Repository path: wiki/base/anchors/A-morri-botchan/A-morri-botchan.md · rendered 2026-09-09

Tier 1 precedent anchor — Yasotarō Morri, Botchan (Master Darling) (1918), against Sōseki's 「坊っちゃん」(1906), chapter 1

The project's second JA→EN precedent cell, built for the reason A-garnett-vanka was built against A-shaw-spider-thread: to find out whether the first cell's findings are about that translator and that story or about the language pair. It is also the shelf's first anchor whose published translator left a written statement of his own register programme, so what he intended and what he did can be read against each other — the configuration this project has, until now, only ever had for its own lead.

The close reading below is the lead agent's and is flagged internal-judgment-only. The register figures are not: they are four non-Anthropic seats rating 288 cells blind, recomputed by an independent verifier (19 checks, 0 failures), and they are reported in RS-20260807b-register-collapse.

What it is

• Source. 夏目漱石「坊っちゃん」, serialised in Hototogisu, April 1906. Chapter 1 whole: 7,499 non-space characters, 22 paragraphs, stored here as botchan-soseki-1906-ch1-ja.txt. Copy-text Aozora Bunko 000148/752_14964, 底本『ちくま日本文学 全集　夏目漱石』筑摩書房, fetched 2026-08-07. Sōseki d. 1916: public domain.
• Translation. Yasotarō Morri (1882–1959), Botchan (Master Darling), Ogawa Seibundo, Tokyo, 1918 — the earliest published English Botchan (S-botchan §Identity, cross-checked against Sato's 2019 bibliography). Chapter 1 whole, 4,461 words, stored as botchan-morri-1918-ch1.txt with the uncorrected scan beside it as botchan-morri-1918-ch1-rawocr.txt.
• Rights, cleared per translator and separately from the author. Morri's translation: US publication 1918, pre-1929, public domain on the publication basis; the life basis (d. 1959) is not relied on. Both texts are therefore storable whole, which is why this is a complete parallel reading rather than an excerpt sketch (charter §7, A8).
• Where it sits in the shelf's design. Structurally far, culturally far — the same cell as A-shaw-spider-thread, deliberately. The two differ on everything else: Akutagawa's 蜘蛛の糸 is a third-person Buddhist parable of 1918 rendered by an American in 1930; Botchan is a first-person comic novel of 1906 rendered by a Japanese journalist in 1918, and its narrator is constituted by his register.

Provenance & access route

Aozora Bunko over HTTPS (tools/fetch_aozora.py, cp932) and Internet Archive botchanmasterdar1918nats djvu text, both 2026-08-07, both free. No copyrighted translation was opened; the five other English Botchans (S-botchan) are in copyright and were not consulted for this page, so nothing here compares Morri with a successor.

The scan is uncorrected OCR and this page treats it as a witness, not as a text. It breaks words across lines (pro vince), sets quotation marks with inside spacing, reads them as u/il/f, and drops running heads and one printed footnote into the body. Every string quoted below was located in the scan by machine before being quoted; where a quotation restores a word break or a quotation mark, the scan's own reading is given in brackets. The furniture-stripping is 14 blocks, all printed by materials/build_loci.py.

What it grounds

1. The chapter is a two-pole register system, and the poles are externally visible

The narration runs in the plain, blunt 俺-register: おやじ for the father, 婆さん for Kiyo, 奴 used of people and of chestnuts alike, ぬ/ん negatives beside slang (無鉄砲, 馬鹿気ている, 二束三文, 瓦落多, 真平ご免だ, 持てあました). Against it stands Kiyo's deferential speech — あなた, お/ご prefixes, なさいます, 下さい, and at its top 遊ばせ — and the honorific by which she addresses him, which is the book's title. The father's 貴様 is a third point, contemptuous-familiar, below the narration.

This is not the lead's impression: four seats reading the Japanese alone, on one 0–6 register scale, order the three sets HIGH > NEU > LOW at 4 of 4, with a pooled HIGH − LOW of 2.325 and mean pairwise Spearman 0.742 (RS-20260807b §3).

2. Morri declared his register programme in his own preface, and it is the programme this cell tests

"It is profusely interladed [interlarded] with spicy, catchy colloquials patent to the people of Tokyo for the equak [equal] of which we may look to the rattling speeches of notorious Chuck Conners of the Bowery of New York. It should be frankly stated that much difficulty was experienced in getting the corresponding terms in English for those catchy expressions. Strictly speaking, some of them have no English equivalents. … The slangs in English in this translation were taken from an American magazine of world-wide reputation editor of which was not afraid to print of "damn" when necessary, by scorning the timid, conventional way of putting it as "d—n."." — Morri, "A Note by the Translator", 1918

Three things are on record in that paragraph that this project reached independently, 108 years later: that the narrator's low Tokyo colloquial is what the book is made of; that English has no corresponding terms for it; and that the translator's answer was to import a domestic low register (Bowery slang) rather than to invent one. The third is Berman's « pseudo-argot » — the move R21 was written to perform (R21 §V5) — proposed in 1918 as a solution by the translator himself.

