The Voynich Ninja

This my table has been prepared for the draft of our submission to the European HCC of this year, but since the submission was declined, the stuff is unlikely to be ever released, especially given that it did not have enough time to progress from the draft version.

The table, obviously, is also a raw draft and is posted into the forum for the sake of discussion, reference, and possible further development. It builds upon Cham's CLS and also incorporates ideas from You are not allowed to view links. Register or Login to view..

Here's how the table is to be interpreted. Each cell shows a combination of two "basic" shapes: the first one taken from the respective row on the left, the second one taken from the respective column on the top. The intersection of a basic shape with itself stands not for a "combination", but rather shows whether this basic shape does occur standalone in the VMS or not. If a combination or a basic shape does not occur in the VMS, the respective intersection cell is grayed out. The order matters. For example, e plus dash yields c, but dash plus e yields h. Some combinations are yet indefinite (as suggested by more than one character in a single cell).

[attachment=1417]

I mostly agree with this chart but I'm wondering why EVA-y isn't in there (or, at least, the straight-legged version of EVA-y which is a c-shape with a long leg).


One of the things that could potentially be interpreted differently from the way you've classified the shape construction is the tails on EVA-r and EVA-s. If you give the tail a more flexible positioning (top of letter or bottom of letter both being acceptable), then one can think of EVA-r as "i" with a tail and EVA-n/EVA-m at the ends of words, where it looks like a V or W, as a tail at the bottom of the letter. It means the idea doesn't fit quite as neatly into a grid, but it could account, in a very simple and consistent way, for the EVA-n/EVA-m shapes.

Oh, I just noticed another one that I had classified differently. Hang on, I'll try to hunt up my version of this (I have a chart too)... back in 10 minutes, I have to forward some files to a client, then I'll and get a screensnap of my version so we can compare notes

So much for 10 minutes (I need to remind myself that nothing related to computer software ever takes 10 minutes), but now that I'm done with that task...


This is something I had planned to post in a blog, but I simply don't have time to write up everything, so here's the organization scheme I put together a while ago.

It's organized a bit differently from yours, Anton, but there are also some commonalities, and maybe the differences can provide some food for discussion:

Anton! Thanks for the visual table.

I quickly made additions and changes that (in my opinion) need to be made. Later, I will make a complete table, including modifiers T1, T2, T3, T4. (T4 - horizontal tail in the gallows P and F.
[attachment=1418]

@JKP

My table is not complete, partly because I don't have the complete picture, partly because the rejection of our submission effectively interrupted the work, and I turned to other (non-Voynich) matters. So I just put it here "as is", not without the idea to push myself to work further upon it.

I forgot to mention that the table does not deal with "aggregate" characters - those which may be constructed by combining three or more base shapes. Like benched gallows and others.

Also, the order of the base shapes in the table is arbitrary. I don't know what the actual order is, if any has been intended at all.

You are right that some glyphs are not included, that's because there are some problems with y, as well as with r/n/s/b which I'm in doubt how to deal with.

Regarding r/n/s/b.  First of all, I am quite far from certain that the character £ is the right component to get r from i (or s from e etc.). In the text, the tails in r, s, n etc. are more "sweepy" and asymmetric than this one. This might be attributed to the certain degree of freedom given to the pen after the alphabet was constructed and approved, but anyway, visually the actual tails in the text and the symbol £ just don't look similar. The only supposed composite in which it fits well is o (that is, e plus £). Maybe also Î.

Next, you raise the important question about the difference between, say, r and n (or s and b). The most straightforward explanation is indeed that the difference is in the point to which the tail is attached. But, as you correctly note, this does not fit the grid. Maybe, while constructing the alphabet, the guy somehow "ran out" of the opportunities presented by the grid, so he needed to introduce an additional principle.


Another explanation is that the actual difference is meant to be in the shapes of the tails, and not in the attachment point. Only that this difference is blurred due to the "degree of freedom of the quill" that I suggested above. Hence the question is if this difference can be traced somewhere (maybe in some more "elaborate" parts of the VMS where the scribe was not in a hurry).

With the s, there is the additional complication that the EVA "s" actually looks quite differently in different places, even when it is standalone.

