New essay: The Framework Told Us Our Test Was Wrong

true
essay
methodology
category theory
type theory
calibration
falsifiability
A new essay on where category theory and type theory paid in the A=1 program, and where they did not. The lesson was a corrected criterion, not a new result: a lattice with only adjacency has no built-in distance, so judging its geometry against a round ruler smuggles in a measurement no inside-observer can make. Only disagreement between sectors is physical. Methodological; it asserts nothing about which geometry the framework will settle on.
Author

Jack D. Menendez

Published

July 29, 2026

There is a new essay on the site — The Framework Told Us Our Test Was Wrong — a retrospective on where category theory and type theory earned their keep in the A=1 program, and where they plainly did not.

It starts personally: a copy of Russell’s Principia bought young and never read, and Type Theory — the tool Russell was forced to invent to escape his own paradox — finally opening that door fifty years late, only to hand back a correction within a week. The correction is the spine of the piece. We had written a test that measured whether the lattice treats every direction the same by asking whether it came out round. But a lattice with only adjacency has no Euclidean distance; its natural “circle” is the taxicab unit ball, which is a diamond. In a world whose only distance is the taxicab, a polygon is a circle — so a round ruler was never ours to hold, and calling the lattice lopsided smuggled in a measurement no observer inside it could ever make. The corrected criterion is sharper: absolute isotropy is untestable, because a distortion shared by light, matter, rulers and clocks is invisible. Only disagreement between sectors is physical.

From there the essay keeps an honest ledger. Category theory paid on exactly one axis — what counts as observable, separating real structure from artifacts of description (it is what forced the correction above). It did not do the geometry, which is ordinary combinatorics, linear algebra and group theory. Along the way it pairs with the calibration discipline: a dimensionless prediction has no dial, so it cannot be tuned to agree — falsifiability immune, by construction, to “you fitted it.”

It is a methodological piece, in the program’s usual lane: a claim about how we work, not about what the universe is — and a companion to the optical-axis no-go, where the last correction came from the sky rather than from mathematics.

Read it: The Framework Told Us Our Test Was Wrong