Repository path: workshop/translations/jeli-il-pastore/register.md · rendered 2026-09-09
Page metadata (front matter)
| type | ledger |
|---|---|
| id | jeli-binding-register |
| status | active |
| created | 2026-07-27 |
| updated | 2026-07-28 |
| senses | consistency, voice, cultural-mediation, style-correspondence |
| internal-judgment-only | true |
| provisional | true |
| links | workshop/translations/jeli-il-pastore/R05-v1/translation.md, workshop/regimes/R05-serial-long-work.md, wiki/arms/ARM-longwork.md, workshop/translations/jeli-il-pastore/R05-v1/span2-fid-prereg.md, wiki/base/sources/S-verga-farina.md |
Binding register — «Jeli il pastore»
What this is. The decisions made in an earlier span that bind every later span of the same work. A short translation does not need one: everything is in view at once, and consistency is a property of a page you can re-read. Over 11,474 source words in five sittings weeks apart it is not, and "keep it consistent" stops being an instruction and becomes an instrument or it does not happen.
Status of the entries. Every row cites the log decision that created it (D<n> in R05-v1/translation.md). A row may be revised, but only by a numbered erratum on the translation page plus a dated line in §Revisions below — never by silently changing the row, because the point of the register is to make drift visible rather than to prevent it.
Unresolved rows are as important as resolved ones and are listed separately. An unresolved row was a decision the next span was required to settle, not one it might quietly inherit. After S047 the work is translated whole and there is no next span, so what remains open (U4) is open permanently, and U6 is closed unsettled with its measurement recorded — see §Closed without settlement.
Voice and register
| # | binding | from |
|---|---|---|
| V1 | Plain modern English, unliterary, rural lexicon. No English regional dialect, no dialect spelling, no archaism. | D1 |
| V2 | Verga's parataxis is preserved; comma-chains and e-chains are not converted to subordination. |
D1, D10 |
| V3 | Proverbs and community idiom are rendered literally and left strange, never swapped for an English proverb. | D9, D13 |
| V4 | Contractions in speech, none in narration. (Adopted by drift at D19; tested at span 2 against 24 speech turns and kept — D24. One rhetorical exception, named there: a full form is allowed in speech where the contraction would lighten the gravest line a speaker has.) | D19, D24 |
| V5 | Imitative sounds respelled for an English ear: iuh! → yuh!. Calls and cries are not respelled and are kept as the source sets them — ohi! ohi! ohi!, ahu! ahu! ahu!. A sound that has become a noun is translated by the corresponding English noun, not respelled — il ron ron del contrabasso → the thrum of the double bass. (Scope clauses added at span 3, D45, and span 4, D59: V5 governs utterances; every available English respelling of a call imports a culture that is not there, and a lexicalised sound-noun is an ordinary word in Italian and must be an ordinary word in English.) |
D11, D45, D59 |
| V6 | No added frame. Where the Italian gives community voice no attributive frame, the English adds none — no as they said, no quotation marks, no italics. | D20 |
| V7 | Verga's guillemets are kept as one mark — single inverted commas, no gloss, no expansion. The marked/unmarked contrast is authorial and is not flattened in either direction. | D20, D21 |
| V8 | Dash-set speech takes English quotation marks, with the attributive tag left where Verga puts it (after the first clause, not fronted). | D20 |
| V9 | A gnomic present or a conditional outside the narrative sequence stays outside it, never harmonised to the narrative past — where English has a form for it. | D20, D22 |
| V10 | Verga's intensifying reduplication is kept as reduplication — adagio adagio → slowly, slowly; bianco bianco → white, white; fredda fredda → cold, cold. Scope, re-drawn at span 5 (erratum E6): kept where the English word is short and doubles idiomatically; rendered singly where it does not — serio serio → quite seriously, mogio mogio → downhearted. ⚠️ The criterion is euphonic and would not replicate for another reader — the same defect U4 carries. D58's grammatical scope (adverbs of manner and degree only) was contradicted by frozen span-1 and span-2 prose it was written over: 24 sites in the work, 6 of them reduplicated adjectives, treated 5 different ways. |
D53, D58, D71 |
Names, address, and the horses
| # | Italian | English | note | from |
|---|---|---|---|---|
| N1 | don |
don — lower case |
as Verga sets it; against English convention | D5 |
| N2 | fra |
Fra |
Fra Cola | D4 |
| N3 | massaro |
Massaro |
