Entry | Poetry translation

DUT Poëzie SPA Poética (Traducción)

Introduction

Poetry translation: a brief history

Poetry has been translated across the world since ancient times. In China, written translations of poetry date back to the 4th century BCE or earlier. In Europe, Greek poetry was being translated into Latin by at least the 1st century CE. And devotional poems were being translated between Indian languages early in the 1st millennium CE. Discussions of poetry translation principles survive, in Europe, from the Roman era – e.g. Horace’s Ars Poetica (19-17 BCE). With the upsurge of translation into and between European vernaculars in the Renaissance, there was also an upsurge of written debate about poetry translation principles, mainly in translators’ prefaces – a debate which has continued worldwide to the present day. As the academic discipline of translation studies took shape during the late 20th century, the principles, debates and models current in poetry translation influenced the discipline’s theory-base. Indeed, several founding figures of translation studies, like James Holmes and André Lefevere, wrote about poetry translation: their views are mentioned in the sections which follow.

What is poetry?

To explore what is special about translating poetry, we should first ask what is special about poetry as a genre – i.e. as a “socially defined cluster of communication acts” (Jones 2011b: 169). Let us start with an example: a Chinese poem, by Ma Zhiyuan (c. 1250–1321). This version is in simplified characters, with a pinyin transcript and an English literal crib (because Chinese does not mark number or tense unless needed, I arbitrarily use singular and present respectively):

秋思	Qiū Sī	Autumn Thought
枯藤老树昏鸦	kū téng lǎo shù hūn yā	withered vine old tree twilight crow
小桥流水人家	xiǎo qiáo liú shuǐ rén jiā	little bridge flowing water person’s house
古道西风瘦马	gǔ dào xī fēng shòu mǎ	ancient road west wind thin horse
夕阳西下	xī yáng xī xià	evening sun west sinks
断肠人在天涯	duàn cháng rén zài tiān yá	broken hearted person is-at world’s end

A genre’s communication acts may be viewed in terms of textual and extra-textual features, relating to language and social context respectively – as explored below.

Poems share certain typical textual features. None are exclusive to poetry (several are also found in advertising copy, for instance), and not all poems use all of them. But the more such features a text has, the more poem-like it appears to readers. And to listeners: poetry is both a written and a spoken genre.

Firstly, poems often highlight or play with language form. Thus sound is often important. This frequently forms patterns, like those of metre or rhyme. Qiū sī, for instance, has five lines of 6–6–6–4–6 syllables; and each has an -a end-rhyme (yā–jiā–mǎ–xià–yá), at least in modern pronunciation. Some poems incorporate word-play, grammatical ambiguity, or reactivated idioms, i.e. idioms used in both their figurative and their literal sense (Jones 2011a: 109-110). Poems may also use parallelism (repeated syntactic structures), as with the repeated dependent + head-noun pairs (“withered vine / old tree / twilight crow”, etc.) in lines 1–3 of Qiū sī.

Emily Dickinson (Source: Wikimedia Commons)

Poems typically use precise, carefully-chosen lexis. They may contain implicit or associative meanings – some culture-specific, and others (e.g. Qiū sī’s association of autumn and evening with melancholy) shared across cultures. Meanings may even be ambiguous, and/or open to different interpretations by different readers (including different translators). Poets may use original or unusual collocations, metaphors, etc., as in Emily Dickinson’s lines describing a storm: “The Doom’s electric Moccasin / The very instant passed – / On a strange Mob of panting Trees” (an “electric moccasin” was a vivid folk-name for a water snake).

Poems may also inventively bend the conventions of orthography or grammar, as in E E Cummings’ poem that begins: “anyone lived in a pretty how town / (with up so floating many bells down)”. Or of style, as with Philip Larkin’s startlingly colloquial lines “They fuck you up, your mum and dad. / They may not mean to, but they do”.

Moreover, poetry tends to use these rich and varied devices in a particularly small space – typically a couple of pages or minutes at most.

