A Classic Readers Thought They Knew Arrives Again

A literary retranslation is a complete new translation, made directly from the original language, of a work that already exists in the target language. A translator begins again from the source text rather than repairing an earlier English version. Publishers commission such work because translations age in ways originals do not, because rights lapse and free a title for rival editions, and because scholarship changes what editors and translators understand the original to say, to imply, and to withhold.

A retranslation is not a correction of the old translation but a new reading of the original, and its value lies less in fixing errors than in restoring choices the first translation was forced to suppress. Every translation resolves thousands of ambiguities in one direction, and a first version, produced for readers with no other access to the book, tends to resolve them toward comfort, clarity, and the taste of its own period. A later translator, working with better editions, fuller scholarship, and a different sense of what English prose can carry, resolves the same passages toward difficulty, strangeness, and density where the original holds those qualities.

The pattern repeats across the European canon. Thomas Mann reached English readers for two generations through H. T. Lowe-Porter, whose Buddenbrooks appeared in 1924 and whose The Magic Mountain appeared in 1927, and later translators rebuilt both books from the German. Tolstoy reached English readers through Constance Garnett and through Louise and Aylmer Maude before Richard Pevear and Larissa Volokhonsky offered a rival War and Peace in 2007. Homer, Cervantes, and Dostoevsky have each accumulated English versions across three centuries, and each new version announces, implicitly or in a preface, a claim about what earlier versions missed.

Three questions govern every case. Versions multiply because the conditions that produced the first translation expire: copyright ends, the receiving language changes, scholarly editions improve, and publishers compete for a backlist title with proven demand. Versions differ in concrete, checkable ways: sentence length, register, dialect, religious and philosophical vocabulary, the handling of passages an earlier age cut or softened, and the treatment of words the original repeats on purpose. Judgment between versions is also a practical skill rather than a matter of loyalty, since a reader can compare signature passages, test how each version handles dialogue and description, and read prefaces as evidence of method rather than as advertising.

A classic, then, is never translated once. The original stays fixed while everything around the original moves, and retranslation is the mechanism by which a receiving literature renews a lease on a book it refuses to surrender. The sections that follow trace that mechanism from its definitions through its economics, its desk practice, and its governing theory, using named translators, dated editions, and parallel cases that a reader can verify on a shelf.

What Counts as a Retranslation and What Does Not

A retranslation starts again from the source text. The translator reads the original language, makes fresh decisions at every sentence, and accepts responsibility for the whole architecture of the book: its chapter rhythms, its recurring words, its shifts of register between narration and speech. Nothing in the earlier English version binds the new translator except the shared original, and passages where the two versions agree usually agree because the source permits a limited range of honest renderings. The test is genetic rather than stylistic: a retranslation descends from the original, not from its predecessor.

A revised edition follows the opposite path. An editor or a second translator patches an existing translation, correcting errors, modernizing spellings, and replacing dated phrases, while the underlying choices survive intact. The sentence architecture remains the first translator’s architecture, the omissions remain omissions unless an editor restores them from the original, and the voice a reader hears remains substantially the voice of the original translation team. Revision has genuine uses, since it preserves an admired version at low cost, but a revised edition cannot deliver what a retranslation promises, because the suppressed choices of the first rendering stay suppressed inside the patched text.

A reprint changes nothing but the packaging. The same translation appears with a new cover, a new introduction, and a place in a redesigned series, and the title page verso records the original date of translation for a buyer who checks. An abridgement moves in the other direction, cutting a long work to a shorter length for schools or for general readers, and the cuts fall predictably on digressions, philosophical chapters, and descriptive passages that carry an author’s method. A relay translation, also called an indirect translation, works from an intermediate language rather than from the original, so the English text stands two removes from the source. Early European versions of Russian fiction frequently arrived through French, because French versions existed first and French was widely read, and the French intermediary’s choices about register, censorship, and style passed silently into the English text.

The distinctions matter to a buyer comparing title pages because marketing language blurs all of them. A cover can announce a classic as newly translated when the edition is a reprint with a new introduction, and a publisher can describe light revision in terms that suggest a fresh start from the original. The reliable evidence sits in small print and in prefaces: the translator’s name and the date of translation, an explicit statement that the version was made from a named edition of the source, and a preface that discusses specific passages rather than general admiration. A genuine retranslation names its source edition, explains its principles, and leaves a trail of decisions a reviewer can test against the original. A repackaged older version offers praise for the author and silence about method. The difference between the two products is the difference between a new reading and a new wrapper, and the title page, read carefully, almost always discloses which product is on sale.

Why Translations Age When Originals Do Not

An original ages only in reputation. The words of Buddenbrooks, published in German in 1901, stand exactly as Thomas Mann published them, and a German reader meets the same sentences across every decade. A translation, by contrast, belongs to two moments at once: the moment of the original and the moment of its own making. The English of a translation is the English of a particular year, with that year’s idiom, that year’s sense of politeness, and that year’s assumptions about foreign places. As the receiving language moves on, the translation stays behind, and the distance between the two becomes visible first as quaintness and later as distortion.

The target language drifts along several tracks at once. Idiom turns over quickly, so dialogue translated into the slang of one decade sounds costumed a generation later. Forms of address shift as social hierarchies flatten, and a nineteenth-century English version that rendered every servant, official, and elder with elaborate deference imposes a social world the original may not contain. Taboo vocabulary moves in both directions: words an early translator softened or deleted lose their charge, while words an early translator used innocently acquire charges the original never carried. Syntax drifts as well, since the long periodic sentence that felt natural in English prose of 1900 reads as a deliberate archaism in English prose of 2000, and a translator must decide whether the original’s long sentences should sound old, or should sound as natural in English as the original sounded in its own tongue.

Knowledge about the source culture grows independently of language. Scholarship recovers the religious vocabulary of a Russian novel, the legal and commercial detail of a German family chronicle, or the oral conventions beneath a Greek epic, and each recovery exposes places where an early translator, working without that knowledge, guessed or generalized. Scholarly editions of the original improve in parallel: editors establish cleaner texts from manuscripts and corrected printings, restore passages that censorship removed, and document variant readings, so a later translator often works from a better original than the first translator possessed. Expurgated passages supply the clearest cases, since passages cut for sexual content, religious doubt, or political offense in an early edition return to the text once the receiving culture stops demanding the cuts, and only a fresh translation can absorb restored material without visible seams.

David Bellos, in Is That a Fish in Your Ear? (2011), argues that a translation is anchored in its own moment because every translation answers the needs and assumptions of the audience that commissioned it. The receiving literature’s taste completes the mechanism: a period that prizes fluency gets fluent translations, a period that prizes strangeness gets versions that keep the foreign texture visible, and each period reads the previous period’s versions as dated precisely where those versions obeyed the earlier taste most faithfully. Ageing, on that account, is not a defect that better translators could have avoided. Ageing is the predictable fate of a text written in a living language about a text written in a language that, for literary purposes, has stopped moving. Retranslation is the scheduled repair for that fate, repeated for as long as readers keep demanding the book.

The First Translation Sets the Terms

The first widely read translation of a foreign classic does far more than introduce a book. The first version defines the author’s image in the receiving language, fixing the tone readers associate with a name, the spellings of characters and places, and often the ranking of the author’s works. A novelist who is ironic and abrasive in the original can become gentle in the first translation, and the gentle version becomes the standard against which reviewers measure every later attempt. Readers defend the author they met first, and critics write about qualities that belong, in part, to the translator. The first translation also sets the commercial benchmark for the title, proving that a market exists and establishing the sales history that later publishers cite when a rival version is proposed.

Constance Garnett supplied the defining case in English. Beginning in 1912, Garnett translated more than 70 volumes of Russian literature, including Tolstoy, Dostoevsky, Chekhov, Turgenev, and Gogol, and for two generations the English idea of Russian fiction was substantially her creation. Garnett worked at extraordinary speed, often completing a volume in months, and her prose carried a smooth late Victorian and Edwardian register that leveled differences among authors. Dostoevsky’s jagged dialogue, Tolstoy’s repetitions, and Chekhov’s plainness all passed through the same fluent surface, so English readers met a unified Russian soulfulness that owed as much to the translator’s habits as to the originals. Her versions also standardized titles and character names that English reference works repeated for decades, so her choices structured syllabuses and scholarship as well as private reading. Scholars have documented omissions and errors produced by that speed, and later translators have made careers by contesting her versions, yet her versions also performed the indispensable first task: the versions made Russian literature central to English reading, created the audience that later translators inherited, and set the terms every successor had to answer.

First translations shape even the table of contents of a foreign literature in English. Translators and publishers choose the works they can sell, and early choices harden into a canon: the novels Garnett translated became the Russian novels, while works she passed over waited decades for English versions. Names fixed in a first translation resist correction with surprising strength, since readers, teachers, and reference works repeat the established spellings and titles long after scholars propose better ones. Even the physical memory of reading belongs to the first version, because a widely adopted translation becomes the text of school editions, reading groups, and quotations in other books.

For those reasons a retranslation never competes against a neutral predecessor. A retranslation competes against the version in which a culture first fell in love with an author, and reviewers judge the challenger by standards the incumbent created. That asymmetry explains the heat of translation controversies and the caution of publishers, and that asymmetry also explains the subject of the next case: an English voice for Thomas Mann so complete, and so contested, that the contest itself became a chapter in Mann’s English history.

Helen Lowe-Porter and the Mann English Readers Inherited

H. T. Lowe-Porter became the English voice of Thomas Mann and held that position for two generations. The American publisher Alfred A. Knopf commissioned Lowe-Porter to translate Buddenbrooks, published in English in 1924, and the success of that edition led to The Magic Mountain in 1927 and to a long sequence of Mann’s novels, stories, and essays in her renderings. For readers in Britain and the United States, Mann’s irony, Mann’s sentence rhythms, and Mann’s philosophical dialogues arrived in Lowe-Porter’s English or not at all. Mann rewarded the arrangement with loyalty, praised her work in correspondence, and valued the reach her versions gave the fiction in the English-speaking market. A single translator mediating an entire major author creates a coherent shelf, and the coherence itself became part of Mann’s English identity.

The criticisms of that mediation accumulated slowly and then converged. Later translators and scholars comparing the English texts against the German documented a consistent set of features: long periodic sentences divided into shorter units, irony softened at the level of diction, difficult or digressive passages compressed or omitted, and a residue of plain errors in detail and idiom. The Knopf schedule demanded enormous output across long novels, stories, and essays, and the documented pattern fits work produced under sustained deadline pressure. Lowe-Porter also translated from a position inside the taste of her period, which prized a smooth and elevated English surface, and Mann’s German, with its deliberate awkwardness, bureaucracy, and parody of official language, resisted that surface at every level.

What did Lowe-Porter change in Thomas Mann?

