Bifolia created using previous as reference

[npcompl33t]

 
Ho do you measure those coordinates? The vertical coordinate seems to be from the first text line on the page.  Is the horizontal one measured from the edge of the Beinecke Library (BL) image?

These are good questions


The horizontal coordinate is not measured from the vellum edge. 

It comes from the Voynichese Project’s recorded word-box coordinate system. That frame is registered separately to the Beinecke scans, so it is not directly the pixel distance from the left edge of the full-resolution BL image either. In the body-height-scaled view, both axes are divided by the page’s median word-box height; these are not millimetres.


You are correct about the vertical origin. 
Each page is translated vertically using a baseline anchor on its opening text line. The plots retain spacing between paragraphs within a page, but they do not preserve where the opening sits on the physical sheet. This was done intentionally, because when using a previous page as a layout reference, you can lay one page on top of the other, essentially removing the top or bottom margin from one sheet.  


Your You are not allowed to view links. Register or Login to view. angle observation is valid, but it is a result of some confusing settings within the app.

There are two types of rotations performed:

• Recorded, true margins: restores the overall text-line angle by moving points vertically only. Horizontal coordinates remain unchanged.
  
• Rigid turn, margins lean: rotates the entire page, changing horizontal coordinates too. This produces the extra margin slant we identified earlier.
  

When comparing angles directly, the rigid turn is useful because it identifies the overall slant on the page, making it easier to compare across bifolia. However this can affect the other visual identifier, the margin, because this rotation will change margin differences.

None of the underlying calculations used this transformation, it was put in place to make it easier to visually compare the angles directly between bifolia, but I can see the issue in that it is distorting the other independent factor, so i am going to disable this by default. It remains an option in a drop-down selector if you would like to more directly compare the angles.

   

This feature was already disabled in the second image i shared. Here is what the first looks like with it disabled, it doesn't particularly change the story, you can still see the angle on the final sinulion page is carried over from the previous bifolia, and is typically different from the angles of the other 4 pages. The margin offsets are also still preserved between the bifolia.  

   


As for why the scribe did this.. I can't say, its beyond the scope of this analysis to determine motive. I'm just putting this data out for others to asses. 

Aren't you falling again for the "Book Sorting" fallacy?  Namely, if you sort a bunch of things by similarity of an attribute X, you will find an apparent gradual evolution in the value of X -- even if (or especially when) the X of each thing is an independent random variable.

Possibly, but the point is that these two measurements, margin and text angle, are independent of any glyph or word statistics, and should be independent of each other. The way to truly determine if the book size 'fallacy' is occurring, is to obtain as many different measurements as possible, and determine if they align. Wouldn't you agree with that?

All 3 aligning could still be coincidence, but the goal here was to attempt to bring some new evidence into the picture. That felt more productive than continuing to debate the same glyph data.  


All of this is publicly available onYou are not allowed to view links. Register or Login to view.. I encourage others to dig into this and welcome any peer review of the data. I'm also happy to rerun things or try different measurements if you think there may be better ways! I do think there is a useful new signal here!