Massaro Agrippino | D4 |
| N4 | gnà |
gnà |
gnà Lia | D4 |
| N5 | campiere, campaio |
field-guard | occupation, therefore translated; two source words, one English word | D4, D29 |
| N6 | vaccaro |
cowherd | " | D4 |
| N7 | lettighiere / lattigliere |
litter-man | source spells the place-name both ways; regularised | D4, D15 |
| N8 | pastore |
herdsman | incl. the title | D2, D3 |
| N9 | pastorello |
herd-boy | D3 | |
| N10 | signorino |
the young master | must be identical in narration and in Jeli's joke | D12 |
| N11 | la Bianca |
Bianca | the bell-mare | D6 |
| N12 | il puledro zaino |
the bay colt | D6 | |
| N13 | la cavalla vajata |
the dapple mare / the Dapple | second form when addressed as an individual | D6 |
| N14 | la cavalla mora |
the black mare | D6 | |
| N15 | mandra, armento |
herd (horses, cattle) / flock (sheep) | two source words, one English word — but the English word depends on the animal, and the row broke on it at span 4. Erratum E5 | D7, D56 |
| N16 | branco |
drove | D7 | |
| N17 | sacca di tela, tasca di tela |
canvas bag | two source words, one English word | D8 |
| N18 | batuffoletto dei cenci |
the bundle of rags | recurs; Jeli's whole property | D6 (class) |
| N19 | zufolo |
pipe (zufolo di sambuco → elderwood pipe) |
— | |
| N20 | onze |
onze | not converted, not glossed | D14 |
| N21 | compare, zia, zio |
compare, zia, zio |
three more address terms, kept; case as Verga sets it — lower-case don, gnà, compare, zia, zio; capitalised Fra, Massaro |
D4, D5, D31 |
| N22 | tumolo / tumoli |
tumoli | Sicilian dry measure; follows N20 | D30 |
| N23 | sarmenti |
cuttings | as span 1; fascio di sarmenti → bundle of cuttings |
D-span1, D28 |
| N24 | ecalibbiso / eucaliptus |
ecalibus / eucalyptus | the pair is the point — the peasants' form in their mouths, the correct form in the narrator's | D28 |
| N25 | fattore |
the overseer | the estate manager, a senior rank; distinct from N5's campiere/campaio → field-guard, and the distinction is load-bearing at ¶79 |
D39 |
| N26 | lo stellato |
the star colt / Star in address | white blaze on the forehead; second form when addressed, as N13 | D40 |
| N27 | il morellino |
the little black / my fine little black in address | " | D40 |
| N28 | la puddara |
puddara |
Sicilian for the Pleiades; kept unglossed with onze and tumoli, because a book-word would not be in Jeli's mouth |
D46 |
| N29 | del tu (the T-form) |
tu, italicised | English has no form; address-terms are kept (N1, N4, N21) and tu is one. Precedent: span 1's excellency for vossignoria |
D47 |
| N30 | paneperso |
bread-waster | a one-word insult, literally lost-bread; the bread thread runs ¶61→¶80→¶88→¶92→¶105→¶124 and no idiomatic English keeps it | D44, D65 |
| N31 | pecoraio |
shepherd | the sheep job Jeli takes at the Salonia. Distinct from N8 on purpose: the source uses two words and so does the English, which is what makes the demotion visible. The feminine will not cross: pecoraia → keep sheep (verb), because English shepherdess is a pastoral word V1 forbids — D82 |
D55, D82 |
| N32 | galantuomini |
the gentry | the men who do not work with their hands; recurs across the span 3 / span 4 boundary | D67 |
| N33 | comare / comari |
comare / comari |
the feminine of compare; joins N21 |
D66 |
| N34 | grano (coin), palmo / palmi (length) |
grano, palms |
Sicilian units, unconverted and unglossed, with onze (N20), tumoli (N22) and puddara (N28). lira is an English word and is not in this class |
D60 |
| N35 | soldo | soldo / soldi | a coin whose value an English reader cannot compute; joins onze (N20), tumoli (N22), grano (N34), puddara (N28) under D14. lira remains an English word and is not in this class | D83 |
| N36 | roba | irreducible — no single English word | Verga's great word, and the register's largest gap, found by census at span 5 and not by noticing: 6 occurrences, already rendered everything he owned (¶9), things (¶51, ¶52), and now everything we had (¶140, property), goods (¶181, wealth), provisions (¶187, food for a dinner). This row records a loss rather than instructing a rendering — the first row of that kind. Any later use must be chosen for its sense and logged | D73 |
| N37 | ballatojo | the gallery (a bell-tower's) / the landing (a house's outside stair-head) | one Italian word, two places; identical in shape to N15. ¶79 is the Tebidi bell-tower, ¶158 and ¶177 are Mara's town house | D74 |
Place-names are kept as in the source and are not listed individually: Tebidi, Licodia, Vizzini, Ragoleti, Passanitello, Donferrante, la Commenda, il Jacitano, Marineo, Resecone, Caltagirone, Poggio alla Croce, Buccheri (il Bucchierese → "from Buccheri"), and the fair of San Giovanni.
