Relation between word types and vocabulary density.

[npcompl33t]

This doesn't mean the proposed order is wrong. It means that MATTR and the seriation agreeing is not evidence of a chronology: a text written the same way everywhere, with sessions that are more or less repetitive, gives the same picture.

I would agree the MATTR lining up isn’t  evidence for the chronology, I don’t think that was koens point. I think his point was more that the mattr declining as sharply as it does over the course of the bio pages, in the same order as two differnt chronological methods, is extremely odd, especially with how large the swing is. The mattr should not be related to section, scribes, etc, and it shouldn’t swing this much over a few pages. The fact that the ordering from 2 methods puts it in sequential decline is a substantial investigative lead. 

I also think you are overstating the implications of your model. It was trained on the very data you are having it predict. The same text written everywhere in the manuscript absolutely does not manifest this way, it is only in bio, so if you are able to replicate that, your model must have trained… on bio. If the state of the words is the result of a defenerate copying process, then the copying process is what generates the word forms. They are not independent.

Ive been quiet on this for a few days, but Koens post here has made me look into a few things, I think I have some substantial findings and hopefully will be able to share them soon. i want to make sure everything is airtight with them.