MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Question: Is the hyperspace of the Hilbert cube $H=[0,1]^\mathbb {N}$ homeomorphic to $H$?

Remarks and definitions:

1) The Hilbert cube $H$ is a compact metric space, where the metric is given by the $\ell_2$-norm of sequences. A classical theorem on metric spaces says that every compact metric space is isometric to a closed subspace of $H$.

2) The hyperspace of a metric space $X$ is the metric space of all non-empty compact subsets of $X$ given by the Hausdorff metric. Another classical theorem on metric spaces says that the hyperspace of a compact metric space is again a compact metric space.

Combining 1) and 2) shows that the hyperspace of the Hilbert cube is isometric to a closed subspace of the Hilbert cube. So my question asks whether we also can get a homeomorphism (can we even get both spaces isometric?).

share|cite|improve this question
Isometric, unlikely. Show that the hyperspace fails the parallelogram law. – Gerald Edgar Nov 3 '12 at 20:12
Actually, the parallelogram law involves the affine structure as well as the metric, so we cannot do that directly. Maybe try this: there are two points in the hyperspace with non-unique midpoint. – Gerald Edgar Nov 4 '12 at 13:03
up vote 5 down vote accepted

The hyperspace of any Peano continuum (locally connected metric continuum) is homeomorphic to the Hilbert cube; this is a result of Curtis and Schori, see here. I learnt about this result in a paper of Torunczyk, where a different proof is given (that paper is also available online)

When it comes to asking about an isometry, I'm not sure which metric you put on the Hilbert cube...

share|cite|improve this answer
@Julien Thanks for the reference. I'm not officially asking the isometric version (using the $\ell_2$-norm, also see the comment of Gerald above), but I wanted to know the topological type of the hyperspace of $H$ in the first place. – Marcus Nov 3 '12 at 21:10
@Marcus You're welcome. I think for the isometric version Gerald Edgar suggestion looks good. – Julien Melleray Nov 3 '12 at 21:53

Rule out isometry ... In the hyperspace of $\mathbb R$, let $A=\{0,1,2\}$, $B=\{0\}$, $C=\{2\}$ and $M = \{1\}$. Then points $A, B, C$ all have distance $2$ from each other; they are the vertices of an equilateral triangle. But $M$ has distance $1$ from each of $A, B, C$, so $M$ is the midpoint of each of the three sides of that triangle. Not possible in Euclidean geometry.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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