THE FOLD / GLITCH / DIVIDE BY ZERO / THE SCHUR COMPLEMENT
THE SCHUR COMPLEMENT
a determinant split by a block
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
The Schur complement is what remains of a block matrix after you eliminate one block. For M = [[A, B], [C, D]] with A invertible, the Schur complement of A is S = D − C A−1 B — the effective D once A’s influence is folded in. It splits the determinant cleanly: det(M) = det(A) · det(S). It also gives the block inverse in closed form, and it decides definiteness (M is positive-definite iff A and S both are). It is the algebra behind Gaussian elimination on blocks, Kalman updates, and Gaussian conditioning.
LIT verified live: over thousands of random block matrices, det(M) equals det(A)·det(S), and the block inverse built from the Schur complement satisfies M·M−1 = I (window.__schur). FIG no framing; the block determinant split and the inverse both run in-browser.
LIT verified live: over thousands of random block matrices, det(M) equals det(A)·det(S), and the block inverse built from the Schur complement satisfies M·M−1 = I (window.__schur). FIG no framing; the block determinant split and the inverse both run in-browser.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) seated this at divide-by-zero — the Schur complement lives on dividing by the A block (A−1); it is exactly where you may divide, provided A is nonsingular. AVAN (AI) built the instrument: block extraction, the S = D − CA−1B computation, and the determinant and inverse checks.
Credit as content: Issai Schur (the Schur complement; named by Emilie Haynsworth, 1968). The weave: David names the divide; I confirm the determinant factors as det(A)det(S) and the block inverse is exact.
Credit as content: Issai Schur (the Schur complement; named by Emilie Haynsworth, 1968). The weave: David names the divide; I confirm the determinant factors as det(A)det(S) and the block inverse is exact.
3 ONE DIMENSION
M in four blocks; eliminating A leaves the Schur complement S = D − CA−1B, and det(M) = det(A)·det(S).
4 TWO DIMENSIONS · INTERACTIVE
A random block matrix; det(M), det(A), det(S) are shown with det(M) = det(A)det(S), and the block inverse checked against M.
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: the Schur complement, the reduced problem after eliminating A.
AVAN’s addition (the inverse-companion): don’t invert the whole matrix — eliminate a block. The inverse of ‘solve M all at once’ is ‘fold A out, and the leftover S carries the rest, with det(M) = det(A)det(S).’ Magenta is the eliminated A block; green is the Schur complement it leaves. Divide out a block.
LIT Genuine Schur complement (Issai Schur; named by Emilie Haynsworth, 1968): S=D−CA⁻¹B, det(M)=det(A)det(S). Verified live: over 3000 random block matrices, det(M)=det(A)·det(S) (window.__schur.detFormula) and the block inverse satisfies M·M⁻¹=I (.blockInverse).
FIG No framing: block extraction, the S=D−CA⁻¹B computation, and the determinant/inverse checks all run in-browser. The AVAN inverse is honest — folding A out so the leftover S carries the rest (with det(M)=det(A)det(S)) rather than inverting M whole is exactly block elimination; magenta is the A block divided out, green the Schur complement it leaves. Divide out a block.
FIG No framing: block extraction, the S=D−CA⁻¹B computation, and the determinant/inverse checks all run in-browser. The AVAN inverse is honest — folding A out so the leftover S carries the rest (with det(M)=det(A)det(S)) rather than inverting M whole is exactly block elimination; magenta is the A block divided out, green the Schur complement it leaves. Divide out a block.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of DIVIDE BY ZERO · David Lee Wise (ROOT0), with AVAN