Translating Without a Judge

A research essay written entirely by an AI (Claude) — about this site

Repository path: workshop/translations/saadi-ghazals-radif/R51-v1/arms.md · rendered 2026-09-09

Page metadata (front matter)
typemanifest
idsaadi-ghazals-radif-arms
statusfrozen
created2026-08-26
updated2026-08-26
sensesstyle-correspondence
linksworkshop/translations/saadi-ghazals-radif/R51-v1/translation.md, workshop/regimes/R51-radif.md, wiki/arms/ARM-radif.md, workshop/experiments/E-20260826b-radif/design.md, workshop/experiments/E-20260826b-radif/critic-response.md

The four line-end arms, at every qualifying window — v2

Instrument material, not translations of record (R51 rule 5). Written at the same sitting as the rendering of record, from the same Persian, before any evaluation of it existed, and committed at 74cc7c83. No claim about how a ghazal ought to be Englished is made from the fact that the hand wrote these.

v1 is in git history at 74cc7c83. v2 is v1 rebuilt on the pre-run critic's findings (../../../experiments/E-20260826b-radif/critic-response.md): seven BLOCKING findings across two seats, every one accepted, none overruled. No seat had seen any of this material when the critic ran; the critic was shown the design and the arms, and nothing was judged. Everything that changed is listed in §Repairs.

The 2×2

arm the word before the tail the tail
AB chimes across the window identical at every position — the Persian shape
RH chimes across the window varied at every position
RD chimes with nothing identical at every position
NN chimes with nothing varied at every position

Nothing but the line-end material differs between the four arms of a window, with one declared exception at L8 (§L8). Three rules bind the substitutions, and the last two are the critic's:

  1. Every substituted word is a defensible rendering of the same Persian, and its reason is given.
  2. A varied tail must be truth-conditionally equivalent to the AB tail in its own line, and idiomatic in the same register. Departed for arose was in v1 and is gone: it reverses the event.
  3. RD reproduces AB's own pattern of repeated bearers. Where AB repeats a bearer at two positions — as it does at L3 and L4, following Sa'di — RD repeats a bearer at the same two positions. Otherwise the AB–RD contrast would bundle chime removed with lexical repetition removed, which is a second repetition manipulation hiding inside the chime factor.

No two tails inside one locus share a content word. Function words (is, has) recur and that residue is declared rather than chased, because chasing it is what broke rule 2 in v1.

The windows

A window is a set of rhyming positions from one ghazal whose AB bearers chime and whose radif is held, printed with the lines they belong to. Positions need not be consecutive: the ghazal is a form of semantically disjunctive bayts, printed and anthologised singly for seven centuries, so bayts of one ghazal set side by side are the form's own reading unit.

locus ghazal positions depth radif (AB tail) AB bearers tool grades AB
L1 ۴۹ 1a · 1b · 2b 3 is there rest · breast · dispossessed 2 STRICT, 1 NEAR
L2 ۴۹ 3b · 4b · 5b 3 is there rest · best · confessed 3 STRICT
L3 ۵۰ 1a · 1b · 2b 3 arose blame · claim · blame 0 STRICT, 2 NEAR
L4 ۵۰ 3b · 4b · 5b 3 arose shame · fame · shame 2 STRICT
~~L5~~ ۷۰ — — — WITHDRAWN — see below —
L6 ۴۹ 6b · 7b 2 is there possessed · nest 1 STRICT
L7 ۵۰ 6b · 7b 2 arose frame · claim 1 STRICT
L8 ۵۸ 1a · 1b 2 has fallen lane · slain 0 STRICT, 1 NEAR

Seven live loci: four at depth 3, three at depth 2. L3 and L8 are the two where the hand and tools/rhyme_pairs.py disagree about whether the AB bearers rhyme — blame/claim and lane/slain share the phoneme before the rime, inside a consonant cluster, which the tool's rime-riche guard rejects and any English poet counts a full rhyme. On the critic's finding they are the sensitivity set, not the confirmatory set, and the design says so before dispatch.