Regarding y. I'm not certain what is the shape that forms its tail. I had the idea that it might be the shape that is top column 5 in your table, especially given that there are such its composites with the dash as EVA &131; and &204; But I could not find it standalone in the Extended EVA. However, your table suggests that such standalone occurrence(s) exist(s)! (Can you please advise where?)

Now, at least for many (though not all) instances of y, the tail looks more "curved" that the supposed base shape. However, once more this could be attributed to the "freedom" of the quill...

One more interesting point related to this base shape (should I call it "poker" for brevity?) is whether it is actually the vertical or the poker that is used in the formation of q. Sometimes the line is definitely straight in q, and sometimes it looks like this poker. Two different versions of q even co-exist in one and the same line (e.g. f80v, line 8). So are they really meant to be different or it is just "freedom" of the pen? I guess the latter, because in some instances of q the line is curved into the other side - which makes it impossible to distinguish between the vertical and the poker components.

JKP, Wladimir and others - I suggest that when discussing rare symbols we provide references to the folio/line where those symbols can be seen.

E.g. JKP's table suggests that there are both vertical and poker standalone. I'd like to look at the context...

(09-06-2017, 12:17 PM)Anton Wrote: You are not allowed to view links. Register or Login to view.@JKP

My table is not complete, partly because I don't have the complete picture, partly because the rejection of our submission effectively interrupted the work, and I turned to other (non-Voynich) matters. So I just put it here "as is", not without the idea to push myself to work further upon it.

I forgot to mention that the table does not deal with "aggregate" characters - those which may be constructed by combining three or more base shapes. Like benched gallows and others.

Also, the order of the base shapes in the table is arbitrary. I don't know what the actual order is, if any has been intended at all.

I'm glad you did, it's thought-provoking and interesting to see different approaches to this aspect of the manuscript, even if they are only in the development stage.


The order of my base shapes is also fairly arbitrary. I was torn between trying to look for a logical "system" and recording what I saw as "components" based on how they were penned. The two approaches yield slightly different results.

Also, I didn't know what to do with the "o". It's not necessarily composed of other characters, it might simply be a unique base shape. I composed it of two cees facing each other (which is cheating, actually), for the sake of discussion, but it might be composed of a c with a lowered tail or stands alone (is not used to construct other shapes). On the 4 x 17 table the "o" stands ahead of the characters that resemble alpha, beta, gamma, delta, which is not the normal sequence of the Greek alphabet. In Greek, omicron and omega are not at the beginning, they follow the others. Sometimes I wonder if it's meant to represent "zero" rather than the letter "o".

Quote:You are right that some glyphs are not included, that's because there are some problems with y, as well as with r/n/s/b which I'm in doubt how to deal with.

Yes, I agree. I wasn't certain what to do with EVA-y either. Even though it's written like a "c" with an extra stroke (they are even sometimes separated), the shape itself could be interpreted as an "o" with a tail" or as a c with a lowered tail (as opposed to a separate descender). There are a couple of places in the manuscript (e.g., to the right on one of the pool pages) where the tail, the swoop shape, stands alone. I cannot remember if the longer more slowly curved tail of EVA-y stands alone, but I'm pretty sure I've seen the straight rod, the long one, stand alone.

Quote:Regarding r/n/s/b.  First of all, I am quite far from certain that the character £ is the right component to get r from i (or s from e etc.). In the text, the tails in r, s, n etc. are more "sweepy" and asymmetric than this one. This might be attributed to the certain degree of freedom given to the pen after the alphabet was constructed and approved, but anyway, visually the actual tails in the text and the symbol £ just don't look similar. The only supposed composite in which it fits well is o (that is, e plus £). Maybe also Î.

I'm not certain either whether a tail on the "i" that is either high or low is what the creator intended, but since it works and is simple, I thought I would start with that idea and see where it leads.

Also, it's far from clear whether there are two classes of tail (or three classes of tail) as there are many variations in the shape and it's hard to know which variations might be meaningful. In Latin, a tail with a slight hook (which also occurs in the VMS) is a different abbreviation from a curved tail and there's at least one example of a perfect Latin hooked tail in the VMS, a shape that doesn't usually occur accidentally, so the scribe was clearly familiar with the distinction in Latin... but whether it was written that way in the VMS out of habit or because it's intended to be interpreted as a different shape, I don't know. Also, a tail with a loop, in Latin, is the "-is" abbreviation (in Greek, it stands for the "o" abbreviation) but again, what it means in Voynichese, is not clear.

