The geometric shape in question is a compound of two polytopes: an 11-hypercube with edge length $2$ and an 11-simplex with edge length $\sqrt6$ whose vertices are a subset of the hypercube’s. What is the structure of this shape’s symmetry group?

Edit: proof that it’s possible: https://www.math.uchicago.edu/~may/VIGRE/VIGRE2011/REUPapers/Markov.pdf

And explicit coordinates:

(–1,–1,–1,–1,–1,–1,–1,–1,–1,–1,–1) (1,–1,1,–1,–1,–1,1,1,1,–1,1) (1,1,–1,1,–1,–1,–1,1,1,1,–1) (–1,1,1,–1,1,–1,–1,–1,1,1,1) (1,–1,1,1,–1,1,–1,–1,–1,1,1) (1,1,–1,1,1,–1,1,–1,–1,–1,1) (1,1,1,–1,1,1,–1,1,–1,–1,–1) (–1,1,1,1,–1,1,1,–1,1,–1,–1) (–1,–1,1,1,1,–1,1,1,–1,1,–1) (–1,–1,–1,1,1,1,–1,1,1,–1,1) (1,–1,–1,–1,1,1,1,–1,1,1,–1) (–1,1,–1,–1,–1,1,1,1,–1,1,1)

  • 3
    $\begingroup$ There is no such simplex. $\endgroup$ Feb 3, 2021 at 1:27
  • $\begingroup$ Are you sure about that? $\endgroup$ Feb 3, 2021 at 2:19
  • $\begingroup$ Wouldn't that hypercube have edges of length $2$, not $1$? $\endgroup$ Feb 3, 2021 at 2:39
  • $\begingroup$ Oops, you’re right. $\endgroup$ Feb 3, 2021 at 2:53
  • 1
    $\begingroup$ In fact, a regular simplex of dimension $n$ can be inscribed in a hypercube of dimension $n$ if and only if $n+1$ is the order of a Hadamard matrix. If we also want every automorphism of the simplex to extend to an automorphism of the hypercube, then the $n=2^m-1$ result is correct. $\endgroup$ Feb 3, 2021 at 13:55

1 Answer 1


The automorphism group of this configuration $C'$ is the Mathieu group $M_{11}$.

Firstly, we construct a larger configuration $C$ consisting of a 12-dimensional orthoplex inscribed in a 12-dimensional hypercube.

In particular, the orthoplex in $C$ consists of the vectors $\{ \pm v : v \textrm{ is a column of } H \}$, where $H$ is a 12-by-12 Hadamard matrix whose first row is $(1, 1, \dots, 1)$. The hypercube in $C$ simply consists of all $2^{12}$ vectors whose entries are $\pm 1$.

The automorphism group of $C$ contains the order-2 centre $\{\pm I\}$, modulo which it is isomorphic to the Mathieu group $M_{12}$. This is described here in terms of the [unique up to isomorphism] order-12 Hadamard matrix:


The two different 12-element permutation representations of $M_{12}$ correspond to permuting:

  1. The 12 'coordinate axes' of the hypercube;
  2. The 12 pairs of opposite vertices of the inscribed orthoplex;

The subgroup $K$ that fixes a particular one of the 12 coordinate axes also has the order-2 centre $\{ \pm I \}$, modulo which it is isomorphic to $M_{11}$. This follows from $M_{11}$ being describable as the stabiliser of a point in the permutation representation of $M_{12}$.

Indeed, $K$ factors as a direct product $C_2 \times M_{11}$, where the first factor indicates whether the first coordinate axis is flipped or not, and the second factor indicates the permutation of the remaining 11 coordinate axes.

Consequently, $M_{11}$ is precisely the subgroup of the symmetry group of $C$ which fixes the unit vector $v_1 = (1, 0, \dots, 0)$.

Finally, note that the OP's configuration $C'$ can be described as the restriction of $C$ to the 11-dimensional hyperplane $\{ x . v_1 = 1 \}$. (The 11-simplex is obtained as a facet of the 12-orthoplex, and the 11-hypercube is a facet of the 12-hypercube.) Another construction of $C'$ is to take the vertices of the simplex to be the columns of the rectangular matrix $H'$ obtained by deleting the first row from $H$; the vertices of the hypercube are again the $\pm 1$-vectors.

It follows, therefore, that the symmetry group of $C'$ is $M_{11}$. This has an 11-element permutation representation (permuting the coordinate axes of the hypercube) and a 12-element permutation representation (permuting the vertices of the simplex).

  • $\begingroup$ Doesn't the orthoplex have dimension 11, not 12? $\endgroup$ Feb 3, 2021 at 14:00
  • $\begingroup$ No, $C$ is a 12-dimensional orthoplex (24 vertices) inscribed in a 12-dimensional hypercube (4096 vertices). The codimension-1 configuration $C'$ is an 11-dimensional simplex (12 vertices) inscribed in an 11-dimensional hypercube (2048 vertices). $\endgroup$ Feb 3, 2021 at 15:20

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.