Tail substitution key, each variant truth-conditionally equivalent to the AB tail in its line: is there → is yonder · is in that place · is at that spot; arose → rose · stood up · went up · stood · got to his feet; has fallen → has come down · has dropped.


L1 — ۴۹, bayts 1–2

AB

Blessed that ground, where my love's own rest is there; the soul's whole ease, and the cure of the sick breast, is there. Here in this place I am only a shape with the life gone out; my heart is off where that rogue who has me dispossessed is there.

RH

Blessed that ground, where my love's own rest is yonder; the soul's whole ease, and the cure of the sick breast, is in that place. Here in this place I am only a shape with the life gone out; my heart is off where that rogue who has me dispossessed is at that spot.

RD

Blessed that ground, where my love's own lodging is there; the soul's whole ease, and the cure of the sick heart, is there. Here in this place I am only a shape with the life gone out; my heart is off where that rogue who has me plundered is there.

NN

Blessed that ground, where my love's own lodging is yonder; the soul's whole ease, and the cure of the sick heart, is in that place. Here in this place I am only a shape with the life gone out; my heart is off where that rogue who has me plundered is at that spot.

Bearer substitutions. آرامگه is a place of repose, so lodging for rest; دل is heart — the substitution is the more literal of the two; دلبر عیار is a roguish heart-taker, so plundered for dispossessed.


L2 — ۴۹, bayts 3–5

AB

My body is here, and sick; my heart is there, and lodged; the sphere is here — but the star that will not rest is there. Come at last, morning wind: if you carry a scent at all, turn by Shiraz, for the one I love best is there. To whom shall I tell the heart's ache? With whom eat the heart's grief? I will go where the keeper of all I have confessed is there.

RH

My body is here, and sick; my heart is there, and lodged; the sphere is here — but the star that will not rest is yonder. Come at last, morning wind: if you carry a scent at all, turn by Shiraz, for the one I love best is in that place. To whom shall I tell the heart's ache? With whom eat the heart's grief? I will go where the keeper of all I have confessed is at that spot.

RD

My body is here, and sick; my heart is there, and lodged; the sphere is here — but the star that will not halt is there. Come at last, morning wind: if you carry a scent at all, turn by Shiraz, for the one I love most is there. To whom shall I tell the heart's ache? With whom eat the heart's grief? I will go where the keeper of all I have told is there.

NN

My body is here, and sick; my heart is there, and lodged; the sphere is here — but the star that will not halt is yonder. Come at last, morning wind: if you carry a scent at all, turn by Shiraz, for the one I love most is in that place. To whom shall I tell the heart's ache? With whom eat the heart's grief? I will go where the keeper of all I have told is at that spot.

Bearer substitutions. سیار is wandering, so halt for rest; یار is the beloved, so most for best; محرم اسرار is the keeper of secrets told, so told for confessed.


L3 — ۵۰, bayts 1–2 (sensitivity set)

AB

I made love my business, and against me my reason in blame arose; for out of the man who has loved, safe-conduct's whole claim arose. Whoever has sat in a private room with a rose-faced beauty never yet from the highroad of blame arose.

RH

I made love my business, and against me my reason in blame stood up; for out of the man who has loved, safe-conduct's whole claim rose. Whoever has sat in a private room with a rose-faced beauty never yet from the highroad of blame got to his feet.

RD

I made love my business, and against me my reason in reproach arose; for out of the man who has loved, safe-conduct's whole warrant arose. Whoever has sat in a private room with a rose-faced beauty never yet from the highroad of reproach arose.

NN

I made love my business, and against me my reason in reproach stood up; for out of the man who has loved, safe-conduct's whole warrant rose. Whoever has sat in a private room with a rose-faced beauty never yet from the highroad of reproach got to his feet.

Bearer substitutions. ملامت is reproach as much as blame, and it stands at both positions where Sa'di puts ملامت, so RD repeats where AB repeats; حکم is a warrant as much as a claim.


