1

Currently, I am studying smooth manifolds and I want to solve some exercises. There is a question that says:

Show that the dimension of a connected topological manifold is defined without ambiguity! Meaning that if $\dim M=n$, with the change of charts, the dimension still is $n$. Then show that this is true for a $C^r$ connected manifold using Inverse Function Theorem.

To prove the first part, let $(U,\varphi)$ and $(V,\psi)$ be two charts with $\varphi(U) \subseteq \mathbb{R}^m$ and $\psi(V) \subseteq \mathbb{R}^n$ and suppose that $U \cap V \neq \emptyset$. Then, since $\psi \circ \varphi : \varphi (U \cap V) \rightarrow \psi (U \cap V) $ is a homeomorphism, $m=n$. Now, using this, if $M$ is a connected topological manifold with atlas $\mathcal{A}$, for all charts $(U_{\alpha}, \varphi_{\alpha})\in \mathcal{A}$, $\varphi_{\alpha}(U_{\alpha}) \subseteq \mathbb{R}^n$ and $n$ is constant for all the charts. Am I right?

But I have a problem showing the second part! I think that since every $C^r$ manifold is also a topological manifold, the answer to this question would be trivial! Why do we need "Inverse Function theorem" to answer it??

Any help is appreciated.

  • 2
    Why do we need "Inverse Function theorem" to answer it?? We don't. That doesn't make much sense to be honest. Also and n is constant for all the charts. requires a proof. This is typically done by considering $f:M\to\mathbb{R}$, $f(x)=\text{local dimension around }x$ and proving that $f$ is continuous (and so constant if $M$ is connected). What you've proven is that $f$ is well defined (i.e. independent on the choice of atlas). – freakish May 31 '19 at 07:59
  • What seems strange to me about the exercise is that proving the statement for topological manifolds seems much harder than the case of differentiable manifolds. I think you need algebraic topology for the topological case, while only linear algebra is needed for the differentiable case. – JWL May 31 '19 at 08:02
  • @JWL Algebraic topology to prove that the dimension is well defined on connected topological manifolds? What are you talking about? (see my previous comment) – freakish May 31 '19 at 08:07
  • @freakish Yeah, you're right. I'll write it down in detail. But do you have any ideas about the second part? – Zeno cosini May 31 '19 at 08:08
  • @JWL No, we don't!!!! The first part is as I've written. – Zeno cosini May 31 '19 at 08:09
  • @Yasinowski No, I don't. I don't see how inverse function theorem applies here. Are we talking about applying IFT to chart map? As in manifold version of IFT? This looks like a big artillery, you need notions of tangent space, etc. Can you even prove generalized IFT without knowing that the dimension is constant? And even if we do have it, I don't see how it brings us closer to the solution. Imo: this is bs. – freakish May 31 '19 at 08:12
  • @freakish I was saying that "Then, since $\psi\circ\phi^{-1}$ is a homeomorphism, $m=n$" is not trivial at all. It is related to something like this – JWL May 31 '19 at 08:13
  • @freakish I don't think so... We need to use the normal version of IFT. Let $(U,\varphi)$ and $V,\psi)$ be two $C^r$ charts, it means that $\psi\circ \varphi ^{-1}: \mathbb{R}^m \rightarrow \mathbb{R}^n$ is a $C^r$-diffeomorphism. I think I have to apply IFt to these chart maps... – Zeno cosini May 31 '19 at 08:19
  • @JWL Ah, I got you now. Ok, so this is the hidden detail. But I still don't see how the inverse function theorem makes that easier or provable at all. Especially since $\psi\circ\phi^{-1}$ already is a homeo/diffeo morphism. – freakish May 31 '19 at 08:20
  • @JWL we don't need algebraic topology for invariance of dimension and invaraince of domain etc. Brouwer did it without. The algebraic topology proofs developed partly from his ideas. – Henno Brandsma May 31 '19 at 15:04
  • Hi could you find any solution for the second part of your question? Please share with me I have also the same assignment: I need to prove with IFT. –  Dec 08 '21 at 02:20
  • @L.G. Is that anything you are not satisfied about the answer below? It's not clear what is the purpose of this bounty. – Arctic Char Dec 09 '21 at 14:09

