This may have been discussed in some form or other before I found my way to this forum, but I came across this item in today's local paper and thought it might be of interest:
"You've got about 600,000 English words to work with. But only 43 of them make up half of everything you say. Only nine go into a quarter of what you say. These nine are:
and, be, have, it, of, the, to, will, you. So contends that language expert Robert Chapman."
I have no information regarding how he arrived at these figures, but it seems to me that a statistical study of "what people say" would be a very difficult undertaking and would require a number of clearly stated parameters.