Generalizing Pappus' Theorem6 September 2000 We define the following binary operation on pairs of points in the projective
plane. Given x=(X1,X2)
and y=(Y1,Y2),
we define STUPID PICTURE It is easily checkable that this operation is non-associative, abelian,
and has the additional property that Now choose three unrelated pairs of points x, y, and
z.
The following applet displays the orbit of points when repeatedly applying The following applet displays orbits of several iterations of this construction.
The red, blue, and magenta colored points signify the pairs (X1,X2), (Y1,Y2),
and (Z1,Z2).
By clicking ``SPHERE'', you can view a hemispherical projection of the
plane onto the sphere (which is then projected onto a circle) that gives
you a global view of the nature of the orbits.
APPLET
It turns out that every such orbit lies on a degree three polynomial, a cubic. This of course needs to be prooven, but perhaps the applet makes it believable. Anyway, this can be prooven with some intense algebra and I am currently looking for a better proof. ![]() We know that 9 points in the plane determine a cubic polynomial runing through them. Let C be a cubic curve (a degree three projective algebraic variety). This gives us the following theorem: Theorem 1 Given a degree three projective algebraic variety C, and an octahedron with vertices lying on C. If any three of this octahedron's opposing edges intersect on C then all six of the octahedron's opposing edges intersect on C. Proof: The six vertices of the octahedron determine three pairs of points in the plane (each point is paired with its opposite). Now we can perform our construction on these pairs, call them x, y, and z. As a consequence of what was discovered above, each point in the pairs Now we have the tools to prove Pappus' Theorem and Pascal's Theorem
quite simply.Theorem 2 (Pappus' Theorem) Given points
X1, Y1, and Z1 on
a line l and points X2, Y2,
and Z2 on a line m, the points Pascal's Theorem can also be shown to follow directly from this result. |
||