Turning now to extra-textual features, poetic communication begins with signals alerting receivers what genre to expect. Thus the website title The Poetry Archive, or Selected Poems on a book cover, primes audiences to expect poetry rather than travel directions, say. As does a gathering advertised as a poetry reading. Poetry also has a specific page layout – typically, short lines. Or, in speech, a certain tone of voice.

As with all literature, communication via poetry is not meant to be informative, but to be appreciated in its own right. And with poetry, the emotional and/or intellectual experience is especially intense. Socially, poetry is a highly-valued art form – probably because of poetry’s linguistic virtuosity plus this intensity of experience.

Attitudes and approaches

So what does this mean for translating? I start by looking at social attitudes towards poetry translating, and translators’ approaches (preferences for source-target relations). As a source of examples, here is Autumn Thoughts, my own unpublished translation of Qiū sī:

Withered vines, old trees, gloaming crows;

Little bridge, flowing stream, people’s homes;

Ancient road, west wind, bony horse;

In the west, sunset glows.

World’s end. Bleak at heart, on one roams.

Lost in translation?

Because poetry’s textual features often rely on language sound and nuance, it is popularly regarded as a difficult genre to translate. This has led to a discourse of loss: for instance, Lefevere’s (1975: 99) view that “most translations […] are unsatisfactory renderings of the source [poem]”. It also means, however, that poetry translators are allowed plenty of leeway to deviate from source semantics (as when I added “glows” to Line 4 of Autumn Thoughts, to rhyme with Line 1’s “crows”).

In extra-textual terms, translated poetry shares the features mentioned above. However, its translators tend to have more visibility than translators even in other literary genres. Their name often appears on book covers, or at the end of individually-cited poems (in an anthology, for instance). And at public readings, they often read alongside the source poet. But audiences may also look at translated poetry more critically than non-translated work (Wechsler 1998, in Boase-Beier 2004: 25) – viewing innovative language-use as translationese, for instance.

Approaches

Poetry translation approaches may be arranged along a literal↔free spectrum, as in the following figure:

crib / interlinear	meta-poem			adaptation / version / imitation	poem inspired by	pseudo translation
	metaphrase	re-creative translation / paraphrase
		source-oriented	target-oriented

<<< more literal freer >>>

Approaches to poetry translation (Source: Author)

At the most literal extreme, a crib is a semantically-close guide to the source which does not claim to be a target-culture poem – like the initial English guide to Qiū sī above. Cribs are sometimes printed alongside or beneath the source in editions of classic poetry aimed at non-native readers of the source language. As a word-for-word interlinear, often with explanatory notes, a crib may be used by a target-language poet as a basis for creating a poetic translation.

Most often, however, translators aim to produce a “meta-poem”: a text that “match[es] the original” closely enough to “be considered a translation” while also being “considered a [target-culture] poem” (Holmes 1988: 24-25,50). Within this, one can distinguish two broad approaches:

The translation is formatted or recited as a poem while remaining fairly literal – Dryden’s “metaphrase” (1680: 145-146). This implies seeing the source poem’s “pure perfect content” as residing in its wording and imagery (Goethe 1814). Thus, for UK poet Ted Hughes, minimally polishing his co-translators’ interlinears without regard to source poetic form was “part of a wider literary tradition of seeking an authentic sounding voice” (Bergin 2013: ii).
“Re-creative” translating (Jones 2011b: 172) aims to convey the source’s overall poetic package (semantics, poet’s intent, imagery, allusion and poetics) rather than prioritising surface wording. Thus, in Autumn Thoughts, Line 2’s “Little bridge” is a close semantic equivalent, but I added “on one roams” in Line 5 to reflect the source’s rhyme structure. This resembles Dryden’s “paraphrase”, where the poet’s “words are not so strictly followed as his sense, and that too [may] be amplified, but not altered” (Dryden 1680: 145-146). Among both poet translators (who regard themselves as primarily poets) and linguist translators (who regard themselves as primarily translators), this is the most common approach (Jones 2011a; Lobejón Santos & Jones 2020).

