The Voynich Ninja

(11-11-2022, 05:22 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.Are you using "ivtt" to make these choices? If not, you may find that it is in fact trivially easy to change these using that tool.

IVTT is a wonderful tool, but I've just been making these adjustments in the same Python code I use for the rest of the analysis, partly because it seems slightly easier that way to change the settings and get a result all in one step.

(11-11-2022, 01:43 PM)pfeaster Wrote: You are not allowed to view links. Register or Login to view.

(11-11-2022, 05:22 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.Are you using "ivtt" to make these choices? If not, you may find that it is in fact trivially easy to change these using that tool.

IVTT is a wonderful tool, but I've just been making these adjustments in the same Python code I use for the rest of the analysis, partly because it seems slightly easier that way to change the settings and get a result all in one step.

OK, just wanted to make sure you are aware it exists.
As soon as I find myself typing the same sequence of commands a couple of times, I create a script, and complicated things become 'one step'. (Also saves a lot of time from typos )

Here I present a study of some of the patterns of LINES by employing a simple method of parsing.

It follows directly on from my recent investigations into a Vord Paradigm. This is an extension of the same direction of research, but now applied to the LINE as a unit or container of text.

The method is:

Vords in the line are examined, in sequence, according to whether their initial glyph is a consonant or a vowel.

We mark a separator (break) where the sequence changes so as to group common initial-glyph vords together.

For example:

<f105r.P2.28;H>

oees.olkeedy.qockhy.raiin.chol.okair.oteedy.qopchedy.odaiin.ypchedy.ykam-

Some vords start with a vowel, some with a consonant. We mark a division in the sequence where this changes.

This line is divided thus:

oees.olkeedy.[BREAK] qockhy.raiin.chol. [BREAK] okair.oteedy. [BREAK] qopchedy. [BREAK] odaiin.ypchedy.ykam-

The pattern of initial glyphs in this line is therefore:

VV - CCC - VV - C - VVV

We depict the line thus divided up with breaks as:

oees.olkeedy.
qockhy.raiin.chol.
okair.oteedy.
qopchedy.
odaiin.ypchedy.ykam-

One of the virtues of the method is that the divisions made to a line of text are easy to see and (generally) incontestable in themelves.


The assumption of this approach is that the vowel/consonant distinction is an important organizing principle at many levels. In the absence of any obvious syntax, how else might we consider the assembly of vords in lines?

To be honest, I can’t find a better way to divide lines up into parts.

This mode of parsing is at least not unnatural and is suggested by the habits of the text.

It is a common observation that similar vords can run in clusters in lines. This method draws out such clusters on the basis of the status of initial glyph. We find clusters (runs) or patterns based on initial vowel or consonant.

Consider this cluster from pg 58r:

oteeoaly.otar.alkar.or.aldar=

This is not a cluster of [o-] vords, it is, more exactly, a cluster of vowel-initial vords. We often see this. Most often among vowels we see clusters of [o] initial vords, but often other vowels will be included in the sequence such that the basis for the cluster seems to be the category ‘vowel-initial vords.’

Clusters of consonant-initial vords are very common.

I am not making claims about the nature and function of lines, except to say that they are discrete collections of vords. We may or may not understand them as ‘sentences’.

This method is one way of looking at these collections and making them more manageable. Much is made of LAFU: line as “functional unit.” Perhaps. But in the first instance these collections of vords have patterns. The purpose of the method is to assist the study of line patterns.


The patterns to be observed when a line is parsed in this way fall between two extremes.

Firstly, there are a great many lines with entirely consonant-initial vords.

Examples:

<f66v.P.4;H>

shokeshy.daiin.cheos.shky.dol.shckhy.dal.kodal.chekal.dal-

<f82r.P2.19;H>

posalshy.qokedy.cphal.shedy.sheol.keeedy.lkaiin.shedy.qoly-

There are cases of lines entirely of vowel-initial vords, but they are far less common. The first is on pg. 19v:

<f19v.P.12;H>

yees.ykchol.oty.ytor.ytar.ytchor.ytaiin=

We might call these lines, FULL CLUSTERS, the whole line being a collection (cluster) of either consonant-initial or vowel-initial vords.

