Repository path: workshop/translations/hafez-darad/R56-v1/translation.md · rendered 2026-09-09

«حافظ» غزل ۱۲۴ and غزل ۱۱۸, rendered whole under R56 in four declared arms

Source: Ganjoor's Hafez (Qazvini–Ghani), /hafez/ghazal/sh124 and /hafez/ghazal/sh118, fetched 2026-08-29; the Persian and the frozen selection rule are in ../source.md. Regime: R56 — the radif carried by inversion, with the order priced and the register crossed. Machine-checked: runs/RS-20260829b-inversion-price/pair_check.py — 204 checks, 0 failures, run before the experiment that uses these arms was designed.

Eighteen bayts, each rendered four times. The four arms are the declared variant set of R56 §3: DP (direct order, plain lexis) · IP (inverted, plain) · DA (direct, archaic) · IA (inverted, archaic). The full four-arm table is runs/RS-20260829b-inversion-price/items.json; the rendering of record — IP, the radif standing last in present-day English — is printed whole below.

The rendering of record

غزل ۱۲۴

He from whose hyacinth-lock the perfume its curl has, for the heart-lost he again coquetry and rebuke has.

Like the wind you pass by the head of the man you have killed. What can be done? It is life, and haste it has.

His moon that shows the sun, from behind the veil of his hair, is a sun that before it a cloud has.

My eye has sent a flood of tears running into every corner so that your straight cypress fresher water has.

Your bold glance spills my blood by mistake. May it be given time: a fine, sound idea it has.

If the water of life is this that the friend's lip holds, then plainly Khizr for his share a mirage has.

Your wine-heavy eye means to take the liver out of my heart: it is a drunken Turk, and perhaps a taste for roast meat it has.

My sick soul lacks the face to ask you anything. How happy the wounded man who from the friend an answer has.

When will she turn a glance toward Hafez's wounded heart— that drunken eye of hers, which in every corner a wreck has?

غزل ۱۱۸

The man who in his hand a cup has the sultanate of Jamshid for ever has.

The water from which Khizr found life— seek it in the tavern, which the cup has.

Leave the thread-end of your life to the cup, for from the cup this thread its order has.

We and wine on one side, the ascetics and piety on the other— to see for which of them the friend an inclination has.

Outside your lip, cup-bearer, there is nobody in this whole turning age who his desire has.

Every one of drunkenness's manners the narcissus on loan from your lovely eye has.

The mention of your face and hair is, for my heart, a litany that morning and evening it has.

On the raw breast of those who suffer your ruby lip salt in full has.

In the well of your chin, like Hafez, my soul, your beauty two hundred slaves has.

The other three arms, one bayt shown four ways

Grades

Radif: HELD for both poems. دارد → has stands last at 10 of 10 rhyming positions in sh124 and 10 of 10 in sh118, in the IP and IA arms. Nothing was bought: pair_check.py verifies that every I line is a permutation of its D counterpart, so no word was added to make the tail reachable.

Chime: NOT ATTEMPTED — R56 §5, and see the log entry D7.

Translator's log

Frozen 2026-08-29 before E-20260829b-inversion-price was designed and before any published English of either ghazal was sought. Decisions and alternatives; the two evaluative sentences are flagged in place and the page carries internal-judgment-only.

D1 — why the object had to move at all, which is not obvious from the Persian. Persian is verb-final, so شتابی دارد (a haste it-has) is unmarked order and the radif falls at the line end for free. English is verb-medial: has haste is unmarked and haste has is not. The whole cost of this device in English is the cost of one syntactic move, and it is a move English makes grudgingly — object-before-verb is available for topicalisation and for contrast, and both readings are wrong here, because in Hafez the object is not topical: it is the rhyme.

D2 — the option list, written before the first English line, per R51 §1. Four ways to get "has" into last position: (i) front the object (a cup has); (ii) front an adverbial and leave the object out of the clause entirely (on loan from your eye has, with the object stranded in the previous hemistich); (iii) convert to an existential or a relative so that the verb is final by right (…the cup there is), which changes the verb and so loses the radif; (iv) use a different English verb whose complement can precede it. (iii) and (iv) were refused, since either abandons the identity of the tail, which is what a radif is. So (i) and (ii) are the whole repertoire, and which one a position permits is decided by the Persian, not by the translator: (ii) is available exactly where the Persian already puts something other than the object before the verb. Two positions in sh118 do that (به وام دارد, صبح و شام دارد), and the item table codes them ADV before any English was scored.

