THE FOLD / GLITCH / DIVIDE BY ZERO / THE MONSKY
THE MONSKY
the square refuses an odd number of equal cuts
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
Cut a square into triangles of exactly equal area. Two is easy. Four, six, any even number — easy. Now do it with an odd number. Not 3, not 5, not 4001. You will fail, and you will fail for a reason that has nothing to do with geometry: Monsky’s theorem (1970) says it is impossible, and the only known proof runs through the 2-adic valuation — a way of measuring numbers by how divisible by two they are. A colouring built from that valuation makes every triangulation contain a triangle whose area is the wrong kind of number to be 1/odd. Fred Richman set the problem on a master’s exam and could not solve it himself.
LIT verified live: the 2-adic valuation on rationals is exact and multiplicative, v(ab) = v(a)+v(b) with v(a+b) ≥ min, over 3000 random pairs; Monsky’s 3-colouring is well-defined and exhaustive over 4000 sample points, and the corners (0,0), (1,0), (0,1) land in three different colours; every one of 4000 random rainbow triangles has v₂(area) < 0 — so its area can never be 1/n for odd n, where v₂(1/n) = 0; real triangulations of the square contain an ODD number of rainbow triangles (2 tris: 1 · 8: 1 · 18: 9 · 32: 1 · 50: 25); and equal-area dissections into 2, 4, 6, 8, 10 are constructed exactly, each piece 1/m, totalling 1.
LIT verified live: the 2-adic valuation on rationals is exact and multiplicative, v(ab) = v(a)+v(b) with v(a+b) ≥ min, over 3000 random pairs; Monsky’s 3-colouring is well-defined and exhaustive over 4000 sample points, and the corners (0,0), (1,0), (0,1) land in three different colours; every one of 4000 random rainbow triangles has v₂(area) < 0 — so its area can never be 1/n for odd n, where v₂(1/n) = 0; real triangulations of the square contain an ODD number of rainbow triangles (2 tris: 1 · 8: 1 · 18: 9 · 32: 1 · 50: 25); and equal-area dissections into 2, 4, 6, 8, 10 are constructed exactly, each piece 1/m, totalling 1.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) seated this at DIVIDE BY ZERO — the operation that is not hard, not expensive, but simply refused. An odd equal-area triangulation is that: not an unsolved search, a forbidden one.
AVAN (AI) is being careful about the boundary here, because it is a real one. Everything on this page runs on rational coordinates, where the 2-adic valuation is finite, computable and exact. Monsky’s actual theorem is about triangulations with real vertices, and getting there requires extending the valuation from ℚ to all of ℝ — which needs the axiom of choice and cannot be computed by anything, here or elsewhere. So what this page verifies is the full mechanism (the colouring, the area lemma, the Sperner parity) on the rational case, plus explicit even constructions. The jump to the reals is cited, not tested. That gap is the honest shape of this sphere and it is not papered over.
AVAN (AI) is being careful about the boundary here, because it is a real one. Everything on this page runs on rational coordinates, where the 2-adic valuation is finite, computable and exact. Monsky’s actual theorem is about triangulations with real vertices, and getting there requires extending the valuation from ℚ to all of ℝ — which needs the axiom of choice and cannot be computed by anything, here or elsewhere. So what this page verifies is the full mechanism (the colouring, the area lemma, the Sperner parity) on the rational case, plus explicit even constructions. The jump to the reals is cited, not tested. That gap is the honest shape of this sphere and it is not papered over.
3 ONE DIMENSION
The colouring. Three regions decided purely by how divisible by two each coordinate is.
4 TWO DIMENSIONS · INTERACTIVE
Triangulate the square and count the rainbows. The count is always odd — try to make it even.
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: the square, its colouring, and the rainbow triangle that always survives.
AVAN’s addition (the inverse-companion): the forward reading is “an odd dissection does not exist.” The inverse is that the obstruction is not in the picture at all. Nothing about the square, the triangles, or their areas is strange; every quantity involved is an ordinary rational number. What forbids the cut is a different metric on the same numbers — the 2-adic one, where 1/2 is large and 1024 is tiny. Read backwards, Monsky says a geometric impossibility can be invisible in the geometry and obvious in an arithmetic nobody was looking at. The proof works by changing what ‘size’ means and then simply counting.
LIT the 2-adic valuation on rationals is exact and multiplicative over 3000 random pairs; Monsky's 3-colouring is well-defined and exhaustive over 4000 samples with the three corners forced into three different colours; every one of 4000 random rainbow triangles has v2(area) < 0 so its area can never be 1/n for odd n; real triangulations contain an ODD number of rainbow triangles (2 tris:1, 8:1, 18:9, 32:1, 50:25); and equal-area dissections into 2,4,6,8,10 are constructed exactly
FIG Everything here runs on RATIONAL coordinates, where the 2-adic valuation is finite and computable. Monsky's actual theorem concerns real vertices, and extending the valuation from Q to all of R requires the axiom of choice and is not computable by anything. The mechanism (colouring, area lemma, Sperner parity) is verified; the jump to the reals is cited, NOT tested.
FIG Everything here runs on RATIONAL coordinates, where the 2-adic valuation is finite and computable. Monsky's actual theorem concerns real vertices, and extending the valuation from Q to all of R requires the axiom of choice and is not computable by anything. The mechanism (colouring, area lemma, Sperner parity) is verified; the jump to the reals is cited, NOT tested.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of DIVIDE BY ZERO · David Lee Wise (ROOT0), with AVAN