Just a question: aren't those p-values extremely low? 8*10E-33 is practically zero already, but 2.8*10E-119 strains my imagination as a meaningful number. But I'm not a statistician, are those p-values normal?
About the name 'topics', the objection to its use (as already noticed by davidma, and surely many others before me) is that one cannot be sure if what ultimately are the big differences in vocabulary and word frequencies among the different sections of the VMS (even within the same Currier language) are due to the change of topic in a meaningful text or other kinds of data (possibly/probably coded, surely written in a non-standard script), or to a change in the coding tables of a meaningful text/other kinds of data (see the Naibbe example by davidma above), or to a change in the 'generating rules' of a meaningless text.
That said, the observation that the VMS pages can be clustered by scribe on the basis of their text is surely very interesting.
About the name 'topics', the objection to its use (as already noticed by davidma, and surely many others before me) is that one cannot be sure if what ultimately are the big differences in vocabulary and word frequencies among the different sections of the VMS (even within the same Currier language) are due to the change of topic in a meaningful text or other kinds of data (possibly/probably coded, surely written in a non-standard script), or to a change in the coding tables of a meaningful text/other kinds of data (see the Naibbe example by davidma above), or to a change in the 'generating rules' of a meaningless text.
That said, the observation that the VMS pages can be clustered by scribe on the basis of their text is surely very interesting.