Repository path: workshop/translations/alfred-preface/R04-v1/translation.md · rendered 2026-09-09

King Alfred, Preface to the Pastoral Care (c. 890), into modern English

Contamination: high, and measured

Longest common run with Sweet 1871: 20 tokens.

"a single one south of the Thames when I came to the throne. Thanks be to God almighty that we"

tools/dependence_check.py against a quarantined extraction of Sweet's English (materials/sweet1871-en-QUARANTINE.txt, 1,044 tokens, never read by the translator), computed on the R06 draft before the revision was written. Shared 7-grams 104, 12-grams 24, 15-grams 10. Raw: ../../../experiments/E-20260729b-graded-drift/analysis/contamination.json.

Where that sits. CLAUDE.md records the lead's measured range across five translations as 0 tokens (Ovid) to 21 tokens (Turgenev–Garnett). This is 20, on a text whose standard English translation is Sweet's and has been for 155 years. The standing rule therefore applies at full strength: this rendering is not an independent third translation of this work and must never be used as one.

Three things about the run, and the middle one is the interesting one.

1. "a single one south of the Thames" — the OE is ānne ānlēne … be sūðan Temese. A proper name and a preposition: close to forced.
2. "when I came to the throne" — the OE is ðā ðā ic tō rīce fēng, literally when I seized the kingdom. This is elective. came to the kingdom, succeeded to the realm, took the kingdom are all available and none is worse. This is the shape RS-20260728b-forced-run-ru found in the Turgenev run — a forced head and an elective interior — reproduced here on a different language, a different century and a different comparator, and it is the first time this project has seen that shape outside Russian.
3. "Thanks be to God almighty" — Gode ælmihtegum sīe ðonc. A fixed English formula; near-forced.

One census site falls inside the run: ðonc → thank (F26), where both renderings take the reflex.

The measurement is a lower bound. The comparator is OCR of a 155-year-old scan, of visibly variable quality, and OCR noise can only break shared runs, never create them.

Priming declaration — what was seen, when, and what it cannot explain

R04 §1 requires this where a parallel rendering sits in the same volume as the source, and the honest version has four items rather than one.

• Before the draft was written, one clause of Sweet's English was seen while locating the passage by grep: "which is called in Latin Pastoralis, and in English Shepherd's Book,". This was declared in the frozen design (§8) before drafting. The draft renders it the Herdsman's Book.
• After the draft was frozen and committed (dfc375c), three further truncated clauses appeared in grep output while the comparator was being located: "there were very few on this si…", "and I believe that there were…", "the Thames when I came to the…". These are after the freeze and cannot have influenced the draft, which is where the 20-token run was measured. They could in principle have influenced the revision; §3 of the log says what was done about that.
• What none of this explains is the run itself, which was already present in the frozen draft.

The translation

King Alfred bids Bishop Wærferth be greeted in his own words, lovingly and as a friend; and I bid it be made known to you that it has very often come into my mind what wise men there once were throughout England, both in holy orders and in the world; and how happy the times then were throughout England; and how the kings who had rule over the people in those days obeyed God and his messengers; and how they kept both their peace and their customs and their authority at home, and also enlarged their territory abroad; and how they prospered both in war and in wisdom; and also how eager the holy orders were about teaching and about learning, and about all the services they owed to God; and how men from abroad sought wisdom and teaching here in this land, and how we now must fetch them from outside, if we are to have them at all. So completely had it fallen away in England that there were very few on this side of the Humber who could understand their services in English, or even render a single letter out of Latin into English; and I think there were not many beyond the Humber either. There were so few of them that I cannot think of a single one south of the Thames when I came to the throne. Thanks be to God almighty that we now have any supply of teachers at all. And therefore I command you to do as I believe you wish to do — to free yourself from the affairs of this world as often as you can, so that you may apply the wisdom God gave you wherever you can apply it. Think what punishments came upon us in this world when we neither loved it ourselves nor allowed it to other men: we loved the name only, that we were Christians, and very few of the practices.