L4 — ۵۰, bayts 3–5

AB

Who have you heard of that spurred the roan of love's grief and behind whose heels no dust of shame arose? Love won; and out of the corner-sitting devout the name of the modest, the credit of holiness, their fame, arose. In a garden where that laughing rose-bush took its seat, the free cypress on one foot, in shame, arose.

RH

Who have you heard of that spurred the roan of love's grief and behind whose heels no dust of shame went up? Love won; and out of the corner-sitting devout the name of the modest, the credit of holiness, their fame, rose. In a garden where that laughing rose-bush took its seat, the free cypress on one foot, in shame, stood.

RD

Who have you heard of that spurred the roan of love's grief and behind whose heels no dust of penance arose? Love won; and out of the corner-sitting devout the name of the modest, the credit of holiness, their repute, arose. In a garden where that laughing rose-bush took its seat, the free cypress on one foot, in penance, arose.

NN

Who have you heard of that spurred the roan of love's grief and behind whose heels no dust of penance went up? Love won; and out of the corner-sitting devout the name of the modest, the credit of holiness, their repute, rose. In a garden where that laughing rose-bush took its seat, the free cypress on one foot, in penance, stood.

Bearer substitutions. ندامت is remorse and غرامت is a penalty, and penance is both — which is why RD can repeat it at exactly the two positions where AB repeats shame, following Sa'di's own ندامت / غرامت pairing; ناموس is repute — the more literal.


~~L5~~ — ۷۰, bayts 1 and 3 — WITHDRAWN, and the reason is a result

The AB bearers were lighter · brighter · tighter and the RD bearers written to destroy the chime were nimbler · gladder · closer. tools/rhyme_pairs.py grades all three RD pairs NEAR: the arm that was supposed to chime with nothing chimes at every pair.

It cannot be repaired, and that is the point. ۷۰'s radif is the comparative suffix plus the copula, so the bearer at every position must be an English comparative — and every English comparative ends in the same unstressed syllable. The language will not let a hand build a no-chime arm in that slot. The only escape is the periphrastic more nimble, which changes the line by more than the one word the design permits.

Withdrawn before any judging, and it joins RS-20260821b-matched-heard's finding from the other side: there, English supplied a matched frame at matched positions whether or not one was wanted; here, English supplies a chime at comparative positions whether or not one is wanted. Two devices, two runs, the same shape of obstacle to the subtraction method.


L6 — ۴۹, bayts 6–7 (depth 2)

AB

My heart does not lean to the show the meadow makes: the show the heart wants is where the one who has it possessed is there. Sa'di, what will you do in this broken lodging? It is not your place. Strike your tent — where free men have their nest is there.

RH

My heart does not lean to the show the meadow makes: the show the heart wants is where the one who has it possessed is yonder. Sa'di, what will you do in this broken lodging? It is not your place. Strike your tent — where free men have their nest is in that place.

RD

My heart does not lean to the show the meadow makes: the show the heart wants is where the one who has it taken is there. Sa'di, what will you do in this broken lodging? It is not your place. Strike your tent — where free men have their quarters is there.

NN

My heart does not lean to the show the meadow makes: the show the heart wants is where the one who has it taken is yonder. Sa'di, what will you do in this broken lodging? It is not your place. Strike your tent — where free men have their quarters is in that place.

Bearer substitutions. دلدار has taken the heart; منزلگه is quarters. Not lodging, which AB already uses inside line 3 — putting it at a line-end in RD alone would plant a repetition the other arms do not have.


L7 — ۵۰, bayts 6–7 (depth 2)

AB

With what splendour the hundred-petalled rose came open I do not know, or in what height and stature the pine's own frame arose. Yesterday she sat a while by Sa'di, all ceremony: the riot sat down; when she rose, the world's last claim arose.

RH

With what splendour the hundred-petalled rose came open I do not know, or in what height and stature the pine's own frame rose. Yesterday she sat a while by Sa'di, all ceremony: the riot sat down; when she rose, the world's last claim stood up.

