I think http://www.cs.man.ac.uk/~schalk/notes/llmodel.pdf might be useful. And John Baez and Matt Stay's math.ucr.edu/home/baez/rosetta.pdf (where I found the citation for the first paper) has a fair amount about this sort of question.

On Tue, Feb 22, 2011 at 7:55 PM, Dan Doel <dan.doel@gmail.com> wrote:
On Tuesday 22 February 2011 3:13:32 PM Vasili I. Galchin wrote:
>        What is the category that is used to interpret linear logic in
> a categorical logic sense?

This is rather a guess on my part, but I'd wager that symmetric monoidal
closed categories, or something close, would be to linear logic as Cartesian
closed categories are to intuitionistic logic. There's a tensor M (x) N, and a
unit (up to isomorphism) I of the tensor. And there's an adjunction:

 M (x) N |- O  <=> M |- N -o O

suggestively named, hopefully. There's no diagonal A |- A (x) A like there is
for products, and I is not terminal, so no A |- I in general. Those two should
probably take care of the no-contraction, no-weakening rules. Symmetric
monoidal categories mean A (x) B ~= B (x) A, though, so you still get the
exchange rule.

Obviously a lot of connectives are missing above, but I don't know the
categorical analogues off the top of my head. Searching for 'closed monoidal'
or 'symmetric monoidal closed' along with linear logic may be fruitful,
though.

-- Dan

_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe