Relation between word types and vocabulary density.

[Koen G]

Or the other way around: TTR is types/tokens.

Not sure what came through first, my hasty correction or your reply  

About Q13: its broad TTR is still within bounds for normal language behavior. Here is another graph I tried to improve a bit with Claude. Note that language dots are medians and they hide a lot of variability and outliers. This is merely an indication, a way to provide a sense of scale.

   


Of course, it is when we also look at other things, like small-window MATTR behavior, strings of similar words and character entropy, that it becomes clear how extreme Q13 is.