The Voynich Ninja

(01-08-2025, 11:30 PM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.Here is a clip of page f1v, spanning the last few words of lines 9 and 10 (the last ones on the page): 
What fantastic things can you imagine seeing on the indicated places?
A)  B) C) D) (this is the hardest)
Honors question: find the s in that image (not near B).

OK, here are my answers to my quiz above.

First, more generally, after looking hard at the images for all these years, I see that the VMS underwent least one major restoration effort, possibly more than one.   A couple of centuries after it was written, its owner at the time saw that many parts had faded almost to the point of invisibility.  So he naturally undertook to restore it before it was too late, by carefully retracing all the faded parts of the text and drawings.  

The Retracer -- either the owner himself, or some hired scribe -- took pains to retrace quite accurately over the original strokes, so that usually the restoration is not noticeable except for the much darker ink.  But he could not avoid leaving plenty of scattered evidence here and there.  

A common kind of evidence is due to speed differences.  Many original strokes -- in particular, the plumes of r, s, n and the tails of y, l, m -- were done by the original Scribe with quick sweeps of the pen while lifting it off the vellum.  That created characteristic "mousetail" strokes: smoothly curved, gradually becoming lighter and narrower, ending in a sharp point.  The Retracer, on the other hand, had to go overt them slowly in order to follow the same path.  As a consequence, plumes and tails that have been completely retraced have become fatter, with more or less uniform minimum width, with a blunt end, often jittery and kinkier.  To make things worse, in many cases the plume or tail was clearly retraced in the "wrong" direction.

The Retracer must have been conscious of this problem, so in many cases he retraced only part of the plume or tail, leaving the faint but sharp original end sticking out ahead of the fatter and blunt retrace.

Another kind of evidence is glyphs that were clearly mangled because the Retracer failed to see some faded details, or misunderstood what he could see.  

And another kind of evidence is places where the Retracer failed to follow the original stroke, so that parts of it remained visible, sticking out from the side of the retraced one.  These miscarriages are fairly rare on the text, but are common on the drawings -- where the Retracer presumably did not think they would matter much. 

Once one admits the possibility that such restoration happened, it is impossible to not see dozens of these clues on almost every page of the VMS. 

But now for that specific image.  Here it is again, with additional arrows:

[attachment=11178]

Items A and C are both cases where the distal half of a plume was retraced, and probably in the wrong direction (down instead of up). Note the rather uniform thickness and blunt ends of the retraced part.  Also the slight jitter. 

By the way, the body of C is probably original.  Since that part of the r was not retraced, the Retracer must have decided that it was still good enough.  One can then imagine the state of the glyphs and strokes that were retraced. 

Item E is the tail of an l that seems to be original, since it has the "mousetail" appearance.  Item F, on the other hand, is a tail of an y that was retraced only for a short distance, leaving the end of the original "mousetail" sticking out beyond the end of the retraced part.

Item B is a nice one.  The original glyph was an ordinary Sh.  However, by the time of the restoration, the right half of the plume had completely vanished, leaving only the left half.  The Retracer did not understand what had happened, and "restored" that glyph by extending the left half of the plume down to meet the first e stroke, even extending the ligature line to the left.  Thus he created a weirdo that does not occur anywhere else: a Ch with a plume that rises from the first C and turns clockwise.

Several people here have stated that they see nothing at D.  With my superior pareidolia, I see less than nothing there.  That is part of a large area on this page where insects apparently sneaked between this page and page f2r, and nibbled extensively at the surface of the vellum.  This "wormscrape" damage seems to be mostly confined to the margins of this page, but it did affect a few glyphs.  I suspect that the chemicals that were applied to page f1r,and seeped through the vellum, made that area more appetizing to the insects.  

It was the damage at D that, in particular, erased the right half of the plume of B.  I believe that the middle part of the ligature may have been erased too, and was (correctly) restored by the Retracer.

