The studies — Image Structure Reader

Study 01 · painting

Matisse, La Leçon de piano

Henri Matisse, 1916

The foundational read: figure/ground instability and single-channel figures. The first calibration pole.

Critique Lab book in study folder

Study 02 · painting

Degas, Dancers, Pink and Green

Edgar Degas, c.1890

Dissolution as order. A measured anomaly, externally confirmed as a documented figure — the instrument's premise demonstrated. Source of the accent-scale corrections (I-5, I-7).

Critique Lab book

Study 03 · painting

Pollock, Convergence

Jackson Pollock, 1952

The all-over field. The phrase-grammar battery, and the resolution-sweep discipline (I-10) tested against dense fill.

Critique Lab book

Study 04 · painting

Caravaggio, The Calling of Saint Matthew

Caravaggio, 1599–1600

Scholarship treated as falsifiable counter-tests. The dark-reproduction noise findings (I-11), and the strongest claim-discipline record in the corpus.

Critique Lab book

Study 05 · calligraphy

Soejima Taneomi, Two-Line Calligraphy

Meiji period, ink on satin

First non-Western, non-painting object. A controlled comparator-corpus design; the substrate and ablation findings (I-12/13/14). Stratigraphy and R_spatial were born here.

Critique Lab book

Study 06 · painting

Klimt, Mäda Primavesi

Gustav Klimt, 1912–13

First high-key work. The address channel was born here, and so was the null-control test built against it (I-15): a portrait whose command apparatus is maximal while its mass and chroma are displaced to the ground.

Critique Lab book

Study 07 · painting

Bruegel (attr.), Fall of Icarus

after Pieter Bruegel the Elder

Distributed order versus dissolution. The displaced-subject question — a figure that is 0.18% of the canvas — answered with I-3/I-4 discipline rather than a size-confounded ratio.

Critique Lab book

Study 08 · photography

Cartier-Bresson, Behind the Gare Saint-Lazare

Henri Cartier-Bresson, 1932

The third domain. Pre-registration held on the first test. The seed-chaos correction to the null tool was earned here (I-16/17).

Critique Lab book

Study 09 · generated

Sora pair, Little Red

Default vs Steered, born-digital

The fourth domain, and the epistemically cleanest case: the pixel array is the artifact, no reproduction chain. One prompt family spanned the corpus's full R_spatial range — arrangement, not engine statistics, sets that axis.

Critique Lab book

Study 10 · generated

Midjourney ensemble, Little Red

Base vs Steered, 2×3×8 draws

The corpus's first draw-variance measurement — claims at cohort level, in the register's first cohort-range rows. Showed the centred-emblem basin is conditional on engine and prompt, not a property of generated images as a class.

Critique Lab book

Study 11 · painting

Velázquez, Las Meninas

Diego Velázquez, 1656

The deepest shadow-key in the corpus, on a composition that buries its nominal subjects. Earned the annular and dark-on-dark zone-read limits (I-18/19) and the third null family — a figure-position shuffle that withdrew a verdict the phase-scramble had passed.

Critique Lab book

Study 12 · painting

Malevich, Black Square

Kazimir Malevich, 1915

The adversarial limit: an edge-and-gradient instrument aimed at a painting built to have near-zero internal structure. Five control rounds retired four first-draft "records" (I-20), separating a clean ~1.7° hand-tilt from the century's decay. Earned the silent-zeroing fault I-23.

Critique Lab book

Study 13 · painting

Rothko, No. 31 & Green on Maroon

Mark Rothko, 1958 & 1961 · in copyright

The corpus's first matched pair, and the exact counter-pole to Malevich: all interior and no edge. No. 31 has thirty countable interior bodies; Green on Maroon has none.

Critique Lab book