According to the Universality Theorem by Mnev (see below theorem 8.6.6 from [1]), for any open semialgebraic variety V there is a uniform oriented matroid of rank 3 whose realization space is stably equivalent to V.

My question is whether one can transfer the result also to uniform oriented matroids of rank 4 and higher, i.e., for a fixed r $\ge$ 4 and for any open semialgebraic variety V is there a uniform oriented matroid of rank r whose realization space is stably equivalent to V?

8.6.6 Universality Theorem (Mnev 1988).

  1. Let $V \subset \mathbb{R}^s$ be any semialgebraic variety. Then there exists a rank 3 oriented matroid M. whose realization space R(M) is stably equivalent to V.
  2. If V is an open subset of $\mathbb{R}^s$, then M may be chosen to be uniform.

[1] A. Björner, M. L. Vergnas, B. Sturmfels, N. White, and G. M. Ziegler, Oriented Matroids. Cambridge University Press, 1999.


Yes, the Universality Theorem works for any rank $r\geq3$. One way to prove this is to consider a rank $3$ oriented matroid $M$ with the desired realization space, then take the dual $M^*$, do a lexicographic extension in general position and then take the dual again. Repeat this $r-3$ times.

Lexicographic extensions preserve the realization space up to stable equivalence, and so does taking the dual. A lexicographic extension increases the number of elements by one and keeps the rank. Hence, doing it to the dual increases the rank of the primal by one.

  • $\begingroup$ Thanks for the answer, Arnau! I have two follow-up questions. Will the procedure you suggested run in polynomial time? Is the procedure a known technique for which there is a literature reference? Thanks! $\endgroup$ – Jae Nov 17 '14 at 14:50
  • 1
    $\begingroup$ Yes, this should run in polynomial time. Lexicographic extensions are explained in section 7.2 of [1]. In Lemma 8.2.1 and Proposition 8.2.2 in the same book it is proven that the realization space of a lexicographic extension is preserved up to homotopy equivalence. The proof for stable equivalence is similar. You can also have a look at arxiv.org/abs/1402.7207. $\endgroup$ – Arnau Nov 18 '14 at 11:51
  • $\begingroup$ Really cool and simple trick, thx @Arnau ! I had something more complicate in mind but i think equivalent. $\endgroup$ – Nikolai Mnev Mar 31 at 14:27

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.