Relation between word types and vocabulary density.

[Koen G]

Looking at broad-window (100+?) MATTR alone, the shift from high-Q20 to Q13 would be like a shift from Latin to an analytic vernacular, or a very varied text type to a repetitive one. 

This was one of the better graphs I got out of Claude yesterday:

   




Interestingly, all plant sections are more or less in the same zone, no matter if they're A or B. This is remarkable: MATTR is clearly variable in the MS, but it transfers across Currier languages for similar pages?

There is a clear shift in Q20, which is showing up in all seriation efforts as well. I'm not sure what kind of conclusions to draw from these measures coinciding. It seems that an increased reliance on core vocabulary is the best explanation.