Further along the spectrum are adaptations, versions or imitations. For Dryden, imitations not only “vary from the words and sense”, but also “forsake them both as [they see] occasion” (Dryden 1680: 145-146). They are generally produced by target-language poets. And, nowadays at least, they are rarely labelled “translations”. They may, however, often be credited largely to the rewriter rather than the source poet, especially if the rewriter is a well-known figure.

Beyond this, there are poems inspired by other-language poems. And further still, pseudo-translations: original poetry written as if it were translated. An English-language example is Christopher Reid’s Katerina Brac – an invented work modelled on translations of Eastern European poetry popular in Western countries during the late Cold-War period (see Societies, cultures, politics below).

Boundaries between these approaches, of course, are fuzzy rather than clear-cut (pseudo-translation excepted, perhaps). And translators may vary their approach within one poem, depending on the challenge being tackled.

Debates in the re-creative approach

Even within a re-creative approach, debates remain. One is how to reconcile the meta-poem’s twin aims, source-poem loyalty and target-poem effectiveness, if – as often happens – they conflict.

Then source-oriented translators and scholars lean more towards source-poem loyalty. This need not mean literalness. Boase-Beier (2004: 29) argues that reflecting the source’s style is crucial, because style encodes the source writer’s point of view, and thus makes the poem literary. Venuti (1995: 23-24) advocates conveying “whatever features of the foreign text abuse or resist dominant cultural values” – idiosyncratic or subversive language, say – though reflecting the act of resistance is more important than cloning the specific source style.

For target-oriented translators and scholars like Folkart (2007: 6-7), how source poets structure their “raw material” is “artistically next to irrelevant”, and a “focus on [source-poem] patternings […] can lead to artistically inadequate results”. Instead, she urges translators to “[write] out your own understanding of the text” (Folkart 2007: 8).

In practice, many re-creative translators try to balance both orientations – though, as Cowper (1791: 184) pointed out, there is a “very delicate line” between the risks of being “servile” versus “hardly […] true” to the source. Holmes (1988: 54) sees this as finding “counterparts” or “matchings” to source items/devices/styles which perform similar functions in the target language/culture. Or finding good-fit analogies: for instance, in Mike Stocks’ translations of sonnets by the 19th-century Rome poet Giuseppe Gioachino Belli, Stocks’ colloquial British English arguably has a similarly subversive tone to Belli’s working-class “Romanesco” original .

Some poems (typically, those using fixed form, unusual diction, wordplays, allusions, intertextual references and/or culture-specific content) are harder to translate than others (typically, poems lacking such features). With the former, it is often impossible to convey every source-poem feature. Then translators have to compromise (Jones 1989: 197) – for instance, between recreating a striking image versus a word-play. Here, translators often have a personal “hierarchy of correspondences” (e.g. prioritising semantics above sound, or vice versa), which is why different translators can make radically different translations of the same poem (Holmes 1988: 86).

For re-creative translators, creativity is a key tool. Creativity scholars define this as tackling problems by exploring solutions which are both “novel” (i.e. original) and “appropriate” (Sternberg & Lubart 1999: 3).. For poetry translators, this means considering solutions which deviate semantically or stylistically from the source while still conveying its poetic intent and imagery (Jones 2011a: 140-142). Translators usually consider creative changes because translating literally or cloning the style is impossible or would harm target-poem quality, but sometimes because creative solutions simply seem more effective (Jones 2011a: 140-142). One may distinguish here between:

Creative “adjustments” within the same semantic or stylistic space as the source item: e.g. changing “water” (水: Qiū sī, Line 2) to “stream”. All re-creative translators use these (Jones 2011a: 140-142).
Creative “transformations” away from relevant source semantic/stylistic spaces, e.g. adding “glows” to Autumn Thoughts (Line 4) to rhyme with Line 1’s “crows”. Translators usually consider transformations only after adjustments have proved unable to solve the translation problem (here, how to rhyme two lines). Some translators, however, avoid transformations altogether, fearing betrayal of source content (Jones 2011a: 140-142).