Lowe-Porter shortened Mann’s long periodic sentences, softened irony and abruptness, compressed or omitted difficult passages, and introduced occasional plain errors, so the English Mann sounded smoother, gentler and more conventionally elegant than the German Mann, whose syntax carries argument, hesitation and social observation inside a single unfolding sentence.

John E. Woods, whose The Magic Mountain appeared in 1995 and whose Doctor Faustus appeared in 1997, spoke in interviews and prefaces to the task of restoring Mann’s sentence architecture, arguing in substance that the long sentence in Mann carries the thinking of a passage and that dividing the sentence changes the meaning, not merely the music. Parallel reading supports the narrower claims directly: a passage of dialogue that runs for a page in German and half a page in the early English version has lost material, and a sentence broken into four sentences has changed the order in which a reader receives qualification and reversal. Scholars editing Mann in English, including Ritchie Robertson for the Oxford editions, have treated the early versions as documents of reception rather than as reliable texts for study, a judgment about function rather than about the translator’s historical importance.

The balanced verdict separates two questions that polemic often merges. Lowe-Porter’s translations made Mann a major author in English, carried the fiction through the decades of Mann’s exile and fame, and gave generations of readers genuine access to books that might otherwise have waited far longer for translation. The same translations also stand between a modern reader and Mann’s prose, filtering difficulty, compressing argument, and normalizing a style whose strangeness is part of its meaning. Retranslation in the Mann case therefore began as an act of recovery rather than rivalry. Woods rebuilt the syntax, later translators followed with rival versions of the early novels, and each new edition invited the comparison that the monopoly of a single voice had postponed for seventy years. The Mann shelf in English demonstrates the general law with unusual clarity: a first translation can succeed completely as literature and still require replacement as evidence.

Copyright Expiry and the Economics of the Backlist

Copyright decides how many translations of a classic may compete at one time. While a work remains in copyright, the rights holder licenses translation rights, and a single authorized translation can hold a market for decades, protected against rivals by law rather than by merit. A publisher that controls the standard version of a famous novel collects steady sales with no incentive to fund a replacement, and a rival publisher with a better translator under contract cannot legally offer an alternative. The monopoly explains why some dated translations survive long past the point where reviewers have documented their weaknesses in print: the obstacle is a property arrangement, not a judgment about quality, and the arrangement ends only when the term of protection expires.

Expiry changes the economics in a single stroke. Once protection lapses, any publisher may commission a rival version, translators may work without permission, and the only barriers become cost, reputation, and distribution. The same mechanism freed a celebrated American novel for a flood of rival editions and adaptations, a process traced in how the Great Gatsby entered the public domain, and the pattern repeats whenever a major author’s term runs out: several houses move at once, because each house knows that the first credible new version can capture the course adoption and reading group market for a generation. Rights in older works also change hands during an author’s afterlife through statutory routes, as how copyright termination rights work explains in the context of grants and reversions, and every transfer reshuffles which publisher may exploit which edition.

Thomas Mann died in 1955, and the lapse of his copyright in Britain at the start of 2026 opened the way for rival English versions from several publishers in a single year, including a Buddenbrooks translated by Mike Mitchell and edited by Ritchie Robertson for Oxford University Press, a The Magic Mountain translated by Simon Pare, and a collection of stories translated by Lesley Chamberlain for Pushkin Press. The cluster was not a coincidence of taste but a predictable response to a legal date that publishers had circled for years, with translators commissioned in advance so that editions could appear as soon as the term expired. A wave of that kind also resets reviewing habits, because critics can compare competing versions of the same chapters in the same season, and comparison, once printed, permanently weakens the assumption that the older standard version simply is the book.

The backlist steadiness of classics lowers the risk of every such commission. A famous novel sells predictably to schools, universities, and reading groups, year after year, with minimal marketing expense, and a new translation gives a publisher a reason to repackage the title, win review coverage, and claim the adoption market from a competitor. Translation fees and advances are modest against that long revenue tail, and a successful version can remain in print for half a century. Expiry therefore does double work in the history of retranslation: expiry supplies the legal opening, and the backlist supplies the commercial confidence, and together the two forces explain why retranslations arrive in waves keyed to legal dates rather than in a steady stream keyed to scholarly need.

How a Publisher Commissions a Retranslation

A retranslation begins as an editor’s judgment about a gap. An editor identifies a classic whose standard English version reads as dated, contains documented omissions, or belongs to a rival house whose edition dominates course adoptions. The editor weighs the title’s steady sales against the cost of a new translation and the risk that reviewers will prefer the incumbent, and the calculation favors famous books with frustrated readers: a canonical novel whose admirers complain about the translation is a stronger candidate than a neglected title with no audience waiting. Series editors hold an advantage in that calculation, because a series with loyal buyers can absorb a slow start for an individual volume.

Sample chapters form the next stage and the real audition. A publisher commissions a translator to render a defined portion of the original, often a famous chapter or a passage of dialogue with a reputation for difficulty, and editors read the sample against the old version and against the original where the house employs readers with the source language. The sample tests the qualities that matter over 800 pages: stamina of register, handling of repeated words, courage with long sentences, and a prose rhythm a reader can inhabit for weeks. A brilliant ten pages guarantee nothing about a whole novel, yet a weak sample ends a candidacy, and experienced editors treat the sample as evidence about method as much as about talent. Competing samples for the same book occasionally circulate within a house before a contract is offered.

The commercial frame comes from the great classics series. Penguin Classics and Oxford World’s Classics supply the design, the distribution, the scholarly apparatus of introductions and notes, and the promise to booksellers that a volume belongs to a permanent library rather than to a season. A retranslation inside such a series inherits the series reputation and sells beside the originals in translation that the series has standardized for decades. The apparatus matters to the translation’s reception, since an introduction can explain the new version’s principles, document the old version’s omissions with examples, and teach readers how to notice the differences that justify the purchase.

The money remains modest relative to the book’s life. Translators of literary fiction commonly receive an advance against royalties or a fee calculated by the thousand words, with royalty participation varying widely by house and by the translator’s standing, and even a successful classic rarely makes a translator wealthy from a single volume. The economics favor translators with academic posts, pensions, or portfolios of work, a fact that shapes who can afford to spend two or three years inside one long novel. Marketing, by contrast, spends freely on a single claim: freshness. Covers and publicity announce a translation for a new generation, quote reviewers on clarity and energy, and invite the comparison with the old version that the house believes the new version will win. Later sections test that claim against evidence, because freshness is a date rather than a quality, and a version marketed as fresh in one decade becomes the dated incumbent that a later editor proposes to replace.

The Translator at the Desk: Method and Materials

Retranslation begins with a decision about which original to translate. The copy text, meaning the specific edition of the source work that a translator follows as the authoritative wording, determines thousands of later choices, because editions of a classic can differ in passages, punctuation, and even chapter divisions. A conscientious translator selects the best scholarly edition of the original, records where that edition departs from the version an earlier translator used, and treats restored passages as part of the assignment rather than as intrusions. Where the first translator worked from a censored or corrupted edition, the choice of copy text alone can add pages of material and change the shape of scenes, and the preface of a serious retranslation names the copy text so that reviewers can audit the foundation.

Glossaries govern the second layer of method. A long novel repeats motif words, forms of address, professional terms, and family nicknames hundreds of times, and a translator builds a working glossary for recurring terms so that a word carrying thematic weight is rendered consistently, or varied only where variation is deliberate and recorded. Consistency matters most where an author repeats a word as an argument: Tolstoy’s repetitions, Mann’s recurring musical and medical vocabulary, and Homer’s fixed epithets all punish a translator who reaches for a fresh synonym at each occurrence. The glossary also fixes names, titles of address, and units of money and measure, the small decisions that readers feel as reliability even when readers cannot name the cause.

Rhythm receives its own test at the desk. Translators read drafts aloud, because the ear catches what the eye forgives: sentences that cannot be breathed, dialogue that sounds written rather than spoken, and paragraphs whose cadence flattens the contrast between narration and speech. Reading aloud also exposes the false fluency of a draft that paraphrases gracefully while drifting from the original’s order of information. Many translators work in staged passes, with a first pass for accuracy against the source, a second pass for English rhythm conducted away from the source, and a third pass that reconciles the two, checking every liberty taken in the second pass against the passage that prompted the liberty.

A decisions log holds the whole enterprise together. The translator records choices about names, recurring phrases, dialect, and disputed passages, with reasons, so that an editor’s query in month eighteen can be answered from evidence rather than from memory. Consultation extends the desk outward: an author’s letters and surviving drafts clarify disputed readings, correspondence with earlier translators occasionally survives and occasionally warns, and specialists in the period’s law, medicine, music, or seamanship check the vocabularies that a novelist borrowed from working life. Edith Grossman, in Why Translation Matters (2010), describes translation as a sustained act of critical reading in which every sentence of the original must be understood before a sentence of English can be risked. Retranslation adds one further material to the desk: the old version itself, present as a rival to consult, to avoid, and at intervals to concede a solution to, because a predecessor’s rendering of a famous line sometimes survives testing against the original better than any alternative the new translator can devise.

Six Works, Twelve Translations, One Pattern

The table settles a concrete claim about scale: which famous works were retranslated, by whom, and what the later version principally changed, across German, Russian, Spanish, and Greek literature in English. The six cases span three centuries of English publishing, from Alexander Pope’s Homer in the early eighteenth century to the Mann editions of 2026, and the span itself carries information. Retranslation is not a fashion of one publishing era but a permanent feature of how English literature maintains foreign classics, recurring whenever language, scholarship, and rights align. The cases also share a commercial trait, since every work in the table held an established place in English reading before the later version appeared, and every later translator therefore worked against an incumbent with loyal readers. The final column of the table records the principal change claimed and demonstrated for each later version, in terms specific enough for a reader to test against parallel passages rather than accept on authority. The pairings also expose a difference in motive that the final column compresses. Some later versions arose from scholarly conviction that the incumbent had misrepresented the original, others arose from a publisher’s calculation that a famous title could support a rival edition, and most arose from a mixture of the two motives in proportions that prefaces rarely state plainly. The dates in the table carry a lesson about timing, since the later versions cluster at moments when copyright, series publishing, and scholarly confidence aligned, and a gap of a single decade in either direction often separates a celebrated replacement from an edition that arrived before readers felt any need for a replacement.