Item G is a case of an original o that the Retracer "restored" to look somewhat like an a This was not that bad; most transcribers would probably still read it as o. But in many other places the Retracer almost certainly turned o into a, r into s, Ch into Ih, or vice-versa ...

Item H is the body of an r that I suspect was retraced in a second restoration episode, some unknown time after the first one.  But that is a topic for another time.

And finally, item I is the answer to the honors question.  Originally there was an s there, but the right half of it was completely wormscraped away, leaving only the e stroke of the body and the tip of the plume.  The Retracer did not see the latter, and thus "restored" that glyph as an o.

What impact does this restoration have on our investigations?  Fortunately, the retracing of the text was for the most part accurate, so the errors that it introduced (like B and I above) are not that many.  Any sound analysis already had to take into account that the text probably contains a non-zero percentage or errors, by the Scribe and by the transcribers who prepared the digital file. The restoration only increased that percentage by some modest amount.  

The restoration should affect transcription efforts, since it changes the interpretation of glyphs that are misshapen or ambiguous.  Rather than deciding whether the current shape of a symbol looks more like an r or an s a transcriber should try to discern what the original shape might have been before it was retraced. And many glyphs that have been listed as "weirdos" or "rare" may be in fact just Retracer mistakes, like B above.

The restoration has a bigger impact on the interpretation of figures, because some details were clearly altered by the Retracer(s) in semantically meaningful ways.  For instance, the crown on the nymph of f72r3 and the cross held by the nymph of You are not allowed to view links. Register or Login to view. may have been added by a Retracer (who gifted many of the nymphs with a characteristic "scalloped showercap").

All the best, --jorge

I see no obvious retouch attempts here, as I said before, but suppose for a moment there were retouch attempts. I tried putting a mm grid over the corresponding images, using the known dimensions of the manuscript. I guess the grid could be off by some 10%, but I think for this argument it's not important. How could one retrace strokes with so high a precision?

[attachment=11185]

(09-08-2025, 01:43 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.How could one retrace strokes with so high a precision

The original characters were traced with strokes 0.2 to 0.5 mm wide, with precision ~0.1 mm.  That is in fact the precision with which most anyone will write or draw with any sufficiently fine pen.  Just try it.  So the retracing did not require any more precision than the original scribing did.

As for the "double strokes" in the second image: if you mean the o, that is not two separate traces, but the normal trace of a quill pen that is running out of ink. 

A quill pen (like 19th century steel pens, and modern fountain pens) has a split nib. The two prongs will spread open as the pen touches the paper, then close when the pen is lifted.  This is what makes ink flow in a controlled amount from the "reservoir" up in the hollow of the quill to the tip of the nib.

But when the "reservoir" gets exhausted, the two prongs create two parallel strokes instead of a single broad stroke.

That o was retraced indeed, but the retracing (even though it was split) covered the original completely.

But maybe you meant the second e stroke in the Sh.  Indeed that is an obvious example of retracing.

In fact, it is almost certain that all the dark writing in that second image was retraced.  

All the best, --jorge

(09-08-2025, 02:19 PM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.The original characters were traced with strokes 0.2 to 0.5 mm wide, with precision ~0.1 mm.  That is in fact the precision with which most anyone will write or draw with any sufficiently fine pen.  Just try it.  So the retracing did not require any more precision than the original scribing did.

While this is humanly possible, I'm not sure it was practically feasible. I can barely achieve 0.15 mm precision with a very good tool and a magnifying glass for a one time job performed with extreme care. I've sharpened a pencil to leave roughly 0.15 mm wide trace and tried tracing a line over mm grid. I managed to get within ~0.15 mm precision. This wasn't easy, and trying to trace the line made a jagged line. To me personally it looks nearly impossible that someone would perform this high precision (think money forgery quality) job over hundreds of pages without leaving very obvious traces, like in the right image in my previous post.

[attachment=11187]

One thing I think could happen, is while the ink was still wet on the page, the scribe added a touch of ink with the tip of the quill to some strokes. In this case I think it's possible that the shape of the stroke would be perfectly replicated with the liquids mixing together, while parts of the letters would remain of different darkness. But to me the simplest explanation for the variations is some oily/gooey component in the ink that would capture more dye particles and create small dark blobs, which maybe weren't even visible until the ink dried out.

