Removing a redundant field in CartesianClosed/Canonical

[MiddleAdjunction]

In the canonical definition of CCCs, the field curry-resp-≈ is redundant and can be defined as a function in the record. I personally cannot see a reason justifying the redundancy but it might be there to improve definitional equality in some way (similar to introducing the redundant sym-assoc as opposed to a function using assoc), in which case there possibly needs to be a comment nearby explaining such reason.

[sstucki]

LGTM! When I wrote this module I didn't realize this assumption was redundant. I also don't see any good reason to keep it.