Repository path: workshop/translations/hafez-sahar-bolbol/R55-v1/translation.md · rendered 2026-09-09

«حافظ» غزل ۱۳۰ rendered whole under R55, in a plain register

Source: Ganjoor's Hafez (Qazvini–Ghani), /hafez/ghazal/sh130, fetched 2026-08-31; the Persian and the eleven rhyming positions are in ../source.md. Regime: R55 — the radif carried or refused, and the refusal enumerated. Register declared before writing: plain modern English. No thou, no -eth, no inversion licensed by archaism. That declaration is the control: RS-20260829b-inversion-price (§7.42) measured that the object–verb inversion at a line end costs 2.407 in a plain register and 1.926 in an archaic one, so holding the register plain is what makes the price visible.

Selection, mechanical and frozen before the study limb of this session existed. From the frozen census of all 495 ghazals (runs/RS-20260829-radif-hands/census.json): the radif-bearing ghazals whose radif is exactly کرد — there are seven — and among them the lowest-numbered one in which the qāfiya of the maṭlaʿ is not the object of the radif verb. That is sh130. Its sister sh131, where the qāfiya is the object at all nine positions, was rendered on 2026-08-29 under this same regime (T-hafez-ghazals-radif-R55-v1). The radif and the rhyme-sound are held constant across the two poems and only the grammatical relation varies.

What it tests. framework/v0.2 §7.41 gained a craft clause on 2026-08-29, on one poem: where the radif is a light verb whose object is the qāfiya, the shape crosses whole into English and needs no inversion at all. sh130 has the same light verb and the object relation at only four of its eleven positions. Neither Leaf 1898 nor Bell 1897 translated it.

Contamination: none, measured — ../dependence-note.md.

This is not an independent second opinion of anything (CLAUDE.md, the contamination rule). It is a first-hand report of what the shape costs to attempt, which no coding of a finished page can give.

The rendering of record

At dawn the nightingale to the east wind had this to say: what love of the rose's face has done to us, and in what way.

That colour drove the blood out of my cheek and into my heart, and out of that same garden made me the thorn's own prey.

I am the bondman of the high resolve of that dear one who did a good deed with no show of face and no display.

I will cry out against strangers now no more: whatever was done to me, my own friend did — not they.

If I looked for profit from the sultan, that was my mistake; and if I sought faith of my love, my love did betray.

Well may it fare with that breeze of the early morning, which did the pain of those who sit up at night allay.

It drew the rose's veil aside, and the hyacinth's tress, and made the fastening of the bud's coat give way.

On every side the nightingale in love is crying out, and the east wind, in the midst of it, took the pleasure away.

Carry the good news to the street of the wine-sellers: Hafez has repented of the abstinence of display.

Faith, among all the grandees of the city, to me Kamal ud-Dawla wa'd-Din Bu'l-Wafa did repay.

Grades

Radif: UNREACHABLE for the poem (R55 §4 — a radif stands at every rhyming position or it is not a radif). The enumeration is below and was written before the rendering.

Chime, position by position — HELD 7, SHORT 4, REFUSED 0.

The four positions where the qāfiya is the radif verb's own object or complement hold the chime at two of four; the seven where it is not hold it at five of seven. The English reuses two rhyme words, way (manner at 2, yielding at 8) and display (both at 4 and 10 in the same sense) — the second reuse mirrors the Persian's own reuse of ریا at exactly those two positions.

Translator's log

Frozen 2026-08-31 before any evaluation of this translation was designed, before any published English of this poem was opened, and before the session's study limb was written. Decisions and alternatives; no quality claims. Any evaluative sentence carries internal-judgment-only, which the front matter already asserts for the page.

D1 — the radif was enumerated first, and it failed for a reason that is not in the framework

R55 §1 requires the options to be enumerated before choosing. کرد is one Persian word. English has no one word that covers what it does across these eleven positions, because English lexicalises کردن as two verbs, do and make, and which one a collocation takes is fixed by the noun, not by the syntax.

So the radif is UNREACHABLE, and the obstacle is lexical, not syntactic. That is a different kind of obstacle from any the framework names: §7.36.1 and §7.40 both reason about what an English clause can end on, and §7.41 replaced their prohibition with an inversion whose price is a register. None of the three would have predicted this poem's failure, because did and made are each perfectly capable of standing last. The failure is that no single one of them collocates with all eleven of the nouns Hafez put there.

The check a translator can run before writing a line is therefore not "is the qāfiya the radif verb's object?" but "does one English verb serve every collocation the radif makes across the whole poem?" — and that is answerable off the Persian.

D2 — why this poem, and what it does to the clause published two days ago

sh131, rendered on 2026-08-29, took the same radif کرد at nine positions of nine because every one of its qāfiyas was the nominal of a light-verb compound — غارت کرد, اشارت کرد, زیارت کرد — and English has the identical construction, so made a plundering / an indication / a visitation crossed whole. The clause that went into §7.41 from that poem does not say it is a whole-poem condition. sh130 shows that it must be: here four qāfiyas are objects or complements and seven are not, and the radif fails outright. A condition that holds at some positions buys nothing at all, because a radif is all-or-nothing by definition.

internal-judgment-only: nothing here says either rendering is good.

D3 — the chime family, and the two fillers I could not avoid

The Persian chime is -ا, a long ā. Two English families were enumerated:

• -ain / -ane (pain, disdain, feign, maintain, complain, vain): it fits four positions exceptionally well — درد → pain is exact, جفا → disdain, وفا کرد → did faith maintain, ریا → feigned. It fails at v1 (صبا), v4 (مبتلا) and v14 (وا) without buying words.
• -ay (say, way, prey, display, they, betray, allay, away, repay): fits nine positions without buying anything, and it is also the family whose vowel is nearest the Persian's.

-ay was taken. The two fillers are named in the grade table and are these:

• v2, and in what way. چه‌ها کرد is quantitative and exclamatory — what things it did!. In what way is manner, which is a shift, and it is bought for the rhyme. Marked SHORT.
• v20, repay. وفا کرد is kept faith; repay imports a requital the Persian has not got. Marked SHORT. The alternatives were alway (archaic, barred by the declared register) and display (a third use of a word already used twice).
• v4, prey. مبتلا is afflicted; the thorn's own prey adds a hunting metaphor. Marked SHORT even though it is close, because R51 §1 says a chime bought with a word the Persian has not got is not a chime.
• v8, not they. The contrast between بیگانگان (strangers) and آشنا (the intimate) is the couplet's own, but it is in the first hemistich and not they restates it at the rhyme. Marked SHORT on the strict reading.

D4 — what was lost that no grade records

بوالوفا is a pun. The patron's title کمال دولت و دین بوالوفا ends in a name that means father of fidelity, and the hemistich before it is about وفا, fidelity: the man's name is the word the couplet is about. English can transliterate the name or translate the pun, not both. The name was kept and the pun was dropped. This is the same class of loss RS-20260831 recorded for Sa'di's unframed Arabic — a device whose motive and whose vehicle go together.

صبا is the east wind of dawn specifically, and is the poem's second character; the east wind keeps the referent and loses the season and the erotic convention that comes with it.

D5 — no inversion was needed, and that is itself a datum

Not one line of the rendering of record inverts. The whole apparatus §7.41 built on 2026-08-29 — an object–verb inversion at the line end, available in an archaizing register and not a plain one — is irrelevant to this poem, because the radif was abandoned before any inversion could be required. Where a hand drops the radif and keeps the chime, the inversion price is never paid. Leaf 1898 does exactly this at eight of his nine radif-bearing odes.