Recreating formal source-poem patterning (e.g. metre, and especially rhyme) is notoriously challenging. Translators need advanced versification skills. It is highly time-consuming. The pattern may have no target-language counterpart, or may be harder to recreate: English, for example, unlike classical Chinese, does not easily allow lines of six monosyllabic content words. Or the pattern may give different cultural signals: thus some (e.g. Sorrell 2000: 74) claim that rhyme and rhythm in English translation sound hackneyed or comical. And because such patterns are language-specific (“yā/crow” and “jiā/house” rhyme in Chinese but not in English, say), recreating them usually requires creative transformations – which, as mentioned, some translators view as betrayal (e.g. Lefevere 1975: 56,59). Therefore many translators abandon or dilute source-poem patterning – e.g. reshaping rhymed poetry as free verse, or replacing full rhymes with half/near/imperfect rhymes.

Other translators, by contrast, see formal patterning as conveying crucial poetic effects or signalling the source poem’s cultural background. Hence they prioritise recreating it in translation. Here, Holmes (1988: 25-27) identifies three main approaches:

“mimetic”: imitating the source form;
using an “analogical” form with a similar cultural function – e.g. translating a Spanish endecasílabo (hendecasyllable, i.e. 11-syllable line) into English iambic pentameter;
“organic”: choosing the form that best fits source-poem content (e.g. Qiū sī’s three 2-syllable feet per line becoming Autumn Thoughts’ three 2-stress feet).

Similar challenges arise with poetic styles (e.g. colloquial, regional, archaic or idiosyncratic) – be these the poet’s personal style, and/or different voices within the poem, and/or the source literary culture’s stylistic conventions (e.g. classical Chinese poetry’s highly condensed, content-word-only diction). These again prompt a similar range of responses by translators, ranging from neutralisationnormalisation (converting into standard modern target language) to mimetic, analogical or organic recreation. Few, however, explicitly advocate neutralisation.

Norms

Translators work within a framework of poetic and translational norms. These change over time and space. Until recent decades, for instance, most English translations of older works were lightly archaised; but nowadays, archaised English translations (even of modern poems using deliberately archaic voices – Jones 2000: 79) tend to be judged negatively, at least by literary critics. Over the last half-century, however, translation norms worldwide appear to have converged to an ethic of reliably representing the source poem’s style and content – though with all the variation described in the previous section. Nevertheless, “versions/‌imitations” by target-language poets retain a strong niche.

Translating processes

Knowledge about poetry translators’ working processes derives largely from translators’ written reports, plus recent real-time recordings of solo and collaborative translating (Jones 2011a; Lobejón Santos & Jones 2020). These processes are examined below.

Before translating

Poetry translation projects may be commissioned by a source poet (as described in Project teams below), a publisher or an editor. Often, however, translators select which poet to translate, or which poems to translate from a wider corpus. Selection is typically driven by the translator’s liking for the poems, or (especially with poet translators) by a sense of affinity with the source poet. Translatability and good fit with target-culture poetic norms can also influence poem choice, but perhaps less frequently.

Before translating, Born (1993: 61) recommends building up background knowledge by reading the source poet’s biography and as much of their work as possible. Most translators pre-read the source poem for meaning and to analyse its poetic form (Jones 2011a: 120,124-125).

