Vocabulary Norms - Continued

[Peyton Todd]

Brian Richards has pointed out to me that even type/type ratios will be a function of sample size. I should have thought of that: we use the same verbs over and over again for an ever-shifting array of nouns in each different context we encounter. 

Nonetheless, even for type/token ratios, which are more obviously a function of sample size, that function must have a shape. With all the studies of type/token ratios, has anyone tracked that function so that, hopefully, one could extrapolate across different size samples? Or is it too much a question of what is being talked about, etc.?

My interest stems from the fact that, impressionistically, I believe my subject's vocabulary was rather low compared to that of other children, and may have been more heavily weighted toward nouns than verbs. (Yes, I do know about the 'noun bias' in toddlers; my subject is older.) I can of course scale down CHILDES samples, or combine several, as needed, to have samples of similar size to my own, but I'm still hoping to avoid that extra work. For a valid comparison, I'd have to do it for lots and lots of samples.

Peyton Todd