At the opposite extreme to FULL CLUSTER lines are lines where one vord after another alternates between vowel-initial and consonant-initial vords, CVCVCV. FULLY ALTERNATING lines are relatively rare. Here is one from page 16r:

<f16r.P2.6;H>

oshaiin.
dyky.
oeees.
deeeod.
aiin.
dtoaiin-

We see that this alternation is based on the consonant/vowel combination [od]. It goes: o.d.o.d.a.d. The more common alternation is: q.o.q.o.qo.

More commonly, lines approximate the strict CVCVCV alternation. This line, for example, only deviates slightly:

<f70r2.I.1;H>

otchsy.ol.
shy.
oteo.
sal.
ol.
keey.
oteol.
choky.
otchoshy.

Another example:

<f104r.P.16;H>

ocheoithey.
qoctheody.
ykeeodey.
qoepchy.
opchey.
qoty.shtey.
yteedy.
shody-

Lines can be assessed in these terms, namely how near they are to strict alternating sequence.

Again: the opposite of the alternating sequence is the CLUSTER. Lines can be described as either alternating or clustering, flowing or clumping. Most lines are a mixture of both tendencies.

Consider this line, for instance:

<f103v.P.25;H>

sain.
olkeeey.
qokan.
oteedy.
qotain.
otal.oty.opar.aram.oteeam-

It is an alternating CVCVCV line with the initial-glyph pattern s.o.q.o.q.o but with a vowel-initial cluster at the end.

Here is the opposite, a consonant-initial cluster followed by an alternating CVCV sequence:

<f107r.P.4;H>

teeody.chedain.qoteey.qokar.deeo.ltedy.
otar.ain.
chady.
otokcho.
qoked.
okchedy-

Or this example:

<f44r.P.10;H>

qokchor.
okchy.
qoto.
ykol.
choky.choky.chol.dam-

Here, the alternating flow of CVCV is broken at the end of the line with a cluster of consonant-initial vords.

In many cases, though, vords seem lacking in alternation and are rather groups of clusters – the clustering tendency dominates the line. This line for example:

<f105r.P2.28;H>

oees.olkeedy.
qockhy.raiin.chol.
okair.oteedy.
qopchedy.
odaiin.ypchedy.ykam-


I propose that strict alternation CVCVCV is, in fact, the paradigmatic formation, with the two initial glyphs of the Vord Paradigm QOKEEDY alternating, thus:

q-
o-
q-
o-
q-
o-

Indeed, we can posit a LINE PARADIGM that is a direct extension of our VORD PARADIGM.


The implications of this will be considered elsewhere. Such a model opens many possibilities as a tool.


To continue the exposition though:

A minimum deviation from the static FULLY CLUSTERED line is either the PREFIXED or the SUFFIXED line. Sometimes, a deviating vord appears before or after an otherwise full cluster.

Here is a PREFIXED line on page 18:

<f18r.P.7;H>

oshor.
shaiin.cthy.sholdy.doldy.doldaiin-

And another on page 47:

<f47v.P.11;H>

oteey.
cho.chdy.chy.key.chyky-dchy.daiin.chy-

Here is a SUFFIXED line on p. 24v:

<f24v.P.3;H>

octhol.odchees.oesearees.okam.
chcth-


As well as vords prefixing or suffixing clusters, there is also the INTRUSION. This is where a deviating vord intrudes into what is otherwise a fully clustered line.

Here is a case:

<f42r.P3.20;H>

shol.chol.shoky.
okol.
sho.chol.chol.chal-

And another:

<f22r.P.3;H>

qokol.dykaiin.
okchy.
daiin.cthol.ctholo.dar.shain=

And two more from p. 102r2.

<f102r2.P.8;H>

shockhy.qockhey.sol.
eeey.
dol.cheol.doaiin.qkeeey.cthey-

<f102v2.P1.1;H>

podeesho.
oteeos.
sheos.qokeeo.chckhhy.shkeey.chekeod-



As well as this there is also what I will call BRACKETING. This is where there is both a prefixing and a suffixing vord and they serve to BRACKET a cluster. It is different from an INTRUSION.

