{
  "id": "C5",
  "by": "critic BC-P1",
  "scope": "sequence",
  "cites": [
    "main:3",
    "cast:1",
    "pier:2",
    "pier:3",
    "pier:5",
    "pier:10",
    "pier:11",
    "pier:12"
  ],
  "read": "Pier:11 took C4 and moved the window's right jamb in to 500 so the shade edge would clear the glass. It did clear the corner, and it landed the jamb on the pavement shadow's vertex instead. On pier:12 the window ends at x 500 and the pavement shade's corner sits at x 498 on row y 630 (xf(624) = 500 by the chunk's own formula), so the window's right edge drops straight onto the shadow's foot through the band of lit stucco below the sill. That is C1's main:3 coincidence (jamb and pavement wedge sharing x 620) re-made with a diagonal in between. Count the contacts: jamb (main:3), her contour and hem (cast:1), the paper's tip (pier:2), the window's corner (pier:10), the jamb over the vertex (pier:12). Since pier:3 the edge has been one ruled line, xf(y) = 660 - (y-96)*160/528, laid unchanged in pier:3, 4, 9, 10, 11 and 12, running from roofline to kerb. Every refine since has moved an object to clear it, and the moved object lands on it. Behind the line the shade is a single field: at x 780 it reads 0.51 to 0.53 from y 120 to y 480, identical on pier:5, pier:10 and pier:12. That wall has never been repainted, only the things set into it. With no light of its own, everything inside the field is a cut-out against one constant. The same trouble has hopped between objects: the paper as a pale card (main:3 0.62, pier:2 0.56), the dress as the darkest spot (pier:10 0.28), and now the door. Pier:11 glazed the door to 0.26 in black bands and took the paper down to 0.41. Pier:12 called both glazes over-strong and re-laid them opaque. The door landed at 0.23 (target 0.30), a near-black slab at a 0.29 step that now holds the darkest 1 percent of the picture (median x 907, y 373). The paper landed at 0.555 against its 0.52 wall, a pale card again on the fourth return. The new strip of wall from 500 to 545 reads 0.82 at chroma 0.23, lighter and more saturated than the lit stucco under the window (0.78, 0.15), so a bright ochre bar stands against the edge and points down to the vertex. The proposition's open question, wall or shadow, is answered as shadow. But the shadow is a stage flat: one straight wedge covering the whole right of the facade, top to street, with pier, figure and door inside it as props. The chair still holds the picture's maximum chroma (0.41 at 273, 565) on 0.84 pavement, and its sun still says only that it is empty.",
  "pressure_points": [
    "Fifth registration of the light edge: jamb x 620 (main:3), her contour and hem (cast:1), paper tip x 615 to 620 (pier:2), window corner at 524/540 (pier:10), and now the window jamb at x 500 directly over the pavement shade's corner at x 498, y 630 (pier:11 and pier:12). C4 asked for 20 units of clearance at the corner; pier:11 bought them by moving the jamb onto the line's foot. The edge has not changed since pier:3 (same xf in every chunk), so each repair is an object moving around a fixed line.",
    "The shade has no interior: 0.51 to 0.53 at every probe on x 780 from y 120 to y 480, unchanged across pier:5, pier:10 and pier:12. Pier, figure and door are set against one constant, so whatever was re-laid last becomes the spot. That was the paper (main:3, pier:2), then the dress (pier:9, pier:10), now the door at 0.23 holding the image's darkest 1 percent (pier:12).",
    "Pier:11 and pier:12 were two sittings spent correcting their own glazes inside that field. The opaque re-lay missed in both directions: door 0.23 against a 0.30 target, paper 0.555 against a 'just under the wall' target of below 0.52. The paper's return above its shade is now the fourth (0.62, 0.56, 0.53, 0.555).",
    "The new wall strip at 500 to 545 (pier:11) reads 0.82 at chroma 0.23, brighter and warmer than the lit stucco under the window (0.78, 0.15). It is a lit bar riding the edge, the brightest wall in the picture, aimed at the vertex.",
    "One ruled wedge from roofline to kerb divides the facade into two zones again, rotated from C1's vertical: lit left with window and chair, shaded right with pier, figure and door. The chair's sun is still only the other zone: maximum chroma (0.41) and hardest drawing on the brightest pavement (0.84), answering nothing but 'empty'."
  ],
  "action": "Destroy the single ruled edge and the flat field behind it, and lay the light again from a caster with a shape. A neighbour's roofline with a step, a cornice and a gap, a post or an awning edge: anything that breaks the boundary so it meets the facade, the sill band and the pavement at different x positions and never as one straight line ending at a vertex on the kerb. Let light into the shade. That might be a slit or patch of sun through the caster's gap landing on the pier wall or the pavement near her but not on her, the paper or the door. It might be warm reflected light from the lit pavement grading up the foot of the shaded wall. Either way the field gets a structure from top to bottom, and the paper, dress and door are judged against their local shade, not a 0.52 constant. Inverting the hour is also open: the pier in sun and the chair in shade, if the painting can say what that costs the pause. Match the chair's shadow to whatever caster you choose. Its sun may finally do work, for example by being the same sun that reaches her side, or it may be taken away. Geometry stands: the window at 40 to 500, the pier, the door, her profile and the chair's position. So the new light must not put an edge, corner or vertex on the x 500 column, on her contour, on the paper or on the door's jambs. Re-lay the strip at 500 to 545 at the value its new light gives it, not 0.82. Do not set the door or paper by glaze or re-lay before the new light is down. Their values follow from it.",
  "mode": "bold",
  "mode_why": "C1 was bold on light and replaced the vertical rule with a thrown diagonal. Since then two refine critiques (C3, C4) have each named one contact between that diagonal and an object, and each repair moved the object or nudged the edge and made the next contact: paper tip, window corner, now jamb over vertex. That is recurring repair aimed at details around something no one has named. The light is one ruled line with a flat 0.52 field behind it, and that field cannot hold anything without making it a spot. Pier:11 and pier:12 show the cost: two sittings spent correcting their own glazes inside the flat shade, ending with the darkest block overshot and the paper above its wall for the fourth time. A third refine would move the jamb or the edge a few units again and keep the wedge. Geometry is working (C2's pier, step and profile hold on pier:12) and is not the layer at issue.",
  "layer": "light",
  "ulcer": "Since pier:3 the light has been a single straight line, xf(y) = 660 - (y-96)*160/528, ruled from roofline to kerb with an unvarying 0.51 to 0.53 shade behind it. Its only free variable is where it touches things, so every repair relocates an object onto it: the jamb over the pavement vertex at x 500 on pier:12, after the paper tip and the window corner. Its flat interior makes whatever object was re-laid last into the spot: paper, then dress, now the 0.23 door.",
  "marrowline_triggers": [
    "recurring_repair"
  ],
  "collapse_tags": [
    "Recursive Echo",
    "False Stabilization",
    "Unspent Gesture"
  ],
  "permission": "Nothing in the light is protected, including what you painted before this critique. Everything else stands. Refuse, contradict or invert it; keep it only if the painting can say what keeping it costs.",
  "t": "2026-10-12T21:54:48+00:00",
  "state_at": "active"
}