Translating

Translators divide the complex task of translating a poem into a set of nesting subtasks (Jones 2011a: 116-117). Individual moves (e.g. checking a word in a dictionary) combine to make a micro-sequence. These usually tackle a unit of text, prosody, etc. small enough to be held in mind at one time: e.g. translating 在天涯 (“is-at world’s end”: Qiū sī, Line 5) as “world’s end”. In a macro-sequence, translators tackle a medium-length text segment (e.g. a line) by combining several micro-sequences. Thus Qiū sī’s Line 5 was translated with one macro-sequence: this comprised two micro-sequences tackling the underlying phrases “断肠人 / broken-hearted person” and “在天涯 / is-at world’s end” respectively, interrupted by a micro-sequence looking for rhyme-pairs with another line (generating “on one roams” to rhyme with Line 2’s “homes”). Macro-sequences combine to form a run-through of the whole poem. A working session lasts from the moment one starts translating till one finishes for the day or takes a long break; it may involve one or several runs-through of one or more poems.

Zapisano

Lines 1–4 of “Zapisano nad jednim stihom”, by Ivan V. Lalić: fair copy of my 1st- plus 2nd-draft English translation plus notes (Source: Author)

Early written drafts are often handwritten (see example above), enabling notes, synonyms and brainstormed solutions to be consulted later. For some translators, the first draft is very literal. In the figure above, I wrote this in black, with notes on the opposite page (e.g. about the Spanish poem quoted in the epigraph). Many sessions, ideally interspersed by at least a few days’ so-called drawer time (when the draft is put aside), are needed to shape the draft into a viable poem; gradually, the focus shifts from source-poem equivalence to target-poem effectiveness, sound and “flow” (Born 1993: 61; Jones 1989; Bly 1983). Thus the pink annotations and deletions show how, in the 2nd draft, I reworked the literal into iambic pentameters with an ABAB rhyme-scheme. (The version submitted to publishers, months later, has hardly changed: “The dilute acids of the rains, my God, / Are eating summer stubbornly away; / The sun a blotch, cloud-bound but for the odd / Patch of bluish lichen, day after day”.) This makes poetry vastly more time-consuming to translate than most other genres (Jones 2011a: 118-119). Most time is spent on issues of lexis and imagery, showing the importance of getting the “right” wording and conveying poetic images; poetic form takes up less time, even with rhymed and rhythmed target poems (Jones 2011a: 129,158).

Hearing and reading translated poetry

Translated poetry is brought to audiences in various ways. Spoken events may happen at festivals, poetry evenings or book launches, before a physical audience and/or live-streamed. Because most source poems are by living poets, the source poet and translator often read in tandem, as mentioned.

Written translations are published in poetry or literary journals, single-poet books or anthologies. Anthologies may feature multiple or single source cultures (or languages). Anthologies are important in shaping a canon of “representative” source-culture poetry within the target culture, though what is included and omitted also inevitably reflects the anthologists’ unconscious biases and conscious selection decisions – under-representing or deliberately highlighting women poets or urban themes, for instance (Barnaby 2002: 86,88). Poems may also be part of a multi-genre work – a journal featuring prose literature plus poetry, for instance.

Most publications print the target text only. Some, however, are bilingual, with source and target poems on facing pages. While allowing source-language readers to experience the source text, bilingual editions could also encourage them to criticise translators’ choices.

Publications are not only paper-based. Poems are short and therefore web-friendly texts: hence there is a lively online scene of journals, blogs, etc. featuring translated poetry. Informal web publications, however, may omit the translator’s name, and may publish without copyright-holders’ permission.

Broadly speaking, readers read translated poetry similarly to original target-language poetry (Boase-Beier 2020: 413). Translated poetry, however, often also has a preface/‌afterword and background notes, and (as mentioned above) may even be presented bilingually. This, Boase-Beier (2020: 413) argues, can encourage readers of translated poetry to read “critically”: for instance, relating the target poem to the source poem, but also to contemporary poetry more widely. Lefevere (1975: 87), by contrast, dislikes such paratextual material, condemning the “completely unjustifiable splitting up of the source text” into the target poem plus information in footnotes. A counter-argument to Lefevere’s view would be that reading poetry is particularly dependent on extra-textual context, including cultural assumptions and references – and that footnotes, prefaces, etc. give readers of foreign poetry the context that they would otherwise lack.