Work	Earlier English translation	Later English translation	What the later version changed
Buddenbrooks by Thomas Mann	H. T. Lowe-Porter, 1924	Mike Mitchell, 2026	Restored sentence length, irony and omitted material, with a scholarly apparatus
The Magic Mountain by Thomas Mann	H. T. Lowe-Porter, 1927	John E. Woods, 1995, and Simon Pare, 2026	Rebuilt the long periodic sentences and philosophical debates in full
War and Peace by Leo Tolstoy	Louise and Aylmer Maude, 1922 and 1923 edition	Richard Pevear and Larissa Volokhonsky, 2007	Kept the French passages visible as French and restored Tolstoy’s rougher repetitions
The Odyssey attributed to Homer	Alexander Pope, 1725 and 1726 edition	Emily Wilson, 2017	Matched the line count of the Greek and used a plain contemporary register
Don Quixote by Miguel de Cervantes	Tobias Smollett, 1755	Edith Grossman, 2003	Recovered the comedy and speed of the dialogue for modern readers
Crime and Punishment by Fyodor Dostoevsky	Constance Garnett, 1914	David McDuff, 1991	Slowed the pace, restored religious vocabulary and roughened the prose

Read as a set, the table shows later versions competing on a single ground: restoring what the earlier age suppressed, whether suppressed syntax, suppressed roughness, suppressed comedy, or suppressed foreignness. The earlier versions in the table were rarely careless, and several remained beloved for generations, which strengthens rather than weakens the pattern. Suppression was the price of admission that a first translation paid to a receiving culture with fixed expectations, and the later versions in the table could charge a different price only because the first versions had already won the works a permanent English audience.

Schleiermacher’s Fork: Move the Reader or Move the Author

Friedrich Schleiermacher delivered the lecture On the Different Methods of Translating to the Berlin Academy in 1813, and the lecture still organizes every argument about retranslation two centuries later. Schleiermacher held that a translator faces exactly two genuine paths. Along the first path, the translator brings the author toward the reader, rendering the foreign work so that the work reads as though the author had written in the reader’s language, with the reader’s idioms, rhythms, and assumptions left undisturbed. Along the second path, the translator takes the reader toward the author, bending the receiving language toward the original’s structures and strangeness, so that the reader meets the foreign work on ground that remains visibly foreign. Schleiermacher dismissed mixtures of the two paths as unreliable, since a mixture produces a text that is neither fully natural nor faithfully strange, and Schleiermacher declared a preference for the second path for serious literature.

Each path charges a characteristic price. Moving the author toward the reader buys immediate intelligibility and pleasure, and sells the foreignness that made the work worth importing: jokes land, sentences flow, and the specific grain of another culture’s thought dissolves into familiar English. Moving the reader toward the author preserves that grain, and charges the reader effort, patience, and a tolerance for constructions that English resists. A version on the second path can also curdle into pedantry, keeping every foreign feature except the one that mattered most, the experience of reading a living book rather than a specimen. Schleiermacher wrote for a German audience building a national literature through translation, and the lecture treats translation as cultural policy, yet the fork survives outside that context because the underlying geometry never changes: every sentence sits somewhere on the line between the author’s home and the reader’s home, and a translator must choose where. The lecture also insists that translation is never a mechanical transfer, since a language carries a way of thinking that resists direct carriage into another tongue, and every rendering therefore interprets at the very moment the rendering presents itself as a simple act of carrying across.

Nearly every later argument about retranslation restates the fork in new vocabulary. Debates about fluency against resistance, about domestication against foreignness, and about translations that read smoothly against translations that keep an accent all redraw the same two paths with fresh examples. Retranslation acquires a characteristic plot inside the scheme: a first translation, written for readers with no other access to a foreign author, tends toward the first path, because a publisher introducing an unknown writer cannot ask readers to travel far toward a stranger. A later translation, written for readers whom the first version already converted, can afford the second path, asking an established audience to accept longer sentences, stranger registers, and untranslated residue as the price of closer contact. The sections that follow examine the theory built on that plot, test the theory against cases where the plot fails, and weigh the evidence that the fork, for all its endurance, describes a tendency rather than a law.

Domesticating and Foreignizing: Venuti’s Charge

Lawrence Venuti gave the central tension of literary translation a pair of names that have shaped discussion ever since. In The Translator’s Invisibility, published in 1995, Venuti argued that English language publishing rewards a domesticating approach, in which the foreign work is rewritten until it reads as if it had been composed in English in the first place. The sentences flow without friction, idioms match familiar usage, cultural references are replaced or smoothed, and the strangeness that marked the work as coming from elsewhere recedes. The translator disappears behind that fluency.

Foreignizing translation, in Venuti’s account, refuses that disappearance. It keeps visible signs of difference in diction, syntax and cultural reference, and it asks the reader to register that the work belongs to another language and another set of assumptions. A foreignizing version may preserve sentence structures that feel unusual in English, retain terms for food, money, rank or kinship that have no exact counterpart, and resist the urge to explain every unfamiliar custom inside the sentence. The result is rarely comfortable in the way a fluent commercial novel is comfortable. That discomfort is the point. It signals distance rather than pretending distance away.

The commercial pressure toward fluency explains why domestication so often wins, even when translators privately prefer a stranger text. Publishers sell translated fiction to readers who mostly read in one language, and a version that requires repeated pauses for unfamiliar constructions carries an obvious marketing risk. Reviewers reinforce the pressure when they praise a translation for reading smoothly and criticize another for awkwardness, without asking whether the smoothness belonged to the original at all. Each individual decision looks small and defensible. Together they move a work a long way from the texture of its source.

Retranslation enters this argument as a promised correction. A publisher announcing a new version of a classic often claims that the older translation domesticated the work and that the new one restores its foreign character. The claim has force where an earlier version cut passages, regularized dialect or turned hesitant, repetitive or deliberately awkward prose into polished English. A retranslator can point to specific sentences, restore omitted material and choose to keep constructions that an earlier generation of editors would have removed. Marketing language alone does not prove that such a correction happened, however. Fluency can return in a new disguise, with contemporary idiom replacing older idiom while the underlying smoothing continues unchanged. The distinction therefore has to be tested in the text rather than accepted from the cover copy.

Venuti’s charge also reframes what a reader loses when a translation succeeds too completely. If a nineteenth century German novel, a Russian epic and a Spanish satire all arrive in the same polished, neutral English, the differences among their worlds start to blur. Voices that were distant from one another in their own languages become neighbours in a single style. Foreignizing choices, used with judgment, preserve some of that separation. Retranslation matters in this framework not because a later date guarantees honesty, but because a later translator, working after a work has found an audience, may be permitted to keep difficulties that a first translator, introducing an unknown author to a wary market, felt compelled to remove.

Berman’s Hypothesis: First Versions Bend, Later Versions Return

Antoine Berman set out the most influential account of why retranslations might differ systematically from the translations they follow. In his 1990 essay on retranslation, Berman proposed that a first translation tends to assimilate a foreign work to the receiving culture. The translator introducing an unfamiliar author faces resistance from publishers, reviewers and readers, and responds by reducing strangeness, clarifying ambiguity and presenting the work in forms the new audience already knows how to read. The result is a version bent toward the target language, valuable for establishing the author but incomplete as a record of the original’s otherness.

Later translations, in Berman’s account, operate under changed conditions. The author has a reputation, the work has a place in syllabi, bookshops and critical discussion, and readers approach it expecting a classic rather than sampling an unknown name. A retranslator can therefore afford to move in the opposite direction, back toward the source. Sentence structures that a first translator broke apart can be kept intact, repetitions that looked like carelessness can be preserved as design, and cultural details that were replaced with familiar equivalents can be restored. The retranslation becomes a return journey, made possible by the very success of the domesticated version it seeks to replace.

The evidence usually offered for the pattern comes from successive versions of German and Russian classics in French and English publishing. Early English versions of Thomas Mann shortened and regularized prose that later translators worked to restore. Successive English versions of Fyodor Dostoevsky and Leo Tolstoy have been described by their publishers and reviewers as recovering abruptness, humour and theological argument that smoother earlier versions softened. French publishing, which Berman knew closely, offers parallel sequences in which a canonical foreign work passed from an introductory version shaped for a general audience to later versions produced for readers assumed to want the work on its own terms. In each case the sequence looks orderly in retrospect: introduction first, fidelity later.

That orderliness is itself a reason for caution. Publishing history rarely moves in a single direction for long, and a pattern drawn from celebrated cases may reflect which versions publishers choose to publicize rather than how translation as a whole behaves. A retranslator commissioned to restore a source has every reason to describe the predecessor as domesticated, whether the description fits in detail or not. A first translator working from a different edition of the source, or for a publisher with unusual tolerance for difficulty, may have produced a version stranger than anything that followed. The sequence of dates alone settles nothing.

For that reason the hypothesis is best treated as a claim to be tested in a later section, not as a settled law of literary history. It gives a reader a set of questions rather than a verdict. Did the first version cut, soften or regularize, and can those changes be shown passage by passage. Did the later version reverse them, or did it substitute a different set of compromises. Did the conditions of publication actually change between the two versions, or does the story of a bold return rest only on advertising. Where the answers support Berman, the hypothesis explains a great deal about why a retranslation was needed and what it changed. Where they do not, the failure of the pattern is equally informative, because it reveals commercial motives, shifting taste and individual temperament at work behind a tidy theory.

Syntax Is Meaning: Sentence Length as a Translation Decision

A long sentence is not a container that can be emptied into several short ones without loss. In much serious prose the shape of the sentence is part of what the sentence means. Delay, interruption and qualification change how a thought arrives, and a reader who receives the conclusion at once has read a different thought from a reader who was made to wait for it, collecting conditions and doubts along the way. Translation forces a decision about that shape in every paragraph, because languages differ in how much suspension they tolerate before the sense resolves.

Thomas Mann makes the problem unavoidable. His periodic sentences hold the main verb back, stack subordinate clauses inside one another, and allow irony to accumulate as qualifications modify what seemed settled a moment earlier. A narrator appears to praise a respectable merchant family, then a subordinate clause introduces a reservation, then a parenthetical remark sharpens the reservation into judgment, and only at the end does the sentence disclose where its weight falls. The thinking happens inside the architecture. Shorten the architecture and the judgment arrives too cleanly, as a verdict rather than as a mind in motion.

Why do translators break up long sentences?

Translators divide long sentences because English publishing conventions reward clarity and pace, because a literal rendering can tangle the grammar, and because each clause must find a place before the main verb appears. Division guarantees readability, yet it also removes delay, subordination and suspense, which in Mann carry irony and thought in motion across a paragraph.

The pressures behind division are practical as well as stylistic. English readers are trained by journalism, commercial fiction and much nonfiction to expect frequent full stops, and a page of German prose rendered with its periods intact can look forbidding before a single idea has been judged on its merits. Translators also face genuine grammatical limits. A construction that German handles through case endings and verb placement may become ambiguous in English if every clause is preserved in order, and some rearrangement is unavoidable. The question is therefore not whether a translator ever divides a sentence, but how often division becomes the default, and what pattern of thought disappears when it does.

John E. Woods offered the most discussed English alternative in his 1995 translation of The Magic Mountain. Reviewers praised Woods for keeping the architecture of Mann’s sentences where H. T. Lowe-Porter had frequently broken them into shorter units. Keeping a long sentence readable in English requires placing clauses so that a reader can hold unfinished business in memory without confusion, and it requires confidence that readers will accept a slower rhythm in exchange for the movement of the argument.

