Sample from page f113v, lines 33-37.
This webpage calls attention to many places in the Voynich Manuscript (VMS) where glyphs, words, or entire lines appear to have been traced over original versions. I distinguish two types of such events, backtracing and retracing, that I believe happened in very separate occasions, for distinct motives, with distinct consequences. I also believe that awareness of those anomalous writing events is important in order to properly transcribe the text and interpret the figures.
The writing on the VMS (and in many other manuscripts) shows great variations of weight, which includes both the width of strokes and the darkness of the ink. There are noticeable weight changes along a single stroke, between strokes of the same glyph, between words on the same line, between lines on the same page. There may be significant variations also from page to page, and from section to section.
|
Sample from page f113v, lines 33-37. |
 |
Sample from page f1r, lines 1-2, middle. |
 |
Sample from page f1r, lines 1-5, middle. |
 |
Sampel from page f1r, lines 12-16 (parag P3), left side. |
 |
Sample from page f1r, lines 13-17 (parag P3), middle part. |
 |
Sample from page f1r, lines 18-19 (parag P3), middle. |
 |
Sample from fage f1v, lines 6-10 (parag P2) right half next to plant. |
 |
Example of retracing on figure from page f2r, rightmost flower. |
 |
Sample from page f2r, lines 8-12 (parag P2), right margin. |
 |
Sample form page f2r, lines 9-12 (parag P2), left margin. |
 |
Sample from page f3r, lines 1-6 (parag P1), left margin. |
Stroke width depends on the size and shape of the pen's nib, the speed and direction of the pen's motion, the pressure applied by the Scribe, the stiffness of the pen's "legs", and the finish of the parch's surface.
Stroke darkness too depends on a multitude of causes. The fators that affect the width of strokes also affect the thicknes of the ink film deposited by the pen, hence its darkness. Apart from that, the ink may be lighter in some strokes than in others because it was more aged or had a different formulation, or because it was applied more thinly, or because it became lighter after it was put down - by chemical decay, by being rubbed off, washed off, softened and wiped or blotted off, destroyed by spilled or intentionally applied chemicals, or eaten away by insects.
In this report we discuss three other causes of weight variations: recharging, backtracing, and retracing. Understanding these processes, the last one in particular, is important for obtaining correct transcription and interpreting all sorts of statistics, such as word and glyph frequencies and occurrence of labels in the text.
A common pattern of weight variation observed in running text (multi-line patagraphs) occurs because the pen must be periodically dipped in ink. This action creates a small reservoir of ink in the lower part of the quill, above the sculpted pen proper, and also wets the tip of the pen. After such re-inking, the first couple of words comes out heavier than usual, because of ink that has adhered to the pen's tip. Then follow half a dozen or more words with relatively uniform weight, as the ink is drained from the quill reservoir by the opening and closing of the ???legs. When that reservoir is exhausted, the weight starts to quickly decrease again. In some case, this decay may continue until the glyphs become almost invisible. Ideally, before reaching that state, the scribe should pause and re-ink again the pen. Thus we should see this normal ink dynamics pattern repeat along the text with an irregular period.
However, we often see in VMS text some isolated glyphs or short glyph sequences that are noticeably heavier than both the preceding and succeeding ones. The most likely explanation for these anomalies is that the Scribe, some time after doing that part of the text, went back and wrote or re-wrote those glyphs with a pen that was more loaded with ink. I will say that those glyphs have been backtraced (not "backtracked")
The Scribe may have backtraced those glyphs for several possible reasons. They may have come out so weak in the first pass that the Scribe felt necessary to reinforce them. An example of this is the word @daiin on page f1r. One can clearly see the fainter version poking out from under the backtraced one.
Another reason wht the Scribe may have decided to backtrace a glyph is that it may have come out incomplete, malformed, or slightly incorrect. For instance, the pen may have skipped or ran out of ink halfway through what should have been an EVA @d, so that it came out as @s. Alternatively, what should have been an @r ended up with a curved body that made it look like an @s.
