Yesterday, 12:15 PM
I find the possibility that certain combinations are avoided because they contradict the construction rules of the script very fascinating.
It also made me think of Lisa Fagin Davis's post about p = ke.
What I liked most, however, was that your approach is different, because you also keep running into the same "bumper". So I started asking myself: Could the approach itself be fundamentally problematic from the very beginning?
Where do I start with my thoughts? With the construction of the glyphs.
Let's leave EVA and all transcriptions aside for a moment and just look at the glyphs and the text itself. What I notice then (as is already known from other studies) is that some glyphs can be broken down into smaller components.
But looking at the running text, we cannot determine exactly how the glyphs should be broken down.
An example: imagine writing "uiiuiinui" in German cursive handwriting. If we leave out the dots above the "i", someone unfamiliar with the script would see a sequence of strokes and curves that they can no longer distinguish. They wouldn't know where one letter begins and another ends.
Perhaps we are experiencing exactly the same thing with Voynich.
Let's get a little more specific and theoretically assume:
a = e + i
We could then assign the following values (lowercase letter = Voynich glyph, uppercase letter = example value):
e = A
i = B
a = C
k = K (I'm simply assuming that k remains k)
This gives us three possibilities:
1. k + a = K + C = D
2. k + a = K + A + B = E
3. k + a = K + C or K + A + B = D or E
What I mean is that these variations could form a hierarchy, they could be decomposable, or they could depend on context. D or E would represent a new value, not the sum of the individual values.
And now we could take it a step further and express it mathematically:
K + (A + B) ≠ (K + A) + B
(Note: The + is only used for illustration.)
This might sound illogical at first, but such a system could produce a huge variety of different values.
The "problem": it would require additional rules. Otherwise, the combinations would become meaningless or unmanageable.
And this brings me right back to your observation, Quimqu. Davis and Stolfi also fit into this system with their findings.
Perhaps q + o / qo must largely belong together. Perhaps ch is simply ee with a connecting stroke, but that stroke might only be possible in this particular combination to produce a certain value.
And regarding the Gallows: perhaps p and ke can only occur in certain combinations because otherwise the content produced by combining the values would no longer make sense?
Are there not only rules for how elements can come together, but also clear prohibitions against certain configurations?
How much context can we allow for reading and writing?
In a natural language, for example, a great deal of context can be involved. Nevertheless, it still needs rules. While we construct sentences in our native language almost instinctively, we are essentially using combinations of values to do so.
What I find interesting is that these values do not necessarily have to follow exactly the same rules. Even within the same native language, we can adapt the rules or arrange the values differently. The written text is still readable because we understand the context. However, a minimum of rules is still necessary.
"Lange Rede, kurzer Sinn" (as we say in German, roughly meaning "long story short"), let's call it a composition of many variations, which may also have an ambiguous structure and must follow certain, perhaps very complex, rules.
If that were the case, would we even have a system that could be captured by the existing transcriptions?
Could this be exactly the crucial point where statistical analysis reaches its limits?
It also made me think of Lisa Fagin Davis's post about p = ke.
What I liked most, however, was that your approach is different, because you also keep running into the same "bumper". So I started asking myself: Could the approach itself be fundamentally problematic from the very beginning?
Where do I start with my thoughts? With the construction of the glyphs.
Let's leave EVA and all transcriptions aside for a moment and just look at the glyphs and the text itself. What I notice then (as is already known from other studies) is that some glyphs can be broken down into smaller components.
But looking at the running text, we cannot determine exactly how the glyphs should be broken down.
An example: imagine writing "uiiuiinui" in German cursive handwriting. If we leave out the dots above the "i", someone unfamiliar with the script would see a sequence of strokes and curves that they can no longer distinguish. They wouldn't know where one letter begins and another ends.
Perhaps we are experiencing exactly the same thing with Voynich.
Let's get a little more specific and theoretically assume:
a = e + i
We could then assign the following values (lowercase letter = Voynich glyph, uppercase letter = example value):
e = A
i = B
a = C
k = K (I'm simply assuming that k remains k)
This gives us three possibilities:
1. k + a = K + C = D
2. k + a = K + A + B = E
3. k + a = K + C or K + A + B = D or E
What I mean is that these variations could form a hierarchy, they could be decomposable, or they could depend on context. D or E would represent a new value, not the sum of the individual values.
And now we could take it a step further and express it mathematically:
K + (A + B) ≠ (K + A) + B
(Note: The + is only used for illustration.)
This might sound illogical at first, but such a system could produce a huge variety of different values.
The "problem": it would require additional rules. Otherwise, the combinations would become meaningless or unmanageable.
And this brings me right back to your observation, Quimqu. Davis and Stolfi also fit into this system with their findings.
Perhaps q + o / qo must largely belong together. Perhaps ch is simply ee with a connecting stroke, but that stroke might only be possible in this particular combination to produce a certain value.
And regarding the Gallows: perhaps p and ke can only occur in certain combinations because otherwise the content produced by combining the values would no longer make sense?
Are there not only rules for how elements can come together, but also clear prohibitions against certain configurations?
How much context can we allow for reading and writing?
In a natural language, for example, a great deal of context can be involved. Nevertheless, it still needs rules. While we construct sentences in our native language almost instinctively, we are essentially using combinations of values to do so.
What I find interesting is that these values do not necessarily have to follow exactly the same rules. Even within the same native language, we can adapt the rules or arrange the values differently. The written text is still readable because we understand the context. However, a minimum of rules is still necessary.
"Lange Rede, kurzer Sinn" (as we say in German, roughly meaning "long story short"), let's call it a composition of many variations, which may also have an ambiguous structure and must follow certain, perhaps very complex, rules.
If that were the case, would we even have a system that could be captured by the existing transcriptions?
Could this be exactly the crucial point where statistical analysis reaches its limits?