Across eight hundred pages such choices accumulate into interpretation. A novel built from suspended, qualifying sentences presents consciousness as layered and self correcting. The same material divided into brisk declarations presents consciousness as decisive. A reader who compares a page of ordinary narration, a family conversation or a passage of reflection will see it at once. Syntax decides how intelligent, how hesitant and how ironic every character and narrator appears to be, and a retranslation that restores the sentence restores a large part of the book that a plot summary would suggest had never been missing.

Register, Dialect and the Voices That Resist Carrying

Every society talks in layers, and a novel records those layers as faithfully as it records events. Formal address, merchant politeness, servant directness, schoolboy slang and the speech of children carry information about rank, region, education and intimacy that a plot summary discards. A translator must decide what to do with that information when the target language divides its speakers along different lines. English has dialects and class markers in abundance, but they map onto English and American histories, not onto the histories of Lübeck, Madrid or a Russian estate. Any equivalent is therefore an approximation, and any refusal to approximate leaves the social structure of the original partly silent.

Buddenbrooks depends on those distinctions. The merchant generation speaks with a confidence grounded in ledgers, civic office and inherited standing, while younger family members, employees and outsiders reveal their positions through hesitation, deference, bluntness or studied imitation. When all of them arrive in the same neutral English, the novel loses a mechanism by which it shows a hierarchy loosening over time. A change in how a character is addressed, or in whether a character dares to interrupt, can signal a shift in power more economically than a paragraph of narration.

Don Quixote poses the same problem from the opposite social direction. Sancho Panza speaks in proverbs, malapropisms and earthy idiom, and his language sets a counterweight to the elevated, book formed speech of his master. A translator who renders Sancho in tidy standard English destroys the contrast that makes their dialogue a contest between two ways of knowing the world. A translator who gives Sancho a strong regional dialect from England or the United States restores contrast at the price of relocation, moving a Spanish peasant into a landscape his proverbs never knew. Edith Grossman, whose Don Quixote appeared in 2003, discussed translation as an act of writing in Why Translation Matters, published in 2010, and her version is widely read for keeping Sancho’s voice distinct without turning him into a stage rustic borrowed whole from English tradition.

Three broad strategies recur, and each exchanges one loss for a gain. Levelling the voices produces clarity and pace, and it loses the social information that the original encoded in speech. Inventing an equivalent dialect produces immediate contrast and energy, and it loses geographical truth, importing associations the author never intended. Marking difference through syntax and vocabulary, through sentence length, formality, forms of address and the choice between plain and elevated words, produces a subtler contrast that survives without a borrowed accent, though it loses some colour and comedy.

Retranslation reopens the bargain because taste in dialect changes quickly. A version that used a recognizable urban vernacular may date faster than the original, whose regional speech belongs to a settled literary history. A later translator may therefore choose restraint where an earlier translator chose colour, or may recover jokes, insults and proverb chains that an earlier version levelled away in the name of decorum. Neither direction is automatically closer to the source. The honest test is whether the English preserves the function of the speech in its scene: who gains authority, who is mocked, who is misunderstood, and who reveals more than intended. Where those functions survive, the voices have been carried, even though no individual word has an exact twin in the other language.

Titles, Names and the Words on the Cover

Judgment of a translation begins before the first sentence, on the title page, where a series of small decisions announces how the new version understands the work. A title can keep a foreign word, translate it, add an article, drop a subtitle or restore one that an earlier edition omitted. Each choice affects what a reader expects the book to be about. A family chronicle, a novel of decline and a portrait of a merchant house may be the same volume under different titles, yet the emphasis placed on the cover guides reviewers, booksellers and readers toward different questions before the argument of the book has a chance to correct them.

Buddenbrooks illustrates how little it takes to shift that emphasis. The German title names a family, and English editions have appeared both with and without the definite article before the name. The shorter form presents the family as a proper name standing on its own, while the longer form presents them as a particular household singled out from a wider world of households. Neither choice can be dismissed as a typographical detail. Articles in English carry habits of reference that German does not reproduce exactly, and a retranslator deciding between the forms is deciding how familiar or how emblematic the family should feel at first sight. Subtitles, series labels and cover descriptions extend the same decision outward.

Names inside the book raise a harder version of the problem, especially in Russian translation, where systems of transliteration have changed across generations. Constance Garnett’s versions, published from 1912 onward, used spellings that English readers of her generation found settled and recognizable. Later translators adopted different conventions for the same Russian sounds, so a character familiar under one spelling returns under another that looks like a stranger. Patronymics, diminutives and forms of address add further layers. A Russian character may be addressed by first name, by first name and patronymic, or by a shortened affectionate form, and each choice signals distance or intimacy. A translation that standardizes all of them to a single English name gains simplicity and loses a social instrument the author used constantly.

Place names carry interpretive weight of a similar kind. A translator must choose between an established English form, a form closer to local usage and a form that preserves a historical name no longer used by residents. The choice can make a setting feel timeless, foreign or politically situated, and inconsistent choices within one volume can confuse the geography of the narrative itself. Retranslation reopens every one of these settled decisions at once. A new edition typically prints a list of principal characters, standardizes spellings and explains its conventions in a note, and those preliminary pages are evidence rather than decoration. They show whether the translator treated names as transparent labels or as part of the meaning. A reader comparing versions can learn a great deal from the name list alone: which relationships are made visible, which ambiguities are resolved before the story begins, and which foreign forms are trusted to become familiar through use rather than being replaced at the door.

What Gets Cut, Softened or Explained Away

The most consequential changes in a first translation are often absences. A passage shortened, a scene summarized, a paragraph quietly removed: none of these announces itself to a reader with no other version for comparison. Suppression takes several distinct forms, and separating them matters because a retranslation can repair one form while repeating another. A cut made by a publisher for reasons of taste differs from an abridgement made to meet a deadline, and both differ from the smoothing of an ambiguity that a translator resolved without noticing that resolution was a choice.

Publisher censorship and self censorship account for the first form. Editors introducing a foreign author to a new market have historically worried about religion, the body and politics, and translators aware of those worries have softened passages before an editor needed to ask. Blasphemous jokes become mild, sexual frankness becomes suggestion, and criticism of institutions becomes generalized reflection on human folly. The changes protect the book’s reception and falsify its record of what its society permitted a writer to say.

Deadline abridgement accounts for the second form. A long novel commissioned on a tight schedule invites quiet compression, especially in descriptive passages, repetitive dialogue and episodes that a hurried translator judges detachable from the plot. The title page may still promise a complete text. Readers then argue about an author’s pacing, structure and powers of description on the evidence of a book the author never wrote in that shape. Retranslation exposes such abridgement only when a later translator works from a fuller edition and a publisher allows the restored length to stand, since restored material raises printing costs and can make a classic look forbidding beside the shorter version generations have finished and loved.

Explanatory smoothing accounts for the third and subtlest form. Originals leave motives unexplained, allow sentences to support two readings and let narrators contradict themselves. A translator under pressure to deliver a clear text resolves those openings, choosing one reading and translating it confidently. The English becomes easier and poorer at the same time. Ambiguity that invited rereading is replaced by a single sense that invites only agreement. Restoring ambiguity is delicate work, because English grammar sometimes forces a choice that the source could postpone, yet a careful retranslator can preserve more openness than a predecessor did, through word order, punctuation and a refusal to add explanatory phrases.

The documented patterns in H. T. Lowe-Porter’s Mann translations belong largely to the first and third forms. Critics and later editors have described her versions as softening Mann’s difficulty and resolving his irony into clearer statement, a record discussed in detail in the sections devoted to her work. Constance Garnett presents a different case centred on speed. She translated Russian literature at remarkable pace across many volumes, and scholarly discussion of her versions treats errors and flattening as a matter of critical record rather than private accusation. David McDuff, himself a translator of Dostoevsky, was among the retranslators who discussed her versions publicly while presenting alternatives.

Restoration has its own excess. A retranslator determined to recover everything a predecessor suppressed can overcorrect into pedantry, expanding every compact phrase, explaining references the original trusted readers to catch and loading sentences with scholarly caution until the narrative stops moving. Completeness on the page is not the same as fidelity in effect. The strongest retranslations therefore restore substance without worshipping it, and they accept that a living sentence in English sometimes serves a suppressed passage better than a museum display of everything it contained.

The Worked Case: Buddenbrooks in 1924 and in 2026

A general argument about retranslation earns its keep only when it survives contact with a single book read closely. Buddenbrooks offers an unusually clean test because one English version carried the novel for generations and a new version, published by Oxford University Press in 2026 in a translation by Mike Mitchell and edited with an introduction and notes by Ritchie Robertson, was prepared in full knowledge of that inheritance. The earlier version made Thomas Mann famous in English after its publication in 1924, introduced his fictional Lübeck to readers who had never heard of the city and established the terms in which English language criticism discussed him for decades. Without a readable first translation there would have been no English audience waiting for a second one.

The documented limits of the 1924 version are the limits examined in the discussions of Lowe-Porter’s method and of suppression in first translations. Long sentences were divided, difficult constructions were regularized and material and nuance were lost in ways that later scholarship documented passage by passage. None of that cancels the version’s historical role, but it defines the ground on which a competitor can claim to offer something real. Mitchell’s version competes on restored syntax, recovered material and scholarly apparatus. The introduction and notes by Ritchie Robertson give a reader context that the earlier edition did not supply in the same form, and the translation itself asks to be judged on whether Mann’s sentences think in English with something of their German movement.

The long family dinner sentences provide the first pressure point. Mann’s dinners are social performances in which politeness, appetite, business anxiety and family hierarchy operate at once, and his sentences hold those forces together through subordination and delay. A version that breaks the performance into short exchanges gains clarity about who speaks and loses simultaneity, the sense that judgment is forming while courtesy continues. A version that keeps the sentences long risks congestion in English, yet preserves the comedy of a household maintaining its forms while its foundations shift. Comparing the two versions at such a scene teaches more than comparing their famous passages, because the dinner table is where the novel’s social mechanism runs at full pressure.

Narration and irony provide the second pressure point. Mann’s narrator stands close to the Buddenbrook family and apart from it at the same time, adopting their vocabulary of respectability while arranging details so that the vocabulary indicts itself. A translation that sharpens the irony turns the narrator into a satirist delivering verdicts from outside. A translation that softens it turns the narrator into a memorialist sharing the family’s illusions. The difference lies in small choices of word and clause order, repeated hundreds of times. Mitchell’s claim on a reader rests substantially on this register, on letting respectability sound convincing in the sentence where it is also being exposed.

