+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Theories & Solutions (https://www.voynich.ninja/forum-58.html)
+--- Thread: The Book Switch Theory (/thread-5035.html)
RE: The Book Switch Theory - ReneZ - 14-03-2026
(9 hours ago)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.But if he asked any Jesuits in Prague (directly of through any contact, Jesuit or not) about a "TepenecWhatever"from around 1600, they would have known that he was referring to Sinapius/Horciky, and would have known about that entry in Schmidl, and more.
Indeed, and that is what he did (archives director rather than Jesuit), and indeed he got a quote from that book in return. That is also how we knew about that source, when we were in the Clementinum.
RE: The Book Switch Theory - Jorge_Stolfi - 14-03-2026
(Today, 03:38 AM)kckluge Wrote: You are not allowed to view links. Register or Login to view.But that's the wonderful thing about conspiracy theories -- we can't prove that he didn't read it before 1920.
I don't think it is appropriate to label either of the competing hypotheses as a "conspiracy theory". The popular meaning of that term is a conspiracy by a large number of people -- like all the scientists at NASA, all airline pilots, all academic Egyptologists, etc. But in both hypotheses under discussion here the only "conspirator" would be Wilfrid -- so there is not even a "co-". The are both "monospiracies"...
Quote:Which is why understanding the concept of burden of proof and who has it and why it's on the person making the positive assertion
I don't think one can always objectively identify which claim is "positive", or decide who has the burden of proof. The Forged Signature hypothesis (F) is that the "signature" on You are not allowed to view links. Register or Login to view. was put there by Wilfrid. Its opposite (G = notF) is that the signature is "genuine", to the extent that it was put there by someone else and was there when he got the book. Which is the "positive" hypothesis, and why? Why doesn't the burden of proof fall on those who believe G?
I prefer to think in the "Bayesian" way. But first, we must understand that a probability is not a real thing, like a temperature or a longitude. It is only a numerical expression of one's belief in some statement -- and therefore inherently personal. There is no "the probability" of some fact, but only "my probability", and "your probability" and "Rene's probability", etc.
Now, offhand, each of us may assign some "prior" probability Pr(F) to hypothesis F, and therefore 1 - Pr(F) = Pr(G) to hypothesis G.
Hypotheses F and G have a number of potential consequences -- statements whose probabilities depend on which of F or G is true. For example, let M be the statement "Marci and/or Baresch mentioned Jacobus in their letters to Kircher", and N = notM be its opposite, "they did not mention Jacobus". If the signature was forged by Wilfrid, then there would be no reason for anyone to think that Jacobus ever owned the book. Then M almost certainly would not happen; that is, the probability of M given F would be essentially zero, or Pr(M|F)=0. Otherwise, if the signature was there in 1911, it surely was there in the 1600 too; in which case, before we looked at the letters, we should have given some probability, say 10%, that they would have believed that Jacobus had owned the book and would have mentioned that fact in the letters. That is, we should set Pr(M|G) = 0.10.
But then, after knowing that M did not happen, a rational person should lower his probability of G and raise that of F. Probability theory gives a formula for this adjustment (Bayes's formula). For example, suppose one started with no preference for either hypothesis -- that is, Pr(F) = Pr(G) = 0.50. Then, after observing that N is what happened, not M, Bayes's formula says that one should revise the probabilities to Pr(F|N) = ~0.53, Pr(G|N) = ~0.47. That is, the fact that Marci and Baresch failed to mention Jacobus should raise one's probability of forgery by ~3% over the prior 50%.
However, if one started with Pr(F) = 0.01 and Pr(G) = 0.99, then after observing that N happened the same formula says that one should revise one's probabilities to Pr(F|N) = ~0.011, Pr(G|N) = ~0.989. That is, if one starts out 99% sure that the signature is genuine, the fact that Marci and Baresch failed to mention Jacobus should raise one's probability of forgery by a mere 0.1%.
A similar discussion applies to many other potential consequences, such as "Jacobus's signature was mentioned by the Jesuits in some catalog" (J), "Wilfrid wrote to Garland inquiring about a TepeneWhat" (L), "Wilfrid applied chemicals to f1r" (Q), "Wilfrid eventually mentioned Jacobus as a previous owner" (W), etc. One should choose conditional probabilities for these events -- Pr(J|F), Pr(J|G), Pr(L|F), Pr(L|G), etc -- then use Bayes's formula to recompute P(F|notJ and L and Q and notW and ...).
"Strong evidence" against G would be some consequence S that was observed but which had very low conditional probability Pr(S|G). That same fact, of course, would be strong evidence for F.
It would be nice if there was strong evidence for F, because then even a skeptic who started with Pr(F) = 0.01 would be convinced -- that is, would get Pr(F|S) = 0.80 or whatever. Same if there was strong evidence for G.
So, rather than say "the burden of proof is on those who claim X", I would say "if you want to convince those skeptics who give X a very low prior probability, you had better present or point out some strong evidence for X". (Although a big enough list of weak evidence may eventually convince them, too.)
But unfortunately we don't have any strong evidence, for or against either F or G.
Consequence notW -- "Wilfrid never mentioned Jacobus as a previous owner" -- should be strong evidence for F, because everybody should assign Pr(notW|G) a value near zero. But it seems that those who are quite certain about G also think that "Wilfrid never mentioning Jacobus as owner" is not strange at all -- that is, they set Pr(notW|G) to 50% or more...
(By the way, of course no sensible person would bother to do all that math, because all the numbers Pr(X) and Pr(Y|X) are rather arbitrary anyway. Yet, if one decides only that they are "large" or "small" or "about 50-50", with enough consequences one may estimate that the result would be "high" or "low", even without doing the math. Because if several of those probabilities are "low", the product of all those probabilities will be "very low".)
All the best, --stolfi