finite projective planes have points and lines
Given a finite projective plane that contains a quadrangle OXYZ (i.e. nothree of these four points are on one line). To prove: the plane has points and lines for some integer , and there are points on each line and lines through each point.
Let and be the lines OX andOY, which must exist by the axioms. By the assumption OXYZ is a quadranglethese lines are distinct and Z is not on them. Let there be points X on other than O, for each of them one line ZX exists, and is distinct(one lines cannot pass through two X unless it is but that’s not aline through Z). Conversely every line through Z must intersect in aunique point (two lines intersecting in Z cannot intersect at another point,and Z is not a point on ). So there are lines through Z (OZ is oneof them). By the same reasoning, using , there are lines through Zso . We also found points (including O) on and the same numberon . Intersecting the lines through Z with XY (on which Z does notlie, the quadrangle again) reveals at least distinct points there andat most because for each point there there is a line through it and Z.
The lines not through O intersect in one of the points X and in one of the points Y. There are possibilities and each of themis a distinct line, because there is only one line through a given X andY. The lines that do pass through O intersect XY in one of the points there, again one line for each such point and vice versa. That’s lines through O and not through O, in all.
There are lines through X (to each of the points of ) and lines through Y (to each of the points of ). Intersect the linesthrough X other than XY with the lines through Y other than XY, these intersections are all distinct because for any P there’s only one linePX and one line PY. Note we did not use the line XY. Conversely for any P noton XY there must be some PX and some PY, so there are exactly points noton XY. Add the points on XY for a total of .
The constructions above already showed lines through some points (X, Yand Z), by the same games as before that implies for each of them pointson every line not through that point. We also saw points on some lines(, , XY) which implies for each of them lines through every pointnot on that line. Such reasoning covers items on first application andrapidly mops up stragglers on repeated application.
Some form of this proof is standard math lore; this version was half remembered and half reconstructed.