MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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 : $C$ is a pure, full-dimensional polytopal complex(a special case of a regular cell complex) in $\mathbb{R}^d$. I know that the boundary of the underlying set is a PL-sphere. Is it true that $C$ is a PL-ball?

Definitions for this question:

  1. Polytopal complex is a finite nonempty collection of convex polytopes in $\mathbb{R}^d$ that contains all faces of its polytopes, and such that the intersection of two of its polytopes is a face of each of them.
  2. Dimension of the complex is the largest dimension of a polytope in the complex.
  3. A complex $C$ is pure if each of its faces is contained in a face of dimension $dim(C)$.
  4. Underlying set is the union of its faces.
  5. PL stands for piecewise-linear
  6. PL-k-ball is something that is PL-homemorphic to a simplex of $k$ dimensions.
  7. PL-(k-1)-sphere is something that is PL-homeomorphic to a boundary of a simplex of $k$ dimension.

(If the answer is yes for the above question) Further Question : Now we look at a non-full dimensional pure polytopal complex $C$. We define the "boundary" of $C$ to be the set of points where there does not exist a neighborhood that is PL-homemorphic to $\mathbb{R}^{dim(C)}$. Then if the "boundary" of $C$ is a PL-(dim(C)-1)-sphere, is $C$ a PL-dim(C)-ball?

P.S. Please give a comment if some parts of the question is not suitably defined. Also, I have no idea how hard this is, so please comment if you know it is obvious or is really hard.

9/19 : Edited so it includes the dimension restriction.

share|cite|improve this question

First question. Your question is equivalent to the Schoenflies problem: does a locally flat PL $(d-1)$-sphere in $S^d$ (and hence in $\mathbb R^d$) bound a PL-disc?

The answer is "yes" for $d \neq 4$; for a proof when $d > 5$ see the book Rourke C.P., Sanderson B.J. Introduction to piecewise-linear topology (Springer, 1972). It is a direct consequence of the Cobordism theorem and the Generalized Poincaré Conjecture in the PL setting. The answer in dimension $d=4$ is however unknown.

Second question. The answer here is "no": take a triangulated surface in $\mathbb R^3$ with one boundary component which is not a disc (for instance, a holed torus). Note also that one can choose a non-manifold example: for instance, a sphere with a disc attached to its equator.

share|cite|improve this answer
Thank you for the answer. 1. I need a bit of help with the answer for the second question: "The triangulated surface in $\mathbb{R}^3$ with one boundary component which is not a disc". Since the complexes I'm looking at are compact, I guess this would look something like a two-dimensional doghnut in $\mathbb{R}^3$. Then doesn't the boundary consist of the inner/outer circles and is not connected? 2. I want to try to modify the second question to make it work, so I hope you won't mind me not accepting the answer yet. – Suho Oh Sep 19 '10 at 13:43
By a "holed torus" I mean a standard torus surface in <math>\mathbb R^3</math> (a 2-dimensional object) with one small open disc removed: it is a compact surface with (1-dimensional) connected boundary. Another example of compact surface in 3-space with connected boundary which is not a disc is a Mobius strip. In fact, every compact surface with boundary embeds in 3-space. – Bruno Martelli Sep 19 '10 at 19:01

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.