Chelsea, from what I can see, it is POSSIBLE that this utterance is best parsed into smaller units. Certainly the last phrase fulfills some definitions as its own utterance, structurally (it's a "sentence" in its own right), but utterances usually have more complex determinants. Criteria vary, but we tend to use "2 out of 3" set of criteria to potentially separate long turns like this into shorter utterances: is there a final contour to some phrases in the middle of this run, are there silences or respiratory intakes, and does something stand as a grammatical "sentence" or phrase. Only 2 need apply for us to segment something like this into smaller units. These sorts of determinations are important for other reasons, since so many kid language measures are proportioned over utterances, such as MLU, IPSYN, DSS, etc. If some transcribers are making "maximal" decisions about where utterances end, and others are more conservative, you will wind up with some very variable estimates of language, not just fluency.
It would be impossible to know if any part of this qualified to be divided out without the audio. If you wanted to share a clip with me privately, I might be able to provide some guidance.
I can definitely see a situation where, if a speaker were sufficiently disfluent, and each disfluency counts as a word as well as the actual words, that setting a limit of 50 per utterance COULD run up against some problems. But the current limit is an attempt to prevent people from placing turns, rather than utterances, on speaker tiers, since so many language variables proportion over utterances, not turns.
best,
N
Nan Bernstein Ratner, F-, H-ASHA, F-AAAS, ABCLD
Professor
Hearing and Speech Sciences
University of Maryland
0100 Lefrak Hall
College Park, MD 20742