Lübeck merchant speech provides the third pressure point. The older generation’s confidence, the clerks’ deference and the younger generation’s loosening grip on inherited forms must be audible without importing a ready made English regional dialect. A translator can mark the differences through formality, vocabulary and rhythm, accepting a quieter contrast, or can reach for stronger colour and accept associations from another place and class system. The same pressures appear wherever a single novel accumulates versions across languages and generations, as how translations handle The Great Gatsby shows for a very different American classic and its translators abroad. In Buddenbrooks the choice decides whether decline is heard as a change in speech before it is understood as a change in fortune. Two versions can agree on every event in the novel and disagree, at that level, about what the novel believes caused the fall.

The Worked Case: Homer, Tolstoy and the Reviewers’ Contest

Retranslation becomes most visible when reviewers turn a publication into a contest between named versions, and two cases show how differently such contests can be argued. Emily Wilson’s translation of The Odyssey, published in 2017, was the first English translation of the poem by a woman, and discussion of it centred on choices that were plainly matters of interpretation. Wilson matched the Greek line for line and worked in iambic pentameter, giving the English poem a regular measure while refusing extra lines for explanation or ornament. Economy of that kind forces decisions everywhere else. Epithets, repeated formulas and descriptions of violence and servitude must fit the line, and the translator’s priorities become legible in what survives the constraint.

Her rendering of the famous opening epithet drew the sharpest debate. Wilson described Odysseus with a plain English word centred on complexity, where earlier translators had chosen terms emphasizing manhood, resourcefulness or wandering. Reviewers who admired the choice argued that it presented a hero defined by twists of character and circumstance rather than by a conventional masculine ideal, and that the plainness corrected centuries of accumulated grandeur. Reviewers who questioned it argued that something of the hero’s stature and strangeness was reduced when the opening settled for a familiar modern adjective. Both sides were arguing about interpretation, not error. The Greek permits argument, and the debate itself demonstrated that a retranslation can reopen a character whom fluent familiarity had closed.

The versions of Tolstoy and Dostoevsky by Richard Pevear and Larissa Volokhonsky produced a longer and more divided contest. Their working method has been described consistently in accounts of their partnership: Volokhonsky, working from the Russian, prepares a detailed draft that follows the source closely, and Pevear then shapes that material into English prose. Their War and Peace appeared in 2007, and their versions were presented as recoveries of roughness, repetition and humour that smoother predecessors had polished away. Admirers credited the partnership with restoring the grain of Russian prose, arguing that awkwardness in the English often corresponded to deliberate effects in the original and that earlier fluency had misrepresented major writers as more conventional stylists than they were.

Detractors reached the opposite verdict and should be given their case in its strongest form. Gary Saul Morson, a scholar of Russian literature, was a prominent critic of the Pevear and Volokhonsky versions, charging them with stiffness in English and with benefiting from marketing that treated established translations as discredited. In that reading, fidelity claimed at the level of the phrase produced infidelity at the level of the sentence a reader actually experiences, and publishers sold novelty as recovery. Supporters answered that comfort is a poor measure of Tolstoy, whose prose makes demands in Russian that a wholly comfortable English version has by definition refused. The dispute has never resolved into a single verdict, and its persistence is instructive. Wilson’s contest turned on visible choices within a strict form that readers could inspect line by line. The Pevear and Volokhonsky contest turned on a larger question that parallel passages alone cannot settle: whether strangeness in translation is evidence of the original speaking or evidence of the translators failing to write English, and who has the authority to tell the difference.

The Case Against Retranslation: When the New Version Is Worse

The strongest argument against the cult of retranslation begins with evidence rather than taste. Kaisa Koskinen and Outi Paloposki examined retranslation in their research published in 2010 and found the pattern proposed by Antoine Berman unreliable as a general description of publishing history. Retranslations often happen for commercial reasons rather than scholarly ones: a publisher needs a text free of a rival’s rights, a series needs a uniform set of covers, an anniversary creates a marketing occasion, or a famous name becomes available to front a new edition. Later versions, examined closely, are not systematically closer to their sources than the versions they replaced. Some move toward the source in one respect and away from it in another. Some modernize idiom while preserving every structural compromise of the predecessor. Chronology, on this evidence, predicts very little.

The same research restores first translations to serious consideration. A first translator working close to a living literary culture, or writing in a period whose prose habits happen to suit the source, may produce the boldest version in a sequence. Later translators, aware of reviews, scholarly scrutiny and the commercial need to justify a new edition, may become cautious where the first translator was free. Boldness matters in literature because many canonical works were bold in their own languages, in syntax, in humour and in their treatment of belief and the body. A careful version that never risks an equivalent boldness can be accurate in detail and false in effect. Age alone tells a reader nothing about which of those qualities a version possesses.

Constance Garnett’s defenders make that argument in its most concrete form. Her versions of Tolstoy, Dostoevsky, Chekhov and Turgenev have been criticized for errors and for haste, and the criticism rests on a documented record. Yet readers and writers who defend her argue that her musical, periodic English prose suits Tolstoy in particular, giving his sentences a swell and continuity that flatter replacements break into plainer units. A translation lives in the ear as well as on the page, and a version whose rhythms carry a reader through hundreds of pages performs work that a more scrupulous version may fail to perform at all. Generations formed their understanding of Russian literature through Garnett’s cadences. Dismissing that achievement because later scholarship found mistakes confuses one kind of value with the whole of value.

Publishers, meanwhile, have a clear interest in declaring old versions obsolete. A new translation can be copyrighted, marketed, excerpted and sold as an event, while a public domain predecessor can be reprinted by anyone at any price. The language of recovery serves that interest whether or not recovery occurred. A famous name on the title page sharpens the problem. A celebrated novelist, poet or public intellectual attached to a new version can sell a difference that parallel reading does not confirm, because buyers are purchasing confidence as much as text. None of this proves that a given retranslation is hollow. It proves that the incentives surrounding retranslation reward exaggeration, and that a reader is entitled to discount claims made by parties who profit from them.

The thesis defended across these sections therefore needs narrowing, and the narrower version is stronger for the concession. A retranslation earns its place only where it restores a demonstrable suppression or answers a demonstrable change in knowledge: a fuller source edition, recovered passages, documented cuts, a register that can be shown to have been levelled, scholarship that changes how a passage should be understood. Elsewhere a new version is one more reading among rivals, to be judged by the same tests as the version it hopes to displace and entitled to no presumption in its favour. Some retranslations meet the narrower test triumphantly. Others fail it while wearing its language. The failure does not condemn the practice. It disciplines the claim, and it returns judgment to the only place where judgment can be exercised honestly, in passages read side by side.

Retranslations That Failed to Displace the Originals

Publication is not succession. A new translation enters a field already occupied by habits, institutions and affections, and many distinguished versions have failed to move the occupant. Louise and Aylmer Maude produced English versions of Tolstoy that carry a distinction no marketing campaign can reproduce: their work was made with Tolstoy’s approval, and Tolstoy knew and endorsed their rendering of his work. Their War and Peace has survived a century of challengers, including versions prepared with resources, scholarship and publicity far beyond anything available to the Maudes. Approval by the author is not a guarantee of stylistic superiority, and later translators have identified choices of theirs to dispute. It is, however, a form of authority that a challenger cannot replicate, and it anchors the Maudes in the history of the book rather than merely in its bibliography.

Constance Garnett presents the larger case. Wave after wave of Russian retranslation has been announced as her replacement, individual volumes have been praised at her expense and publishers have built series on the premise that her era had ended. She remains in print through all of it. Her versions continue to be reissued, assigned, quoted and loved, not only because they are inexpensive or familiar, but because they work as English books. The phenomenon repeats at smaller scale across literatures. Announced retranslations appear with strong reviews, sell through an initial printing and vanish within a decade, while the version they targeted continues in a modest, durable edition that never leaves the catalogue. Literary history remembers the announcements. Bookshops remember the survivors.

Four mechanisms explain that incumbency. School and university adoption comes first, because a text set for a course generates steady sales, teaching notes, companion guides and page references that all favour the adopted version, and changing versions imposes costs on every teacher who uses them. Quotation habits come second, because scholars, journalists and other writers quote the version they know, and each quotation renews that version’s claim to be the standard text. Price comes third, because an older translation free of rights payments or long amortized can be sold cheaply in formats a new translation, still recovering its costs, cannot match. Affection comes last and may matter most. A first loved version becomes, for an individual reader, the original itself. The voices heard in adolescence or in a first encounter with a great book acquire an authority that no apparatus of notes can dislodge, because they are bound to the memory of discovery.

None of these mechanisms measures accuracy, and together they can protect a flawed version long after its flaws have been documented. They can also protect a great one from a fashionable displacement it never deserved to suffer. Incumbency is therefore neither a verdict nor an injustice in itself. It is the environment in which every retranslation must operate, and it explains why the history of translation is a history of coexistence punctuated by occasional succession, rather than a ladder in which each generation climbs past the last. A retranslator who understands that environment writes for readers who already possess a version of the book in memory, and the task is not to erase that possession but to earn a place beside it.

How Criticism Judges a Translation It Cannot Fully Check

Reviewing a translation presents a structural problem that reviewing an original does not. A critic may read the target language with complete assurance and the source language only partially, or may know the source well and lack a fine ear for prose rhythm in the language of publication. Either imbalance shapes the verdict before a single passage is compared. An honest review therefore begins by acknowledging its own limits, stating the languages the reviewer commands and the basis on which judgement rests, rather than presenting an impression of the English text as a measurement of fidelity to a work the reader of the review cannot inspect unaided either.

Responsible practice works by sampling, and the sample is chosen where translation choices concentrate. The opening page reveals how a translator handles exposition, proper names, and the establishment of narrative voice. A scene of dialogue exposes register, since speakers who sound identical in translation may be sharply differentiated in the source by class, age, or region. A passage of sustained description tests whether imagery survives as imagery or arrives flattened into summary. A famous crux, a pun, a proverb, a line of scripture quoted inside the fiction, or a sentence critics have disputed for decades, shows what a translator does when no equivalent exists and a decision cannot be avoided.

The translator’s stated method supplies a second check. Prefaces, interviews, and notes declare intentions about dialect, archaism, sentence length, and the handling of terms without equivalents. A reviewer can test the finished text against those declarations. A translator who promises to preserve long, periodic sentences and then delivers short, uniform units has failed by an internal standard, whatever an outside reader thinks of the result. A translator who announces deliberate modernisation and carries it through consistently invites a different argument, about the wisdom of the policy rather than the competence of its execution.

The characteristic failure mode is praise for fluency alone. A review that calls a version readable, smooth, and invisible, and stops there, rewards exactly the habit Lawrence Venuti criticised in The Translator’s Invisibility (1995), in which domesticating choices erase the foreign character of a work and present the translator’s labour as though no mediation had occurred. Fluency is a real virtue, since unreadable fidelity serves no reader, but fluency purchased by deleting difficulty is a loss described as a gain. Stronger reviewing asks what a smooth passage cost, compares the difficult places across rival versions, and treats awkwardness as evidence to be weighed rather than as an automatic defect. Criticism of translation is therefore best understood as triangulation among parallel passages, declared method, and the reviewer’s disclosed competence, with certainty reserved for the few points where all three lines meet. A review that follows that discipline teaches its readers how to distrust single impressions, including the impression the review itself creates, and that instruction outlasts any verdict on one edition.