Translators

Here we look at poetry translators as people: their cognitive skills, emotions and identities.

Skills

Re-creative poetry translators arguably need four overlapping expertises:

source-poem reading skills and literary knowledge
target-language poetry-writing skills, ideally in a variety of styles and forms
source-target language transfer skills
the strategic ability to balance the demands of source-poem loyalty and target-poem effectiveness

The crucial second requirement means that print-published poetry translators are virtually always target-language native writers. This usually means they are also source-language non-native readers (except for a few balanced bilinguals, who are equally at home in the source and target poetries). In informal web publications (such as blogs), however, source-language native writers may upload their own translations of poems they admire.

Motivation and emotion

The motivation to translate poetry is rarely financial (pay is often low, as explained in Communities, fields and habitus below). Motivation is almost always intrinsic: the belief that a certain poetry or poet needs a wider audience, plus the intellectual and aesthetic pleasure of translating itself. Moreover, publishers’ or editors’ deadlines are usually generous or flexible, in recognition of poetry translation’s time-consuming nature. This lack of external stress has a flipside, however. Working for two years of one’s free time, say, to translate a poet not known in the target culture can feel a long and lonely task: hence translators also need qualities of persistence and self-belief, though source-culture-native enthusiasts of the poet can bolster the translator’s motivation (“Derek”, in Jones 2011a: 88). This high personal investment in the project, plus the fact that most reviews of translated poetry pass judgement on the translation’s quality, means that poetry translators are especially sensitive to feeling validated or “swiped off” by reviewers’ positive or negative verdicts respectively (“Derek”, in Jones 2011a: 99).

Loyalty and identity

Nord views translator ethics as determined not so much by the textual concept of “fidelity”, but by the interpersonal concept of “loyalty” toward the translator’s “partners”: source writer, translation commissioner, and target readers (2001: 185). A poetry translator’s prime sense of loyalty is typically to the source poet (except perhaps in some projects commissioned by an editor, where the prime loyalty may be, or may also be, to the editor). This often extends to a sense of loyalty towards the wider source culture, or towards a grouping within that culture (those promoting a non-sectarian model of Bosnian society during and after the 1990s war, for instance – Jones 2011a: 81). Whether translators are source-culture diaspora/heritage actors or target-culture natives appears to make little difference (Jones 2011a: 76). The motivation to communicate an admired source work to target readers, however, also means that translators can regard themselves as intercultural ambassadors (Jones 2000).

The borderline between the identities of linguist translator and poet translator is fuzzy: many who translate directly from foreign languages also write their own poetry, and vice versa. But poet translators often find that their translating and original poetry writing cross-fertilise each other. UK poet Ted Hughes, for example, collaboratively translated various European poets. His translations bear the imprint of his own rough-hewn poetic style, but the literalist style of his translations (see Approaches above) also influenced his own poetry (Bergin 2013).

A poet’s voice, as expressed in their stylistic choices, is arguably a key expression of their poetic identity. In reshaping a work for target readers, there is a fine line between finding a fitting target-language analogy for the source poet’s voice, and appropriating or distorting that voice. Was Edward Fitzgerald’s 1859 English Rubáiyát, for example, simply an “imitative” response to Omar Khayyam’s medieval Persian original, or did it peddle Fitzgerald’s orientalist Victorian visions of a mysterious, hedonistic East?

More recent discussions have addressed issues of translator diversity. When white writer Marieke Lucas Rijneveld was commissioned to translate black US poet Amanda Gorman’s presidential inauguration poem “The Hill We Climb” into Dutch, some commentators asked why a black Dutch performance poet had not been approached instead. It would be dangerously reductionist to insist that translators should translate only poets who match their own identity. But this riposte would have more traction if the ethnicity and gender profile of poetry translators, and of key actors in the cultural/arts sector (like translation commissioners), reflected that of wider society – and if a young black woman were therefore as likely as a middle-aged white man to be commissioned to translate Homer, say. This is unfortunately not yet the case.

