3 ms·
Interestingly, in physics covariance and contravariance of vectors are basically on equal footing. This is pretty different the OP usage for which covariance a
by jessriedel 11y ago
Interestingly, in physics covariance and contravariance of vectors are basically on equal footing. This is pretty different the OP usage for which covariance and contravariance denote either toward or away from the root of a tree (the object hierarchy).
Like you, I'd be interested to know whether there's a deeper connection in the terminology.
- tel 11y agoAs others have said, you can recover the notation exactly through category theory
- jessriedel 11y agoCould you link me? Ctrl-F-ing for 'vector' didn't help.
- tel 11y agoMy understanding, though I've never personally made it explicit, is that you view "contra/covariance" as arising not as a property of vector spaces exactly (concretely) but instead from tensors-as-multilinear-maps and you talk about them as functors between transformation groups on vector spaces which either preserve (co-) or reverse (contra-) transformations. The inner product lets you view vector spaces as spaces of transformations already so the same sort of idea comes out now viewing vector spaces as maps in their own right.
- vilhelm_s 11y agoSo I'm just a computer scientist, not a mathematician or phycist, but I don't think there is a direct connection. The words co/contravariance is used for two different things in mathematics, vectors in physics/differential geometry, and functors in category theory. (https://en.wikipedia.org/wiki/Covariance_and_contravariance https://en.wikipedia.org/wiki/Covariance_and_contravariance). The computer science meaning is inspired by category theory (and can be seen as a simple instance of it). But the vector/functor meanings seem to be unrelated. There is a bit of a connection, because the operation ٭ which takes a vector space V to the set of 1-forms V٭ is a contravariant functor. (I.e., linear maps V→U give maps of 1-forms U٭→V٭). But in physics, we say that the members of V are contravariant and the members of V٭ are covariant, i.e. the physics notion is about individual spaces, not about mappings between spaces.