Universal Kernel

reduction

Which absences are the same fact seen twice?

An absence reduces to another when the book proves it from the other without reproving it; the reductions sort the absences into three clusters that meet only through bridges.

the group of order 168the octonionsHilbert spaceat 7its dimension partDickson’s listsign of μKlein’s listGalois’s window{5, 7, 11}spinor window{3, 5}binary stabilizers2T, 2O, 2Ispin bundle nontrivialfor p = 7, 11PSL(2,7) not in SO(3)no real formbelow dimension 6PSL(2,7) ≅ PSL(3,2)Sylow structureof PSL(2,r)plane and line meetonly at (2,7)Fano’s axiomFano plane onlyin characteristic 2Hurwitz boundsolvable torus groupsminimal genus 3doubling lemmaHurwitz’s theoremno octonionic projectivespace beyond the planeJordan stopat h3(O)AdamsBott–Milnor–Kervaireoverlap lemmaGleasonKochen–SpeckerPSL(3,4) ≇ A8alone: element orders only1344 → 168 automorphisms, builtGleason’s 3 and the 3 of PG(2,2), nameBloch sphere = celestial sphere, built
Plate 3.11The absences and the reductions between them, in three clusters, with the bridges that are their only meetings.
Definition(Reduction, common source)

An absence AA reduces to an absence BB, written B⇒AB\Rightarrow A, if the book proves AA from BB by an argument that does not reprove the content of BB. Two absences have a common source CC if both reduce to CC. When no reduction in either direction and no common source is known, the two are unrelated here.

In classical logic every theorem implies every other, so the relation has to mean something finer. These relations describe proofs, not truth, and unrelated here is a statement about present knowledge, not a theorem of independence.

Remark(Clusters)

The reductions fall into three clusters. The group of order 168: Galois’s window, the spinor window and the spin bundle, the real form, the plane and the line, Fano’s axiom and the Hurwitz bound, with common sources in Dickson’s list and the sign of μ=2h−2+∑(1−1/mi)\mu=2h-2+\sum(1-1/m_i). The octonions: Hurwitz’s theorem, the stop at the plane and the Jordan algebras, with common source the doubling lemma. Hilbert space: Gleason and Kochen–Specker, with common source the overlap of contexts. The separation of PSL⁡(3,4)\PSL(3,4) from A8A_8 stands alone.

Between clusters no reduction is known. They meet through bridges, each with its status: the 1344 automorphisms of the octonions that permute the units ±ei\pm e_i induce all 168 automorphisms of the Fano plane of the units, with a kernel of 8 sign changes, a built bridge; the Bloch sphere and the celestial sphere are one P1(C)\Proj^1(\C), a built bridge; and the three of Gleason’s threshold and the three of PG(2,2)\mathrm{PG}(2,2) share only a name.

Proposition(Gleason implies Kochen–Specker)

For each d≥3d\ge3, Gleason’s theorem implies the Kochen–Specker theorem.

Proof

A valuation would be a frame function, hence of the form e↦⟨e,ρe⟩e\mapsto\langle e,\rho e\rangle, which is continuous on the unit sphere. The sphere is connected, so a continuous function with values in {0,1}\{0,1\} is constant; but on any context it takes the value 1 once and the value 0 at least once.

Example

Two reductions meet at the group of order 168 and are complementary: Galois’s window at p=7p=7 produces the double life PSL⁡(2,7)≅PSL⁡(3,2)\PSL(2,7)\cong\PSL(3,2), and the Sylow structure of PSL⁡(2,r)\PSL(2,r) excludes every other isomorphism between a plane group and a line group. Neither uses the other.

Fano’s axiom reduces to the statement that the Fano plane lives only in characteristic 2. So the double life joins characteristic 2, the plane, to characteristic 7, the line, and neither side can be moved: the plane side by Fano’s axiom, the line side because r(r2−1)/gcd⁡(2,r−1)=168r(r^2-1)/\gcd(2,r-1)=168 only for r=7r=7.

In the volume
XIVWhy Octonions