Working with others

Translators may work solo, where one person reads the source poem and writes the target poem, or collaboratively, where the translating is shared among two or more people. Even solo translators do not work alone, however. How poetry translators work with others is explored in the sub-sections below.

Project teams

Translators, collaborative or solo, belong to a wider project team – along with, for instance, an editor and/or publisher, and maybe a graphic/web designer. Also, because solo poetry translators need particularly high source-language reading and target-genre writing expertise (see Skills above), they often use input from text helpers: source-culture/‌poetry informants, other translators, and/or target-language readers (Jones 2011a: 40; Flynn 2004: 276).

The source poet is frequently a key actor in the team. S/he may commission the translation, often by approaching the translator directly. This may involve offering payment on completion of the translation – or, more usually, proposing that the translator finds a publisher, with the source poet and translator sharing royalties. The source poet is arguably the ideal source-poem informant. However, if s/he knows the target language, s/he may advocate (or occasionally insist on) replacing the translator’s idiomatic renderings with literal but poetically clumsier equivalents (Phillips 2001: 30).

Poetry translators typically resent such intervention because it breaches the norm that translators are responsible for target-poem wording. More widely, it breaches the assumption by other team members that translators “know what should be done” (Flynn 2004: 280) in any project, i.e. operate autonomously with respect to translating approach and decisions. Similarly, copy-editors’ suggestions after submission of the manuscript are usually minimal. Some teams – especially with anthologies – consist of an editor, who assembles the overall text, plus multiple translators. Even here, editors rarely suggest changing the submitted translations – except perhaps with novice translators who may benefit from expert feedback.

Publishers typically have to apply for translation permissions, and sometimes for translation grants from source-country bodies (Sampson 2001: 81). Commissioning translated poetry is higher-risk than publishing original poetry, because publishers can rarely read the source language and hence cannot judge the source work’s quality (Sampson 2001: 81), unless its poet is already known in the target culture. Hence publishers often trust translators to recommend new poets (Sapiro 2019: 37). This means that translators typically have a mix of self-motivated “personal initiative” projects (works they are enthusiastic about, which they pitch to publishers), and commissions (Flynn 2004: 275). With more “experienced and reputable” translators, personal-initiative projects tend to outweigh commissions (Flynn 2004: 276).

Collaborative translating

In collaborative poetry translation, key participants are:

the source poet
and/or a target-language poet
and/or a language advisor / bridge translator, who has source-poem reading and source→target language-transfer skills.

Conventionally, the target-language poet writes the target poem, while the others assist him or her (Raffel 1988: 129). If the language advisor / bridge translator is a target-language native writer, s/he may occasionally wish to co-write the target poem: while some target-language poets welcome this, others do not. Sometimes only the target-language poet is credited as the translator in publications, concealing the other participants’ work. Conversely, Calleja and Collins (2019) view the language advisor / bridge translator as the only true translator. I would argue, however, that all participants in a collaborative group should be called “(co-)translators”.

Collaborative poetry translation happens for various reasons. In less-translated language pairs, it circumvents the bottleneck where there are too few solo translators to take on all poets who need translating. It can also enable established target-language poets to translate other-language poets with whom they feel an affinity, and early-career poets to engage actively with work beyond their own poetic heritage.

There are many possible collaborative configurations. For example:

A language advisor produces an interlinear, which a target-language poet reworks into a poem. This works best if the interlinear is the basis for discussions between both participants throughout the drafting process.
A source poet works with a target-language poet. They may even translate each other’s work (e.g. Octavio Paz and Charles Tomlinson’s Airborn / Hijos del aire: 1981).
If source and target-language poets do not know each other’s language well, a language advisor may join them, forming a so-called poettrio.
Translation workshops involve several target-language poets plus a source poet and/or a language advisor – and perhaps even solo translators who work in the relevant language pair.
Sometimes, two or more poetry translators with similar skill-sets may collaborate – e.g. the Greek–English translators Edmund Keeley and Philip Sherrard, both of whom knew Greek well and were published English-language poets.