Unresolved — and the work is now complete, so nothing further will settle them
| # | question | why it is open | raised at |
|---|---|---|---|
| U4 | How the narrator's own lyric voice is to be told apart from the community's, once a span contains both at length. Span 3 made this worse, not better (RS-20260727d §3): an independent critic produced eight further span-3 sites meeting the criterion as worded, showing that criterion (iii) has no boundary at all. Span 2's "9 of 21" is a count of one reader's list, not of a set. |
V6 tells the translator not to frame community voice and does not say how to recognise it — and the project now knows that declaring the criterion soft describes it without bounding it. No boundary was found; the row stays open and its difficulty is now measured rather than asserted. | prereg §1(iii); RS-20260727d §3 |
Closed without settlement
| # | question | how it ended | at |
|---|---|---|---|
| U6 | Whether D62's prior-attribution rule survives its one remaining site — ¶171's «chè le corna sono magre, ma mantengono la casa grassa!», an unframed proverb never previously attributed, and the only site in 11,474 words where the proverb and prior-attribution accounts come apart. |
NOT SETTLED, by the experiment's own pre-registered gate. E-20260728d put all eight sites to three independent models; the instrument gate failed on both legs, so the test site's verdict is not reported as bearing on D62 (RS-20260728d §4). Descriptively the site returned a unanimous fixed expression reading, identical to the framed proverb at ¶5 — and the gate forbids using it. The row closes unsettled because the work is finished and there is no further span; what would settle it is a second work with more than one unframed proverb, and an option scheme not written by the translator. |
S047 |
Why this row could not simply be left open. An unresolved row is one the next span is required to settle, and after S047 there is no next span. A row left open on a completed work is not a live question, it is an unmarked loose end — so U6 is closed with its result recorded as a null rather than carried.
Settled
| # | question | settled how | at |
|---|---|---|---|
| U1 | How to mark free indirect discourse. | Settled from Verga's own statement of method, not from the translator's ear: rows V6–V9, derived in R05-v1/span2-fid-prereg.md from S-verga-farina and frozen before span 2 was written. |
D20, span 2 |
| U2 | Whether V4's contraction split survives sustained dialogue. | Survives, unchanged, against 24 speech turns including two extended exchanges; one rhetorical exception recorded rather than a rule changed. | D24, span 2 |
| U5 | Whether V7 survives an unframed guillemeted clause — the class span 2's strain was thought to belong to. | Settled in V7's favour, and the variable turned out to be a third one. ¶125's «che sembrava un manico di saliera» is unframed and takes single inverted commas without strain, because the reader met these exact words in Mara's mouth 3,900 words earlier (span 3, ¶79). The marks are not introducing a speaker, they are recalling one. So the discriminating variable is neither bare-NP-ness nor unframedness but prior attribution, and span 2's three failures were failures of first appearance. V7 unchanged; the new question is U6. |
D62, span 4 |
| U3 | Whether V7 survives a guillemeted phrase that is neither a proverb nor a bare noun. | Settled in V7's favour, narrowly. ¶65's «come San Giovanni fosse arrivato sotto l'olmo,» is a clause with Verga's own attributive frame («diceva il contratto») and takes single inverted commas with no strain: the frame supplies the speaker, so the marks are free to be ordinary quotation. The settlement is narrow and it opens U5 — the class that strains may be unframed, not bare-NP. | D43, span 3 |
Revisions
Dated entries, each pointing at the erratum on the translation page that authorises it.
- 2026-07-27 (S039), erratum E3.
V4status changed from adopted by drift, flagged to tested and kept (D24);N5extended to covercampaio(D29). RowsV6–V9andN21–N24added, not revised. - 2026-07-28 (S045), erratum E5.
N15made referent-sensitive —mandra/armento→ herd of horses and cattle, flock of sheep (D56). The row was set when every referent in the story was equine and broke silently when span 4 moved Jeli to sheep; spans 1–3 are unaffected.V5gains a second scope clause (D59), and the reduplication rule that had lived only in the log becomes rowV10, scoped to manner and degree (D58). RowsN31–N34are additions.U5settled with a variable neither of its two candidates named (D62), andU6opened in its place. -
2026-07-28 (S047), erratum E6.
V10re-scoped — from a grammatical boundary (adverbs of manner and degree only) to a euphonic one (kept where the English word is short and doubles idiomatically), because the grammatical boundary set atD58is contradicted by frozen prose from spans 1 and 2 that nobody had checked it against:bianco bianco→ white, white is an adjective predicated of a face and is doubled. No frozen prose changes. The row now carries an explicit flag that its criterion would not replicate for another reader (D71).N31gains the note that its feminine will not cross (D82). RowsN35–N37are additions. A second finding of the same shape is recorded on the translation page and repairs nothing:V7was created in span 2 and never applied backwards, so the two span-1 guillemet sites carry double quotation marks, not V7's single ones — measured rather than tidied (E-20260728d). Two rows written late and never checked against the spans they describe: one failure mode, found twice in one session. -
2026-07-27 (S042), erratum E4.
V5scoped — from onomatopoeia respelled to imitative sounds respelled, calls and cries kept (D45). This is the row's first substantive revision; it narrows what V5 binds and adds nothing. RowsN25–N30and unresolved rowU5are additions, not revisions.U3settled (D43);U4restated with its difficulty measured rather than asserted (RS-20260727d§3).