19-07-2016, 02:57 AM
Basically, the auto-copying should put us on the lookout for potential encoding schemes that allow for many ways to encode a message, while allowing for only one way to then unambiguously decode any given encoded string (as my scheme above satisfies).
Sadly, that does expand the potential pool of encoding schemes. If we knew that the encoding scheme had to be a strict 1-to-1 procedure of unique input to unique output back to unique input, then that would severely constrain the possible schemes (as it turns out, it constrains it too severely. Maybe that has been our problem so far—assuming that it has to be a strict 1-to-1 encoding scheme).
Such encoding schemes could potentially rely upon mechanisms at the sub-character level (curve vs. line, type of tail, etc). I hope I'm not the only one who is suspicious about the fact that the Voynich script happens to be so systematically constructed. Naturally-evolved scripts don't look that neat and tidy (it's not just the main curve/line dichotomy, but how the gallows characters are cleanly grouped along 3 dimensions of variation (left loop vs. no left loop, crossing right descender vs. no crossing right descender, and bench vs. no bench), there's the fact that "r" and "s" have the same tail, as do "g" and "h", etc.). It makes me doubt that the Voynich script could be, for example, an extant example of a previously unattested script from Central or Eastern Europe.
Sadly, that does expand the potential pool of encoding schemes. If we knew that the encoding scheme had to be a strict 1-to-1 procedure of unique input to unique output back to unique input, then that would severely constrain the possible schemes (as it turns out, it constrains it too severely. Maybe that has been our problem so far—assuming that it has to be a strict 1-to-1 encoding scheme).
Such encoding schemes could potentially rely upon mechanisms at the sub-character level (curve vs. line, type of tail, etc). I hope I'm not the only one who is suspicious about the fact that the Voynich script happens to be so systematically constructed. Naturally-evolved scripts don't look that neat and tidy (it's not just the main curve/line dichotomy, but how the gallows characters are cleanly grouped along 3 dimensions of variation (left loop vs. no left loop, crossing right descender vs. no crossing right descender, and bench vs. no bench), there's the fact that "r" and "s" have the same tail, as do "g" and "h", etc.). It makes me doubt that the Voynich script could be, for example, an extant example of a previously unattested script from Central or Eastern Europe.
![[+]](https://www.voynich.ninja/images/collapse_collapsed.png)