If this is a useful exercise from a Voynichological standpoint, I suppose it would be useful for two reasons. First, we might be able to figure out how this particular cipher works, which I believe is what Mark is interested in learning. But second, presuming it has the expected features (e.g., nulls, multiple cipher characters for some or all plaintext letters, cipher characters for letter combinations, nomenclator-style word signs, etc.) it also gives us a nice opportunity to explore methods for cracking such a cipher. That second reason is why I think it would be more valuable to work through the puzzle here step by step than to try to solve it in isolation and then just report the results. So: If we've correctly identified a pair of cipher characters representing nc, then we should be able to leverage that observation to analyze the cipher characters to either side of it, since the underlying plaintext to the left should be ...citosperohi, and the underlying plaintext to the right should be mitteread....
Let's start on the left with the cipher character that looks like a plus sign [+].
Since this is (tentatively) part of the word hinc, it can't be a nomenclator-style word-sign. So unless this cipher is behaving in some unusual way, it seems to me that the [+] must be one of the following:
Depending on which of those three values the [+] has, the [o] cipher character one position to the left of it must in turn have one of these values:
There are other possibilities, such as [o] representing the combination oh across a word break, but the probability of that seems vanishingly small.
We can refer to the foregoing hypotheses as 1, 2, 3, 1a, 1b, 1c, 2a, 2b, 3a, 3b, 3c, 3d.
I believe we can easily rule some of them out. For example, the [o] character appears numerous times in the ciphertext, but the word spero occurs only once in the decipherment, so 3d doesn't work.
Others are more difficult to assess. For example, does [+] represent hi? In the image below, I've marked tokens of hi in red and tokens of [+] in blue. Not all tokens of hi would necessarily be encoded as [+], even if that's what [+] represents; but all tokens of [+] should presumably represent hi if any of them do -- again, presuming this is a "normal" cipher of its era.
But tokens of [+] are trickier to identify than tokens of hi. Some may be illegible due to water damage. Others are visible but may contain extra marks that could conceivably transform them into other, unrelated characters. The most ambiguous case, I think, is the [+] seen here closely followed by a dot:
Given its placement, that particular token happens to be one that appears especially unlikely to match any token of hi in the plaintext decipherment. Still, it seems more likely to be a token of [+] than the following cases where a cross actually appears to be attached to another stroke (and I didn't mark these in blue):
The initial nihilominus, at least, definitely doesn't correspond to a ciphertext [+]. Maybe nihil was written as a word-sign, but it seems unlikely that there would have been multiple cipher characters provided for the hi combination specifically. But more importantly, if we remove the two tokens of nihilominus from the list of tokens of hi, there just don't seem to be enough hi's to cover all the unambiguous tokens of [+]. So I'm inclined to rule out hypothesis 3, which would leave [+] as either a null or one of (probably) several options for encoding i.
Does this logic seem sound so far? Is anyone else playing around with this puzzle?
Let's start on the left with the cipher character that looks like a plus sign [+].
Since this is (tentatively) part of the word hinc, it can't be a nomenclator-style word-sign. So unless this cipher is behaving in some unusual way, it seems to me that the [+] must be one of the following:
- a null
- a cipher character for i
- a cipher character for hi
Depending on which of those three values the [+] has, the [o] cipher character one position to the left of it must in turn have one of these values:
- a= another null; b= a cipher character for i; c= a cipher character for hi
- a= a null; b= a cipher character for h
- a= a null; b= a cipher character for o; c= a cipher character for ro; d= a nomenclator symbol for spero
There are other possibilities, such as [o] representing the combination oh across a word break, but the probability of that seems vanishingly small.
We can refer to the foregoing hypotheses as 1, 2, 3, 1a, 1b, 1c, 2a, 2b, 3a, 3b, 3c, 3d.
I believe we can easily rule some of them out. For example, the [o] character appears numerous times in the ciphertext, but the word spero occurs only once in the decipherment, so 3d doesn't work.
Others are more difficult to assess. For example, does [+] represent hi? In the image below, I've marked tokens of hi in red and tokens of [+] in blue. Not all tokens of hi would necessarily be encoded as [+], even if that's what [+] represents; but all tokens of [+] should presumably represent hi if any of them do -- again, presuming this is a "normal" cipher of its era.
But tokens of [+] are trickier to identify than tokens of hi. Some may be illegible due to water damage. Others are visible but may contain extra marks that could conceivably transform them into other, unrelated characters. The most ambiguous case, I think, is the [+] seen here closely followed by a dot:
Given its placement, that particular token happens to be one that appears especially unlikely to match any token of hi in the plaintext decipherment. Still, it seems more likely to be a token of [+] than the following cases where a cross actually appears to be attached to another stroke (and I didn't mark these in blue):
The initial nihilominus, at least, definitely doesn't correspond to a ciphertext [+]. Maybe nihil was written as a word-sign, but it seems unlikely that there would have been multiple cipher characters provided for the hi combination specifically. But more importantly, if we remove the two tokens of nihilominus from the list of tokens of hi, there just don't seem to be enough hi's to cover all the unambiguous tokens of [+]. So I'm inclined to rule out hypothesis 3, which would leave [+] as either a null or one of (probably) several options for encoding i.
Does this logic seem sound so far? Is anyone else playing around with this puzzle?