Re: AW: Voynichese = Old/unknown/extinct kind of Chinese dialect

[Brian Eric Farnell <bfarnell@xxxxxxx>]

Jorge Stolfi wrote:

>     > Another possibility is that they are grammatical markers for
>     > person and case. Along that line of thinking, if a list were
>     > compiled of all words in a certain class like the gallows w/o
>     > tables words, then removing all of the gallows characters from
>     > the words should shorten the length of the unique word list by a
>     > disproportionate amount. Potentially it might even shorten it by
>     > a factor equal to the number of characters removed.
> 
> In fact more. Consider removing all "t"'s and "d"'s rom English words.
> That not only merges "tip" with "dip", but also merges "at", "tad",
> "dad", "add", etc. with "a".
> 
> --stolfi
Yes, but t's and d's are extremely common, change that to the
five vowels in English and you could get some fantastic results.
However, I think if you compared the frequency of what you
dropped to the resulting changes, a set of letters that were a
class unto themselves (tone markers or grammatical markers)
would give you unexpected results.  In completely toned Mandarin
romanization dropping the 4 tones tones would change the unique
word list by a factor of four.  Dropping person markers would
drop it by a factor equal to the number of unique person
endings.  I think with the possible exception of vowels, nothing
else would do that (drop n letters, unique word list is divided
by n).  Perhaps this could occur by accident, but it would be
expected if you had a unique class of characters that looked
like letters but didn't function that way.
Regards,
Brian