How a Reader Compares Two Translations in Ten Minutes

Comparison does not require fluency in the source language, because the most revealing evidence sits inside the rival English texts themselves. Two versions placed side by side disagree in thousands of small decisions, and a reader can learn to read those disagreements as a record of interpretation. The method works for any classic available in competing editions and asks only for patience, a pencil, and willingness to trust patterns over first impressions.

How can a reader compare two translations without knowing the original language?

Choose three short passages in each version, place them side by side, and study where the English texts diverge. Repeated divergence reveals method. One version consistently shortens sentences, softens insults, or explains allusions that the other leaves standing. Patterns, not isolated lines, expose the reading each translator made.

Begin with three test passages. The first page shows how each translator establishes voice, handles names, and sets a tempo for narration. A heated exchange of dialogue shows whether speakers retain distinct levels of formality, whether insults keep their force, and whether interruptions and incomplete sentences survive or are tidied into orderly replies. A quiet descriptive passage shows whether specific nouns stay specific, whether metaphors remain strange where strangeness belongs, and whether the prose can slow down when the scene slows down. Three passages cannot settle every question, but they sample the three kinds of work a novel asks of its translator, and a version that falters in all three samples rarely recovers elsewhere.

Sentence length offers the simplest diagnostic. If one version habitually splits a single long sentence into three short ones while the other preserves the longer architecture, a decision about rhythm and subordination has been made on every page. Neither choice is automatically superior. Preservation can produce congestion, and division can produce clarity, but a reader should know that the tempo being experienced belongs in part to the translator. The same attention applies to jokes, forms of address, and repeated motif words. A joke replaced by a neutral statement signals caution. A formal pronoun flattened into uniform familiarity erases a social signal that the plot may later depend upon. A motif word rendered by five different English words across a chapter breaks a pattern the author built deliberately, while consistent rendering lets the pattern accumulate.

Reading a page of each version aloud exposes what silent reading conceals, since breath follows syntax and monotony becomes audible within minutes. A translator’s note then helps interpret the evidence. A version whose choices are visible and explained allows a reader to disagree intelligently and to locate the source of discomfort in a stated policy. A version that merely reads smoothly may be excellent, but smoothness alone gives no account of what was surrendered. Preference formed by comparison, passage against passage, rests on firmer ground than preference borrowed from a reputation. Ten minutes of testing will not settle which version a library should canonise, but the testing replaces inherited opinion with observed habit, and observed habit is the ground on which a serious choice between translations has to stand.

Editions, Notes and the Scholarly Apparatus

A translation rarely reaches a reader as bare text. Material surrounds it, introducing, explaining, and quietly certifying it before the first chapter begins. That surrounding material is called the paratext, a term used by Gérard Genette for everything that frames a text and mediates between the text and its audience: introductions, notes, glossaries, maps, chronologies, cover copy, and the name of a series printed on a spine. Paratext is not decoration. Paratext directs attention, settles questions a reader has not yet learned to ask, and establishes the authority with which a particular version will be received by libraries, teachers, and reviewers.

An introduction argues a case. An introduction places a work in a literary tradition, summarises the author’s career, and usually explains why a new translation was needed and what principles governed its making. Such framing primes interpretation. A reader told that a novel is principally a satire reads its tender scenes differently from a reader told that the same novel is principally a family chronicle. Notes perform a different service. A note can identify an allusion, explain a unit of currency, describe a legal custom, or record that a phrase carries a double meaning in the source language that English cannot hold in a single rendering. Each note removes a small obstruction without altering the sentences of the translation itself.

The apparatus prepared by Ritchie Robertson for the 2026 edition of Buddenbrooks, published by Oxford University Press, illustrates the division of labour at its most productive. Notes explaining Lübeck customs, local institutions, commercial practices, and Thomas Mann’s allusions allow the translated prose to retain foreign names, forms, and habits instead of dissolving them into familiar English equivalents for the sake of instant comprehension. Help sits in the margin, where a reader may consult or ignore it, rather than inside the sentence, where explanation would flatten the fiction into a guided tour. Glossary and chronology extend the same principle, gathering reference material at the edges of the book so that the centre can remain a novel.

Apparatus also carries risks. Excessive annotation can intimidate a first reader, turn fiction into coursework, and imply that enjoyment without scholarship is illegitimate. A partisan introduction can harden into orthodoxy, discouraging readings the evidence would support. The best scholarly editions therefore keep their scaffolding visible and separable, allowing the translation to be read clean, consulted when difficulty arises, and argued with when the framing overreaches. Authority earned in that manner differs from authority merely asserted by bulk. A reader who understands the apparatus can also discount the apparatus, skipping a tendentious introduction, sampling notes only at points of genuine difficulty, and treating the edition as a set of tools offered for inspection rather than as a verdict delivered before the evidence has been heard.

Translation Prizes, Series and the Signals Publishers Send

Before a page is read, a translation arrives already certified by signals that most readers absorb without examining them. A prize sticker on a cover, the uniform design of a classics series, a famous imprint, and an endorsement quoted from a newspaper all promise that somebody competent has inspected the book and found it worthy. Each signal certifies something real, and none of them certifies the quality readers most often assume, which is fidelity to the source at the level of the individual sentence. Understanding what each signal can and cannot guarantee turns a reader from a consumer of reassurance into a judge of evidence.

A prize certifies the taste of a jury. Jurors read submissions, argue from preferences formed by wide reading, and reward the book that commands agreement under the rules of a particular competition in a particular cycle. The International Booker Prize recognises author and translator jointly, a structure that at least acknowledges translation as creative labour rather than invisible service, and that directs attention and money toward the translator’s share in a book’s success. A prize can certify that informed readers found a version powerful, memorable, and well made as an English book. A prize cannot certify that those readers compared every chapter against the source, because juries rarely work that way and competitions are not audits.

A series certifies an editor’s standards. Uniform classics series maintain lists over decades, commission new versions according to house principles, and stake a collective reputation on consistency of presentation, scholarship, and production. Inclusion in such a series indicates that a publisher judged a work canonical, a translator employable, and an edition commercially durable. Cover design certifies something narrower still: a marketing budget and a theory about who should buy the book. Design aimed at students signals apparatus and price discipline. Design aimed at general readers signals readability and gift appeal. Neither signal reaches inside the sentences.

None of these signals is worthless, and dismissing them would waste genuine information about editorial care and institutional commitment. The error lies in treating certification of process as certification of equivalence. Sentence level fidelity is established only by comparison, quotation of parallel passages, and the slow accumulation of scholarly review over years. Signals decide which version gets stocked, displayed, taught, and reviewed first, and that distribution of attention shapes which translation a generation actually reads, regardless of the verdict that closer inspection might eventually reach. Teachers assign the version with usable notes, booksellers reorder the version with visible momentum, and reviewers request the version a publicity office sends first, so signals compound into availability, and availability quietly hardens into the appearance of consensus about merit.

What a Retranslation Cannot Recover

Every retranslation is commissioned in the shadow of an implied promise, that a version closer to the original can be produced if errors are corrected, scholarship is applied, and the receiving language is handled with greater courage than earlier translators were permitted. Much of that promise can be honoured. Mistakes can be repaired, censored passages restored, and distortions of register corrected. A residue remains that no competence can remove, because part of what an original possessed belonged not to its words alone but to the moment, audience, and linguistic world in which those words first circulated, and that world does not travel.

Can a translation ever be definitive?

No translation is definitive, because the source holds more relations among sound, sense, custom, and allusion than any single target text can carry at once. Each version selects a defensible subset of those relations. Selection changes with scholarship and taste, so later versions revise the choice rather than completing it.

The first audience’s horizon is the deepest loss. Original readers brought assumptions about honour, religion, money, family obligation, and literary convention that required no explanation and left no trace in the text itself. A translator can annotate those assumptions, or can substitute modern equivalents, but cannot make a later audience unknow them into instinct. Explanation produces knowledge about a world, which is a different experience from inhabiting that world without noticing it. Footnotes close the informational gap while widening the experiential one, since every consultation reminds a reader that the fiction was addressed to somebody else first.

Wordplay, metre, and cultural resonance survive partially at best. A pun depends on accidental kinship between sounds inside one language, and that kinship rarely repeats in another. Metre can be imitated by choosing a comparable English measure, yet comparable measures carry different associations accumulated through different poetic traditions. A word for a food, a garment, or a form of address arrives trailing the history of its own culture, and the nearest English term trails a different history behind it. Skilled translators compensate, inventing effects at one point to balance effects surrendered at another, but compensation is creation under constraint, not recovery of what was lost.

Edith Grossman, in Why Translation Matters (2010), argued that a translation is a reading, and readings multiply rather than converge. The claim explains why each version documents its own receiving culture as faithfully as the version documents the source. Vocabulary choices date a translation, assumptions about explicitness date a translation, and ideas about what a classic should sound like date a translation most of all. Definitiveness is therefore a marketing word, useful on covers and misleading everywhere else. The honest goal is narrower and more demanding: a version that knows its own choices, states them plainly, and accepts responsibility for the portion of the original that those choices set aside.

The Author’s Afterlife in Another Language

An author who writes in one language and is read in another lives a double literary life, and the second life is built largely by translators. Reputation in the receiving culture follows the versions available there, their tone, their prestige, and their presence in classrooms, so that an author can be solemn in one language, comic in another, and difficult or approachable depending on decisions taken far from the desk where the original was composed. Retranslation matters beyond the page because each new version revises that second life, confirming or displacing the figure that readers in the receiving culture believe the author to be.

Thomas Mann’s English reputation was constructed through one translator’s voice and then rebuilt by her successors. For decades, English readers who spoke of Mann’s irony, Mann’s formality, and Mann’s distance were describing qualities filtered through H. T. Lowe-Porter’s renderings, which set the terms in which his fiction was reviewed, taught, and imitated. Later translators, working under different assumptions about syntax and explicitness, produced a Mann who reads as a different novelist in measurable ways, funnier in places, harsher in others, and less uniformly elevated. The German texts underneath both reputations stayed constant while the English author changed.

The Russian novel in English is partly Constance Garnett’s construction in a similar sense. Her versions introduced Tolstoy, Dostoevsky, and Chekhov to generations of English readers and to English writers who learned fictional technique from them. A shared, slightly Edwardian voice across authors who differ greatly in Russian gave the receiving culture a unified idea of Russian literature, earnest, spacious, and spiritually charged, that owed as much to the mediator as to the mediated. Cervantes in English presents the opposite pattern, four centuries of accumulated versions in which no single voice monopolised the author, so that Don Quixote has been successively a burlesque, a romance, a philosophical novel, and a comedy of language, each emphasis licensed by a translation and then preserved in criticism written around it.

