I was reading Fulton and Harris' discussion of exterior and symmetric products as quotient spaces of tensor products in their rep theory book when I noticed that they made this claim (the emphasis is mine):

The exterior powers $ \bigwedge^n V $ and symmetric powers $ \operatorname{Sym}^n V $ can also be realized as subspaces of $ V^{\otimes n} $, assuming, as we have throughout, that the ground field has characteristic 0.

Why must the ground field be characteristic zero for this to be true? In particular, I don't see any problem with the quotient construction of the exterior and symmetric products for a vector space with nonzero characteristic; if there is no problem with this construction, then the inclusion of these products into the tensor product ought to be well-defined as well. Am I missing something obvious here?


The usual construction of embedding the symmetric power into the tensor power involves dividing by $n!$. This can't always be done in positive characteristic (in fact there are only finitely many $n$ where it is valid).

  • $\begingroup$ Thank you. It's pretty apparent there were some silly assumptions I made - in particular, composing the natural inclusion and the natural projection does not necessarily create a projection map into the image of the inclusion. Also, the projection map into the exterior and symmetric subspaces must be idempotent by definition, which necessitates division by $ n! $. $\endgroup$
    – yepikhodov
    May 31 '15 at 23:52
  • $\begingroup$ @yepikhodov You're welcome. $\endgroup$ Jun 1 '15 at 2:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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