3 Answers3

4

Your argument about overlapping chart sets uses invariance of domain: the fact that a homeomorphism between open sets of $\mathbb{R}^m$ and $\mathbb{R}^n$ exists implies $n=m$ is quite non-trivial. It's actually not quite invariance of domain, but the related fact that topological dimension (there are several equivalent definitions for separable metric spaces) is a topological invariant and $\dim(\mathbb{R}^n)=n$ (where $\dim$ is one of those equivalent topological dimension functions; topologically speaking the main fact we need for that is Brouwer's fixed point theorem.) This can be applied for simple topological manifolds, as all we need is homeomorphisms in the charts.

Now the chart open sets of the manifold form an open cover and so if $M$ is connected we can apply the chain characterisation of connectedness to conclude that the local dimension is constant: take any points $x$ of local dimension $m$ and $y$ of local dimension $n$. So there are finitely many chart open sets $U_1, \ldots, U_k$ such that $x \in U_1$, $y \in U_k$ and $U_i \cap U_{i+1} \neq \emptyset$ for all $i \in \{1,\ldots,k-1\}$. By the "overlapping chart sets implies equal dimension" argument you also gave and which is, as said, justified by non-trivial topological results on Euclidean spaces, we conclude that the local dimension of all $U_i$ is $m$: $U_1$ because of $x$, $U_2$ because it intersects $U_1$, $U_3$ because it intersects $U_2$ up to $U_k$ because it intersects $U_{k-1}$ (or make it a finite induction). So $m=n$ because local dimension is well-defined.

For $C^r$ manifolds we don't need the inverse function theorem, we know all charts maps are homeomorphisms by definition (or maybe you need the inverse function theorem to see that if it's not part of the definition?). You just fall back to the fact on topological manifolds.

Henno Brandsma
  • 250,824
  • The first part is right. But if we need to prove the second part without the topological manifolds, we need IFT. I will post my answer ASP! Thanks any way... – Zeno cosini May 31 '19 at 13:39
  • @Yasinowski What is the definition of $C^r$ manifolds that it is not a topological manifold? – Henno Brandsma May 31 '19 at 14:39
  • $C^r$ manifolds are definitely topological manifolds. But I need proof using IFT. – Zeno cosini May 31 '19 at 14:55
  • @Yasinowski don't get fixated on that. A simpler proof is often a better proof. – Henno Brandsma May 31 '19 at 15:02
  • 1
    In the differentiable case, one can construct the 'tangent space' and maps between manifolds induce vector space maps between tangent spaces. This easily implies the well-defined-ness of the dimension since dimensions of vector spaces can be defined in an elementary way. This argument might use IFT in some implicit way, as many arguments in basic differentiable manifold theory does, but I don't think so. – JWL May 31 '19 at 15:29
4

Show that the dimension of a connected topological manifold is defined without ambiguity. Meaning that if $\dim M=n$, with the change of charts, the dimension still is $n$. Then show that this is true for a $C^r$ connected manifold using Inverse Function Theorem.

Let us start showing the dimension of a connected topological manifold is defined without ambiguity.

First note the following lemma.

Lemma 1: If $U$ is an open set of $\Bbb R^n$, $V$ is an open set of $\Bbb R^k$ and there is a homeomorphism $h : U \to V$ then $n=k$.

This lemma is a consequence of the Invariance of Domain Theorem (which is a theorem in algebraic topology).

Now let $\{(U_\alpha,\phi_\alpha)\}_\alpha$ be a topological atlas of $M$. For each $\alpha$, $\phi_\alpha : U_\alpha \subset M \to \phi(U_\alpha) \subset \mathbb{R}^{n_\alpha}$ is a homeomorphism.