There are plenty of examples, usually with two vowel-initial vords bracketing a consonant-initial cluster:

<f3r.P.12;H>

okadaiin.
qokchor.qoschodam.
octhy-

<f3v.P.10;H>

ochor.
daiin.qokshol.daiim.chol.
okary-

<f5r.P.2;H>

ochey.okey.
qokaiin.sho.ckhoy.cthey.chey.
oka*or.otol-

<f21v.P.5;H>

osho.
deey.ctho.l.sho.cthy.daiin.dait.
oky-

<f28r.P.5;H>

oksho.otor.
chy.kchoror.chodaiin-sho.cthody.
okoy-

Here is a case where the vowel-initial cluster is bracketed:

<f40v.P.15;H>

qokchey.qody.
or.aiiin.okaiin.o.
ckhy-sheod.faimy-


We could also report cases of INTERNAL BRACKETING:

<f108r.P.6;H>

ol.
cheol.qo.qokeey.qokeey.qokeedy.sheoky.
otedy.
qotey.qokchey.chdar.
aiin.y-


Sundry other interesting things emerge. Consider the SYMMETRY of this line, for example:

<f112v.P.2;H>

oar.osal.okeeshy.
qokeey.
okain.
qokal.
okeol.oty.oraiin-

This parsing of lines reveals a certain order of symmetries.

Look at how this cluster of [sh] vords is bracketed:

<f44v.P.4;H>

yokalod.
shaiin.shor.shorody.shky.sho.
ytchy.opchod.opy-

And in some cases the clusters get progressively smaller as the line proceeds:

<f108v.P.20;H>

polkeedal.sheo.kchey.lotedaiin.
otedy.opchedaiin.otshedy.
qotey.raiin.
ol-



This quite straightforward method allows us to take any line of Voynich text and describe some of its properties. We can classify them and describe them with a set of consistent criteria. For a start, they are alternating or clustering. Or prefixed, or suffixed, or bracketed, and so on.

We might describe the following line as ‘Alternating with a final V-cluster.’

<f103v.P.25;H>

sain.
olkeeey.
qokan.
oteedy.
qotain.
otal.oty.opar.aram.oteeam-

Or this as ‘Alternating with an initial C-Cluster’:

<f107r.P.4;H>

teeody.chedain.qoteey.qokar.deeo.ltedy.
otar.ain.
chady.
otokcho.
qoked.
okchedy-

This is a case of an alternating sequence being bracketed by two C-Clusters:

<f111r.P.29;H>

salkeedy.lkeedy.chckhy.rky.chey.
okeeo.
chr.
al.ar.
qokeedy.qokedy.qoteedy.qod-


Just as consonant/vowel alternation (CVCVCV) is a structural foundation of the default vord QOKEEDY, so it is structural to lines. We expose this by the simple device of following where vords alternate or cluster in lines according to initial glyph.

Thanks Hermes777, for this thread and a new way of thinking about word patterns.

I had a quick look at the distribution of cluster lengths.  Two histograms are superimposed below:  In blue is the distribution of consonant-initial word-cluster lengths (paragraph text from the IT transcription, respecting line breaks and illustration breaks, but not uncertain spaces).  Yellow represents the same text sample, but with word order randomized within each line.  In the grey areas they overlap:

[attachment=6960]

A case could be made that fewer short clusters occur in the native manuscript.  But its overall distribution is very close indeed to statistical expectations, which might count against a functional interpretation of the cluster patterns.

(12-11-2022, 07:29 PM)Hermes777 Wrote: You are not allowed to view links. Register or Login to view.Here I present a study of some of the patterns of LINES by employing a simple method of parsing.

I appreciate the organizational categories you're proposing: the cluster, the alternation, the intrusion, bracketing, symmetry, and such.  It's frustratingly difficult to take certain kinds of pattern I think many of us intuitively sense unfolding within lines (whether they're really "there" or just eye-catching coincidences) and reduce them to models that can be talked about and tested and potentially verified as statistically significant.  But it seems well worth trying, and I'm glad you're doing it.