Quote:Next, you raise the important question about the difference between, say, r and n (or s and b). The most straightforward explanation is indeed that the difference is in the point to which the tail is attached. But, as you correctly note, this does not fit the grid. Maybe, while constructing the alphabet, the guy somehow "ran out" of the opportunities presented by the grid, so he needed to introduce an additional principle.


Another explanation is that the actual difference is meant to be in the shapes of the tails, and not in the attachment point. Only that this difference is blurred due to the "degree of freedom of the quill" that I suggested above. Hence the question is if this difference can be traced somewhere (maybe in some more "elaborate" parts of the VMS where the scribe was not in a hurry).

Very good points and yes, I agree that the shapes of the tails may be significant.

Quote:With the s, there is the additional complication that the EVA "s" actually looks quite differently in different places, even when it is standalone.

Yes, and I haven't studied EVA-s in any detail yet, but I have noticed that some appear to be a "long-c" shape rather than a short-c and whether a long-c is a shape on its own or a combination shape of short-c with a crossbar, I don't know.

Quote:Regarding y. I'm not certain what is the shape that forms its tail. I had the idea that it might be the shape that is top column 5 in your table, especially given that there are such its composites with the dash as EVA &131; and &204; But I could not find it standalone in the Extended EVA. However, your table suggests that such standalone occurrence(s) exist(s)! (Can you please advise where?)

When I first saw the VMS, I thought there were three forms of EVA-y—one with a long tail, one with a short tail, and one with a straight tail.

Now, I'm not so sure. I think the length of the tail may not matter, but I still feel that a curved tail and a straight tail MIGHT be intended to be different because there are a number of places in the VMS where it's almost impossible to distinguish a straight-tail EVA-y from a rounded EVA-q. Whether the tail on EVA-g is a component shape, I don't know. I charted it that way because the approach in the chart was to try to discern if there was a simple rationale to the components, but I'm not sure the EVA-y  tail is the same as other tails or an entity on its own.

Quote:One more interesting point related to this base shape (should I call it "poker" for brevity?) is whether it is actually the vertical or the poker that is used in the formation of q. Sometimes the line is definitely straight in q, and sometimes it looks like this poker. Two different versions of q even co-exist in one and the same line (e.g. f80v, line 8). So are they really meant to be different or it is just "freedom" of the pen? I guess the latter, because in some instances of q the line is curved into the other side - which makes it impossible to distinguish between the vertical and the poker components.

Yes, exactly, the ambiguity is bedeviling!

I wonder what might be the places in the VMS where it is most reasonable to expect the ambiguities to be minimal. Looking e.g. at You are not allowed to view links. Register or Login to view. (the "title" folio which one may reasonably expect to be put down with special care), I think that the tails of r and n look pretty the same (check e.g. the first line of the second paragraph). The same tail, I think, applies to the e to make s of it. While the apostrophe placed above ch (turning it into EVA sh) is of a different shape.

Anton,

the idea to look at pages that were probably done with special care is an interesting one.

Contrary to expectation, You are not allowed to view links. Register or Login to view. is not such a page. The very first word is already hard to read, and the earliest transcribers (Friedmans group and Currier) made something completely different out of it. I mentioned this point a long time ago, and the suggestion was made that the scribe was still 'getting used to' the script.

If this is true, it has important consequences on what may have happened before this scribe dipped his pen and wrote
fachys or fyays (etc). 


There are a few pages in the MS where the text seems to have been written with extra care:
very straight lines and straight right margin. You are not allowed to view links. Register or Login to view. and You are not allowed to view links. Register or Login to view. are examples.

It might be worth looking at these in more detail.

I agree with René that You are not allowed to view links. Register or Login to view. may not be a good example.

To me it has a tentative "not used to writing the script yet" feel to it. The caps on the bench chars are particularly variable, the connections between the two elements of the bench char are variable and sometimes not always connected. FWIW, I You are not allowed to view links. Register or Login to view. about this in 2013.


It may be a good example of how one writes an unfamiliar alphabet before one gets comfortable with it, but I wouldn't consider it a good example of "fluent" use of the script.