Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
Drexel University or The Math Forum.
|
|
|
|
Re: in what sense is a limit a functor?
Posted:
Apr 13, 2006 9:00 AM
|
|
The point about axiomatic foundations is not at all deep, and not
special to category theory. You can extend the language of Zermelo
Frankel set theory by adding a "choice function" which applied to any
set of sets returns a function assigning each of those sets a unique
selected member. That is a form of global choice. But then, if you do
not want choice in general, but only some limited amount of it you
might be able to use some operator more germane to your needs -- e.g.
for any ideal in a Boolean ring it selects a larger prime ideal. This
is weaker than the whole axiom of choice. You can do the same kind of
thing in many ways in categorical foundations -- e.g. an operator
which to any diagram in a complete category assigns a selected limit.
The general summary you give is quite right: Once limits are chosen
the arrows induced by cones under them are unique. The only subtlety
is to keep in mind that a limit for a diagram is not just an object, it
is an object plus a cone to the diagram. A product for sets S and T is
not just a set SxT but the set plus its projection functions
p_1:SxT-->S and p_2:SxT-->T. That is essential to the functoriality
since that is how the induced functions to SxT are determined.
|
|
|
|
