The Voynich Ninja

Assuming the Voynich was written in cipher and assuming that you knew precisely what kind of cipher it was and presuming you knew the underlying language that it was written in then I am wondering how one might go about deciphering the manuscript. Obviously to some significant extent this will depend on what kind of cipher it is. When I talked about "kind of cipher" I mean of course the class or type of cipher. I don't mean that you know the exact implementation of the cipher. For example, if you knew it was a very simple substitution cipher then you might not know the precise letter substitutions.

There seem to be some general problems irrespective of type of cipher. We seem to have no crib or at least no crib that we can be confident in and we don't seem to have many clues as to how we might identify part of sentence or other clues that we can use to break the cipher open, although I could be wrong about that.

My question comes from the fact that I have learnt from some diplomatic ciphers that even when you know the precise nature of the cipher it can still be hard work to break the cipher. This comes from things like the sheer number of distinct substitutions, so every specific substitution needs to be worked out. It comes from the situation that when dealing with substitutions for people or places only knowing the historical, political and geographic context will help you work that out.

The Voynich is different from a diplomatic letter as you are unlikely to have many political substitutions, so you would have to look for completely different kinds of substitutions. The Voynich won't have the standard linguistic phraseology that can be found in diplomatic correspondence that decipherers of diplomatic ciphers are familiar with and can spot to help in the decipherment.

To reiterate I think even if one knows the type of cipher one is dealing with it could still be very very difficult to decipher.

The thing that occurs to me would be some kind of stochastic descent computational search of the solution space to explore the many different combinations.

I don't know how to tackle such a problem and I would be curious as to what others think.

I have had some contact with Dr. George Lasry and he has developed a piece of software using Simulated Annealing with human assistance to solve various ciphers. I have never used it, but the premise seems sensible.

Even if one knows the cipher being used then one has to consider the numerous different possible implementations of that cipher to find which one fits. This is a little bit like finding the right permutation or combination. Depending on the type of cipher there may be a small or large or huge number of possible combinations. So even if one knows the type of the cipher finding the correct combination can be very difficult. This seems to be a problem which would require some computational element in order to solve, so as to explore the solution space to find the one solution that fits precisely.

Well, that's the problem, then. Simple substitution has been eliminated, and after simple substitution comes complex substitution, but how complex is the VMs solution? Methodology is clearly a prime aspect of VMs language investigation, but it is not the only significant factor. Methodology must operate on a subject.

What is the subject? What segment of text is selected for examination? Something random, just start at the beginning, try to find a familiar plant? They haven't produced anything useful.

If the author did not have strong encryption and was cognizant of that reality, still, needing to record a message, what tactics could be used? Isn't it better to split the message into parts and hide the text segments in the illustrations? Stolfi's markers are clearly adequate to the task, but were they created with purpose? If that 'purpose' is determined by the intent in the structure of the illustrations, then purpose is clearly there.

A reassembled text may have unique properties. Perhaps it might be more open to translation. Looking at the markers in various circular bands of text, there is a fair variety of different examples. So, this really is the garden of forking paths as to what to include and what to omit. Consequently, this is an investigation that should be guided by consensus rather than individual preference.

Some cryptanalysis will help. Maybe these links can give some idea of what to do:

You are not allowed to view links. Register or Login to view.

You are not allowed to view links. Register or Login to view.

You are not allowed to view links. Register or Login to view.

You are not allowed to view links. Register or Login to view.