The datum of a group is strictly more than the datum of a group up to isomorphism

124 Views Asked by At

In a discussion regarding Mochizuki's work on the ABC conjecture which can be found here, Peter Scholze stated that

"the datum of a group is strictly more than the datum of a group up to isomorphism"

Explicitly, in what way is this true? Can someone give an example of a group satisfying the statement and explain precisely how one loses information up to isomorphism?

1

There are 1 best solutions below

2
On BEST ANSWER

Here is an example from some old research of mine. I was (habitually) identifying groups up to isomorphism, and it made me miss a very obvious fact. The idea, out of context, is this:

If we think of $\mathbb{Z}^n$ as the group of integer lattice points in $\mathbb{R}^n$, then a sublattice $\Lambda$ will actually be isomorphic to $\mathbb{Z}^n$ as a group, but will represent something entirely different!

Here is a conrete example:

Consider $\mathbb{Z}^2 \leq \mathbb{R}^2$, shown in black. Then $\mathbb{Z}^2 \cong \Lambda \leq \mathbb{Z}^2$ is a sublattice (shown as red circles) carries different geometric information, even though the groups are (abstractly) the same.

two isomorphic but nonequal lattices

Long story short, me identifying the two different embeddings of $\mathbb{Z}^2$ in $\mathbb{R}^2$ led to a headache that was quickly resolved when I finally realized that there's more to a group than its isomorphism class.


I hope this helps ^_^