4
$\begingroup$

I'm currently reading "Bordism of Elementary Abelian Groups via Inessential Brown-Peterson Homology" by Hanke (arXiv:1503.04563) and have come across some notation that I'm not familiar with. Let $\phi: \mathbb{Z}/3 \rightarrow (\mathbb{Z}/3)^2$ be a group homomorphism and $S^5$ the five dimensional sphere in $\mathbb{C}^4$. $\phi$ induces an action of $\mathbb{Z}/3$ on $(\mathbb{Z}/3)^2$ and $S^5$ comes with a natural $\mathbb{Z}/3$ action, the standard action from the lens space construction. The paper then writes $$(\mathbb{Z}/3)^2 \times_{\mathbb{Z}/3} S^5$$ How is this defined? Furthermore, it is stated that this comes with a $(\mathbb{Z}/3)^2$ action.

As a check, it should be defined so that $$((\mathbb{Z}/3)^2 \times_{(\mathbb{Z}/3)^2} (S^{2m_1+1}\times S^{2m_2+1})/(\mathbb{Z}_3)^2 \cong L^{2m_1+2}_3\times L^{2m_2+1}_3$$ is a product of lens spaces.

$\endgroup$
17
  • $\begingroup$ A blind guess from category theory: it is the pullback of $(\mathbb{Z}/3)^2$ and $S^5$ along the relevant morphisms into $\mathbb{Z}/3$. $\endgroup$
    – Alec Rhea
    Commented Jul 25, 2023 at 22:02
  • 2
    $\begingroup$ This is a guess, but I think it's a sort of "tensor". You have groups $H \subseteq G$ and $H$ acting on a set $X$. The object you want is the Cartesian product $G \times X$ modulo the relations $(gh, x) \sim (g, hx)$ for all $g, h, x$. The $G$–action acts on the first coordinate. $\endgroup$
    – Vik78
    Commented Jul 25, 2023 at 22:06
  • $\begingroup$ @Vik78 This is what I was thinking too. Then we would think of $(\mathbb{Z}/3)^2\times_{\mathbb{Z}/3} S^5$ as a disjoint union of lens spaces? $\endgroup$
    – Noah B
    Commented Jul 25, 2023 at 22:17
  • 2
    $\begingroup$ This means you take the Cartesian product and mod out by the diagonal action. i.e. it is three copies of $S^5$. The action of $(\mathbb{Z}/3)^2$ basically just permutes the three factors, and does the lens space action, independently. If you look at Tom Dieck's book on transformation groups (equivariant homotopy theory, etc) you will see this is standard notation in the field. $\endgroup$ Commented Jul 25, 2023 at 23:23
  • 3
    $\begingroup$ It might also mean mod out by the anti-diagonal action, so that you are identifying $(gn,p)\sim (g,np)$, rather than $(gn,np)\sim (g,p)$; the former is more like a tensor product than the latter, and it is what one would use for making an associated bundle. Here the associated bundle is one with fibre $(\mathbb{Z}/3)^2$ over the lens space, rather than the $\mathbb{Z}/3$-bundle given by the quotient map. $\endgroup$
    – David Roberts
    Commented Jul 26, 2023 at 0:50

1 Answer 1

3
$\begingroup$

This has been more or less said in the comments, but it seems like someone should write an answer.

If $H$ is a subgroup of $G$ and $H$ has a left action on a space $X$, then a space $G\times _HX$ is defined, a quotient space of $G\times X$, by declaring $(gh,x)\sim (g,hx)$. Note that we are using the right action of $H$ on the space $G$. (If we like, we can create a left action on $G\times X$ using the right action on $G$ and the left action on $X$, writing $h(g,x)= (gh^{-1},hx)$. But this is unnecessary.)

Let's denote the equivalence class of $(g,x)$ by, say, $[g,x]$. The quotient space gets a left $G$-action, defined by $g_1[g_2,x]=[g_1g_2,x]$. (This is well-defined because the left action of $G$ on (the space) $G$ commutes with the right action: $[g_1(g_2h),x]=[(g_1g_2)h,x]=[g_1g_2,hx]$.

In categorical terms: The subgroup $H\subset G$ (or the homomorphism $H\to G$) is giving us a functor $X\mapsto G\times_HX$ from the category of spaces with left $H$-action to the category of spaces with left $G$-action. This functor is left adjoint to the forgetful functor.

This is very similar to how a homomorphism $R\to S$ of (associative) rings gives a functor $X\mapsto S\otimes_R X$ from $R$-modules to $S$-modules, again left adjoint to the forgetful functor.

$\endgroup$
3
  • $\begingroup$ Thanks, Tom. I was just coming here to say I should write something out. $\endgroup$
    – David Roberts
    Commented Jul 27, 2023 at 22:46
  • $\begingroup$ Maybe I should have let you do it. $\endgroup$ Commented Jul 28, 2023 at 0:13
  • $\begingroup$ I don't mind ! :-) $\endgroup$
    – David Roberts
    Commented Jul 28, 2023 at 4:10

You must log in to answer this question.

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