[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: VMs: Has anyone been down this route before?
On Tue, 31 Aug 2004, Koontz John E wrote:
> I guess I can google on stroke harmony + Voynich.
Looks like the main references are from Jorge Stolfi:
Exposition: http://www.voynich.net/Arch/2001/01/msg00112.html
Correction: http://www.voynich.net/Arch/2001/01/msg00113.html
====================================================================
Quoting the main section extensively:
Stolfi:
What I found out recently is that there seems to be some logic in the
left-right asymmetry after all. As you all know, most Voynichese symbols
are written with one or two two pen strokes, which we can label "left" and
"right" in writing order. In particular, the dealers and circles are the
result of combining either an <e> or <i> stroke on the left, with one of
eight strokes (plumes, flourishes, ligature, or empty) on the right.
Indeed, all but one of the 16 possible combinations are valid characters:
0 1 2 3 4 5 6 7 8
+----------------------------
i | i ii * l r j m n C
|
e | e a o y s d g b I
The four gallows are similarly composed of two left halves and two right
halves, in all combinations. (Benches and platform gallows are
combinations of three or more strokes; let's leave them aside for the
moment.)
Now, the funny thing is that the [prev-symbol and next-symbol]
distributions seem to be determined in large part by the symbol's left
stroke and right stroke, respectively. In other words, symbols on the same
row of the table tend to be prev-similar, while symbols on the same column
tend to be next-similar.
Thus, for example, <a> and <y>, which have an <e> stroke on the left, are
fairly prev-similar but very next-different; while <a> and <i>, which are
have an <i> stroke on the right, are very next-similar but quite
prev-different.
In fact, this rule seems to explain most of the inconsistencies and
tantalizing hints that used to drive me crazy, such as the relationship
between the dealers { <d> <l> <r> <s> }, the finals { <n> <m> <g> }, and
the circles { <a> <o> <y> }. Now it seems almost logical, for example,
that <l> and <y> can be prefixed to gallows, but <r> and <s> can't; while,
on the other hand, <d> and <s> may follow <e>, while <l> and <r> may not.
Also, it makes sense that all the four gallows <k> <t> <p> <f> are
prev-similar, and yet <k> <t> are quite next-different from <p> <f> (the
latter are never followed by <e>).
To be sure, the above "stroke harmony" rule is not entirely consistent.
For instance, <n> and <m>, which both have an <i> stroke on the left, are
fairly prev-different --- <n> is almost always preceded by <i>, while <m>
(like <r> and <l>) are more often preceded by <a>. But there is evidence
anyway that <in> is a single letter, in which case the inconsistency may
well not exist. Also, <d> and <s> are not as prev-similar to <a> as the
theory would predict.
End quotation from Stolfi. One correction in [] made as suggested in
Stolfi's postscript to the original.
====================================================================
Note that it seems that stroke harmony is intended to apply only to a
glyph immediately following a given glyph, and that some glyphs reverse
the harmonic domain. This is a bit different from vowel harmony, where
typically vowels fall into two paired sets, and a leading or otherwise
dominant vowel forces vowels in a word which are not from the same set as
the dominant vowel to switch to a harmonizing vowel - their opposite
number in the same set as the dominant vowel. Under that approach words
would have all e-like strokes or all i-like strokes.
______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list