3
$\begingroup$

I'm looking for a name (or references or search terms) for the following construction.

In a category $\mathcal C$, let a doodle mean a pair $\langle D,\Omega\rangle$ where $\Omega$ is an object and $D$ is a set of morphisms with codomain $\Omega$ (but potentially different domains).

Let a whatsit over a doodle $\langle D,\Omega\rangle$ mean a triple $\langle(\varphi_d)_{d\in D}, A, \psi\rangle$ where $A$ is an object of $\mathcal C$, $\varphi_d$ is a $D$-indexed family of morphisms with codomain $A$, and $\psi$ is a morphism $A\to\Omega$, such that $\varphi_d\psi = d$ for each $d\in D$.

Given a doodle, the whatsits over that doodle are the objects of a category where a morphism from $\langle (\varphi)_d,A,\psi\rangle$ to $\langle (\varphi'_d),A',\psi'\rangle$ is a $\mathcal C$-morphsim $\beta:A\to A'$ such that $\varphi_d\beta=\varphi'_d$ for all $d$ and $\psi=\beta\psi'$.

An initial object of the category of whatsits is a greatest common factor of the doodle. Intuitively, the $\psi$ part of the initial whatsit is the "longest" arrow that is a right factor of every morphism in the doodle.

"Doodle" and "whatsit" are (obviously) my own temporary names.

Question: Do these things have names already? Is it possible to describe them within the framework Wikipedia presents for a universal property?

It would be possible to say that a doodle is simply a whatsit whose third component is an identity morphism (modulo some new way to index the first component of each whatsit), but it not clear to me whether that is useful.

$\endgroup$
  • 1
    $\begingroup$ @goblin: In this analogy a pullback would be a "smallest common multiple" -- the arrows we start with are all factors of the pullback. I need the converse: something that is a factor of each of the initially given arrows. $\endgroup$ – Henning Makholm Jan 24 '15 at 3:27
  • 1
    $\begingroup$ @goblin: Hmm... I think that whatsits may be objects in a slice (not coslice) category over the category of cocones of a discrete diagram. $\endgroup$ – Henning Makholm Jan 24 '15 at 3:34
  • 2
    $\begingroup$ Yep, I think I get you. Given a pair of arrows $f,g : A,B \rightarrow \Omega$ with common codomain, I think you're interested in the coproduct of $f$ and $g$ viewed as objects of the slice category of $\Omega$. When $f$ and $g$ are injective functions, this is the "union" of $A$ and $B$. $\endgroup$ – goblin Jan 24 '15 at 3:34
  • 1
    $\begingroup$ @goblin: Hm, yes, that sounds about right, and less convoluted than my attempt to wrap my head around it. I'd be inclined to accept that as an answer. $\endgroup$ – Henning Makholm Jan 24 '15 at 3:43
  • 1
    $\begingroup$ Okay, I've posted my understanding of the construction. Perhaps something better will come along though. I would Google "colimits in slice categories" in the search for further information. $\endgroup$ – goblin Jan 24 '15 at 3:50
4
$\begingroup$

If I understand correctly:

Fix a category $\mathbf{C}$ and an object $\Omega$ in this category. A $\mathbf{J}$-shaped doodle to $\Omega$ is just a $\mathbf{J}$-shaped diagram in the coslice category $\mathbf{C} \downarrow \Omega$. A whatsit to a doodle $D$ is a cocone to $D$; hence the "greatest common factor" of $D$, being the initial whatsit to $D$, could also be described as the colimit of $D$.

$\endgroup$
  • 1
    $\begingroup$ Okay, so what I'm looking for is simply the coproduct of all the arrows in $D$, viewed as objects of $\mathcal C\downarrow \Omega$. That's much simpler than I could have come up with myself. Thanks, $\endgroup$ – Henning Makholm Jan 24 '15 at 12:36
  • 1
    $\begingroup$ @HenningMakholm, glad I was able to help you for a change :) $\endgroup$ – goblin Jan 24 '15 at 12:44

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.