5
$\begingroup$

There is a notion of 'oidification' in category theory which characterises many object versions of mathematical objects. For example:

  • magmas $\rightarrow$ magmoids

  • loops $\rightarrow$ loopoids

  • groups $\rightarrow$ groupoids

  • rings $\rightarrow$ ringoids

And in reverse, one object magmoids are magmas and so on. This is also referred to as horizontal categorification in contrast to vertical categorification.

NLab mentions that heaps have a many object oidification: heapoids. But do not give an explicit characterisation. It's not immediately obvious to me what this is, unlike the mathematical objects mentioned above, heaps are characterised by a ternary operation.

Hence I'm just looking for pointers to a definition, preferably accessible online as I don't have access to a math library.

$\endgroup$
5
  • 5
    $\begingroup$ You missed monoids --> monoidoids :) Sorry I couldn't resist. $\endgroup$ Commented Aug 20, 2022 at 20:31
  • $\begingroup$ @Benjamin Steinberg: I've heard there are things called andr->androids! ;-). $\endgroup$ Commented Aug 20, 2022 at 20:53
  • $\begingroup$ planetmath.org/heap1 --- see remark number 2 $\endgroup$ Commented Aug 20, 2022 at 21:27
  • $\begingroup$ @CarloBeenakker: Nice find. But I think this is an example of a concept being used in two different ways. What PlanetMath is calling a heapoid is what Nlab calls a Mal'cev Operation and its also called this by Springer's Encyclopedia of Maths. For Nlab, a heapoid should be a many object generalisation of a heap and which is what I'm interested in. But the generalisation that PlanetMath is naming by the same concept is merely a generalisation that comes about by dropping a condition. Obviously if you drop a condition then specialising the object will not reintroduce that condition. In other ... $\endgroup$ Commented Aug 20, 2022 at 21:41
  • 1
    $\begingroup$ @CarloBeenakker: ... words I'm not convinced that this is what I am looking for. Still, thanks for going to the trouble of locating it. $\endgroup$ Commented Aug 20, 2022 at 21:42

1 Answer 1

8
$\begingroup$

The claim that heapoids exist was added to the nLab page in revision 3 by Toby Bartels, so you could ask him what he had in mind.

I can speculate that a heapoid would have

  • a set of objects.
  • families of morphims $f:x\to y$.
  • for any $f:x\to y$, $g:z\to y$, and $h:z\to w$, a ternary composite $t(h,g,f) : x\to w$ (note the reversal of direction in $g$).
  • for any $f:x\to y$ and $g:x\to z$, we have $t(g,f,f) = g$.
  • for any $f:x\to y$ and $g:z\to y$, we have $t(g,g,f) = f$.
  • for any $f:x\to y$, $g:z\to y$, $h:z\to w$, $k:u\to w$, and $\ell : u\to v$, we have $t(\ell,k,t(h,g,f)) = t(t(\ell,k,h),g,f)$.

Then any groupoid would become a heapoid with $t(h,g,f) = h\circ g^{-1}\circ f$, just as any group becomes a heap in a similar way.

$\endgroup$
3
  • $\begingroup$ Thanks. I can now add that to my collection of oids! $\endgroup$ Commented Aug 21, 2022 at 6:19
  • $\begingroup$ This look quite reasonable. But it wasn't me who added it to the nLab; that was Zoran Škoda in revision 2. What I did in revision 3 was to fix a related spelling typo. $\endgroup$ Commented Jul 1 at 21:49
  • $\begingroup$ By the way, not only is every groupoid a heapoid, but another sign that this works is that every object in a heapoid has an endmorphism heap. And another sign is that every heapoid has a structure groupoid, although this one is trickier. (A groupoid morphism from $ x $ to $ y $ should be a pair of morphisms $ f \colon x \to z $ and $ g \colon y \to z $, but two such morphisms should be equal if mediated by an isomorphism $ z \sim z ' $; yet such isomorphisms are what we are trying to define!) $\endgroup$ Commented Jul 1 at 21:55

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .