Repository path: workshop/experiments/E-20260728j-classb-marking/forced/renderings.md · rendered 2026-09-09

Stage 1 — the forced re-translation, and every lead judgment this design uses

Frozen before fetch_dole.py was written and before any word of Dole 1896 was read. Design frozen at 7968abe, amended after the critic pass at 1582101; this page is the whole of Stage 1, and amendment B1 is why it contains the grading tables and not only the prose: "Dole can contaminate the lead's grading rules, forced-rendering alternatives and rationales … not merely the prose."

Contamination: suspected, inherited. This is lead prose derived from T-jeli-il-pastore-R05-v1, whose declaration is suspected on span 1's measured priming accident (RS-20260727-longwork-selection §4). Nine clauses are too short to carry a run measurement of their own. A secondary measurement is declared here, before it can be run, and no prediction is attached to it: the forced ¶42 will be checked against Dole in Stage 3 alongside the frozen ¶42, which asks whether forcing a policy moves a translation toward or away from a published rendering — a cousin of the S050 backlog row about self-revision.

These are probes of availability, not proposed improvements. The frozen span-2 prose is not revised; R05 is append-only and the translation page changes only by erratum. No quality claim is made about any rendering below, mine or the frozen one (charter §5: the lead does not judge its own translation).

1. ELECTIVE / FORCED — amendment A3, applied to all nine sites before Dole is opened

A site is ELECTIVE if the past-tense counterpart of the clause is grammatical English, FORCED if it is not. The critic's objection was that a present tense in Dole might be ordinary English simile idiom rather than a carried mark. The test says whether the translator had a past-tense option and declined it.

All five gnomic-present sites are ELECTIVE, and so is every other site: 9 of 9. The critic's worry does not bite on this material — there is no site where English forced the present, so at every one of the five the translator had a grammatical past-tense alternative and declined it. P4's denominator stays at 5 and the "at least 4 of 5" threshold is unchanged.

2. The frozen S039 renderings, graded by the amended §3

Graded first, so the forced renderings are graded by a rule already applied to prose the lead cannot change.

6 MARKED, 3 UNMARKED, and the three UNMARKED are exactly the three conditionals. The mechanical definition reproduces D22's own "six of nine carried; three of nine flattened" without being told it — which is the check that the definition describes the frozen prose rather than a story about it.

3. The forced attempts

Six sites need none. B1, B2, B3, B5, B8 and B9 are already MARKED in the frozen prose, so there is nothing to force; they are recorded here as requiring no attempt. The work is at B4, B6 and B7.

B4 — «che anche un ragazzo saprebbe curarla»

frozen: …because the illness was one of those plain and well-known ones that even a boy would know how to treat… forced: …because the illness was one of those plain, well-known kinds that even a boy knows how to cure…

MARKED, m1 (knows, present, matrix past, not dash-set speech). Grammatical.

Also tried and rejected: "that even a boy could have cured" — UNMARKED, could have is not in m2's list and the form is the ordinary one. "that even a boy was to know how to cure" — m2's was to is available as a string but yields a prospective sense the Italian does not have; recorded as attempted and declined on sense, not on form.

What the marking costs. The Italian saprebbe is a hypothetical: even a boy would know how, if it came to that. English knows asserts it. And the assertion has a second effect the frozen page did not anticipate: an unframed gnomic present in English narration is readable either as the community's maxim or as the narrator's own, and English has no way to decide between them. So the form that carries Verga's mark is also a form that can hand the sentence back to the narrator — which is the exact thing rows V6–V9 exist to prevent.

B6 — «col solfato si sarebbe guarita subito»

frozen: …and if the fever was not of the kind that kills you whatever you do, with the sulphate it would have been cured at once. forced: …and if the fever was not of the kind that kills you whatever you do, sulphate cures it at once.

MARKED, m1 (cures). Grammatical.

Also tried: "were the fever not of the kind that kills you whatever you do, the sulphate cures it at once" — also MARKED (m1 twice: were-subjunctive inversion and cures), and it marks the protasis as well; recorded as a second successful attempt, and the one that carries most.

What the marking costs, and it is the largest of the three. si sarebbe guarita is a counterfactual about this fever: it would have been cured — and it was not. sulphate cures it is a general truth about sulphate, and the counterfactual is gone entirely. The community's reasoning ("the doctor's money was wasted, because this was curable and this one wasn't cured") becomes a pharmacological remark.

B7 — «pensando che il ragazzo avrebbe saputo aiutarsi, quando fosse rimasto solo»

frozen: …thinking the boy would know how to fend for himself when he was left alone. forced (a): …thinking the boy would shift for himself well enough, when he was left alone. forced (b): …thinking the boy would shift for himself, were he left alone.

(a) MARKED, m3 (well enough, closed-list member) — lexical. (b) MARKED, m1 (were-subjunctive with inversion) — grammatical.