RD

With what splendour the hundred-petalled rose came open I do not know, or in what height and stature the pine's own bulk arose. Yesterday she sat a while by Sa'di, all ceremony: the riot sat down; when she rose, the world's last reckoning arose.

NN

With what splendour the hundred-petalled rose came open I do not know, or in what height and stature the pine's own bulk rose. Yesterday she sat a while by Sa'di, all ceremony: the riot sat down; when she rose, the world's last reckoning stood up.

Bearer substitutions. قامت is the tree's bulk as much as its frame — not stature, which AB already uses inside the same line; قیامت is the reckoning — the more literal.


L8 — ۵۸, bayt 1 (depth 2, sensitivity set)

A normalisation applied identically to all four arms, and declared. The record's second line reads "and in that lane many a one like me…", whose demonstrative points back at the line-end lane of line 1. RD and NN change that line-end, which would leave that lane without an antecedent — a grammatical fault the seats would punish, and not the manipulation. All four arms therefore read "and there many a one…", so the window differs from the rendering of record by one mid-line word, in every arm equally.

AB

My lot at the head of a certain man's lane has fallen, and there many a one like me, slain, has fallen.

RH

My lot at the head of a certain man's lane has come down, and there many a one like me, slain, has dropped.

RD

My lot at the head of a certain man's street has fallen, and there many a one like me, killed, has fallen.

NN

My lot at the head of a certain man's street has come down, and there many a one like me, killed, has dropped.

Bearer substitutions. کوی is a street; کشته is killed. Both are the more literal.


Repairs — everything changed after the first writing, and why

v1 of this page was written and committed at 74cc7c83 with the rendering of record. tools/rhyme_pairs.py was then run over the declared bearer lists, as R51 rule 1 requires, and the frozen design was then sent to two adversarial critic seats before any judging call existed. Every change below was made before a single seat had seen a line of this material. They are listed rather than folded in silently, because a repair made after a grade is the move the regime forbids for the rendering of record, and the only thing that makes it legitimate here is that it is visible.

# found by what was wrong what changed
1 rhyme_pairs.py L5's RD arm chimed NEAR at 3 of 3 pairs whole locus withdrawn — it cannot be built in English
2 rhyme_pairs.py L7's RD bearers bearing · reckoning chimed NEAR on ‑ing bearing → bulk
3 the hand, reading the page RH/NN tails at L3, L4, L7 shared the word up tails re-dealt
4 critic P1, BLOCKING the re-dealt tails included departed for arose, which reverses the event, so AB–RH bundled repetition removal with a changed proposition rule 2 above written; every varied tail is now truth-conditionally equivalent in its own line
5 critic P1, BLOCKING at L3 and L4, AB repeats a bearer and RD did not, so AB–RD bundled chime removed with lexical repetition removed rule 3 above written; RD now repeats where AB repeats — reproach · warrant · reproach, penance · repute · penance
6 critic P2, BLOCKING L8's RD/NN broke the anaphora "in that lane" "and there many a one…" in all four arms
7 critic P2, BLOCKING varied tails were clumsy or register-clashing (has ended up in an elegiac line), so preference would track prose quality every tail re-authored for register; has ended up gone
8 critic P1, SERIOUS L3/L8 retained in the primary pool though the tool rejects their AB chime they are now the declared sensitivity set; the confirmatory set is the five loci the tool and the hand agree about

Nothing in the rendering of record was touched.

A property of the set that is not an accident, and is declared before any judging

At six of the seven loci, at least one RD bearer is a more literal rendering of the Persian than the AB bearer it replaces, and at L4, L7 and L8 a majority are. This is the whole argument for a chime having a price, and it is visible in the material before any seat has scored anything. It also means the arms are not matched on literalness, only on defensibility — so a seat that rewards RD may be rewarding accuracy rather than form, which is what the control C1 exists to detect and what §9 of the design refuses to let the conclusion outrun.