Yet another possibilty is that the Scribe deliberately skipped over the glyph in a first pass, because he could not understand the draft and had to ask the Author what it should be. This may be the case for passages where multiple instances of the same glyph, like @d, were backtraced.
Whatever the reason, the backtracing usually happened seconds, minutes, or maybe hours after the surounding text was written, presumably by the same Scribe with the same pen and same batch of ink. Thus, while a backtraced glyph may be darker and wider than usual, it will be in the same ink seen in the heavier parts of the normal ink dynamics; and the glyph shape and relative weight variation between its strokes will be the same.
However, we sometimes see in the VMS glyphs, words, or entire lines of text whose heavier than normal appearance cannot be easily explained as backtracing. Typically the problem is that the ink of those glyphs seems to have a different hue or other properties that are unlike those that result from mere re-inking the pen. That ink difference often comes together with anomalies in glyph shape, or in the relative weight of the strokes that make up the glyph. The shape anomalies may be subtle, or so drastic that the glyph becomes unrecognizable, or results in quite unusual glyph combinations, like @re; or @{Io}.
The most likely explanation for those anomalies is that those glyphs or words were re-written at a much later time (decades or centuries) by a later owner of the manuscript, or a different Scribe hired for the task. The motivation for such action seems to be that the original writing had become so faint that it was hard to read and in danger or being lost completely. So it was natural, indeed expected, that the owner would try to stop and reverse the damage that way.
I refer to this second type of re-writing as a retracing event, a term that I define as distinct from the backtracing described above. The key difference is that the original Scribe must have been trained by the Author to write the alphabet, and eventually acquired a sense of its "orthography", even if he may not have known the language, encoding, or contents; whereas the person(s) who did this late restoration work, the Retracer(/s/), did not know the Voynichese alphabet, much less the ortography. Thus, while backtracing generally results in valid glyphs, retracing may create invalid ones; while bactracing generally corrects malformations, retracing may magnify them instead.
Both backtracing and retracing can occur on figure outlines. Backtracing is harder to detect since there is usually no "normal" stroke weight that can serve as reference. Retracing however can still be spotted, since its telltale signs -- including ignorance of the nature of the illustration -- can occur in figures as well as in text.
There are three kinds of evidence that make a character suspect of backtracing or retracing:
Besides the ink color and occasional invalid glyphs and glyph combinations, there are other bits of evidece that point to retracing, rather than backtracing, as the cause of anomalous weight.
Sometimes one can see bits of the fainter original strokes poking out from under the retracing ones. These cases are rare in the text, because the Retracer took pains to carefully follow the original strokes, whenever they were still visible, and completely cover them. But he often chose to retrace only part of an original stroke; so one can often see what should have been a single stroke that has both heavy and light sections. These original "ghost" traces are more commonly seen on figures, since apparently the Retracer did not feell that it was necessary to reproduce the smallest kinks and squiggles of the original outline.
Another compelling evidence of retracing is the anomalous speed and direction of certain strokes. In the original writing, the tails of @y, @g, @m, and @l are supposed to be traced from the top down, in a single quick movement of the pen, while the pen is being gradually lifted. As a result, the width and darkness of the trace tapers gradually down to zero, creating a smooth trace with a faint but sharp needle-like point. The plumes of @r, @s, and @n, on the other hand, are supposed to be written from the bottom up. The original Scribe then would have to reduce the pressure on the way up, to reduce the chance of the pen snagging on the rough parch surface. This usually results in a very thin and light trace on the way up. The trace would become broader and heavier at the top, and then it would end in the same way as the tails above, with the pen being lifted out of the parch while still moving at normal speed -- creating the same faint but sharp point.