Canon formation through translation follows the receiving culture’s needs more closely than the source culture’s rankings. Works are retranslated when a receiving culture finds a use for them, when curricula demand teachable editions, when scholars dispute an inherited reading, or when a publisher senses that an established title can bear a new identity. Authors marginal at home can become central abroad, and central authors can wait generations for a version that makes them legible. The shelf of versions is therefore a history of readers as much as a history of an author, recording what each era wanted from the foreign book it chose to rebuild. An afterlife assembled in that fashion can also be revised in that fashion, since a single distinguished new version can shift syllabuses, prompt biographies to be rewritten around different evidence, and return neglected works by the same author to print on the strength of a reputation remade abroad.

Why the Cycle Will Keep Turning

Retranslation recurs because the conditions that produce it are structural, and structural conditions do not expire with any single edition, however accomplished that edition may be. Language drift alone guarantees the cycle. A translation is written in the receiving language as that language stands at a particular moment, and the receiving language then moves, altering the force of slang, the politeness of pronouns, the neutrality of terms for work, gender, money, and rank. A version that sounded transparent to its first readers gradually acquires a period flavour of its own, until readers begin hearing the translator’s decade instead of the author’s century, and a publisher can plausibly promise a clearer encounter.

Rights lapse in waves, and each wave reopens the commercial question. Copyright in an author’s works expires according to legal terms that have themselves changed across jurisdictions and generations, so bodies of work become available for competing editions in clusters rather than in a steady stream. Expiry does not compel anyone to commission a new version, but expiry removes the permission barrier, lowers the cost of entry, and invites publishers who could not previously compete with an incumbent edition to try. Where a market proven by decades of sales meets a sudden opening of rights, proposals follow as a matter of ordinary publishing arithmetic.

Scholarship revises the copy text as well. Editors establish new editions of canonical works from manuscripts, corrected printings, and letters, restoring passages, overturning attributions, and reordering material that earlier translators rendered from defective sources. A translation made from a superseded text inherits every defect of that text, and no excellence of style can repair a chapter that should not have been present or supply a paragraph that the source edition omitted. Each significant revision of a scholarly text therefore creates a standing argument for revisiting the translations derived from its predecessors.

A final driver sits inside readership itself. Each receiving generation wants contemporaries as mediators, translators whose prose habits, critical assumptions, and cultural references belong to the same world as the readers being asked to trust them. Mediation by a living sensibility carries a credibility that even a superb older version cannot reproduce, because trust in a guide depends partly on recognising the guide’s voice as a voice of one’s own time. Publishers understand that desire, translators embody that desire, and scholars supply fresh reasons to act on that desire. Language, law, scholarship, and taste turn together, and the wheel they form has no natural stopping point. A freeze would require all four to halt at once: a receiving language that stopped changing, a body of rights that never lapsed again, a scholarly text that no evidence could revise, and a readership content to be mediated forever by the voices of a single generation.

The Original Stays Still and the Bridge Is Rebuilt

A durable way to understand retranslation is to separate what changes from what does not. The original text, once established by scholarship, stays still. Sentences keep their order, images keep their places, and ambiguities keep their double edges, indifferent to the reputations built upon them abroad. Everything that changes gathers on the bridge between languages: the receiving tongue itself, the scholarship that explains the far bank, the publishing economics that finance the crossing, the critical taste that inspects the workmanship, and the readers whose expectations decide whether a crossing feels faithful or strange. Retranslation is the name given to rebuilding that bridge when the old structure, sound in its day, no longer matches the ground on either side.

Rebuilding is not an admission that earlier builders failed. A bridge raised under the constraints of one era, with its censorship, its dictionaries, its ideas of decorum, and its market for foreign fiction, could carry only the traffic those constraints allowed. Later builders inherit wider permissions, better maps of the source territory, and readers prepared to tolerate more foreignness in the crossing. Their structure restores spans the first builders were obliged to leave out and removes ornament that the first builders were obliged to add. Judged fairly, each version answers the same standing question under different conditions, and the sequence of versions, read together, shows the question itself changing shape across generations.

For readers, the practical consequence is freedom rather than confusion. A classic held in several translations is a classic whose interpretation has not been surrendered to a single mediator, and disagreement among versions marks the places where the original is richest and least reducible. Keeping more than one version within reach, perhaps beside a companion reference shelf, turns those disagreements into a map of the work’s depths instead of an obstacle to choosing a copy. Classics live by being carried across, and they stay alive because the carrying is never finished, only handed on to builders who will work from a different bank, with different tools, toward the same still text waiting on the other side. A literature that stopped rebuilding would not preserve its classics in purity; a literature that stopped rebuilding would strand those classics on a far bank, intact, honoured, and gradually unvisited, which is the one form of loss that no fidelity of wording can prevent.

Frequently Asked Questions

Q: What is a retranslation in literature?

A retranslation is a complete new rendering of a work into a language in which a translation already exists, produced by starting again from the source text rather than by adjusting the earlier English version. The retranslator reads the original as an independent interpreter, makes fresh decisions about syntax, register, dialect, and terms without equivalents, and accepts that the result will compete with a predecessor that readers may love. Retranslation differs from a new edition of an old translation, which repackages familiar wording, and from a revision, which patches that wording at selected points. The existence of a retranslation signals that a publisher, a scholar, or a translator judged the distance between the original and its established rendering large enough, in accuracy, in tone, or in completeness, to justify rebuilding the whole structure rather than repairing parts of a standing one.

Q: Why do classic books get translated more than once?

Classic books attract repeated translation because several pressures converge on the same titles. The receiving language changes across generations, so wording that once sounded neutral acquires period colouring that interferes with the author’s voice. Scholarship improves the source text, restoring passages and correcting errors that earlier translators had no chance to see. Legal barriers fall when rights in an author’s work expire, inviting competing editions. Publishers value proven titles whose readership is established, and a new version gives booksellers a reason to display a familiar work prominently again. Translators themselves seek the assignment, since rendering a canonical author is both a professional distinction and a chance to contest an inherited interpretation. Finally, readers change: expectations about explicitness, dialect, foreign names, and difficult syntax shift, and a version shaped for one set of expectations gradually stops matching the expectations of the audience that follows.

Q: Is a newer translation always better than an older one?

A newer translation is not automatically a better one, because date of publication measures opportunity rather than achievement. A retranslator benefits from improved dictionaries, established scholarly editions, freedom from older censorship, and knowledge of mistakes identified in earlier versions, and those advantages are real. Against them stand qualities that age cannot confer: an older translator may possess a finer prose style, a deeper intimacy with the author’s period, or a willingness to preserve strangeness that a market conscious successor avoids. Some retranslations are commissioned chiefly to give a publisher a proprietary edition, and such commercial origins show in hurried or cautious work. Sound judgement therefore compares versions passage by passage, weighing accuracy, register, completeness, and readability together, instead of treating chronology as a verdict. The strongest claim a new version can make is a specific one, naming what the version restores, corrects, or renders differently, and inviting readers to test that claim directly.

Q: What is the difference between a retranslation and a revised edition?

A revised edition begins from an existing translation and alters it in places, while a retranslation begins from the source text and rebuilds the rendering throughout. Revision suits a translation whose foundations remain sound but whose details have dated, whose errors are localised, or whose source text has been corrected at specific points; an editor or the original translator adjusts wording, updates spelling, and repairs identified mistakes while leaving the underlying interpretation intact. Retranslation suits a deeper disagreement, about voice, register, completeness, or method, that patching cannot express. The distinction matters commercially and legally as well as critically, since a revision stays close to a predecessor’s copyrighted wording, whereas a retranslation creates a substantially independent text. Publishers sometimes blur the boundary in marketing, presenting extensive revision as a wholly new translation or the reverse, so the title page, the translator credit, and any statement of method deserve attention from a reader deciding what a particular edition actually contains.

Q: Who was Constance Garnett and why do her translations still sell?

Constance Garnett (1861 to 1946) was the English translator whose versions, beginning in the early decades of the twentieth century, introduced Russian literature to a wide English reading public. Working at remarkable speed across an enormous body of fiction, Garnett translated Tolstoy, Dostoevsky, Chekhov, Turgenev, and Gogol, among others, and her renderings shaped how English writers themselves understood what a novel could attempt. Her versions still sell for several connected reasons. The prose retains a fluent, dignified readability that many readers continue to prefer. The translations are widely available, frequently inexpensive, and entrenched in classrooms and reading groups through long use. Critics have documented haste, omissions, and a tendency to smooth differences among authors into a single elevated voice, yet familiarity itself is a force in publishing: a version through which generations first met the Russian novel keeps readers who trust that first encounter and recommend the same doorway to others.

Q: Why did Thomas Mann rely on one English translator for so long?

For roughly three decades, nearly every Thomas Mann book reached English readers through H. T. Lowe-Porter, whose versions Alfred A. Knopf published from the 1920s onward. The arrangement suited everyone involved. Knopf held a valuable author and preferred a single trusted hand across a growing list, because consistency of voice helped build a recognisable English Mann. Mann himself, who read English and followed his reception closely, valued her loyalty, her speed and her standing with reviewers, and he supported her position even as scattered criticisms of detail reached him. Lowe-Porter, for her part, carried an extraordinary workload across his novels, stories and essays, and familiarity compounded: each new book arrived in an idiom her readers already associated with the author. Monopoly of this kind was common in the period, when publishers treated a translator as part of an author’s household rather than as an interchangeable supplier. Its cost only became fully visible later, when rival versions showed how many decisions one sensibility had settled for an entire literature of readers.

Q: Why are so many new Thomas Mann translations appearing?

A cluster of new Thomas Mann translations has appeared because the legal, scholarly, and commercial conditions for retranslation converged on his work at the same point. Mann died in 1955, and copyright in his works in the United Kingdom expired at the start of 2026 under the life plus 70 years term, removing the need to license the German texts from a rights holder and opening the field to competing publishers. Mann’s standing made that opening valuable: a canonical author with a large, teachable body of fiction promises durable sales for any house able to establish a distinctive edition. Scholars had also spent decades documenting the compromises of the older English versions through which most readers knew him, so a critical argument for fresh renderings stood ready when the legal argument arrived. Oxford University Press, Pushkin Press, and other publishers commissioned established translators, among them Mike Mitchell, Simon Pare, Ritchie Robertson, and Lesley Chamberlain, and the resulting wave reflects timing more than coincidence.

Q: How does copyright expiry lead to new translations?

