"All knowledge is either identification or classification" - II

So I continue.

When David MacKay finished his Dasher talk, we had a chat with him in the Wireless Foundations' lounge. Earlier in the talk, MacKay had used a light on his hat and a sensor on his computer to write using dasher, at a pretty good speed.

Venkat made a point that this was similar to a sci-fi book he had read, where the people use their eyes to communicate ideas. Not words. Ideas. He asked whether such a system was possible using Dasher.

"Absolutely!" said MacKay, pointing out that symolic language allows him to convey the idea of an aeroplane, using the symbol of a machine, and of a bird.

Underlying this scheme of communicating ideas is the assumption that all ideas can be generated by combinations of previously known ideas. I do not particularly like the assumption. It is not saying that ideas are finite, even countable. For the basis set of ideas can be infinite. Even with a finite set of symbols, much like music, an infinite number of ideas can be generated. But I believe there is much more to ideas than the combinations of some known ones.

Regardless, this is an example of how language, a tool to express and communicate ideas and knowledge, allows us to construct ideas and knowledge as well. And language is about identification, by a term for any new element, and classification, by observing similarities. (There is a structure to language as well, but that is not central to the point I am making). Again, topology (and, indeed, algebra) is a good illustration of this general principle. Whatever we learn, we store it as knowledge using a language, by identification and classification using specific terms.

I have one more issue to explore here, shall continue next week!