3

1

Are there higher-dimensional versions of the concept of rotationally symmetric Venn diagrams, with closed curves replaced by closed surfaces or higher manifolds ?

flag

2 Answers

0

"four intersecting spheres form the highest order Venn diagram that is completely symmetric and can be visually represented"

http://en.wikipedia.org/wiki/Venn_diagram#Extensions_to_higher_numbers_of_sets

link|flag
1 
The same page shows a two-dimensional Venn diagram of five ellipses that is rotationally symmetric, so I think that sentence is referring to a stronger condition. – Jonah Ostroff Sep 27 2011 at 22:22
The two-dimensional diagram is only symmetric up to a reflection, no? – Will Sawin Sep 27 2011 at 23:48
@Will I think Jonah was talking about this one which has some rotational symmetry: upload.wikimedia.org/wikipedia/commons/1/10/… – psd Sep 28 2011 at 22:20
3

The n-simplex as a subset of ${\mathbb R}^{n+2}$ can serve as a stand-in for a Venn diagram. Or if you like you can fatten the vertices until you have over-lapping $n$-balls.

More precisely, consider the convex hull of ${e_1, e_2, \ldots , e_{n+1}}$. This is the set $\{ \vec{x} \in R^{n+2}: \sum_i x_i = 1, \ \ \& \ \ 0\le x_i \}$ where $e_i$ represents the $i$th standard unit vector $[0,\ldots, 0,1,0,\ldots, 0]$ that has a $1$ in the $i$th position. Each vertex represents one of your sets. Each edge represents the intersection between two sets. Each triangle represents the intersection among $3$ sets, and each $k$-simplex (which is determined by a subset of size $k+1$ chosen from ${1,2, \ldots , n+2}$ represents the intersection among $(k+1)$ sets.

To imagine a model use a protractor and draw the complete graph on $(k+2)$-vertices that are represented by the roots of unity $e^{2\pi i j/(k+2)}$ (The $i$ here is different than the index above).

link|flag
1 
Specifically, you can choose a ball of radius $\sqrt(2)$ around each point in the hyperplane that goes through $e_1,e_2,...,e_{n+1}$ – Will Sawin Sep 27 2011 at 23:50
1 
The trick is to do this with $2^{n+1}$ distinct regions (counting the exterior region) in a dimension (potentially much) lower than $n+2$. At least that would be the analogue of the well-known 2-dimensional problem that has been solved in the case of $n$ prime by Griggs-Killian-Savage. – Patricia Hersh Jul 3 at 1:25

Your Answer

Get an OpenID
or

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