2
$\begingroup$

In all introductory texts I'm aware of, convex polytopes are dealt with strictly within $\mathbb R^n$. Being used to quite larger levels of generality elsewhere, I'm wondering if the same can be done here. Specifically, I'm looking for a generalization of convex polytopes in other spaces such that their face lattices still satisfy most of the properties they do in $\mathbb R^n$, like being graded lattices and satisfying the diamond property.

An obvious setting in which convex polytopes work are topological vector spaces of finite dimension. Unfortunately all of these are isomorphic to $\mathbb R^n$, so nothing new.

Interestingly, it seems like most if not all of the basic definitions of the theory can be made in terms only of a betweenness relation. Convex sets/hulls, affine subspaces/hulls, and closed or open half-spaces can be defined without even a single axiom attached to this relation. Of course, very little can actually be proved about these notions, and I don't know what axioms would be needed to get "expected behavior".

I believe hyperbolic space might satisfy my constraints. However I see no obvious way to generalize this example.

$\endgroup$
2
  • 3
    $\begingroup$ There is a large body of literature on this. Look for "convex structures". For instance: elsevier.com/books/theory-of-convex-structures/van-de-vel/… $\endgroup$ Feb 18, 2022 at 15:50
  • 1
    $\begingroup$ I think ther should be quite some literature on spherical and hyperbolic polytopes. All you need is the notion of a halfspace and you can define a polytope as an intersection of finitely many such, and you can add boundedness or finite volume (which is not the same) if required. As far as I know, spherical and hyperbolic polytopes are not richer in combinatorics than usual convex polytopes. But people study e.g. hyperbolic simplices with certain interior angles and ask when these are bounded etc. $\endgroup$
    – M. Winter
    Feb 18, 2022 at 16:27

1 Answer 1

1
$\begingroup$

Some time ago I was having the same question and I found these two:

  • Robert Morelli "A theory of polyhedra" (Adv. Math. 97, No. 1, 1-73 (1993)). He works in modules over ordered rings (and vector spaces over ordered fields). This was very helpful personally to understand valuations on polyhedral algebras in a more general context. zbMATH review here.
  • Walter Prenowitz and James Jantosciak "Join geometries. A theory of convex sets and linear geometry" (Springer UTM). zbMATH review here. I haven't looked at this book in detail, but seems interesting (and perhaps related to your "betweenness relation"). The second paragraph from its introduction says:

The familiar figures of classical geometry $-$points, segments, lines, planes, triangles, circles, and so on$-$ stem from problems in the physical world and seem to be conceptually unrelated. However, a natural setting for their study is provided by the concept of convex set, which is comparatively new in the history of geometrical ideas. The familiar figures can then appear as convex sets, boundaries of convex sets, or finite unions of convex sets. Moreover, two basic types of figure in linear geometry are special cases of convex set: linear space (point, line, and plane) and halfspace (ray, halfplane, and halfspace). Therefore we choose convex set to be the central type of figure in our treatment of geometry. How can the wealth of geometric knowledge be organized around this idea? By definition, a set is convex if it contains the segment joining each pair of its points; that is, if it is closed under the operation of joining two points to form a segment. But this is precisely the basic operation in Euclid. Our point of departure is to take the operation of joining two points to form a segment as fundamental, and to throw the burden of unifying the material on the consistent and relentless exploitation of this operation.

$\endgroup$
1
  • $\begingroup$ The latter book seems really interesting. It does indeed confirm my suspicion that you can get quite far by taking betweenness (actually, a "join" operation) as a primitive and tacking on some simple postulates. $\endgroup$
    – ViHdzP
    Feb 19, 2022 at 8:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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