The Retracer could not reproduce these stroke shapes, because he had to carefully and slowly move the pen along the original path. But a slowly moving pen, no matter how sharp, will make a trace that is at least 0.2 mm wide, with normal darkness. Thus, whenever he tried to retrace a whole plume or tail, the resulting stroke became thicker and darker, with a blunt or round end instead of a sharp point. Moreover, the trace oftem became slightly jittery or even discontinous, instead of smooth and even. And the Retracer often visibly traced the plumes in the wrong direction, from the top down.
The strokes by the original Scribe, whether "first pass" and backtraced, are often peppered with tiny "white" spots, 0.05 mm or less, which I call whitegrain. Another notable difference between original and retraced strokes is that there is no whitegrain in the latter.
(I first thought that those spots -- which must not be really white, just ink-free -- were pockets in the parch that did not get any ink because they were not touched by the nib of the Scribe's pen. However, if that were the case, they should show up in retraced strokes as well. Now I have two other possible explanations. One, the spots are actualy little bumps in the parch; they were originally inked, but the ink was abraded in the following decades or centuries from rubbing against the facing folios. Two, the ink used by the original scribe included a binder, like gum arabic, that insects found edible; and the spots are pockets where there was enough ink for insects to chew it off. Either theory would seemingly explain why whitegrain is present only in the original strokes.)
Retracing is visible on the flowers of the plant on page f2r. Many petals in the central part of the the three corollas were partly retraced, not very carefully. Many of the original strokes, whose ink is lighter than the retracing one, are clearly visible sticking out from under the the retraced ones. Apparently the original ink in those areas was partly washed away when the salmon-pink paint was applied there.
In the text of f2r, clear evidence of retracing are the glyphs @ld; in the word @daiildy; at the end of line 8 (head line of P2). Those two glyphs are not only darker than those around and below them, but are also free of the whitegrain that pervades the latter. The @da; in that same word must have been retraced as well.
The word @{Sh}eey; too (like many others) also may have been retraced almost entirely, leaving the original ink only in the lower part of the plume on @{Sh} and of the tail on @y.
In the words @dls; and @q&136ky; at the start of lines 9 and 10, it appears that the whole words were retraced once, leaving out only the tip of the tails of the @l and the @q, and maybe a short section (~0.3 mm) at the end of the horz arm of the @q, where it connects to the next glyph. Compare, for example, the @q of the second word with the initial @y on line 11, just below it. The tail of the @y in @q&136ky;, on the other hand, was clearly retraced with a slow and hence thicker (~0.2 mm wide) all the way to the tip.
This first retracing apparently included the unique plume on the @o (that turned it into &136). It is not possible to tell whether it was present in the original.
But then those two words were affected by a second retracing event, with even darker ink, that reinforced the @d (maybe leaving out the bottom of the lower loop), the plume of the @s (with two separate very thin strokes, ~0.2 mm wide), the (incomplete) @o (without the plume), the legs of the @k, and the @y (but only the head and the first ~1 mm of the tail).
From the example of @{Sh}eey; on line 9, one may conjecture that most of the text of this page was in fact retraced.
General retracing of the text could explain several malformed glyphs and the absence of whitegrain on most of it. However, it would require extreme painstaking care by the Retracer. (Or washing away the the remains of the original ink, if it was not iron-gall.)
On retraced words and glyphs, the only parts of the original writing that are still visible are usually the ends of tails of glyphs like @l and @y, which end by tapering down to a long point. The explanation is tha those original tails were drawn by a single quick movement of the pen, with decreasing pressure. But the Retracer had to move the pen always quite slowly, in order to follow the original path precisely. If he had traced the full extent of those tails, they would have become thicker and would end with a blunt end, not with a long tapering point. That is in fact what happened when the @y in the @q&136ky; was retraced (or re-retraced) down to the end of its tail. Therefore, in order to avoid that "obviously retraced" look, he had to stop the retracing of a long tail as soon as it became narrower than the narrowest slow trace that he could produce (apparently, at least 0.2 mm wide).
??? Two rounds of retracing?
??? f2r most of the text is retraced.
Last edited on 2025-08-17 21:57:27 by stolfi