Copyright expiry leads to new translations by removing the permission barrier that protects an established edition from competition. While an author’s work remains in copyright, a publisher seeking a fresh translation generally needs a licence from the estate or rights holder, who may refuse, charge heavily, or favour the existing version in which the estate holds an interest. Once the term ends, any publisher may commission a rendering of the source text without that licence, paying the translator rather than the estate for access. The underlying economics then change quickly for a canonical work with proven demand: several houses can enter the same market, differentiate their editions through translators, introductions, and design, and compete on interpretation instead of permission. Expiry also frees scholars to quote, annotate, and edit without restriction, strengthening the apparatus around a new version. The effect is strongest where an older translation has long held a monopoly, since pent up competition arrives the moment the gate opens.

Q: What is the retranslation hypothesis?

The retranslation hypothesis is the claim, associated with Antoine Berman in a 1990 essay on retranslation, that translations follow a directional pattern across time. A first translation, on this account, tends to domesticate a foreign work, adapting the work to the receiving culture’s norms in order to secure acceptance for an unfamiliar author, and only later translations dare to move closer to the source, restoring foreignness, difficulty, and formal features that the introduction suppressed. The hypothesis gave scholars a framework for comparing sequences of versions and gave retranslators a dignified rationale for their work. Subsequent research, notably by Kaisa Koskinen and Outi Paloposki in studies published from 2010 onward, found the pattern unreliable as a general law: some first translations are strikingly source oriented, some later versions domesticate more aggressively for commercial reasons, and local publishing conditions explain individual cases better than sequence alone. The hypothesis survives as a productive question to ask of any pair of versions rather than as a rule that settles the answer in advance.

Q: What do domesticating and foreignizing mean in translation?

Domesticating and foreignizing name the two broad directions a translator can move between a source text and a receiving audience, a distinction Lawrence Venuti developed in The Translator’s Invisibility (1995) from ideas reaching back to Friedrich Schleiermacher’s 1813 lecture on translation method. A domesticating rendering adapts the work toward the reader, replacing foreign customs, measures, names, and idioms with familiar equivalents and smoothing syntax until the text reads as though written in English. A foreignizing rendering moves the reader toward the work, retaining source language patterns, cultural specifics, and deliberate strangeness, and accepting occasional awkwardness as the price of encounter. Neither direction is a guarantee of quality: domestication can produce graceful books that betray their originals, and foreignization can produce pedantic books that mistake obscurity for fidelity. Most serious translations mix both strategies passage by passage, and criticism becomes useful when it identifies where a translator chose each direction and what the pattern of choices costs and gains.

Q: How long does it take to translate a long novel?

Translating a long novel is measured in years rather than months, as an order of magnitude that translators and scholars consistently report for major works of fiction. A translator must produce a complete draft, revise that draft against the source passage by passage, resolve thousands of local problems involving idiom, rhythm, and continuity of motif, and then respond to editors, copy editors, and sometimes scholars reviewing specialised material. Work proceeds alongside research into period vocabulary, historical setting, and the author’s other books, and a rendering of a stylist as demanding as Thomas Mann or Tolstoy slows further wherever syntax itself carries meaning. Many literary translators also hold teaching posts, editorial work, or other commissions, so calendar time stretches beyond the hours the manuscript alone would require. Publishers plan series and publicity around such timelines, and a rushed schedule is visible in the finished book as unevenness: strong opening chapters, thinner resolution of difficulties later, and inconsistencies that careful revision would have caught.

Q: How are literary translators paid?

Literary translators are paid through combinations of structures that vary by country, publisher, and the bargaining position of the translator. A common arrangement is a fee calculated per thousand words of source or target text, paid in stages at signature, delivery, and publication. Some contracts substitute or add an advance against royalties, giving the translator a share of sales after the advance earns out, and subsidiary income from paperback, digital, or foreign editions may be divided as well. Translators of canonical retranslations commissioned by major houses generally command stronger terms than translators of first books, reflecting reputation and the commercial certainty of a proven title. Grants, residencies, and prizes supplement contract income for many practitioners, and professional associations publish guidance on fair terms without fixing a single market rate. The structure matters artistically as well as financially: payment by speed rewards haste, while staged fees and royalty participation reward the slow revision on which durable versions depend.

Q: Can a translation be copyrighted separately from the original work?

A translation can hold copyright separately from copyright in the original work, because many legal systems treat a translation as a derivative work: a new creation built upon an existing one, in which the translator’s original choices of wording, rhythm, and structure attract protection of their own. Separate protection explains ordinary publishing practice. An ancient or long dead author’s text may be free for anyone to translate while a particular modern rendering of that text remains protected, so a rival publisher may issue a competing translation but may not simply reprint the protected rendering. Conversely, translating a work that remains in copyright generally requires permission from the holder of rights in the original, and the resulting translation then carries layered rights belonging to different parties. Protection also expires on its own timetable for the translation, distinct from the timetable of the source. The mechanism varies in detail among jurisdictions, and specific situations turn on the law and contracts involved rather than on any general principle stated here.

Q: How can readers tell whether a translation is abridged?

Readers can detect abridgement through a combination of signals printed in an edition and comparisons made outside it. The title page and copyright page of an honest abridged edition say so, using terms such as abridged, shortened, or adapted, and the absence of a translator credit altogether often signals a heavily edited older text. Physical proportion offers a rough test: a version of a famously long novel that arrives in a strikingly slim volume invites suspicion, especially when rival editions are far longer. Comparison settles the question more firmly. Checking a well known episode, a late chapter, or a passage discussed in criticism against the table of contents and the text reveals omissions quickly, and scholarly reviews of major translations routinely report cuts where cuts exist. Introductions sometimes confess the policy in courteous language about removing digressions or easing the reader’s path. Abridgement is a legitimate choice when declared, since declared cuts let a reader consent to the bargain; undeclared cuts convert a translation into a summary wearing a translation’s authority.

Q: Why do translators write prefaces explaining their choices?

Translators write prefaces because a translation embodies thousands of decisions that the finished text cannot explain from inside, and a preface gives the reasoning a place to stand. In a preface a translator can identify the source edition used, acknowledge passages where the original is ambiguous or corrupt, and state policies on dialect, archaism, names, measures, and terms left in the source language. Such statements convert private judgement into public evidence: reviewers can test the book against the declared method, scholars can cite the reasoning, and readers can decide whether a discomfort felt while reading reflects incompetence or a deliberate policy pursued consistently. Prefaces also serve a defensive purpose, distinguishing the new version from predecessors without disparagement and explaining why familiar renderings of famous lines have changed. A candid preface does not prove that the choices were right, but the absence of any account leaves a reader with a smooth surface and no way to interrogate the mediation that produced that surface.

Q: What makes a translation sound dated?

A translation sounds dated when the receiving language visible in the prose belongs recognisably to an earlier generation rather than to the author’s period or to a neutral literary present. Slang and colloquial dialogue age fastest, since informal speech changes quickly and yesterday’s lively equivalent becomes a period marker within decades. Forms of address, pronouns of politeness, and terms for money, work, gender, and rank also shift in force, so dialogue that once sounded natural can read as stiff, quaint, or unintentionally comic. Assumptions date a version as surely as vocabulary: passages softened for an earlier market’s decorum, allusions explained that readers no longer need explained, and foreign terms replaced by domestic equivalents from a vanished everyday world. Even punctuation and sentence fashion contribute, as preferences for semicolons, exclamations, and periodic structure rise and fall. The original, written in a historical form of its own language, is expected to feel old; a translation that feels old in the wrong way advertises the mediator instead of the author.

Q: Do retranslations change how a book is interpreted?

Retranslations change interpretation because interpretation in a receiving culture is built from the translated text that critics quote, teachers assign, and readers remember, and a new rendering alters that evidence at thousands of points. A restored passage can redirect argument about a character’s motives. A sterner rendering of dialogue can convert a figure long read as gentle into one read as evasive or cruel. Consistent translation of a repeated motif word can reveal a pattern of imagery that scattered renderings concealed, giving critics a structure to analyse that earlier readers could not see. Even the framing shifts: a new introduction, drawing on later scholarship, instructs readers to notice questions of class, empire, faith, or gender that an older apparatus passed over. Interpretation does not change arbitrarily, since the source text constrains what a defensible rendering can claim, but the range of readings a culture actually produces widens or narrows with the version in circulation, and retranslation is one of the few events capable of redrawing that range quickly.

Q: Why are there so many English translations of Don Quixote?

Don Quixote has accumulated English translations since the early seventeenth century, beginning with Thomas Shelton’s version of 1612, because the book sits at a rare intersection of canonical prestige, commercial durability, and permanent translational difficulty. Cervantes wrote prose that mixes high and low registers, proverbs, songs, parodies of chivalric romance, and dialogue shaped by class and region, and every generation of translators finds passages where predecessors visibly struggled, softened jokes, or flattened Sancho Panza’s speech into uniform rusticity. A work taught in universities guarantees steady sales for a distinctive edition, so publishers keep commissioning attempts, and celebrated translators keep accepting the challenge as a test of craft. Edith Grossman’s version of 2003 demonstrated that a fresh rendering could win wide critical attention centuries after the first attempts. The accumulation is therefore self reinforcing: fame attracts translators, difficulty justifies their effort, and each new version renews the argument about how comic, how archaic, and how Spanish the English Quixote should sound.

Q: What should a reader do when two translations of the same book disagree?

A reader facing disagreement between translations holds useful evidence rather than a problem requiring rescue, since disagreement marks a point where the source text permits more than one defensible rendering. The first step is locating the kind of difference involved: a factual matter such as a name, date, or object; a matter of tone, where one version sounds ironic and the other earnest; or a matter of completeness, where a sentence present in one version is absent from the other. Translators’ notes and prefaces often address exactly such cruxes, and scholarly commentary on major works discusses famous disputed passages at length. Comparing a third version can show whether one rendering stands alone or belongs to a family of interpretations. Where the difference affects plot or characterisation, a reader may reasonably choose the version whose stated method inspires greater confidence and read on, treating the variant as a marginal warning that certainty is unavailable. Disagreement remembered preserves the original’s ambiguity as part of the reading experience.

Q: Will machine translation replace literary retranslation?

Machine translation is unlikely to replace literary retranslation in the sense that matters to readers of canonical fiction, although machine tools already assist translators with drafting, terminology, and consistency checks, and their role in the workshop will keep growing. Literary translation turns on judgements that extend across an entire work: whether a motif word must remain constant for 900 pages, how far a narrator’s irony should be sharpened or left dormant, which archaism signals period and which signals parody, and where a deliberate awkwardness in the source deserves an answering awkwardness in English. Such decisions require a governing interpretation of the book, defended in prefaces, argued over by reviewers, and revised as scholarship moves. A system that produces fluent sentences without holding an interpretation can generate a serviceable rendering of informational prose, and can even produce readable fiction at the level of the paragraph, but readability at paragraph level is the standard that criticism of translation has long treated as insufficient. Publishers sell named translators precisely because a person stands behind the choices made.