Repository path: workshop/translations/tini-budz/README.md · rendered 2026-09-09

«Тіні забутих предків» — the milking and the making of the budz (chs. X–XI)

The second unit taken from this work. The first is tini-polonyna (S147, pages 363–365 of the same index); this one continues it, four pages further on, and is built for the same purpose: a passage unusually dense in one coherent class of culture-bound items, with those items repeating inside single blocks.

The work

Mykhailo Kotsiubynsky (1864–1913), «Тіні забутих предків» (Shadows of Forgotten Ancestors), written 1911, first published 1912. Hutsul highland novella. Public domain: the author died in 1913.

The copy-text

Ukrainian Wikisource, Твори (Коцюбинський), Vol. II, New York: Книгоспілка, 1955, transcribed from the DjVu scan under the ProofreadPage extension. Pages 368–371 of the index Коцюбинський М. Твори в 2-х т. Т. 2 (1955).djvu, fetched 2026-08-10 as action=raw from the Сторінка: namespace. All four pages carry pagequality level="4" (proofread). Wiki markup was stripped mechanically; no hyphenation repair was needed — all three page joins inside the unit fall at word boundaries («…важко спочиває | в дерев'яній посуді…», «…поки там твориться | щось…»), and each was rejoined and checked by printing the seam.

Single witness. No collation against a second edition was performed, and nothing in the run built on this text turns on its constitution.

The editorial footnotes were seen, and that is declared rather than denied

The 1955 edition glosses four of the words in this unit — струнка "намет для доїння овець", козар "козячий пастух", подра "полиця під дахом", жентиця "сироватка". They were printed to the console while the class items were being counted, i.e. before the passage was translated. They are dictionary-level denotational information of exactly the kind R04 §1 permits ("the translator reads the source text and whatever non-translational apparatus it needs"), they are not a rendering into any target language, and the four glosses are recorded here so that the record does not have to be taken on trust. tini-polonyna's two footnotes were handled the other way — noted as existing, not read — and the difference is recorded rather than smoothed over.

The unit

source-unit.txt — 769 Ukrainian words in three blocks, running from «Вже третю добу сіє на полонині…» (the third day of rain) to «…струнки чекають на нове молоко…» (the trembita at sundown).

One elision, declared. Between BE and BF the copy-text prints the spuzar's song — two <poem> blocks of Hutsul verse. It is omitted, because verse is a different craft and a translated song inside a prose unit would be noise in any design built on the unit's prose. BD→BE is continuous; BE→BF skips the song and nothing else.

Why this unit. Selected by counting before it was read closely: class items were tallied per page across pages 363–377 first, and pages 368–371 came out densest after the pages the first unit had already used. The within-block repetition — which is the property DRIFT needs and the only property the unit was chosen for — is:

tini-polonyna offered four to five such items per block; this unit offers four to seven, and BD is the densest block either unit contains.

Translations filed here

Contamination

none, and it is a reachability statement, not a measurement — identical in kind and in basis to tini-polonyna's. The only professional English translation of the work is Marco Carynnyk's (1981), which is in copyright, is not freely reachable, and was not consulted. No English rendering of this work is obtainable by this project, so there is no comparator and tools/dependence_check.py could not be run. Per the standing rule (CLAUDE.md, 2026-07-26) a contamination: declaration without a measurement is a placeholder that can hide either end of the measured range, and this one is declared as such.

Why the design built on it does not turn on that, and it is the same argument as before: every arm of E-20260810t is a variant of one lead rendering, and every comparison is within-run and within-translator. A remembered Carynnyk phrase, if any survived in the lead, would sit in all arms alike and cannot produce the contrast under test.

The one thing this unit has that tini-polonyna does not, and it is a limit

The translator knew, before translating, that the text would carry a terminology-drift design. tini-polonyna was translated before E-20260809i existed; this unit was not, and the shape of the successor design was already written down (RS-20260809i §8) when the draft was begun. The consequence is bounded but real, and is stated on both artifacts and in E-20260810t/design.md §8: the MIXED arm is the lead's own class handling and could have been influenced by knowing what it was for, so MIXED is descriptive here and carries no weight in any registered test. The primary contrast — MATCH against DRIFT — is between two arms generated from the marked sites, and is invariant to which form the base chose at any site.