[split] Does Voynichese evolve from A to B?

[npcompl33t]

Moreover, the extent of the whole dataset in the MDS1 direction seems to be only a bit larger than its extent in the MDS2 direction.  Which means that using the MDS1 axis (or any other axis) as the independent variable for curve fitting is an arbitrary choice.  Which means that the order in which the curve visits the clusters is meaningless.

I just want to be clear this is not what is happening in any of my fits. The curve is fitted in all 256d, I am just plotting it consistently with MDS1 on the x axis to make it easier to read. 

Actually I do have one a priori argument, admittedly not very strong, against the existence of "temporal" gradients within each section.  The VMS was surely composed by the Author as a draft on paper which was then transposed to vellum by the Scribe

Then the gradient would appear on the draft and would have been copied into the MS by the scribe.

Frankly, I think that your "gradients" are just pareidolia; and the fact that you need a 3rd-degree curve to get the 4 major clusters "aligned" shows that they are not aligned -- not even along a parabolic "horseshoe".  

What is the opposite of pareidolia?