When I remembered all this, I remembered also how I saw, before it was all ravaged and burnt, how the churches throughout all England stood filled with treasures and books, and a great multitude of God's servants besides — and they got very little good of those books, because they could understand nothing of them, since they were not written in their own language. As if they had said: 'Our elders, who held these places before us, loved wisdom, and through it they got wealth and left it to us. Here one can still see their track, but we cannot follow after it, and so we have now lost both the wealth and the wisdom, because we would not stoop to that trail with our mind.'

When I remembered all this, I wondered very greatly at those good wise men who once were throughout England, and had learned all those books through and through, that they would not turn any part of them into their own language. But then at once I answered myself, and said: 'They never thought that men should ever become so careless, and learning fall away so; it was of set purpose that they left it undone, and they wished that there should be the more wisdom here in the land, the more languages we knew.'

Then I remembered how the law was first found in the Hebrew language, and afterwards, when the Greeks had learned it, they turned it all into their own language, and all the other books besides. And afterwards the Latin people did the same: once they had learned them, they turned them all, through wise interpreters, into their own language. And all other Christian nations too turned some part of them into their own language. Therefore it seems better to me, if it seems so to you, that we too turn certain books — those that are most needful for all men to know — into the language that we can all understand, and bring it about, as we very easily can with God's help, if we have peace, that all the youth now in England of free men, those who have the means to give themselves to it, be set to learning, so long as they are fit for no other employment, until the time when they can read English writing well: let those be taught further in Latin whom one wishes to teach further and to advance to higher orders. When I remembered how the knowledge of Latin had fallen away throughout England before this, and yet many could read English writing, then I began, among the other various and manifold cares of this kingdom, to turn into English the book that is called in Latin Pastoralis and in English the Herdsman's Book, sometimes word for word, sometimes sense for sense, as I learned it from Plegmund my archbishop, and from Asser my bishop, and from Grimbold my mass-priest, and from John my mass-priest. After I had learned it, so as I understood it and could set it out most intelligibly, I turned it into English; and I will send one to every bishop's see in my kingdom; and in each there will be a pointer worth fifty mancuses. And I command in God's name that no man take the pointer from the book, nor the book from the minster: it is unknown how long there may be such learned bishops as now, thanks be to God, there very nearly everywhere are. Therefore I would have them stay always at that place, unless the bishop wishes to have them with him, or they are out on loan anywhere, or someone is making another copy from it.

Translator's log

Frozen at the commit that files this artifact, before any evaluation of it was designed (charter §3, A4; R04 §Procedure 3). Decisions and alternatives, not quality claims. Any evaluative sentence here is internal-judgment-only, and the front matter carries that flag.

1. Two errors the revision caught, and one of them is the experiment's own subject matter

(1a) lēfdon — the draft took a reflex and the reflex was wrong. ¶1 ends nē ēac ōðrum monnum ne lēfdon. The verb is līefan/lȳfan* "to permit, grant", notlǣfan` "to leave, bequeath". The draft wrote "nor left it to other men." The revision reads "nor allowed it to other men."

This is worth more than a correction. The census (design/sites-alfred.md, F52) records exactly one site for lǣfan → leave, at ¶2 lǣfdon "bequeathed", and does not record līefan at ¶1 at all — because the lead, writing the census from the Old English, ran the two verbs together as one lemma. So the census missed a false friend, and then the translator fell into it, and only the revision pass caught it. RS-20260729-drift-window-verify §7 measured the lead's Beowulf census against an independent one at Jaccard 0.4375 and concluded that a stated rule applied by two readers lands 44% together; this is the same finding from the inside, and the miss is not a random one — it is at precisely the kind of site the census exists to find.

(1b) fiorme. ¶2's ðā swīðe lȳtle fiorme ðāra bōca wiston — feorm is "food, provision, entertainment; profit, benefit, use". Draft: "they knew very little of the use of those books." Revision: "they got very little good of those books", which keeps wiston doing less work than the draft made it do. The reflex, farm, is unavailable in any sense and was never in play.

