Statement: IF Voynichese does not represent natural language...

[Linda]

Emma: to be clear, "deception" is low on my list of possibilities. I think there are several indications that the text is linguistic. But if it is not, wouldn't it automatically mean we are dealing with something that's made to look more language-like than it is? The only third option is that the language-like properties we see in script, statistics and layout are an UNintended side effect of whatever else is going on. But that feels so unlikely.

What are you envisioning as a non linguistic text, does this include enciphered or encoded linguistic text? Because i could see various ways these would still resemble language, depending on the method.

[bi3mw]

...

- it is intentionally deceptive
...

I would not assume that it is an intentionally deception. It is probably more the result of an unknown method that was motivated differently. It is even possible that the authors have simply created it because they were able to. I would rather assume a combinatorics that should make a statement for itselves but is very hard to reproduce for "uninitiated ones".

[Koen G]

Linda: I mean that it somehow contains a text, language.... so yes it can be enciphered. My statement is that if it does not somehow encode a text, it is still made to look as if it does. So either it hides linguistic contents we could theoretically unveil, or it pretends that it does. I found it strange to realize that those two options cover the vast majority of possibilities.

[Linda]

Linda: I mean that it somehow contains a text, language.... so yes it can be enciphered. My statement is that if it does not somehow encode a text, it is still made to look as if it does. So either it hides linguistic contents we could theoretically unveil, or it pretends that it does. I found it strange to realize that those two options cover the vast majority of possibilities.

But it could also hide meaningful non linguistic content that then looks like it is linguistic, like for instance one could encode a table of astronmical position data in ways that could end up looking like words. Not sure why you would but you could.

[Koen G]

Yes! But in that case it would be deceptive.

[nickpelling]

"IF Voynichese does not represent natural language (etc)"

Errrm... it would help if you said what you meant by 'represent' here. Are you trying to include or exclude ciphers, steganography, nulls, etc?

Insofar as a cipher encodes language (cryptography), it represents it: and insofar as it conceals language's presence (steganography), it misrepresents it. It's a bit of a mix.

[Koen G]

I mean encodes in the broadest possible sense, from normal writing to steganography.

[Linda]

Yes! But in that case it would be deceptive.

But then everything except uncoded natural language would be non deceptive?

[Koen G]

But then everything except uncoded natural language would be non deceptive?

With "deceptive" I mean that efforts have been made to make it look like something it's not. 

Say it's glossolalia, and the person really wrote what they thought were "tongues" coming to their minds. In that case it's not deceptive because it ended up looking "language-like" because this is the way it came to the person's mind. The lnaguage-like properties are a result of what it is, not because of what it's made to look like.

If, however, the VM text would turn out to encode a whole string of numbers, then my statement is that it is deceptive since we are given the impression that we're dealing with language, the kind you'd find in paragraphs accompanying herbal images.

[nickpelling]

With "deceptive" I mean that efforts have been made to make it look like something it's not. 

For what it's worth, one related argument I've put forward for some years relates to the Voynichese glyphs 'o' and 'a': which is that even if you put everything else about Voynichese to one side, 'o' and 'a' simply do not behave in a language-like way. From this, I conclude that they are not 'undeceptive' vowels, but are instead artificial constructs performing some other function entirely (e.g. they are verbose ciphertext shapes that need to be parsed as a pair to make sense).

Labelese makes this artificiality particularly visible, but the sequential repeats of 'or' (e.g. on folio 15?) are much the same. And so forth.

And if 'o' and 'a' are "deceptive", then so must the rest of Voynichese be (to a large degree, at any rate).