> Far more interesting to speculate about, I think, is if we could create a non-linear written language taking advantage of the two dimensional property of pape

> Far more interesting to speculate about, I think, is if we could create a non-linear written language taking advantage of the two dimensional property of paper/screens to write things which aren't forcibly constrained to the 1D nature of speech. I think the problem is that human language is inherently one-dimensional. Spoken language involves a temporal succession of sounds, words, sentences. Even sign languages, while the signs may be 2D, there is still a 1D temporal succession of distinct signs. Real texts involve certain limited escapes from one-dimensionality – footnotes, hypertext (which exists even in printed books, with the table of contents, index, "see page 123"), etc – but, still they are only limited deviations from the essentially 1D nature of human language. SignWriting is interesting as a way of writing sign language by using pictorial symbols for each sign, as opposed to using the written form of a spoken language text – but SignWriting is like the CJK languages, they may be 2D at the micro-level of individual characters, but at the macro-level texts are still fundamentally 1D. If you could invent a truly 2D language, I doubt the human brain would be capable of processing it. Maybe some AI might become fluent in it, but no human ever could. But even software is predominantly 1D – memory is fundamentally a 1D array of numbers, and in some ways it has been becoming more 1D over time (witness the triumph of flat address spaces over memory segmentation) – which makes me think it might be less than entirely natural even for an AI.