(12-11-2022, 07:29 PM)Hermes777 Wrote: You are not allowed to view links. Register or Login to view.The method is:
Vords in the line are examined, in sequence, according to whether their initial glyph is a consonant or a vowel.
We mark a separator (break) where the sequence changes so as to group common initial-glyph vords together.

I notice you like to refer to Voynichese "vords," which I assume you mean as an acknowledgment that glyph sequences separated by spaces might not function as or correspond to "words," even if they look superficially like them.  But is it worth being equally circumspect about "consonants" and "vowels"?  Or should I say "qonsonants" and "voyels" (or maybe "owels," which more appropriately starts with the class of glyphs it would refer to)?

Maybe your [q-o_k-ee_d-y] vord paradigm really does reflect a literal CVCVCV structure.  But it seems to me that the value of the vord paradigm as a tool doesn't hinge on the correctness of that hypothesis.  As far as the vord paradigm itself goes, it doesn't really matter whether the prefix [q-o], stem [k-ee], and suffix [d-y] each represent separate syllables, or a single syllable CVC collectively, or a number hundreds-tens-units, or whatever else.  Its main analytical claim would seem to be that the [q-o], [k-ee], and [d-y] elements each have an internal structure in which [o], [a], [e], and [y] occupy a somewhat consistent position at or near the end.

With your extension of the vord paradigm to a line paradigm, though, you lock into an interpretation that considers the [q], [k], and [d] parts to be interchangeable Cs, and the [o], [ee], and [y] parts to be interchangeable Vs.  Maybe that will turn out to be correct -- who knows?  But I wonder whether there would be a way to keep your line paradigm as flexible with regard to interpretation as your vord paradigm is.

Do your line-structure observations require drawing no distinction at all among the glyphs you identify as "vowels" and "consonants," or would many of them stand even with narrower groupings of related glyphs -- or maybe with glyphs that occupy specific columns in your paradigm tool?

(12-11-2022, 07:29 PM)Hermes777 Wrote: You are not allowed to view links. Register or Login to view.oteeoaly.otar.alkar.or.aldar=
This is not a cluster of [o-] vords, it is, more exactly, a cluster of vowel-initial vords.

It could also be interpreted as a cluster of vords beginning with the two specific glyphs [o] and [a], which seem to share certain other similarities.  The first three vords also begin not just with "vowels," but specifically with selections from your V1 column (the last two seem as though your tool could parse them in more than one way).

(12-11-2022, 07:29 PM)Hermes777 Wrote: You are not allowed to view links. Register or Login to view.yees.ykchol.oty.ytor.ytar.ytchor.ytaiin=
We might call these lines, FULL CLUSTERS, the whole line being a collection (cluster) of either consonant-initial or vowel-initial vords. 

But alternatively, if we were to draw a distinction between [y] and [o], this would instead look suspiciously like what you call an intrusion, akin to:

(12-11-2022, 07:29 PM)Hermes777 Wrote: You are not allowed to view links. Register or Login to view.shol.chol.shoky.
okol.
sho.chol.chol.chal-

Similarly, [otar.alkar.or.aldar] in the earlier example could be interpreted as alternation rather than clustering if we choose to distinguish [o] from [a].

This leads me to wonder how one might test/confirm/falsify hypotheses such as:

(1) [y*] and [o*] fall into the same category and can form a cluster together
(2) [y*] and [o*] fall into different categories and can form intrusions or alternations relative to one another
(3) [sh*]/[ch*] and [o*] fall into the same category and can form a cluster together
(4) [sh*]/[ch*] and [o*] fall into different categories and can form intrusions or alternations relative to one another

If we don't assume a priori what the glyph categories are, would there be any way to infer them reliably from the line patterns?  Or are line patterns only detectable and differentiable on the basis of categories already inferred through vord-level analysis?