(09-08-2025, 02:19 PM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.But maybe you meant the second e stroke in the Sh.  Indeed that is an obvious example of retracing.

I think t and ch were likely retraced, could be some failure of the writing utensil in other letters.

Edit: with a quill there is an additional complication, one would have to match not only the position, but the width of the stroke as well. I have some limited experience with the quills (both making and writing), and I have practical understanding of how they work, I think it would have been extremely hard to pull off this kind of retouching/retracing even for a professional scribe.

By the way, I used to be in "there is retouching in the MS" camp, but I changed my mind after this thread: You are not allowed to view links. Register or Login to view.
There are some images of ink variation in other manuscripts there and some examples of retouch jobs.

Hypothetically, the original writer looked at yesterday's dry pages and added a few improvements. Yes, there are ink variations, certainly there can be retouches and corrections, but what is the chronological evidence to show a timespan of months, years or centuries?

(09-08-2025, 07:57 PM)R. Sale Wrote: You are not allowed to view links. Register or Login to view.Hypothetically, the original writer looked at yesterday's dry pages and added a few improvements. Yes, there are ink variations, certainly there can be retouches and corrections, but what is the chronological evidence to show a timespan of months, years or centuries?

I assume if "the retoucher" produced "invalid" glyphs, this would imply that "the correct" stokes faded out completely, which takes quite some time, I think. I don't think these are invalid glyphs, just relatively rare variants.

Sure, but if the ink was light in the first place, the original writer may have though it was too pale to last the coming years / centuries and gave it a second stroke, perhaps the next day or a decade later.

Any "invalid" glyphs are an assumption and a quandary.

From initial experimentation with a home-brewn OCR implementation I can say that the minim height of the text on many pages is less than 4 points, which is 1.4 mm. Minim height is the height of characters that have no ascenders or descenders.

It clearly follows that character details have sub-mm dimensions. 

The following image was created during the forensic analyses in 2009, and confirms these dimensions.

[attachment=11197]

(09-08-2025, 06:07 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.While [retracing with 0.1 mm precision] is humanly possible, I'm not sure it was practically feasible. I can barely achieve 0.15 mm precision with a very good tool and a magnifying glass for a one time job performed with extreme care. [...] with a quill there is an additional complication, one would have to match not only the position, but the width of the stroke as well. I have some limited experience with the quills (both making and writing), and I have practical understanding of how they work, I think it would have been extremely hard to pull off this kind of retouching/retracing even for a professional scribe.

OK, maybe not everybody can do it.  But the original Scribe did that routinely.  He controlled the position and width of his strokes with ~0.1 mm precision, at normal writing speed, without a sweat.

When writing a Ch,  for example, he would write an e and then an h with the tip of the ligature precisely touching the tip of the e.  Ditto when writing s, t, etc. And the original figure outlines are generally ~0.2 mm wide, independently of direction.

In the first example of You are not allowed to view links. Register or Login to view., while the retracer went off-course on some glyphs, he precisely retraced many others.  Had he used a broader pen (instead of a narrower one) and ink of matching color, it would be hard to tell that those glyphs had been retraced.

Quote:One thing I think could happen, is while the ink was still wet on the page, the scribe added a touch of ink with the tip of the quill to some strokes.

That certainly did happen;

Another common cause for variations in stroke weight (darkness and/or width) is the normal ink flow dynamics.  After the pen is recharged by dipping it into the inkwell, the first word or two come out heavier than normal, because there is some ink sticking to the very tip of the nib.  But in that brief interval the weight of the strokes drops to the "normal" level.  From then on, the flow of the ink -- from the little reservoir up into the quill's hollow down to the tip -- is regulated by the opening and closing of the two tines of the nib, as the pen is is pressed down and lifted off.  That steady-state regime continues for a line or two, until the reservoir gets exhausted.  Then the strokes quickly become fainter and fainter.  At that point the Scribe would recharge the pen, and the cycle would repeat.

