Brainpower

One of the comments at Maggie's Farm reminded me of some experiments Richard Feynman did with himself: he could count and read, but not count and talk. "... when Feynman told mathematician John Tukey about this, Tukey could do the reverse — talk but not read. The reason was that Feynman would talk to himself in his head, while Tukey would see an image of a clock ticking over. Feynmann suggests this could be because people think differently".

The author of the article suggests that the brain has "modules" like a "sketch pad" or a "phonological module" for words and sounds.

I don't know about how other people use visual thinking in math, but I find that I often do. For example, matrix multiplication I visualize as an action. (I should redo this to slow it down). When I try to figure out the framework for a problem I draw pictures. Equations are partly sentences and partly blocks like pictures.

FWIW, I generally read by "see and say" because I'm already trained to spot the blocks in English words. Sometimes I scan too quickly, and hilarity usually ensues. But I learned phonetically. And when I hit non-English words, I work phonetically. If I have time. Train stops in Germany were a nuisance. The name was often half a block long, and I couldn't read it fast enough--so I read the first and last chunks of the name and hoped that was unique. Problem is, the last chunk was usually "strasse."