[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: VMs: RE: Another method different from Cardano Grilles



On Sat, 29 Jan 2005, mjmurphy wrote:
> Lets assume the existance of some kind of grill-type machinary, but also
> assume that the author employed it unsystematically in creating the hoax
> text; for example, discarding the created strings that they didn't "like"
> for whatever reason (like playing cards with yourself: you can "cheat" with
> impunity).  Perhaps producing text this way would give an answer as to why
> some of the words you might expect statisticially do not appear.

Good point in any event.  If we assume random generation we can still
allow the random generator to discard forms he doesn't like the look of.
We can more or less take it for granted.

However, we can't really say what words we expect statistically to appear
except by referring to the VMs to see what words do appear and how often.
Nobody, including Rugg, has offered a particular device, e.g., a Rugg
Grill (set of matrixes and windows or equivalent) that they claim to be
equivalent to a set used to produce the VMs or any part of it.  So there
is nothing to fix by cheating.

We do have a somewhat analogous problem.  We do have grammars like Jorge
Stolfi's that generate sets of forms including most of the attested words.
They also generate forms that don't appear, or, at least, aren't attested
in some corpus of attested forms.  This is normal for generative grammars
that get detailed enough to generate actual sequences as opposed to
generate templates like prefix-midfix-suffix or subject-verb-object.  We
can assume that such grammars overgenerate (as they say) because the
particular sequence of random generations used didn't happen to turn up
"rteesol" (or "John generate spam"), though they might eventually do so,
or we an assume that some sort of esthetic filter was applied to the
result.

Note that all VMs word grammars proposed so far also fail to generate some
attested forms because they aren't capable of doing so without further
refinement.

______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list