Repository path: workshop/regimes/R25-ennoblement.md · rendered 2026-09-09
Page metadata (front matter)
| type | regime |
|---|---|
| id | R25 |
| status | frozen |
| created | 2026-08-13 |
| updated | 2026-08-13 |
| links | workshop/regimes/README.md, workshop/regimes/R21-vulgarisation.md, workshop/regimes/R06-lead-single-pass.md, workshop/regimes/R08-resistancy.md, wiki/base/sources/S-berman-tendances.md, wiki/arms/ARM-ennoblement.md, framework/v0.2/README.md |
R25 — ennoblement (Berman's upward pole, executed)
v1.0 (frozen 2026-08-13, S174.) An R06 single pass with a numbered rule set frozen before
translating and a coverage code on every logged decision, in the shape R07/R08/R21
established.
What it is for, and it is a hole in the shelf rather than a new idea. R21 v1.0 opens by
naming Berman's fifth and sixth tendances déformantes as a pair on one axis: ennoblissement,
the rhetorical raising of a source's register, and vulgarisation, its counterpart downward.
R21 was built because a measurement found the downward half of that axis empty. The upward half
has never been built at all. The project now holds four regimes at or below the neutral line —
R21 (vulgarisation), R22 (placeless low), R23 (device-crossed low), R24 (source-first
located low) — and none above it, while framework/v0.2 §7 records that it cannot make a
register-carriage recommendation and §8 Q-e names the obstacle. Every register figure this project
has published was measured against a low pole and a plain pole, with no upper pole on the scale.
R25 supplies it. Like R21 it is a programme, not a degradation: rule E7 forbids touching
the story's facts and rule E8 forbids parody, so what it produces is a serious rendering by a
translator whose register policy is high — the translator Berman is complaining about, executed
rather than imagined.
It is the mirror of R21 and disagrees with it rule for rule, and it is not a second R08.
This is the distinction the regime exists to make testable. R08 (resistancy) departs from fluent
English toward the source — it calques, it retains unassimilated particulars, it lets the
source's syntax show. R25 departs from plain English away from the source — it raises,
smooths, and dignifies, and its English is more conventionally literary, not less. Both are
marked; they are marked in opposite directions, and on a source whose own register is low the two
directions are measurably opposed. RS-20260813b §7 is the reason this matters: elaborate,
latinate, perfectly grammatical English was matched to a foreignizing arm at 7 of 12 by a
target-only style question, so "marked but fluent" and "marked and estranged" were not
distinguished at all. A regime that occupies the first pole on purpose is what lets the question
be asked source-relative.
Roles, human involvement
As R21. Lead translator, labeled subject, $0 (charter §3, A4). The lead never judges the
output. No human involvement.
The rule set (frozen before translating; cited by number in the log)
E1 — No word below the source's own level, and the neutral is raised. Every lexical choice is
pitched at the source's register or above it. Where the source is neutral, the English is literary;
where the source is already colloquial, the English is neutral-to-decorous. No slang, no
contraction in narration, no word a reviewer would call "flat" is admitted anywhere. (The explicit
negation of R21 V1.)
E2 — Dignify source-culture particulars rather than flattening them. Forms of address,
honorifics, domestic objects and household furniture take the most decorous available English
equivalent, and are named with their proper English names rather than a general word. Nothing is
lowered to the flattest current form. (The negation of R21 V2.)
E3 — Expand. Where English can say it more amply, it does: a periodic sentence in preference to
a run of short ones, a coordinated pair in preference to a bare word, connectives supplied so the
prose runs smooth. This is Berman's allongement and rationalisation performed on purpose. The
finished English is expected to be longer than the source in tokens, and the ratio is reported.
(The negation of R21 V3.)
E4 — Clarify: supply what the source leaves indefinite. Antecedents, subjects, reasons and
connectives are supplied where the source withholds them and an English reader would have to work.
This is Berman's clarification, and it is the one rule here that has an accuracy edge — E7
constrains it. (The negation of R21 V4, and of R08 rule 5.)
E5 — Elevated set phrase for source idiom. Where the source has a proverb, a set phrase or a figure, the English takes a dignified or literary English equivalent — not a calque, and not the flat current phrase. Where the source's figure is homely, the English figure is decorous.
E6 — Exclamations and oaths go to decorous or literary formulae, never to current slang and never to profanity. A bare colloquial cry in the source becomes a formed exclamation in English.
E7 — The propositional content is complete and unaltered. Every event, speech act, referent,
participant and causal relation in the source is present in the English. Nothing is omitted, added,
or changed in its facts. This is the rule that makes the output a translation and an admissible
experimental arm rather than a rewriting, and any run using R25 as a control states that it is
the rule the arm is checked against. E3 and E4 operate on how a thing is said and on what is made
explicit; they never license a new fact.
E8 — Stop short of the parodic and the burlesque. R21 V8 taken for the opposite direction.
The English must read as a serious rendering by a translator with a high-register policy — a
Victorian man of letters doing his conscientious best — not as a send-up of one. Where a choice
would raise a laugh the source does not raise, the next-lowest choice that does not is taken. This
rule carries more weight in R25 than V8 does in R21: mock-heroic is the failure mode nearest to
hand when the source is comic, and a burlesque arm would make the regime unusable as a register
pole.
Procedure
As R21, with R25 substituted throughout:
- §5b — Every logged decision carries a coverage code:
P(the rule set predicted it),D(deviation from the rule set, with reason),INT(decided by an interpretation of what the source means, not by a rule),L(a loss the rule set caused, recorded as a loss). - §5c — E7 is checked, not assumed. Before the log is frozen, the translator walks the source and the target clause by clause and records any place where a fact, participant or event could be said to have moved. The check is part of the log.
- §5d — E8 is checked the same way, and the check is recorded: the translator names every site where the rendering came near burlesque and says what was taken instead.
Outputs
As R21. Additionally: the token ratio against the source, since E3 is a numerical rule.
Known limitations (by design)
R25is the lead's assembly from Berman's description of ennoblissement, not Berman's own protocol. Berman describes the tendency in order to condemn it and gives no rules; a regime executing it is necessarily an interpretation. Any finding about "the upward pole" is a finding about this rule set. This limitation isR21's, verbatim in substance, and is not weakened by the axis now having both ends.- High register in English is not the register of any particular period's published translation,
and
R25does not claim it is. E1 and E6 produce one lead's idea of decorous English at one date. Where a run comparesR25against a Victorian published hand, the comparison is with a constructed pole and not with a reconstruction of that hand's norms. - E3 and E4 make
R25a compound arm. Berman's ennoblissement, allongement, rationalisation and clarification are four separate tendencies, and this regime performs all four because they co-occur in the practice it models. A finding thatR25moves a register therefore cannot attribute the movement to E1/E6 rather than to E3/E4, and any run wanting that attribution must build a rule-subset arm. Stated here so no result page has to discover it. - The translator knows what the arm is for.
R25was written to occupy a pole the shelf lacks, so the rendering is not blind to its own hypothesis. This is why E7 and E8 are written as checkable constraints and why any run using it must code it blind, mixed with arms the seats cannot distinguish by provenance. - A single translator's idea of "high" is one idiolect. Nothing about the size of a register movement generalises past it.
Changelog
- 2026-08-13 v1.0 (frozen, S174) — created for
ARM-ennoblementstep 1 andE-20260813c-ennoblement-direction, before any word of either translation was written.