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

Let $E=\xi-\eta$ be a virtual vector bundle over a compact base $B$, which we may assume is a CW complex. A quick and dirty way to define the total Stiefel-Whitney class $w(E)\in H^\ast(B;\mathbb{Z}/2\mathbb{Z})$ would be to say that, morally, $E\oplus\eta\cong \xi$, and so we can use the Whitney product formula $w(E\oplus\eta)=w(E)w(\eta)$ to get $$w(E) = w(\xi)w(\eta)^{-1},$$ the element $w(\eta)\in H^\ast(B;\mathbb{Z}/2\mathbb{Z})$ being invertible since $B$ is compact.

I am interested in a more righteous definition, along the lines of Whitney's original definition for honest bundles in terms of obstructions to finding linearly independent sections over skeleta. I think that such a definition exists (maybe in terms of sectioning a certain Hom-bundle) but I wasn't able to find a reference. Hence I ask:

Can the Stiefel-Whitney classes of a virtual vector bundle be defined obstruction theoretically, and if so, is there a good reference in the literature describing this construction?

share|cite|improve this question

Given $\eta$, there is a trivial bundle $N$ such that $\eta\subset N$. Then $-\eta\sim N/\eta$, and the classes of $-\eta$ can be defined as the loci where a certain number of constant sections of $N$ stops being independent modulo $\eta$. This can certainly be restated more formally in terms of obstructions.

share|cite|improve this answer
Thank you for your answer! Still, I would like to know how this restatement goes. Do you know of any references? – Mark Grant Jan 14 '14 at 12:23
No, sorry; I take it as folklore :) You can use the ordinary Stiefel bundle associated with $N/\eta$, or you can probably define some kind of relative Stiefel manifolds (frames in $\mathbb{R}^N$ independent modulo a subspace), or something else like that. I guess this depends on the problem you are trying to solve. – Alex Degtyarev Jan 14 '14 at 13:13

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.