Let us define a function $f: M \to \Bbb N$, in the following way: given any $p\in M$ there is $\alpha$ such that $p \in U_\alpha$, define $f(p) = n_\alpha$.

If there is $\beta$ such that $p \in U_\beta$, then $\phi_\beta \circ \phi_\alpha^{-1}$ is a homeomorphism from $\phi_\alpha(U_\alpha \cap U_\beta) \subset \Bbb R^{n_\alpha}$ onto $\phi_\beta(U_\alpha \cap U_\beta) \subset \Bbb R^{n_\beta}$, so, by the lemma above, $n_\alpha=n_\beta$. So $f$ is well defined.

Now, for all $n \in \Bbb N$, if $p \in f^{-1}(\{n\})$ then there is $\alpha$ such that $p \in U_\alpha$, $n = n_\alpha$ and $U_\alpha \subseteq f^{-1}(\{n\})$. So, for all $n \in \Bbb N$, $f^{-1}(\{n\})$ is open.

Suppose there are $n, m \in \Bbb N$ such that $n\neq m$ and $n, m \in \text{Image}(f)$. Then $f^{-1}(\{n\})$ and $f^{-1}(\Bbb N \setminus \{n\})$ are non-empty disjoint open sets covering $M$. But this contradicts the connectness of $M$. So $f$ must be constant.

The only value of $f$ is the dimension of $M$. This completes the proof of the first part.

Now, since a $C^r$ connected manifold is a connected topological manifold, it follows immediately that the dimension of a $C^r$ connected manifold is defined without ambiguity.

The second part consists in showing that in the case of a $C^r$ connected manifold, we can prove that the dimension of the manifold is defined without ambiguity, without using the lemma 1 above (and so without using the Invariance of Domain Theorem and algebraic topology).

The proof is very similar to the proof above. We are going to use the following lemma.

Lemma 2: If $U$ is an open set of $\Bbb R^n$, $V$ is an open set of $\Bbb R^k$ and there is a $C^r$ diffeomorphism $h : U \to V$ then $n=k$.

Lemma 2 is a consequence of the Inverse Function Theorem. In fact, by such theorem, we know that at each point of $p\in U$, $h'(p)$ is an invertible linear transformation from $\Bbb R^n$ onto $\Bbb R^k$. So, $n=k$.

Now let $\{(U_\alpha,\phi_\alpha)\}_\alpha$ be a $C^r$ atlas of $M$. For each $\alpha$, $\phi_\alpha : U_\alpha \subset M \to \phi(U_\alpha) \subset \mathbb{R}^{n_\alpha}$ is a $C^r$ diffeomorphism.

Let us define a function $f: M \to \Bbb N$, in the following way: given any $p\in M$ there is $\alpha$ such that $p \in U_\alpha$, define $f(p) = n_\alpha$.

If there is $\beta$ such that $p \in U_\beta$, then $\phi_\beta \circ \phi_\alpha^{-1}$ is a $C^r$ diffeomorphism from $\phi_\alpha(U_\alpha \cap U_\beta) \subset \Bbb R^{n_\alpha}$ onto $\phi_\beta(U_\alpha \cap U_\beta) \subset \Bbb R^{n_\beta}$, so, by the lemma 2 above, $n_\alpha=n_\beta$. So $f$ is well defined.

Now, exactly as above, we prove that, for all $n \in \Bbb N$, $f^{-1}(\{n\})$ is open and then we prove that $f$ is be constant.

The only value of $f$ is the dimension of $M$. This completes the proof of the second part.

Ramiro
  • 19,684
0

If $M$ is a smooth manifold of dimensions $m$ and $n$ then for each $p\in M$, we can find charts about $p$ with images open in $\mathbb{R}^n$ and $\mathbb{R}^m$. Hence $T_pM$ is isomorphic to $\mathbb{R}^n$ and $\mathbb{R}^m$