[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: LSC and the VMS
Meaningful vs. meaningless texts... I think we are missing the point.
Meaningful texts can only be written with valid words allowing for a
few words that one may not know the meaning PLUS some kind of
So monkey character texts can never be meaningful (at least for
relatively low orders) because most of the words are not valid
(although statistically indistinguishable) and there is no underlying
On the other hand, word-monkey texts are meaningless texts to
which one can assign some "meaning" which arises from 2 facts.
All words are valid and the n-order word transition probabilities are
in fact re-constructing in statistical terms the building blocks of the
grammar structure that can be derived from the original texts.
Of course these are not "meaningful" in the sense of being the
result of a thinking process, they are just (let's say) statistically
So where does this leave us? I think that the construction of
concordance tables is the first attempt to find some structure larger
than the words level (of course this is very much related to n-order
word monkey probability tables). Jorge did this table, but I am not
sure about how to carry that further on.
Perhaps we can attack the problem from 2 points of view.
1. Look at it statistically at the character level, LSC, long range
correlations, Hurst exponent, etc. Very nice, but we always seems
to be getting the same result: the vms is not random, but we're not
sure whether there is a language. More, the problem we face is that
we do not have a clue about how language-like gibberish could
2. Try to find the building blocks of the underlying grammar.
I think that this is crucial to make sure that there is a language to be
And we also have "meaningful texts" that are not logical, or just
wrong meaning, but that is another story.