A Cryptographic Language

File: 6.jpg (21KB, 450x299px) Image search: [Google]

21KB, 450x299px

As far as I understand it, it seems like tons of crypto algorithms end up being vulnerable to various methods (including brute force) primarily because of the fact that our languages cannot be randomly jumbled up and still expected to yield a meaningful message.

So once they start noticing that they're able to produce grammatical English then the crackers know they've got something.

But what if there was a language in which every single possible combination of letters/words/heiroglyphs/whatever produced a valid grammatically correct statement within that language.
Then brute forcing would become useless because you could never tell when you've got something. Of course they could still maybe crack it by brute forcing entire chains of conversation at once and just manually verifying if each message makes contextual sense with each other. But that would be extremely difficult, and you could probably even avoid that by just talking in codes/riddles whatever and never reusing any of the same lingo the other person uses.
>But anon, you can do that with any encryption method
Yes, but if it's a natural language that's encrypted (say English) even if you talk in a confusing way they can still know that they've got the actual transmissions you've been sending back and forth. So then all that's left to do is interpret the messages.
With the type of language I described above however they could never truly know they have the right messages and they'd always know that their interpretation of the conversation might just be them interpreting random noise and finding meaning.

Of course this wouldn't work for sending any kind of data or files, because good luck creating an image encoding format where every single possible set of data produces a visually coherent image.
But still, it could work for brute-force-proofing plain text encryption (or at least making it much much more difficult).