Repository path: workshop/translations/max-havelaar-i/R07-v1/translation.md · rendered 2026-09-09

T-max-havelaar-i-R07-v1 — Multatuli, Max Havelaar ch. I ¶1–4, under R07 (fluency)

• Source. Multatuli (Eduard Douwes Dekker), Max Havelaar, of de koffiveilingen der Nederlandsche Handelmaatschappy (1860), chapter I, first four body paragraphs. 716 Dutch words. Public domain (author d. 1887). Provenance at ../../../canon/max-havelaar-i/manifest.md.
• The project's first Dutch and its sixteenth source language.
• Regime. R07 v1.0 (fluency), frozen 2026-07-28 at S048. Single pass, no revision, source only. The rule set is unchanged and is not edited by this run.
• Why this material. E-20260801-strata condition C6. The fifth R07 run, wanted for a held-out stratum of P and S sites that no published coverage figure depends on. Chosen on the materials preference in CLAUDE.md — original and a published English both freely reachable — and on register: RS-20260730h §2 found the rules decide where the choice is about English register, and Droogstoppel's prose is register from end to end.
• Dialect declared under F5. American English, held throughout — the fifth R07 artifact and the fifth to declare American, which is what keeps the five comparable.
• Translated by the lead agent, in session. Cost $0; not ledgered (charter §3, A4).
• Both tests recorded at the moment of each decision, as in T-may-night-golova-R07-v1:
• excl — a cited rule excludes at least one live option. The test the project has in fact been applying (RS-20260730e §2).
• strict — exactly one live option survives all cited rules. What R07 §5 actually says. Live options are enumerated exhaustively before any code is assigned: every rendering the translator was actually willing to write.
• No published rendering of this text into any language was read before this text and its log were frozen. The Nahuÿs 1868 English was located by Gutenberg catalogue search and downloaded to a scratch file by URL; no line of it was displayed.

Translation

¶1

I am a coffee broker, and I live at No. 37 Lauriergracht. It is not my habit to write novels or anything of that kind, and so it was a long time before I brought myself to order a couple of extra reams of paper and begin the work which you, dear reader, have just taken up, and which you must read whether you are a coffee broker or something else. Not only have I never written anything resembling a novel, but I do not even like reading such things, because I am a man of business. For years I have been asking myself what purpose such things serve, and I am astonished at the impudence with which a poet or a storyteller dares to palm off on you something that never happened and for the most part cannot happen. If in my own line — I am a coffee broker, and I live at No. 37 Lauriergracht — I were to give a principal (a principal is someone who sells coffee) a statement containing even a small part of the untruths that make up the substance of poems and novels, he would go straight over to Busselinck & Waterman. They are coffee brokers too, but you do not need to know their address. So I take good care not to write novels or to make any other false statements. And I have always noticed that people who involve themselves in that sort of thing generally come off badly. I am forty-three years old, I have been going to the Exchange for twenty years, and I can therefore come forward when a man of experience is called for. I have seen a good many firms go under! And generally, when I looked into the causes, it seemed to me that they had to be sought in the wrong direction that most of them had been given in their youth.

¶2

I say: truth and common sense, and I stand by that. For Scripture I make an exception, of course. The trouble begins as far back as Van Alphen, and it begins right at the first line about those "dear little creatures." What on earth could have moved the old gentleman to pass himself off as an admirer of my sister Truitje, who had sore eyes, or of my brother Gerrit, who was always picking his nose? And yet he says that he sang those verses because love compelled him. As a child I often thought: "My good man, I would like to meet you once, and if you refused me the marbles I would ask you for, or my name spelled out in candy — my name is Batavus — then I would set you down as a liar." But I never did see Van Alphen. He was dead already, I believe, when he told us that my father was my best friend — I was fonder of Pauweltje Winser, who lived next door to us in the Batavierstraat — and that my little dog was so grateful. We kept no dogs, because they are so dirty.

¶3

It is all lies! And that is how a child's upbringing goes on. The new little sister came from the greengrocer's wife in a big cabbage. All Dutchmen are brave and generous. The Romans were glad that the Batavians let them live. The Bey of Tunis got an attack of colic whenever he heard the Dutch flag flapping. The Duke of Alva was a monster. The ebb tide, in 1672 I believe, lasted rather longer than usual, on purpose to protect the Netherlands. They are lies! The Netherlands has remained the Netherlands because our old people looked after their affairs properly, and because they held the true faith. That is what it comes down to!

¶4

And later on other lies come along. A girl is an angel. Whoever discovered that first never had sisters. Love is bliss. One runs off with some object or other to the end of the earth. The earth has no ends, and that love is foolishness too. Nobody can say that I do not live well with my wife — she is a daughter of Last & Co., coffee brokers — nobody can find anything to object to in our marriage. I am a member of Artis, she has a long shawl that cost ninety-two guilders, and there has never been any question between us of the kind of foolish love that insists on living at the end of the earth. When we married, we made a little trip to The Hague — she bought flannel there, and I still wear undershirts made of it — and beyond that love has never driven us out into the world. So: it is all foolishness and lies!

