The other day I happened upon the article “Non-Linear Fully Two-Dimensional Writing System Design: A Manifesto” by fellow conlanger Sai Emrys, with whom I’d taken some classes at De Anza a few years back (a total coincidence―I did not know till now that we share this secret vice!). The basic idea of his paper is simple: our speech is linear, due to the time-based nature of sound. But our thought is not [well, not entirely], nor is the medium of printed paper. So why not devise a writing system which can express ideas in a nonlinear fashon, more closely matched to the way we actually think?

It made for some fascinating reading. According to Marvin Minsky in his book Society of Mind, our thought processes consist of a connected web of frames. In computer terms, we can consider a thought to be a multigraph of connected subthoughts, where the vertices represent base concepts (or more realistically, abstract summaries of a more complex concept), and the edges between them express relationships. Take, for example, the sentence

John buys the bike for Mary.

This sentence seems straightforward enough, the meaning clear and easy. When we hear this sentence spoken, we can’t help but picture a linear sequence of events: John buying a bike for the purpose of giving it to Mary. But inside our minds, things are quite a bit messier:

Of course the network of thoughts in our minds is far more complex, with many more branches helping to define each concept (as well as parallel structures for the parts which are ambiguous). I have only drawn the most relevant portions, the pieces which are directly stated or implied in the original sentence.

Reading a nonlinear document like this may seem hard at first, but with proper training it can be done. The reading “John buys a bike for Mary” is nothing more than a depth-first traversial of the graph above, starting with the vertex “JOHN.” Of course other readings are possible too―nonlinear language is inherently more expressive than its natural language cousins.

Surely I am not the only one who is thinking Sapir-Whorf?

It’s common knowledge that you can’t “think” two thoughts at the same time. Now this is in part an illusion (our minds do many things in parallel), in part a result of the closure loop (which eliminates simultaneously contradictory thoughts), but also I believe in part due to the time-based and therefore linear nature of the language centers in our brains. Could this last one be nothing more than a limitation brought on by the linear nature of natural language? If that’s the case, learning a nonlinear language might break these linguistic chains, opening our minds up to ways of thinking not previously possible.

In a future post I’ll talk about a way we can bring (weak) nonlinearity to spoken language by using multiple channels. But for now, Sai Emrys’ essay on written language is a great place to start.