D3 — the selection was made for a reason and the reason is RS-20260829 §7.41.5. A کرد ghazal cannot price an inversion because it does not need one: where the radif is a light verb whose object is the qāfiya, English has the same compound and the shape crosses whole. دارد with an indefinite-noun object at every qāfiya is the opposite case and the one Payne's hath actually carries. Both poems were chosen from the census on the frozen rule in ../source.md and neither was read before selection.

D4 — what the archaic arm is, and what it deliberately is not. The map applied is lexical and morphological only: has→hath, your→thy, you→thou/thee, pass→passest, shows→showeth, spills→spilleth, means→meaneth, lacks→lacketh, holds→holdeth, it is→'tis, is→be (subjunctive after if), the man who→whoso, for→unto, in→within, for ever→evermore, from which→whereof, perhaps→belike, anything→aught, nobody→none, those who→them that, by mistake→in error, idea→thought, morning and evening→morn and even, toward→upon, on→upon, can→may. No entry moves a word. That constraint is the point: if archaism is doing the work §7.41.3 says it does, it has to do it without touching the order, or else the two factors are the same factor and the framework's sentence is not a claim about register at all.

D5 — the two hardest positions, and both are MIXED. G1-8 (How happy the wounded man who has an answer from the friend) and G1-9 (which has a wreck in every corner) have both an object and a PP after the verb, so carrying the radif fronts both, which is a heavier move than fronting an object alone. G2-1 is the same shape (has the sultanate of Jamshid for ever). I did not lighten them — dropping the PP would have bought the tail with a cut, which R51 §1 forbids — and they are coded MIXED so the item table can see them.

D6 — one place where the translation was bent for the regime, and it is recorded as a loss. sh124 bayt 6 line 1 has دارد inside the hemistich: آب حیوان اگر این است که دارد لب دوست. Rendered as has, the first line would end on the radif word by accident and the manipulation in line 2 would no longer be the only occurrence of the tail in the item. I rendered it holds. That is a real, if small, loss: Hafez repeats the same verb inside the line and the English does not. The same substitution was made at G1-4/G1-8 line 1 (sent … lacks) for the same reason. These are the only places where the experiment's requirements reached into the translation, and they reach into the unmanipulated line, never the manipulated one.

D7 — why no chime, and what is thereby not known. R55 would require the rhyme to fall on the word before the tail: ten rhymes in -ābī and ten in -ām. On this project's own record (T-gulistan-R42-v1, six refusals of eight because the chime needed an end-filler) a monorhyme held ten deep is bought with words the Persian has not got — and a bought word appears in both arms of a pair, which changes what the pair is a pair of. So the chime was refused by regime rather than attempted and graded. What is not known as a result: whether a carried rhyme makes the inversion easier to take or harder. There is an argument each way and no measurement, which is why this is in the log and not in the finding.

D8 — an observation I am registering and not testing. (evaluative; internal-judgment-only.) Writing the IP arm, the fronting felt cheapest where what moved was short and heavy — a cup has, the cup has, its order has — and dearest where what moved was a long phrase with its own internal structure — a taste for roast meat it has, the sultanate of Jamshid for ever has. That is end-weight, the ordinary English preference for putting the long thing last, and it is orthogonal to the object/adverbial distinction the item table codes. The item table does not code phrase length, and I am not adding it after the fact; it is named here so that a later design can code it in advance.

D9 — what a reader of the Persian loses in the D arms. In DP and DA the word has is still there at every rhyming position, just not last. So the D arms are not "the radif dropped"; they are the radif present and unaligned, which is a real translator's option and, on the coded record, is not what Leaf chose — he dropped the word entirely and monorhymed instead (RS-20260829-radif-hands §4). The arms here therefore price the alignment, not the repetition. That is narrower than the framework's sentence and the result page must not widen it.