Start a new discussion

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

Site Tag Cloud

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

• CommentRowNumber1.
• CommentAuthorml
• CommentTimeJan 21st 2021
Hi to all.

In the real numbers field, when you know a function (from the reals to the reals) on the rationals and that this function is continuous (over the reals), then you know the function. In vector spaces, when you know an application on a basis and that this application is linear, then you fully know the application over the whole vector space.

My question(s) : is there a categorical explanation to this fact ? Can this be generalized ?

Marc
• CommentRowNumber2.
• CommentAuthorjesuslop
• CommentTimeJan 21st 2021

I am also interested.

• CommentRowNumber3.
• CommentAuthorUrs
• CommentTimeJan 21st 2021

For instance, in the category of Hausdorff topological spaces, inclusions of dense subspaces are epimorphisms. This captures your first example.

A similar example is: In the category of rings, the inclusion of the integers into the rational numbers is an epimorphism.

In order to capture your second example from a category-theoretic perspective, one would have to first cook up a suitable category where both vector spaces as well as plain sets are objects, in the first place. Once some such choice is made, we could again speak about injections that are epi.

In speaking of injections here, I am implicitly talking about concrete categories. More generally one could say that you are asking for examples of morphisms that are monomorphisms and epimorphisms, but not isomorphisms.

• CommentRowNumber4.
• CommentAuthorRichard Williamson
• CommentTimeJan 21st 2021
• (edited Jan 21st 2021)

From a different point of view, one could simply regard these as examples of universal properties: a vector space is free on its basis, the real numbers are the completion of the rationals, and both of these are universal properties (of a kind fitting into an adjunction, one half of which is a forgetful functor).

• CommentRowNumber5.
• CommentAuthorUrs
• CommentTimeJan 22nd 2021
• (edited Jan 22nd 2021)

That’s of course a good point and maybe closer to what “ml” might be looking for.

In any case, I have taken the occasion to start a list of examples of monos that are epi but not iso, here. Maybe it inspires somebody to add more examples.

• CommentRowNumber6.
• CommentAuthorml
• CommentTimeJan 24th 2021
Thank you for your replies. Richard, I think that your answer is what I was looking for.