While leafing through The Elegant Universe, I have been thinking about metaphors. In theoretical physics (and probably in other fields of science as well), it seems as if they are hunting metaphors – trying to find metaphors that are simple and intuitive, and that succeed in describing the universe in every thinkable way.

To me it seems that the scientists do a lot of thinking, considering any metaphor they can come up with, and then conduct experiments to see whether it holds. Then they describe it mathematically and it is launched as a New Theory.

In software, I think it’s of great value to work in much the same way, although here you try to come up with a metaphor that seems to be suitable for the problem you are solving, and then you create a reality to match that metaphor. So it’s the other way around.

Having skimmed the book very lightly the first 190 pages, I have picked up that string theory seems to hold promise of being a theory of everything. In other words: it might be one heck of a metaphor. But even though it would explain everything, we would never be sure that it’s real – it just would be a perfect way of describing the universe.

I remembered an entry on Gustav Holmberg’s weblog (in Swedish) where he mentioned that Evelyn Fox Keller has studied metaphors in science. Then, I found two books of hers: Making Sense of Life: Explaining Biological Development With Models, Metaphors, and Machines and Refiguring Life. I will check out those later. Also, I ordered The Structure of Scientific Revolutions by Thomas S. Kuhn, which I suspect will deal with this, too.