2

I am reading An Introduction to Algebraic Topology by Rotman. After proving

Theorem 2.7: For every $k\geq 0$, euclidean space$\mathbb R^n$ contains $k$ points in general position,

the book remarked: There are other proofs of this theorem using induction on k. The key geometric observation needed is that $\mathbb R^n$ is not the union of only finitely many proper affine subsets . I want to prove that observation.

1 Answers1

2

Every proper affine subspace of $\mathbb{R}^n$ is contained in an affine hyperplane.

Now, let $C=\{(t,t^2,t^3,\ldots,t^n):t \in \mathbb{R}\}$.

Since every nonzero polynomial with degree at most $n$ has at most $n$ roots, an affine hyperplane of $\mathbb{R}^n$ contains at most $n$ points in $C$.

So if you have a covering of $\mathbb{R}^n$ by proper affine subspaces, you need at least $|\mathbb{R}|$ of them.

Alex Ortiz
  • 26,211
Aphelli
  • 37,929
  • If we fix an affine hyperplane $H$, what is the polynomial you are referring to? – Alex Ortiz Aug 06 '24 at 15:15
  • 1
    @AlexOrtiz: the affine hyperplane $H$ is defined by the equation $a_1x_1+a_2x_2+\ldots+a_nx_n=b$. Then $C\cap H$ consists of the points $(t,t^2,\ldots,t^n)\in C$ such that $-b+a_1t+a_2t^2+\ldots +a_nt^n=0$. This is in bijection with the set of (real) roots of the polynomial $-b+a_1X+a_2X^2+\ldots+a_nX^n$. – Aphelli Aug 06 '24 at 20:57
  • Thank you for clarifying! – Alex Ortiz Aug 06 '24 at 23:03