The Voynich Manuscript Part 8 : Some Notes On Writing Systems

Since I wrote Part 7 of this occasional series of articles about the Voynich manuscript I have been  heavily engaged in writing programs to analyse the Voynich text.  This will take quite a bit more time, as it requires the creation - and debugging1 - of multiple programs to test each linguistic theory component from various perspectives.  For example, I propose the theory that the Voynich symbols - I prefer to call them glyphs - are non-obvious representations of sounds and do not necessarily comprise a simple alphabet.

The problem of reading the VM script is greatly complicated by the fact that the set of VM glyphs is a set of arbitrarily chosen symbols used to represent sounds or ideas.  It is not just the VM glyph set, but every set of writing symbols that is arbitrary.  There is no universal rule by virtue of which a symbol is required to correspond with a given sound or idea in any language whatsoever.  The symbology in all writing systems may have originated in a desire to have a logical correspondence between a symbol and some physical entity, but that correspondence has long since been lost in developed writing systems.  Every alphabet, syllabary, abjad, abugida, consonantary, futhark; every phonographic or logographic writing system uses arbitrary symbols.

The fact of an arbitrary choice on the part of the inventor of the VM glyphs means that no symbol need actually represent the same sound or idea as any similar symbol in another written language.  Any visual resemblance should be treated with caution; the resemblance is probably coincidental.  For example, a closed double-loop 'a' might be written as '8', and the single-loop symbol which looks like an 'a' in such case might represent 'd'.  Something that resembles iiiv may represent mu, niv, inus, ivus.  There are countless other possibilities.  There are so many symbol permutations that simple mathematical analysis suffers from the problem of geometric increases in program running time.

I refer to the Voynich symbols as glyphs so as to make clear the fact that, in my opinion, any Voynich symbol may represent any phoneme or phonemes. Any VM glyph may thus be represented by one or more  Latin style letters.  It is only by in-depth analysis of the patterns of word and glyph usage in the entire VM that we have any hope of discovering an unambiguous pattern of language.   I am conducting such an analysis at present and am making no assumptions about the underlying language.

A Note Of Caution

It is tempting to assign computer codes to the Voynich glyphs and then see patterns in the transcript.  Almost any ASCII transcription system will produce an appearance of systematic pattern in the transcript.  This is an emergent property of the transcription method and it provides no worthwhile insight into the actual underlying language of the VM.   The EVA set of ASCII representations of Voynich glyphs is useful for generating a computer-readable transcript.  However, the appearance of 'readability' in that transcript is probably deceptive.  Any 'vowel-consonant' sequences in the transcript are almost certainly an artifact of transcription, since the EVA codes were chosen specifically to make the transcript 'human readable'.
The EVA alphabet was designed in the framework of a more recent transcription effort. It is analytical, like Frogguy, but it uses only lower-case alphabetical characters. These have also been chosen in such a way that the transcribed text is almost pronouncible.
Source: Voynich.nu

I will return to the topic of the glyphs when I have some results to report.

Continued in The Voynich Manuscript Part 9 : An Amateur's Work?


[1] Just as the author is his or her own worst proofreader, so the programmer is his or her own worst debugger.  I work entirely alone as author and programmer, so I get a double dose of this.  The odd typo in a blog doesn't matter, but a single typo in a program can so distort the output as to prevent replication of results by another researcher.  I would very much like to avoid that problem, hence the desire not to rush my analysis of the VM text.