2. The reflex decisions, taken one at a time

The 80 sites were frozen and scored before a word was drafted, so the translator knew, at every one, that its choice would be counted. That is a priming and it is declared. What the choices were:

• Refused, where the reflex is the obvious trap. gesǣlig → silly (how happy the times); tīd → tide (times); rīce → rich (the throne, my kingdom); wita → wit (wise men); spēd → speed (the means); ǣmettig → empty (free yourself); sellan → sell (gave); feorm → farm; stōw → stow (places, that place); stōl → stool (see); hierde* → *herd* (**Herdsman's Book**);uncūð→ *uncouth* (**it is unknown**);ārǣdan` → read — no: taken, see below.
• Taken, and the modern sense is right. word, king, day, folk (people — refused, see below), wisdom, learning, land, book, church, name, love, see, write, find, know, youth, free, elder, hold, leave, wonder, soon, stillness (peace — refused), send, priest, minster, loan, manifold, understand, believe, thank, few.
• Taken where the modern sense is arguably wrong, and taken deliberately. ārǣdan → read ("can read English writing well"). The Old English verb here is "read out, peruse, interpret", and modern read covers it. This is the same lemma at which A-beowulf-ingeld's 0 of 12 became 1 of 12 — in Beowulf rǣdan means "possess, rule" and the reflex is a trap; in Alfred it means "read" and the reflex is correct. One lemma, two texts, opposite verdicts, which is a fact about the site rather than about either translator, and is the sharpest single argument this session found against treating drift as a property of a word.
• The one refusal I most nearly reversed. clǣne, adverbial, in Swǣ clǣne hīo wæs oðfeallenu. "So clean had it fallen away" is available modern English (clean forgot, clean gone) and is closer to the Old English adverb than anything else on offer. It was refused — "So completely had it fallen away" — because the surviving adverbial clean is colloquial in a way this sentence is not, and the register cost outran the semantic gain. This is the two-axis problem in a single decision: the semantics said take it and the currency said don't.
• folc → folk: refused (the people). stilnes → stillness: refused (peace). Both are cases where the reflex survives with high semantic retention and the register is wrong for ninth-century royal prose.

3. What was done about the post-freeze exposure

Three truncated clauses of Sweet's English appeared in tool output after the draft was frozen (see the priming declaration). Nothing in the revision was changed toward them or away from them, and the 20-token run was deliberately left standing. Rewriting "when I came to the throne" to avoid a comparator the translator has now partly seen would produce a text that is deliberately different from Sweet, which is a worse artifact than one that is independently similar — and it would destroy the measurement, which is the only thing here that is evidence.

4. Cruxes in the source

• ānne ānlēne (¶1). Bright's reading; commonly emended ānlēpne "single, individual". Rendered "a single one", which is what either reading yields; flagged because it is an emendation the translation silently depends on.
• æstel (¶4). The most famous crux in the text: an object worth fifty mancuses accompanying each copy, of unknown referent. Rendered pointer, the conventional guess, with no confidence. bookmark, clasp and leaving the Old English word untranslated were all live.
• hwīlum word be worde, hwīlum andgit of andgiete (¶4). Rendered sometimes word for word, sometimes sense for sense. This is the one phrase in the text with a fixed English form, and choosing anything else would have been an affectation. andgit has no modern reflex at all and is not a census site; the sentence in which this project's whole subject is named survives into modern English only by loanword.
• wealhstōd "interpreter, translator" (¶4) — likewise no modern descendant. Rendered interpreters.
• geðīode "language" — the word the preface uses eight times for the thing it is arguing about. No modern reflex. Rendered language throughout, from Latin.

5. Register

Plain modern English, formal but not archaised: this is a chancery document, and the Old English is businesslike rather than ornamental. The one deliberate archaism is the epistolary third person of the opening formula (King Alfred bids Bishop Wærferth be greeted), which the Old English has and which collapses into first person in the next clause — the source does that, and the translation follows it rather than tidying it.

6. What the log does not contain

Decisions made without noticing them. R04 §Known limitations says so, and on 874 words of prose the count of unnoticed decisions is certainly larger than the count above.