Product is unique up to isomorphism

Written by Patrick Stevens last updated

Recall the universal property of the product:

Given objects and , we define the product to be the following collection of three objects, if it exists: with the requirement that for every object and every pair of maps , there is a unique map such that and .

We wish to show that if the collections and satisfy the above condition, then there is an isomorphism between and . [1]

Proof

The proof follows a pattern which is standard for these things.

Since is a product of and , we can let in the universal property to obtain:

For every pair of maps there is a unique map such that and .

Now let :

There is a unique map such that and .

Doing the same again but swapping for and for (basically starting over with the line "Since is a product of and , we can let …"), we obtain:

There is a unique map such that and .

Now, is a map which we wish to be the identity on ; that would get us halfway to the answer, because it would tell us that is left-inverse to .

But we can use the universal property of once more, this time looking at maps into :

For every pair of maps there is a unique map such that and .

Letting and , we obtain:

There is a unique map such that and .

But I claim that both the identity and also satisfy the same property as , and hence they're equal by the uniqueness of . Indeed,

  • certainly satisfies the property, since that would just say that and ;
  • satisfies the property, since we already found that and that .

Therefore is left-inverse to .

Now to complete the proof, we just need to repeat exactly the same steps but with and interchanged throughout. The outcome is that is left-inverse to .

Hence and are genuinely inverse to each other, so they are both isomorphisms and respectively.

The characterisation is not unique

To show that we can't do better than "characterised up to isomorphism", we show that the product is not characterised uniquely. Indeed, if is a product of and , then so is , where and . (You can check that this does satisfy the universal property for a product of and .)

Notice, though, that and are isomorphic as guaranteed by the theorem. The isomorphism is the map given by .

  1. ^︎

    I'd write and instead of and , except that would be really unwieldy. Just remember that and are both standing for products of and .