The most accurate characterization of Voynichese, in my view, is quasi-linguistic. It is not, in fact, linguistic, but it appears to be so.
This must be intentional. The text is not linguistic, but it is presented as if it is.
This is what requires an explanation.
There is every appearance that the "text" is indeed a text and consists of letters, words, lines and paragraphs.
By appearance, it invites us to read it. It certainly looks as if we should be able to read it and pronounce it.
Moreover, the "text" has distinct linguistic features. Many statistical aspects of the text - including Zipf's Law - are (just) within the range of natural languages.
There are very good reasons to suspect that the Voynich text is a rendering of some natural language plaintext.
Indeed, there are justifiable reasons to adopt it as a working assumption from the outset.
The rendering of the natural language plaintext might have been accomplished in any number of ways:
*encryption
*steganography
*a system of abbreviations
*phonetics
*a vowel-less text
*a syllabry
*a lost language rendered to strange glyphs
*an obscure dialect rendered to strange glyphs
*a foreign language rendered to strange glyphs
And so on.
Many fine researchers have, and still are, pursuing these possibilities - all with the working assumption of an underlying natural language plaintext.
The quest is to work out what has been done to the plaintext, and undo it.
* * *
But there are those who, delving deeper, or just of a different disposition, are more impressed by the abundance of non-linguistic features of the text.
There are any number of ways in which the text does not seem linguistic.
Entropy, for a start. Glyph positions and combinations are extraordinarily constrained and predictable.
There is every appearance of combinatorics: strings of combinations of the same and similar glyphs generated in some automatic or mechanical way.
Many observable patterns and much of the behavior of the glyphs defies any obvious, or even creative, linguistic explanations.
Nor - as was well-established long ago - do they have the obvious fingerprints of known and expected methods of encryption.
It is possible, however, to give an account - by demonstration - of how such a text could have been generated from some mechanical, systematic method such as volvelles or a card-slot system (or even marked dice.)
There have been effective demonstrations by Timm, Rugg and others, showing just how non-linguistic and artificial the Voynich text is.
In part, they have shown that systematic generation can produce linguistic-like features as epiphenomena at the higher levels.
Rebutalls of such studies have shown that the modelling is too systematic, and that the actual Voynich text is far more organic than the generated models.
All the same, scepticism that the text has a meaningful plaintext of any kind is entirely warranted.
To take the text as an artifact, some sort of generated text, and as non-linguistic in nature (despite appearances), is also a justifiable working assumption to adopt from the outset.
We can look at the text and be immediately impressed by its mechanical nature, rather than the appearances of a linguistic text, and take that as our starting point.
It is an essential divide in Voynich research. Is there a natural language plaintext to be recovered or not?
* * *
If not, then what explanation can we offer?
We then have a copiously illustrated manuscript with an 'artificial' text.
Timm, and others, leap to the conclusion that the work must therefore be a hoax. The text is bogus.
The work has been prepared to look like a meaningful text but in fact it is nonsense, with the motive being pecuniary.
It is a fake text posing as a mysterious and valuable alchemical book, made to be passed off at great price to the gullible.
There are many objections to this proposal, not least the care and system evident in the cosmological sections of the work, and eccentricities that do not play to an early Renaissance book market in any conceivable way.
And the text, the language, itself shows not just system, but deliberate, apparently meaningful, system, far beyond anything necessary to pass off a hoax.
In any case, hoax scenarios should, properly speaking, and as a point of method, be a last resort.
The alternative is a generated, but not empty, text.
That is another dividing line in the research.
The text is artificial, not natural. Is it empty?
If it is empty, then the work is probably a hoax.
If it is not empty, then the text must be something else altogether.
Arguably, this could mean there is a plaintext, but the plaintext is not in a natural language.
For example, the Voynich text might be musical notation. That is a non-linguistic possiblity pursued by some researchers. In that case, there is a plaintext but the plaintext is a piece of music.
Or the text might be numbers. Perhaps some sort of indexing system? There is a plaintext, but the plaintext is numbers, not words and sentences in a natural language.
That is, we may have the very nature of the text wrong. There are other types of text beyond those made up of words and sentences.
There are meaningful (non-empty) types of non-linguistic text.
We ought to consider these before we turn to hoax scenarios in despair, and against a wealth of evidence saying that the Voynich ms. is not trivial.
* * *
My own conclusion (after several years of directed study now) is that the text is non-linguistic, or more accurately quasi-linguistic.
I seek to explain its quasi-linguistic nature.
It is not enough to demonstrate that it is non-linguistic; its demonstrable linguistic features - including its obvious appearance - requires a full explanation.
But I do not think the text is empty. The whole text is replete with meaningful design.
But designed to do what? Designed to record and present what information in what way?
