116v

[-JKP-]

...
Thank you Vladimir, indeed you have already thought of sailors. A remark by Marco on the mixture of Greek and Latin in the same paragraph of a medieval manuscript makes me think that the failure of purely statistical methods can be explained precisely by a mixture of languages.

...

Mixing Greek and Latin would skew certain kinds of computational attacks but would have less impact on others. That's why we come at the data from different directions.

The ways in which the VMS differs from natural languages is not in terms of which language it might be, it's in terms of how it differs from ALL natural languages (particularly in its positional characteristics) regardless of whether they are blended or not.

In other words, the statistical differences are not explained by polyglot theories.