Also tried and rejected: "thinking the boy knows how to fend for himself" — a present under an explicit past participial tag thinking; standard English backshifts here and the line reads as an error rather than as a shift. Recorded as attempted and rejected as ungrammatical, which is the only such rejection in the set. "thinking the boy should know how to fend for himself" — should is in m2, but only in its non-deontic use; here it reads as ought to, which is a different claim. Declined on sense.

What the markings cost. (b) is the cheaper-looking one and is the worse: quando fosse rimasto solo is temporal — the father is dying and the boy will be left alone — and were he left alone makes it hypothetical. It converts a certainty into a supposition in the one sentence where the father's whole thought depends on the certainty. (a) costs a word the Italian does not have and nothing else.

4. P3′ — the definiteness table (amendment A2)

P3′. At every site among B4/B6/B7 where a grammatical marking is achieved, the marked clause contains at least one noun phrase that is bare-plural or generic-indefinite where the frozen rendering's corresponding phrase is definite or pronominal. Falsified if at any of the three a grammatical marking is achieved with every noun phrase's definiteness unchanged.

P3′ is FALSIFIED, at B4 and at B7. A grammatical marking is achieved at both with every noun phrase's definiteness unchanged.

And the falsification is worth more than the prediction was. The design's original P3 said the cost was episodic → generic; the critic tightened it to a determiner test that can be read off the page; the determiner test fails at two sites of three. The cost is not referential. It is modal and temporal — B4 turns a hypothetical into an assertion, B6 turns a counterfactual into a general truth, B7 turns a certainty into a supposition — and none of those three moves shows up in a determiner. The loose original P3 would have been confirmed on the same three renderings by calling B4's "boy" generic and B6's "sulphate" generic and waving at B7; the tightened one fails. That is what the tightening was for.

5. The continuous forced rendering of ¶42

Under the same brief: every Class B site in the paragraph marked, using the strongest available marking at each. Class C sites are out of scope and keep their frozen renderings. Register rows bind as they do in the frozen prose (D28's ecalibus/eucalyptus pair, V4's contraction split, V8's quotation marks, N3's Massaro Agrippino).

The doctor was called in too, but it was all money thrown away, because the illness was one of those plain, well-known kinds that even a boy knows how to cure, and were the fever not of the kind that kills you whatever you do, sulphate cures it at once. Compare Menu spent the eyes out of his head on that much sulphate, but it was like throwing it down the well. "Take a good decoction of ecalibus, it costs nothing," Massaro Agrippino advised, "and if it does no good, no more than the sulphate, at least you don't ruin yourself spending." The eucalyptus decoction was taken as well, and still the fever always came back, even stronger. Jeli looked after his father as well as he knew how. Every morning, before going off with the colts, he left him the decoction ready in the bowl, the bundle of cuttings under his hand, the eggs in the hot ash, and came back early in the evening, with more wood for the night, and the little flask of wine and some scrap of mutton he had run all the way to Licodia to buy. The poor boy did everything neatly, like a good housewife, and his father, following him with tired eyes in his little errands here and there about the hut, smiled from time to time, thinking the boy would shift for himself well enough, were he left alone.

P6′ — the count. Occurrences of closed-list m3 members in the forced ¶42: well enough × 1. N = 1. It falls inside a Class B conditional clause (B7).

P6′ is FALSIFIED on its count (N ≤ 1), and confirmed on its clustering (the one m3 item is inside a conditional clause). The design predicted that English pays for the mood with a word; on this paragraph it mostly pays with a tense, which costs no tokens at all. Four Class B sites are marked in the paragraph above and only one of the four needed lexical material.

What the continuous version shows that the site-by-site table cannot. Three tense shifts land inside two sentences, and the paragraph acquires a narrator who generalises — even a boy knows how to cure, sulphate cures it at once — where Verga's paragraph has a community reasoning about one sick man. The policy that was supposed to remove the writer's lens installs a maxim-making one. Whether that is a gain or a loss is a quality judgment and is not made here.

6. Lexical anchors for Stage 2 — named from the Italian, before Dole is opened

Amendment B2: a site is AMBIGUOUS (not excluded) when the anchor is present in Dole but the clause structure does not correspond, and AMBIGUOUS counts against the lead's prediction.

7. Predictions standing at the end of Stage 1

D22's first clause survives Stage 1 and its second does not. "English has no marked conditional to reach for" is true and remains true — no rendering above marks by producing an English conditional. "Three of nine flattened with no available alternative" is false: all three sites admit a marked rendering, and two of the three admit a grammatical one. What English lacks is a marked conditional; what it has is a marked tense, and a translator searching for the first will correctly find nothing while the second is available at every site. The frozen log named the right absence and drew the wrong conclusion from it.