1
$\begingroup$

Consider an index set $I = \{1, 2, \ldots, n\}$. Let $V$ be a vector space and $v_I$ be a basis that is, $\{v_1, v_2, \ldots, v_n\}$. Let $\phi: I \rightarrow V$ be a map that maps the index set to a basis vector $i \xrightarrow[]{\phi} v_i$. Let $W$ be another $n-$dimensional vector space and $f: I \rightarrow W$ be any map. Then we have the following commutative diagram:

Commutative diagram

This looks like a universal property since: for all $W$ and $f$, there exists a unique linear transformation $T: V \rightarrow W$. This is true since $T$ is specified by where the basis elements $v_i$ go into $W$.

Now my question: if there is a universal property, there needs to be a certain category and a certain terminal object. What are these in this case?

It seems like the morphism $\phi$ (or really the pair $(I,\phi)$ is the initial object in a category where the objects are morphisms in the ambient diagram and morphisms are the commutative diagrams.

If the ambient category is $\text{Vect}$ where $V,W$ are objects but $I$ is not... I am confused.

$\endgroup$
2
$\begingroup$

This is indeed a universal property, and we can indeed find a category where $(V, \phi)$ is initial.

The category that you're looking for is a certain kind of comma category.

Here's a direct definition.

Objects are pairs $(U, g)$, where $U$ is a vector space and $g$ is an ordinary function $g : I \to U$.

An arrow from $(U, g)$ to $(W, h)$ is a linear map $T : U \to W$ wuch that $h = T \circ g$.

Composition is just ordinary composition of functions. It's easy to verify the axioms of a category from here.

Then we see that the claim "$V$ is the free vector space on $I$" is simply stating that $(V, \phi)$ is the initial object.

This is related to notions of the representability of a functor and of adjoints.

Edit: if you insist that a universal property be stated in terms of a terminal object, just take the opposite category of the one I have described, which makes $(V, \phi)$ into a terminal object.

$\endgroup$
4
  • 1
    $\begingroup$ And if the OP insists that a universal property should always be interpreted as a terminal object in a certain category (this is a silly convention, but it seems to be implicit in the question), you can just consider the opposite of the category you described in your answer. $\endgroup$ Sep 28 at 20:08
  • 1
    $\begingroup$ Thanks.. got it.. very helpful.. just for clarity I used terminal to mean either initial or final.. but was useful to help with the notion of the opposite category. $\endgroup$ Sep 28 at 20:56
  • $\begingroup$ @MathDilettante Terminal object and final object are synonyms. If you use the word terminal to describe an initial object, you'll just confuse people. $\endgroup$ Sep 28 at 23:44
  • $\begingroup$ @AlexKruckman I see that now.. thx $\endgroup$ Sep 29 at 10:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.