3. What he actually did: the low pole does not arrive, and the high pole does not either

Measured on the same six low-marked and five high-marked sentences in all four texts (RS-20260807b), the spread between the poles collapses:

Morri retains four per cent of the source's register architecture. The narration's lowness is gone — C_low is exactly 0.000 — and Kiyo's deference is very nearly gone with it. The single sharpest cell is the novel's opening sentence, 「親譲りの無鉄砲で小供の時から損ばかりしている」, which the seats put at 2.5 in Japanese and at 5.25 in Morri's "BECAUSE of an hereditary recklessness, I have been playing always a losing game since my childhood" — a 2.75-point rise on a seven-point scale, on the sentence that establishes the voice.

4. The raising is done by an available English shelf, and it is deep

Latinate abstraction and literary syntax, used in ordinary narration where the Japanese is at its plainest: 弱虫 → "a mollycoddle"; 弱虫の癖に → "he had the temerity to come over the fence"; 足搦をかけて → "by a dexterous foot twist"; 苦しがって → "no longer able to stand the stifling combat"; 古川が真赤になって怒鳴り込んで来た → "Furukawa, fiery red with anger, burst into our house with howling protests"; たしか罰金を出して済んだようである → "I believe the affair was settled on our paying for the damage"; 少々気味がわるかった → "I felt even chilled through my marrow"; 先祖代々の 瓦落多 (junk) → "all the bric-a-bric [bric-a-brac] which had been handed down from ages ago in our family". This corroborates RS-20260806g's craft finding on a second language pair and on a published hand: English's raising shelf is deep, available in narration, and reachable by lexis alone.

5. The lowering shelf exists in his English too — and he uses it at one site in six

Morri's Bowery seam is real and is exactly where his preface said it would be: 真平ご免だ → "Nix on languages or literature!"; 大変な遠くへ行かねばならぬ → "I must go darn far away, that it beats Kamakura by a mile"; 小言はただの一度も聞いた事がない。喧嘩もせずに済んだ → "with no kick coming or having no rough house"; 清はおれの自慢を… → "Kiyo would boost me sky high"; 二束三文に売った → "sold for a song"; こんな婆さんに逢っては叶わない → "Nothing can beat an old woman for this sort of thing, I tell you".

So the null in §3 is not a null about English and not a null about Morri's repertoire. It is about distribution: of the six sentences the source marks low, his English sits lower than its own neutral at one — 妙なおやじがあったもんだ → "A funny dad he was" (source 1.25, Morri 1.75, the only strongly negative cell in the run at −2.00). At the other five he raises. The slang is present in the chapter and absent from the places the Japanese marks.

That is A-shaw-spider-thread's finding, on a different property and a different translator: there, "the realia-handling range used inconsistently in one story"; here, the register range used inconsistently in one chapter. Two of the shelf's two JA→EN cells now show the same shape, which is what makes it a candidate generalisation rather than an observation about Shaw.

6. Deference dies by a lexical move, and the syntactic move is a description of what he did

Heading corrected 2026-08-07 (S131), and the correction is the point. This section originally read "Deference dies by a syntactic move, not a lexical one" and asserted the conversion as the mechanism. RS-20260807g tested that claim in a controlled ladder — 17 deferential utterances from two works, content held fixed, discourse form varied, the manipulated arms written by an independent model and rated blind by three seats — and the mechanism claim does not survive. Removing the deference words while keeping the quotation costs 1.9259 on a 0–6 scale; reporting the speech instead of quoting it, keeping every word grammar allows, costs 0.2963 — a ratio of 0.154 against a registered bar of 0.50, failing at 2 of 3 seats and 6 of 9 utterances, and stable at 0.128–0.154 under every exclusion. On Ōgai's courtroom items the conversion costs nothing: "I have come with a petition for His Honour the Magistrate" reports as she had come with a petition for His Honour the Magistrate, and the honorific phrase, the word petition and the bow all survive, because none of them is aimed at a listener.

And Morri does not track his own choice. Across all twelve of chapter 1's deferential utterances he scores 2.381 on the seven he converts and 2.600 on the five he keeps in quotation — level, and level with an arm built by deliberately stripping the deference words (2.431). His lowest single item is one he kept in quotation.

What stands, as observation: the conversion is real, it is his commonest move, and the count is now exact — 7 of 12, registered before the run and confirmed mechanically. What is withdrawn: that it is why the register collapsed. The run's own primary is WITHHELD (its parity gate failed at 27 of 34 against 0.80, by one item), so this correction rests on descriptive figures — but they point one way at every cut, and the claim they replace rested on inspection of four sites.

