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

Hi all. If n>m and I have a continuous map from an n-sphere to an m-sphere, is there a simple way to see whether the map is homotopic to a constant? (For example, is it if and only if the map is not onto? -- this would be probably too simple). For n=m, the answer is "nonzero degre" -- is there a reasanable generalisation of "degree" to different dimensions? Can you recommend me some literature about it? Thanks, Peter

share|cite|improve this question
As you suggest, if the map misses a single point then its image lies in a disc inside the sphere, and contracting the disc to a point gives a null-homotopy. The converse is not true - even if n=m, there are degree zero maps that are surjective. This is a very difficult and interesting problem. In general, calculating the homotopy groups of spheres is one of the central problems of algebraic topology. – Jeffrey Giansiracusa Dec 30 '10 at 12:24
Hi Jeffrey. Thanks for your comment. I know that homotopies of spheres are not known in general, but I thought that there might be some tool for recognizing whether the map is homotopically trivial. – Peter Franek Dec 30 '10 at 12:40
For $n<m$ there are continuous onto maps $S^n \to S^m$, yet all such maps are null-homotopic. What would you consider a reasonable answer in this case, Peter? i.e. would it be acceptable to just say all maps are null homotopic so you don't need to know anything about a particular map $S^n \to S^m$? The problem with that is you've accepted computing the homotopy groups of spheres as a prerequisite to your problem, if you go that direction. So there's nothing simple like degree. – Ryan Budney Dec 30 '10 at 17:48
There is something similar to degree theory but it rarely helps you decide whether or not a given map $S^n \to S^m$ is null, see:… – Ryan Budney Dec 30 '10 at 18:02
My problem comes from informatics. I want to create a computer algorithm that decides, whether f=0 has solution. The function f is assumed to be "nice", from an n-cube (product of n intervals) to Rm, and we can over-approximate its image on small cubes arbitrary precise. I'm interrested in "robust" solution, i.e. such that f+g=0 has still solution for g small. I proved that f=0 has a robust solution iff the map f/|f| from the boundary of a neighborhood of a component of {f=0} to the sphere is homotopically nontrivial. Should I give up any hope, for n>m, to find such an algorithm?(probably yes) – Peter Franek Jan 27 '11 at 23:26

The set of maps $S^n \to S^m$ is better known as $\pi_n(S^m)$. The one other (than n=m) relatively easy case is when we tensor by the rational numbers (and thus answer the question "when is f homotopic to g after adding each to itself some number of times?). The answer in this case is they are always homotopic if m is odd or if m is even and $n \neq 2m - 1$. If $m$ is even and $n = 2m-1$ then $f$ and $g$ are homotopic in this weak sense if and only if they both have non-trivial (or both have trivial) classical Hopf invariant.

share|cite|improve this answer

Every map from a sphere to a finite connected CW complex is homotopic to a surjective one. Collapse one half of the sphere to a line segment, use your original map on the sphere created by the collapse, and map the segment surjectively onto the complex (which can be done by the Hahn-Mazurkowicz Theorem).

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.