{
  "id": "C4",
  "by": "critic BC-P1",
  "scope": "sequence",
  "cites": [
    "main:3",
    "cast:1",
    "pier:2",
    "pier:3",
    "pier:5",
    "pier:8",
    "pier:9",
    "pier:10"
  ],
  "read": "The C3 plan held through pier:3 to pier:5. The shade edge (pier:3 chunk, xf(y) = 660 - (y-96)*160/528) crosses open wall well clear of the paper, she stands with wall showing on both sides, and the paper sank to the wall (0.52 against 0.52 on pier:5). Pier:8's own squint read the door as the darkest block. Coat E (pier:9) then moved the accent it was not aimed at. It set out to separate the glass and quiet the pavement, and it re-laid the figure on top. That took the dress from 0.37 to 0.27 (0.28 on pier:10). Against the door's 0.39 she is now the darkest mass in the picture: the darkest 1 percent of pier:10 sits on her (median x 695, y 511). In the shade wall (0.52) she is a cut-out at a 0.24 step, where the door makes only 0.13. The proposition's order (lit window / shaded pier / dark door) now reads lit stucco / shade / darkest figure / mid door. So the door no longer closes the interval, and the eye stops on her silhouette. That is the spotlight main:2 named, inverted from light to dark. The paper went the same way on a smaller scale. Pier:10 restated the fold and print, and the paper rose from 0.521 to 0.532, back above the wall it sits in (0.523). On pier:10 the print grid makes it a pale outlined card in front of her. This is the third return of the paper above its shade (main:3 0.62, pier:2 0.56, pier:10 0.53). Each return is smaller, but it always comes back through a re-lay that was not aimed at it. Coat E also bought the glass by darkening it from 0.73 to 0.62. That is the trade C1 named on main:3 (0.82 to 0.64). The window is now a mid slab against 0.78 to 0.80 stucco, and the lit end of the interval is the stucco, not the window of coats. The edge pushed left by C3 now meets the window's bottom-right corner (xf(544) = 524 against the jamb at 540). It leaves a grey tongue of shaded glass about 18 units wide at x 522 to 545, y 497 to 547: another light edge registered to an object, this time a corner. In the lit half the chair carries the picture's only full chroma (0.40, the image maximum, at the chair) and its hardest drawing (legs 0.49 on 0.82 pavement). With her a dark cut-out, the picture reads as two spots, red chair and blue figure, across the pavement. The pier between them does no work. Pier:10's contact dark under her hem (a 5-unit stroke from 670 to 738 at y 659 to 663) shows as a broken dashed line across the pale-streaked hem, not as a shadow she stands in.",
  "pressure_points": [
    "The figure is the darkest thing in the picture since pier:9: dress 0.28 against the door 0.39 and the shade wall 0.52 (pier:10). The interval's dark end has moved from the door to her, and a 0.24 step cuts her out of the wall. That makes her a spot, the condition the proposition refuses, now by darkness instead of light.",
    "Recurring return of the paper above its shade: main:3 0.62, pier:2 0.56, pier:10 0.53 against a 0.52 wall. Pier:10's restated fold and three print columns outline it as a card. Each re-lay of the figure group re-lays the paper's value too, so the C3 repair does not survive a later coat.",
    "Glass bought by dimming the window: 0.73 (pier:5) to 0.62 (pier:9 and pier:10), coats 0.53 to 0.47. In squint the 'wide lit window of coats' is a mid-grey slab inside lit stucco at 0.78 to 0.80. This is C1's main:3 trade repeated, and the lit side of the interval is now carried by bare wall.",
    "Fourth registration of the light edge: jamb (main:3), her contour (cast:1), paper tip (pier:2), and now the window's bottom-right corner. Xf(544) = 524 against the jamb at 540 leaves an 18-unit grey tongue of shaded glass at x 522 to 545, y 497 to 547, visible on pier:10 as a nick at the window's corner.",
    "The chair holds the image's maximum chroma (0.40) and its hardest contrast in the lit half (legs 0.49 on 0.82 pavement). Set against a dark cut-out figure, the picture reads as two emblems facing each other across the pavement, and the pier between them is skipped. Pier:10's hem contact reads as a dashed line, so she does not yet stand on the pavement the chair stands on."
  ],
  "action": "Inside the existing geometry and light, without another full coat: lift the dress back to about 0.36 to 0.40 and take the door down to about 0.30, so the door is again the darkest block and she is a dark upright held within the shade, not cut out of it. Glaze the paper and its print down to 0.48 to 0.50, below the wall, so the fold and print are made with darks, not a pale card, and do not re-lay the figure group in any later coat. Raise the glass back toward 0.68 to 0.72 and make the separation from the stucco with a darker band at the window's foot or a reflection, not by dimming the whole pane. The coats stay below the glass. Move the shade edge's crossing off the window corner: either clear it by 20 or more units of lit stucco, or let it cut well into the glass so the shaded part of the window is a shape, not an 18-unit tongue. Replace the dashed hem contact with a soft, slightly wider shadow that joins her to the pavement shade. Lower the chair's chroma toward the coats' red earth, or narrow its contrast, enough that it answers her across the pavement without outbidding the interval. What the chair's sun is for can stay open; only its loudness is at issue. Leave the frontal sign band and the full pavement band as the stated costs of S09; this critique does not ask to reopen them.",
  "mode": "refine",
  "mode_why": "The geometry (C2) and the thrown light (C1) still hold on pier:10: the edge falls on open wall clear of her and the paper, and she stands in a pier with wall on both sides. The new failures came from coat E (pier:9) and the accents (pier:10). That pass re-laid the figure group and the glass with new values: the dress went darker than the door, the paper rose above its wall, and the window was dimmed to make glass. These are value and placement errors inside a working conception, fixed by setting four values and moving one edge crossing. The returns (paper above shade, edge on an object, window dimmed for glass) are real and named as recurring. But each is now a few hundredths of value or a few units of edge, and the cause is visible in one sitting, not a layer no one has named. Destroying a layer with seven sittings left would trade a working interval for the brief's next default.",
  "marrowline_triggers": [
    "recurring_repair"
  ],
  "collapse_tags": [
    "Recursive Echo",
    "Symbolic Overcharge",
    "Unspent Gesture"
  ],
  "t": "2026-10-12T21:48:07+00:00",
  "state_at": "active"
}