The four sites the original observation was made on. Morri's high pole is lost at all of them, and at these four he converts Kiyo's direct speech to indirect, deleting the morphology outright:

And he moves the boundary the other way for the men: where the Japanese runs the father's and the classmate's speech unmarked inside the narration, Morri lifts them out into quotation marks — "Say, you big bluff, I'll bet you can't jump down from there! O, you chicken heart, ha, ha!", "This fellow will never amount to much". So the speech/narration boundary is redrawn in both directions, and in both the movement runs against the source's register: the men gain a dramatic frame the Japanese withholds, and the servant loses the grammar that is her whole social position.

7. 坊っちゃん → Master Darling, and the narration then has to gloss it

The title word is a coinage: Master (English domestic-service address to a boy of the house) plus Darling (endearment). It is the one place Morri builds a device where English had none. The cost is visible inside his own text: at 「そんなにえらい人をつらまえて、まだ坊っちゃんと呼ぶのはいよいよ 馬鹿気ている」 the narration has to hold the coinage at arm's length — "But then how absurd to call such a 'great man' 'Darling.'" — splitting the compound and quoting half of it. A device that has to be explained by the narration is not carrying the register on its own.

8. Ennoblement in Berman's sense, at one site, by omission

The three-yen purse. 「その三円を蝦蟇口へ入れて、懐へ入れたなり便所へ行ったら、すぽりと後架の 中へ落してしまった」 — he took it to the privy and dropped it down the pit. Morri: "I put it in a bag, and carried it in my pocket. Whiie [While] about the house, I happened to drop the bag into a cesspool." The going-to-the-privy is deleted and replaced with a vague locative; the destination survives. The mimetic すぽり goes too. This is S-berman-tendances' ennoblissement occurring in a published JA→EN translation — and unlike the register raising in §4, it is done by omission, which RS-20260806b's scales could not have seen.

9. Cultural mediation: three settings in one chapter

• Transliterate and footnote — 越後の笹飴 → "sasa-ame" with the chapter's one printed note: "Sasa-ame is a kind of rice-jelly wrapped with sasa, or the bamboo leaves, well-known as a product of Echigo province." The single foreignisation, and it is paratextual.
• Flatten — 金鍔や紅梅焼 → "some cakes or sweetmeats"; 靴足袋 → "socks"; 蕎麦湯 → "hot noome [noodle] gruel"; 鍋焼饂飩 → "a bowl of steaming noodles from the peddler" (the peddler is added); 瓦解 → "the Restration [Restoration]".
• Explicitate with an added frame — 「どういう因縁か」 → "by what affinity, as the Buddhists say, I don't know". The attribution is not in the Japanese. Morri supplies a cultural explanation the source does not offer, in a chapter where he elsewhere deletes one.

10. Two factual departures, recorded because an anchor is a precedent and not a paragon

• 「芝居の真似をして女形になるのが好きだった」 → "he had a fondness for taking the part of an actress at the theatre." An 女形 is a male actor playing female roles; the English makes the brother play an actress, which is a different fact about the Meiji stage.
• 「腰を抜かした」 → "laid up", then "get your bones dislocated", then "I'll see I don't get dislocated next time." 腰を抜かす is losing the use of one's legs from shock, not a dislocation. Morri does hold the repetition across the three occurrences, which is the harder half of that sentence and which he gets right.
• One garbled sentence is the scan's or the printer's and is recorded as unresolved: 「ただ懲役に 行かないで生きているばかりである」 → "I am living with out [without] becoming but a jailbird."

Storage discipline

Both texts are public domain and are stored whole in this directory, per the README's rule. The raw OCR is kept alongside the working copy so that every restoration on this page can be checked against what the scan actually says. No copyrighted translation of Botchan is stored or excerpted here.

What this anchor does NOT ground

1. It is one translator. Every §3 figure is about Morri 1918. The generality claim — that the low pole does not cross — rests on the lead's independent rendering failing the same way (C_low(R04) = +0.167, at 4 of 4 seats) and on RS-20260806g's Italian evidence, not on this page.
2. Period is not controlled. A 1918 translator's English is formal by 2026 standards; "raised" here cannot be separated from "period", and no successor translation was consulted.
3. The instrument is four language models. Tier D is NOT PASSED and nothing here is a jury verdict: the seats rated register height, never quality, and no arm was scored for goodness. One of the eight English calls recognised the work and said so in its reasoning; dropping that seat moves no conclusion (RS-20260807b §6).
4. R21 is a control, not a rival rendering, and its presence in the table says only that the low pole is reachable in English narration — at the price of the high pole, which it also destroys.
5. §§4–10 are the lead's close reading. The quotations are machine-located in the stored files; the classifications are not checked by anybody.