a given word can be found by searching for a pattern
that includes a space before and after the word.  Searching for " la "
will yield all cases of the female definite singular article,
including utterance initial, medial, final, and spoken in isolation
(whereas a search for "la" would find these plus all words that
contain these two letters).

To look specifically for utterance-initial occurrences, include an
initial comma in the search pattern (", la "); to find utterance final
cases, include an end-of-line designation after the second space (in
sed or perl: " la $"); utterance medial cases are those that do not
match these two patterns.