www.alphadictionary.com

Here's a piece on how men and women use fillers in their speech. Interesting, though I wish it had gone much further in its analysis.

http://www.theatlantic.com/health/archi ... um/375729/

My wife and I have discussed my frequent pauses. I have always tended for my brain to run off in another direction while speaking, distracted by something I said or something that caught my attention. I have to pay attention while speaking, because even there I am diagramming my sentences, choosing my words, redefining them if I think the audience may not understand.i used to practice sermons out loud, pacing the living room on sat night, so they would flow smoothly on Sunday morning. I think this is why speakers and lecturers of all stripes chase rabbits rather than catching themselves to stay on topic. Sometimes the rabbit trail is more fun!

Go ask Alice.

I'm dubious about the validity of the observations. If you click through to a sample of the transcriptions that Liberman used to obtain his data, you will see the transcriptions are not phonetic but normal English with numerous typographical errors. We can't determine how "um" and "uh" were actually pronounced.

Some clear errors I found in the transcripts:
"Your in New York" instead of "You're"
"wouldn't get a long" instead of "get along"
"ye do people watch reality t.v." What is "ye"? How is it pronounced?
"she's sweat though" when the person probably said "she's sweet . . . ."
"yes a-s-s whole" (typed with hyphens to avoid the Alpha Dictionary PC monitor) instead of "hole"
"people just tiered of it" instead of "tired" (I suppose)
"rush and roulette" (referring to a t.v. program. I'm not familiar with u.s. t.v. shows but google suggests there is/was a program called "Russian Roulette") This transcription error occurs twice.

It seems to me to be quite a jump to make the generalizations Mr. Liberman made on the basis of the data he appears to have used.