You tentatively frame these patterns as line patterns tied to LAAFU.  But it would be interesting to explore whether clusters seem in practice to respect line boundaries or not.  I'm thinking of "incontestable" clusters (i.e., ones that don't require any assumptions about glyph categories) such as:

(f37v): daiin.dy / dshor.dytory.dshor.daiin /dchor

(f111r): qokeedy.qokedy.qoteedy.qod / qocKhedy.qokeechy.qotchey

Another thought: would similar patterns (clusters, brackets, alternation, etc.) emerge if you grouped words by their ending glyphs rather than their beginning glyphs?  I believe nearly all vords end in what you would treat as "consonants" in column CF, which may have nudged you away from this approach.  But insofar as you take an interest in the contrastive vord-break combinations [n.o] and [y.q], it might be useful to distinguish [n] from [y], at least -- and if we do, they sometimes seem independently to display what looks like impressive "clustering" behavior:

(f76v): 
salkchdy.chey.lcheedy.lchedy.qoteedy / 
sol,shey.qotedy.chey.dytey.teedy.lchey.qokedy.chedy.lal,chedy.lchedy /
dchedy.qokeedy.qotchy

(f25r): 
chan.chaiin.qotchain / 
qotcheaiin.dchain.cthain.daiin.daiin.cthain.qotaiin

Unless I'm mistaken, these same sequences would also count as clusters based on your classification of their initial glyphs as "consonants."

Anyhow, this is all nicely thought-provoking, and I hope I can look forward to reading about the implications you see in this "extension" of your vord paradigm into a line paradigm.

(13-11-2022, 03:30 PM)pfeaster Wrote: You are not allowed to view links. Register or Login to view.If we don't assume a priori what the glyph categories are, would there be any way to infer them reliably from the line patterns?

This is something that Emma and I investigated in our You are not allowed to view links. Register or Login to view.. See in particular p.9 and 10. While we focussed on a specific phenomenon (how the last glyph of words tends to correlate with the first glyph of the following word), it is possible that the issue could be analysed in a more general way, including start-start correlations and end-end correlations. Torsten pointed out You are not allowed to view links. Register or Login to view. that shows an interesting set of tables about this.

An entirely different "Voynich grammar" line of research I tried is You are not allowed to view links. Register or Login to view.: basically, an unsupervised classification algorithm that maximizes the possibility of inferring the class of the next word in a sentence on the basis of the class of the current word. This method does not depend on glyph categories at all. My research did not lead to much progress: it mostly confirmed the fact that Voynichese words that behave similarly tend to be morphologically close (much more so than in English, for instance).

(12-11-2022, 07:29 PM)Hermes777 Wrote: You are not allowed to view links. Register or Login to view.Here I present a study of some of the patterns of LINES by employing a simple method of parsing. . . . 

Just as consonant/vowel alternation (CVCVCV) is a structural foundation of the default vord QOKEEDY, so it is structural to lines. We expose this by the simple device of following where vords alternate or cluster in lines according to initial glyph.

Hi, Hermes:

I join the others in thanking you for sharing these thoughts.  It is very refreshing to have an alternative way to think about the ways to parse the glyphs.  

Although I may be just repeating, in a more simplistic manner, an observation that Patrick makes in his posting,  I think it is worth emphasizing the concern with the general classification of each glyph as having an independent and, most importantly, consistent "identity" in your parsing set ups -- even something as simple as "consonant" and "vowel."  For example, a big problem in my mind is that I don't think each "y" is the same "thing" and any parsing that assumes this is going to have real problems, in my opinion.

Given the highly unusual low conditional character entropy of the text, this seems to necessitate some sort of approach as to how the glyphs are interpreted (or parsed) that helps to remedy the entropy gap between this unusual characteristic and other language.  Of course, I am making a myriad of assumptions here with that statement, but I am concerned that any usefulness in your parsing scheme could be totally subsumed if the incredible amount of glyph-glyph (or glyph-glyph-glyph) association that is needed to produce "valid" Voynichese is ignored.

I'm sure you've read Koen's attempts to push this unusual value upward (if not, You are not allowed to view links. Register or Login to view. is his final posting on the topic).  Whatever you think about ciphers or not -- any parsing set up will look enormously more possible in my mind if it takes this huge issue into account.

Of course, you are perfectly able to state you think that keeping the "units" of the glyphs at a single glyph level (whatever that might mean to you) is the "right" thing to do but I do think that taking this fundamental entropy issue into account may glean more applicable results.

Please do not see this as a criticism -- but perhaps as a suggestion for moving your beginning model into something that is more likely to give you language applicable patterns?  Thank you for considering this possibility and I'll think about it as well.

