This is a question of mathematical nature. Let the Lorentz group $O(1,3)$ be defined as a matrix group.

$$\text{O}(1,3) =\{\Lambda\in M_4 (\mathbb{R})| \Lambda^T \text{diag}(+---) \Lambda = \text{diag}(+---)\} $$

One defines the supremum norm on $M_4 (\mathbb{R})$ as:

enter image description here

which turns $M_4 (\mathbb{R})$ into a topological space in the metric topology induced by the norm. Can this supremum norm be made particular for Lorentz matrices? I guess it would, right? It would immediately follow that the Lorentz group is a topological space in the subspace topology of $M_4 (\mathbb{R})$.


1 Answer 1


Yes, $\mathrm{O}(1,3)$ is a topological subspace of $M_4(\mathbb{R})$ under this construction.

That said, there's nothing special about the topology induced by supremum norm in this regard; any subset $S$ of a topological space $X$ inherits a natural subspace topology from its "parent" space. You could define some weird topology on $M_4(\mathbb{R}) \sim \mathbb{R}^{16}$ which is a product topology between $n$ copies of $\mathbb{R}$ with the indiscrete topology and $16-n$ copies of $\mathbb{R}$ with the discrete topology, and it would still define a topology on $\mathbb{O}(1,3)$.

[EDIT: my original answer discussed whether the supremum norm might imbue the resulting topological space with non-physical structure; but as Valter Moretti pointed out in a comment, all norm topologies on a finite-dimensional vector space such as $\mathbb{R}^{16} \sim M_4(\mathbb{R})$ are equivalent, so this point is somewhat moot.]

  • $\begingroup$ Thank you for your answer/confirmation. The trick is that the (west-coast) metric tensor of Minkowski space is quite poor, or let us say not intuitive at defining a meaningful topology by open sets. At least this is what I make of this: physics.stackexchange.com/q/228669 $\endgroup$
    – DanielC
    Oct 25, 2017 at 14:42
  • $\begingroup$ "By using this norm, you may be imbuing the space of matrices with a structure that isn't physically meaningful. This may or may not be problematic depending on what you do next in your derivation. " Related to this part, once the topology is defined, I can show that the Lorentz group is a non-compact topological group, by having a further subset of it (the set of Lorentzian boosts in one direction), endowed with the subspace topology, homeomorphically mapped into $\mathbb{R}$. $\endgroup$
    – DanielC
    Oct 25, 2017 at 15:07
  • 1
    $\begingroup$ All norms on $\mathbb R^n$ produces the same topology. So it happens in particular on $M(m, \mathbb R)$ and their subsets. Using either the operator norm as in OP's post or the standard Euclidean norm leads to rhe same topology... $\endgroup$ Oct 25, 2017 at 18:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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