[attachment=11196]

(However, a really good scribe would recharge well before the strokes started to fade, and he would also scribble a bit on a separate scrap paper after recharging to get into the "normal" regime before resuming the work -- so that the weight would be uniform through the whole document.)

Another cause of weight variation is the Scribe "back-tracing" -- going back and retracing some glyph or stroke that he just wrote, because it came out malformed or too faint. 

And then there may be cases where the Scribe could not figure out a glyph from the draft, so he left that position blank, later asked the Author, and then went back and filled the gap -- with a freshly recharged pen.

And it may also have happened, as you say, that a stroke became darker than normal by accident --- the pen hit some imperfection of the vellum, or the Scribe sneezed, or whatever...

BUT STILL, besides all those "original" weight variations, there are many cases where the only explanation is that the strokes were retraced by someone who knew even less about the Voynichese script and figures than we do now, and did not have access to the draft or the Author.  You have two such examples in that You are not allowed to view links. Register or Login to view. clip (B and I), within a few square cm...

(09-08-2025, 06:30 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.By the way, I used to be in "there is retouching in the MS" camp, but I changed my mind after this thread: You are not allowed to view links. Register or Login to view.  There are some images of ink variation in other manuscripts there and some examples of retouch jobs.

Funny, my reading of that thread is that it rather supports my pareidolia.

In You are not allowed to view links. Register or Login to view. from the VMS, the "oCta" was obviously retraced.  In the original, the last glyph of the word was a y, but the Retracer (like us now) did not see its tail, and thus "restored" it as a.  Also the glyph before the t may have been an e, which the Retracer mistook for a C... 

In tYou are not allowed to view links. Register or Login to view., from another manuscript, those abrupt weight changes may have indeed been quirks of the ink flow... but they may also be a case of the whole text having been retraced, to fix the poor handwriting of the scribe who first did the job. 

For instance, in the p198b fragment, note that the left half of the "9" glyph is narrower and lighter than the right half, and its shape is rather "lame".  Imagine the whole text being written with such strokes.  I think it would make sense that another scribe was charged with retracing the whole thing with a slightly wider pen and more decisive strokes.

But, as in the VMS, the Retracer could not redo the tails and plumes, because the slower retracing would have made them fat, with blunt ends.

Anyway, that example was indeed fully retraced, it probably happened shortly after the original scribing, and using the same ink.  Which is not the case of much retracing in the VMS.

(09-08-2025, 07:57 PM)R. Sale Wrote: You are not allowed to view links. Register or Login to view.Hypothetically, the original [VMS] writer looked at yesterday's dry pages and added a few improvements. Yes, there are ink variations, certainly there can be retouches and corrections, but what is the chronological evidence to show a timespan of months, years or centuries?

The evidence is that the VMS Retracer made many mistakes that can be explained only if parts of the original had faded so much that he failed to see them, and he had no access to the draft or Author.

In that You are not allowed to view links. Register or Login to view. clip that I posted, mistakes B and I happened because insects had completely eaten part of the original glyphs.  You can see the scraping clearly in other parts of that page, and even more of it on page f1r.  The damage was such that the Retracer mistook the remaining half of one plume as being attached to the wrong glyph by the wrong end (the B "weirdo") and failed to see the surviving tip of the other plume (the I case).  It is unlikely that the damage happened while the original Scribe and Author were still around.

(09-08-2025, 09:05 PM)R. Sale Wrote: You are not allowed to view links. Register or Login to view.Sure, but if the ink was light in the first place, the original writer may have though it was too pale to last the coming years / centuries and gave it a second stroke, perhaps the next day or a decade later.

Again, there are indeed cases were the best explanation is that the original Scribe himself retouched glyphs, because they came out crooked or too faint.  But that cannot explain cases B and I,
or even the other cases, in the clip I posted.

All the best, --jorge