The Tin Flute: Draft, Edit and Proof

The record preserves more than a finished title

The National Library bibliography names Hannah Josephson’s 1947 English translation of Bonheur d’occasion.

The 1989 finding aid separately describes Alan Brown’s 1980 typescript before revision, an edited master, publisher suggestions and page proofs. The typescripts are listed as 427 pages; the proofs as 384.

The descriptions identify possible comparison points. They do not show every changed word. Nor does a smaller proof-page total establish that 43 pages of content were deleted: type size, line length and other layout choices can alter pagination. The inventory is a route into evidence, not the evidence of a particular stylistic decision itself.

Sources for this section: [1] · [2]

Original exercise: a register change

Use a French sentence created for this exercise: “Elle attendait près de la porte.” Stipulate the plain gloss “She was waiting near the door.” An invented translator’s draft reads “She waited by the door.” An edited master changes waited to loitered; the proof retains loitered. None of these lines comes from Roy, Josephson or Brown.

Identify the first supplied stage containing the new word. Explain how it changes the suggestion about the person’s presence. Can a catalogue label saying translated by the same person prove that the translator proposed the change? What would you consult next?

Sources for this section: [1] · [2]

Reference response: locate before attributing

The change first appears in the supplied edited master. Waited states an action without supplying much judgment about its purpose. Loitered can suggest lingering without a clear purpose and may introduce a less neutral attitude. The plain gloss alone does not require that attitude; a real translation decision would need the surrounding text and intended register.

The version sequence locates an introduction of wording, not its individual proposer. Track changes, annotated sheets or correspondence could help distinguish a translator’s revision from an editor’s intervention. A later proof preserving the word does not independently establish who first chose it. This is a bounded comparison of invented records, not a verdict on either historical translation.

Sources for this section: [1] · [2]

Keep pagination separate from textual extent

Imagine the same 1000 words set at 250 words per page in one mock layout and 200 in another. The layouts occupy 4 and 5 pages even though no words were added or removed. Actual proofs may also contain illustrations or preliminary matter, so a real extent comparison needs more than this simplified calculation.

A useful research note records the exact version, its physical description and the passage being compared. It preserves the named translator’s credit while leaving the authorship of an uninspected alteration unresolved. The next step would be to examine the documents identified by the finding aid, not invent their contents.

Sources for this section: [1] · [2]