My working assumption, though, is that there is information to be recovered.
The text is artificially generated in some manner - illustrations in the manuscript itself suggest a system of volvelles - but it is not empty: there is content.
The illustrations, certainly, are meaningful, and the accompanying text is likely to be something more than carefully designed gibberish.
But it is not, I can only conclude, some esoteric rendering of a natural language text, nor an elaborate encryption of the same.
* * *
My own investigations - and surveying the 75+ years of statistical and other studies - leads me to the conclusion that the model for the text, the basis for its design, is the YEAR.
Specifically, the YEAR as understood in zodiacal and astrological terms.
The language itself - the glyphs and their combinations - are an expression of the same astrological system set out in the illustrations of the work.
In this view, the language, Voynichese, was specifically created for this manuscript. The text and the illustrations and diagrams are intrinsically linked and are part of a single scheme.
By my account, the glyphs have astronomical or astrological or cosmological or calendrical significations - at least in the first instance, and perhaps in the last.
The foundation for the text, unsurprisingly, is the system of so-called gallows glyphs which, I propose, mark the quarters (and half quarters) of the year, solstices and equinoxes.
The primal text, I propose, is a continuous stream (or cycle) of [o] glyphs, the omicron, then punctuated by the gallows.
The text unfolds by a process of division in exactly the same way - in parallel with - in imitation of - the way the ecliptic is divided to create seasons, zodiacal signs, decanates and other divisions.
The text is thus essentially solar and cosmogonic.
It appears mechanistic, generated - the product of wheels within wheels - because that is the nature of the Ptolemaic and astrological cosmos.
It appears repetitive - yet no exact phrases reoccur - because the Ptolemaic and astrological cosmos moves in repeating but ever-new cycles.
It is generated from a system of volvelles: the volvelles of the heavens.
The glyph set - which is an eclectic assembly of letters, numbers and typographical marks - has been assembled specifically for this purpose: to facilitate the rendering of astrological cycles into a written record, a written text.
I hold out hope, therefore, that if we can reconstruct the design of the language in this light the content recorded with it can be recovered.
But it will not be a meaningful linguistic text along the lines that many expect, such as "Apply this herb to a septic wound under the full moon..." nor anything of the sort.
Nor will it be a long-lost text of Avicenna or anything of that sort.
In my understanding of it, Voynichese glyphs, words, lines and paragraphs all reference astrological cycles.
From the illustrations, we can discern that this ready-built "language" has been applied (albeit in two "dialects") to the growth patterns of plants, the seasons, zodiacal decads (cast as the nymphs of Helios), a landscape and geography (map), a catalogue of fixed stars and solar and lunar cycles, including the eclipse cycle.
But the text is not linguistic. (Nor empty.)
It perhaps presents celestial and terrestrial co-ordinates: certainly overlapping cycles.
It remains to work out exactly of what it consists and exactly how it has been deployed, and why.
I have recently firmed in the view that the direct inspiration for this creation was the copy of Ptolemy's Canones, now Vaticanus graecus 1291, featuring the cryptic miniature of Helios.
Our author has extracted a comprehensive solar (ecliptical) symbolism from that source and applied it to botany (plants belong to Helios) and other related fields and, remarkably, has designed a system of notation, a system of glyphs (numbers and letters etc.) as an expression of that symbolism.
* * *
We cannot, for all of that, escape the conclusion that the author has intended his non-linguistic "system" of co-ordinates, or whatever, to look like a natural and readable text.
That is the problem. If it is not a linguistic text, why does it look like one? It looks for all money like one! Some linguistic-like properties might be epiphenomena, but the presentation is not an accident.
Our author has created a flowing, open, continuous writing system, evidently with scribal ease in mind. It is easy to write and easy to read (if we but could.)
If it is astrological co-ordinates, or whatever - why not set them out in tables? Ptolemy did.
The answer I provide is that the text has been deliberately cast as the "language of the nymphs."
The information contained in the text - of an astrological nature - is presented as knowledge communicated from the celestial nymphs (of the zodiac) and the terrestrial nymphs (of mountain hydrology etc.)
It is simply a case that the divisions of the zodiac (year) are personfied as nymphs (as per Ptolemy's Helios minature) and so the data is transformed into speech as an extension of that personification. Nymphs speak (or sing).
The underlying, implied claim of the work is that the author knows the language of the nymphs and so knows secrets of the nymphs - that is, the cycles of the heavens and the seasons.
That is what makes the text quasi-linguistic. That is why it is presented in that way.
I doubt there is a serious claim being made that the author knows the "language of the nymphs". More likely, the entire device is literary.
Yet, deeper still is the (premodern) idea of cosmos as text. That, in my reading, finally, is the plaintext here.
R.B.