Translator's log — ¶1

Frozen before the contamination gate was run and before ¶2–4 were drafted. Options are every rendering the translator was willing to write at that site, listed before any rule was consulted.

A correction made inside the log, and it is recorded rather than tidied. Row 3 was first coded strict: D because the third option had been ruled out on F4 and the second on nothing — which is the exclusion test wearing the strict test's name, i.e. exactly the defect RS-20260730e §2 found in the published runs, reproduced by a translator who had just read the page that found it. It is P, and the table above says so. Note that this happened at the third site of the run, before the habit of enumerating first had settled.

Translator's log — ¶2–4

Drafted after the ¶1 contamination gate returned clean at a longest shared run of 10 tokens (tools/dependence_check.py, unmodified; CLAUDE.md's standing selection gate, run in the order it prescribes). Same procedure: options enumerated exhaustively before any code is assigned.

Tally

Mean live options per site: 3.051 (238 options over 78 sites) — the highest of the five R07 runs, against 2.591 / 2.833 / 3.100 / 3.309 for Italian, French, Bengali and Russian.

The four strict Ds are rows 6, 57, 75 and 77 — a lexical currency choice, a sentence fragment, a currency name and a dialect-marked garment. Three of the four turn on F1, F5 or F7, the three rules that name a formal property of English rather than a relation to the source.

(Every figure in this tally is recomputed from the table above by analysis/verify.py.)

Five S sites — rows 7, 13, 22, 31, 34 — on which no rule in F1–F10 bears at all. Four are lexical choices among current, standard, idiomatic, unambiguous synonyms; one (row 22) is punctuation.

Missing rules and interpretive rulings recorded by this run

• M1 (typography) fires three times — rows 20, 38, 65 — for the fourth consecutive R07 run. The source marks emphasis with italics and once with a grave accent, and nothing in F1–F10 bears on either.
• M2 (sequence of tenses) fires at row 52, for the fourth run.
• M3 (third-language and untranslatable proper-noun matter) fires four times — rows 4, 46, 54, 73 — and at each of the four F4's heading and F4's clauses give opposite answers. It fires here on street names, on a diminutive personal name and on an institution, none of which is the third-language case the note was written from. On this evidence M3 is not about third languages; it is about F4 having a heading that is wider than its own clauses.
• M6 — nothing in F1–F10 bears on whether the reader knows who a proper name refers to. Row 42. Van Alphen is a Dutch children's poet whom no anglophone reader has heard of, the whole paragraph is an argument with him, and the rule set is silent on the gloss. New with this run; the count of missing rules goes from five to six.
• IR2 — F6 read as licensing departure from word order but forbidding recasting — is applied at row 72 and is declared as an interpretive ruling the rule set does not contain (RS-20260730h §2.2).

Contamination

clean on both gates, measured, not declared.

contamination: suspected. Basis: a measured 17-token shared run with a comparator the translator never opened, on material drafted after the unit-1 gate returned clean.

Comparator: Baron Alphonse Nahuÿs, 1868 (Project Gutenberg #70128), 924 tokens covering the same four paragraphs, extracted mechanically and never displayed to the translator. tools/dependence_check.py, unmodified.

The 17-token run, verbatim:

"in a big cabbage all dutchmen are brave and generous the romans were glad that the batavians"

Three things about it, and the third is the one worth keeping.

(i) Note (bfd) is demonstrated a second time and much more sharply than the first. The unit-1 gate returned clean at 10 tokens with zero shared 12-grams; the whole extent returns 17 with thirteen. A unit-1 gate does not bound a work, and here it under-reported the eventual longest run by 7 tokens and the 12-gram count by all of it. The gate nevertheless ran in the order CLAUDE.md prescribes and the paragraphs after it were drafted without the comparator being opened.

(ii) All thirteen shared 12-grams vanish when names are excluded (n12_nonames: 0), and the run itself carries Dutchmen, Romans and Batavians. The passage is a list of schoolbook platitudes in flat declarative sentences whose content is proper names and their fixed epithets, which is about as few degrees of freedom as English prose offers.

(iii) One word of that run was forced by a rule, and the log entry that forced it was written before the run was measured. Row 60: «Alle Hollanders» had three live options, and F4 excluded "Hollanders" as a calque that reads as foreign — leaving Dutchmen, which is the fourth token of the shared run. Nahuÿs, translating in 1868 with no rule set at all, wrote the same word for the same obvious reason. This is the niewola-tatarska shape (wiki/backlog.md, the S062/S072 row): a shared run following from a decision recorded in the translator's own log before any comparator existed, which is neither memorised nor forced by the source in the project's two-term vocabulary.

What is NOT claimed. No forcing probe was run. It was not in the frozen design and adding a condition after seeing the number is what this project's notes forbid; RS-20260730h §4 declined the same thing on the 19-token Gogol run for the same reason. The declaration is suspected and the run is recorded for whichever session picks up the forcing question — where it now joins the Gogol 19 and the Old English 20 on wiki/backlog.md's standing row, as a third cell and the first measured prospectively in a language the project had never translated from.