Michelle

I find the approach very interesting, but the main problem I see is that it is based on whether the Eva representation of the character is a vowel or a consonant. Given that the Eva representation is essentially arbitrary, this is likely to end up with a result that does not differ much from 'chance'. I believe that this is what @obelus has demonstrated.

It might be worth doing the same thing, but classifying the initial characters by one of the attributes of its shape.

TLDR: Voynchese word ordering is dominated by line effects and repetition. A model based on the alternation of two classes 0,1,0,1,0,1... cannot be accurate, because lines are typically made of an "anomalous" line-initial word, a sequence of words with repetitions and line-positional preferences and a differently anomalous line-final word. In my opinion, a model is likely to require at least 4 classes; lines could be something like S,C,C,N,C,C,C,N,E (S=start, C=occurring consecutively, N=non-repeating, E-end) or maybe S,L,L,L,R,L,R,R,R,E (L=left-preference, R=right-preference).

---

After the ideas put forward in this thread, I ran a series of experiments along the lines discussed in You are not allowed to view links. Register or Login to view.. I used Alexander Clark's implementation of the Ney-Essen algorithm, which classifies words on the basis of their behaviour only, ignoring morphology.

I used Quire 20 from the Zanbergen-Landini transliteration, where benches and benched-gallows were replaced with upper-case characters and 'z' correspond to unreadable or rare characters. Uncertain spaces were ignored. As a first experiment, I only considered the first character of each word and configured the algorithm to find two classes. In this graph, the thickness of the arrows corresponds to the frequency of transitions (only the most frequent words for each class appear in this and the following graphs).

[attachment=6965]

Class:1 includes the line separator (nlnl) and the  characters that typically appear at the beginning of a line. This class is typically followed by a word belonging to the other class. Class:0 includes all the most frequent word initials and this class tends to loop on itself, meaning that words belonging to this class tend to appear consecutively.

So a typical line could be something like:

nlnl 1 0 0 0 1 0 0 0 nlnl

This is the parsing of
<f103r.10,+P0>    daiin.ol.oain.okeol.chol.okam.chety.shedy.otaiin.shedy.teolshy.oteedy.sor,ain
1:nlnl nlnl
1:d daiin
0:o ol
0:o oain
0:o okeol
0:C Col
0:o okam
0:C Cety
0:S Sedy
0:o otaiin
0:S Sedy
1:t teolSy
0:o oteedy
1: s sorain
1:nlnl nlnl

So this result is dominated by the line-initial effects that make the first character of line initial words unusual in other positions.

---

I then tried a similar experiment considering both the first and last character for each word.

[attachment=6967]

These results are more difficult to interpret. Class:0 only includes the following 11 words. It could be that they share the property of only rarely repeating identically, but I am not sure:

nlnl 1086
Cy 850
Sy 445
an 220
ar 167
Sl 85
Co 42
So 24
Sd 23
Cd 23
lo 19

---

Things get more interesting if the number of classes is increased. With 4 classes, the line-break takes a whole class (C:0); C:3 contains line-initial words; C:2 contains the most frequent words that also tend to occur repeatedly; C:3 contains words that do not repeat and all -m words (which typically are line-final).

[attachment=6968]

---

Four classes only based on the first character of words cannot clearly identify line-final words (which have peculiar endings). But sh- and q- words tend to avoid the right-side of lines (see You are not allowed to view links. Register or Login to view.) and here they get their own class (C:1) which is not likely to directly lead to a line-break. C:3 groups line-initial characters and C:0 is the bulk of ordinary characters.

[attachment=6966]

This is a further attempt to present lines of Voynich text in meaningful or at least suggestive ways. A slight detour from the matter of Vord Paradigms, but part of the same quest. 

I proposed some time ago that the Voynich text, although presented as running prose, might in fact consist of lists of vords in the manner of an inventory of some kind. Others have suggested the same.

Do we find patterns of organisation in the text that might be consistent with this?

A possible model might be the following:

Let us make a basic distinction between vords that contain a GALLOWS glyph, and those that do not. That is, the four elevated, conspicuous glyphs, including in their ‘benched’ forms.

This is surely a justifiable distinction. The gallows glyphs are important to the text – whatever that importance might be – and it must surely be of some importance whether a vord contains such a glyph or not.

We can assume that vords without such a glyph are in some way secondary.

We will leave aside the question of [ch] and [sh]. They are not gallows glyphs as such, and so are not counted. Only the four elevated glyphs.

A better distinction might be those with and without a CORE, by the Stolfi model, but any gallows glyph will be a core anyway, and in the first instance their status is clear to the eye and I want to keep the method simple for demonstration purposes.


Now, with any line of Voynich text, make a hard [BREAK] at every vord containing a gallows glyph.

Take this line from pg 107, for instance:

<f107v.P.39;H>

okain.cheor.olkaiin.oain.cheary.raiin.okaiin.odaiin.okaiin.y-

We apply breaks wherever we encounter a vord with a gallows glyph:

okain.cheor.[BREAK] olkaiin.oain.cheary.raiin.[BREAK] okaiin.odaiin.[BREAK] okaiin.y-

We depict it thus:

okain.cheor.
olkaiin.oain.cheary.raiin.
okaiin.odaiin.
okaiin.y-

And we can present it as a list thus:

1. okain.cheor.
2. olkaiin.oain.cheary.raiin.
3. okaiin.odaiin.
4. okaiin.y-

The suggestion here is that each vord with a gallows is an ITEM and vords without a gallows glyph are subsiduary or additional to the item to which they are appended.

The text is remarkably amenable to such a reconstruction. Only rarely do we get unweildly clusters of non-glyph vords. Usually we generate a list of gallows vords appended by one or two non-gallows vords in each case, if at all.

In this model we would assume non-gallows vords - [daiin] is one of them - to be in some way FUNCTIONAL.

Here are some lines presented as lists in this manner, selected at random:

<f29v.P.7;H>

1. qotcho.
2. kchor.daiin.
3. ykaiin.dy-shdy-cho.
4. cthy.
5. sheky-

<f85r1.P.24;H>

1. tedair.
2. ykeedy.dain.cheedy.
3. qokor.chedy.
4. okar.
5. sheekchy.
6. qokedy.olchedy-

<f108r.P.14;H>

1. polchal.shol.
2. qokor.shedy.
3. pcholy.
4. qokal.
5. opchdy.
6. ofal.shor.
7. qokaiin.
8. otalod-

<f16v.P.10;H>

1. qokchy.chydy.
2. ykchy.
3. chckhy.
4. otain.
5. cthor.
6. cthy-

<f111r.P.44;H>

1. pchedar.
2. oteeol.
3. lkeedain.
4. okeam.sheo.
5. teolkedain.shal.
6. keey.
7. lteeal.cheal.
8. oty-

<f13v.P.2;H>

1. oko.
2. qokol.chodal.
3. otchol.
4. cphol.
5. choty-

It is certainly not hard to reconfigure line 14 on page 85r2 as a list:

<f85r2.P.14;H>

1. qtchedy.qodar.
2. qotedar.
3. qokar.
4. qotchd.
5. qotom-

It is arguable that [qodar] might be an ITEM here as well, but we will keep to a strict and simple rule: only gallows glyphs.


The text looks like running prose, but it is much more list-like than it appears. This is a simple method of demonstrating how it might be so. What the vords signify and what the gallows letters are remains an open question. What information is encoded in the ITEMS and how is it modified by FUNCTION vords?

It is enough at the moment to show that our apparently prose text might be a set of lists. The assumption the text is prose might be wrong and might lead to many misguided studies.

Recently, for instance, I encountered a study of word pairs by Mark Fincher: Word Pair Permutation Analysis of natural language samples and it’s value for characterizing the ‘Voynich Manuscript’. It is a study of structures between words and Fincher concludes from it:

‘Voynichese’ is not a natural language in it’s own right. If the VMs text is derived from a plaintext in a natural language, it must have undergone some disruption of word order.

But in fact what the study shows is that - in its word order - the text does not behave like running prose.  The "disruption of word order" suspected by Fincher might simply be that the plaintext is a set of lists.