Social context

This section outlines the wider social context in which poetry translators work.

Communities, fields and habitus

Beyond the project team, translators work within a wider poetry-translation community. In Bourdieu’s terms, this is a field. Fields are “historically constituted areas of activity with their specific institutions and laws of functioning”, whose members have a shared habitus, i.e. an agreed set of working norms, practices and attitudes (1990, in Inghilleri 2005: 134-135). It is debatable whether this community’s features, as outlined below, qualify it for the status of profession (Jones 2011a: 187-188). A better term might be vocation: “a socially regulated activity of high social worth, where high-expertise players can gain cultural and symbolic capital, but not necessarily high pay” (Jones 2011a: 188).

Indeed, one feature of the poetry translators’ habitus is that they rarely earn a living wage. In the UK, say, an 80-poem volume could take a few months’ full-time equivalent to translate, but would earn just a few hundred pounds in royalties – the amount a technical/‌commercial translator would earn in less than a week. Hence there are few full-time poetry translators: poetry tends to be translated by academics, poets, or other enthusiasts in their free time, for the intrinsic reasons mentioned under Motivation and emotion above. Indeed, Sapiro (2019: 24) describes translating poetry as “a disinterested act by the translator, since it is often undertaken for free, as a labour of love”. A few source countries, however – almost all with non-globalised languages plus healthy economies – have institutions which subsidise the translation of their poetry (e.g. the Dutch Foundation for Literature). Translators’ remuneration is then comparable to prose translation.

Another feature that helps define poetry translation as a vocation rather than a profession is that no formal qualifications or training are required to join the field. Accreditation criteria are informal. With novice translators, gatekeepers such as editors and publishers simply consider whether the translators’ output is acceptable; with more experienced translators, gatekeepers look at the evidence of previous translations plus word-of-mouth recommendation by other editors/etc. Indeed, few formal training opportunities exist. Many translation MA programmes offer literary translation modules, but these are often optional, typically with one session on poetry. Literary organisations, however, offer a growing number of mentorships in literary translation, often with poetry translators as mentors.

In compensation, poetry translators enjoy high public prestige – because of poetry’s high social value, but almost certainly also because poetry is regarded as a uniquely difficult translation genre, which requires creative rewriting (see Lost in translation? above). Hence translators of book-length editions of poetry often write a preface or afterword introducing the source poet, but also explaining their own translation decisions – a practice which is less frequent in other genres.

Some poetry translators (“dedicated experts”, like James Holmes in late 20th-century Dutch→English poetry translation: Jones 2018) translate so prolifically over decades that they become central actors in their field. They translate, and therefore become networked, with many source poets. They may also boost the careers of other translators by acting as mentors, translation workshop leaders, and commissioning editors for multi-translator anthologies.

Poetry occupies only a small share of the literary publishing market: of literary translations into French from 1985 to 2002, for instance, the percentage of poetry works ranged between 2% and 15%, from English and Spanish respectively (Sapiro & Bokobza 2008: Graphiques 8.1, 8.3). Poetry publishing is dominated by small, poetry-specialist publishing houses (rather than by the large companies which dominate the fiction market). These are often led by one or two poetry enthusiasts, which can make them readier to follow hunches or take risks – for instance, considering unknown foreign poets pitched by translators, as mentioned above. Profit margins are low, however: hence the low remuneration to translators, especially as royalties are usually split between source poet and translator. And print-runs are typically small (Sapiro 2019: 32). Nevertheless, poetry translations confer high cultural prestige on publishers (as on translators): they are “an investment that yields symbolic capital” (Sapiro 2019: 24).

Literary and poetry-specific journals, both print and online, are also key players in the poetry-publishing ecosystem. Book publishers, for instance, will often only consider new poets whose work has already appeared in journals.

Giuseppe Gioachino Belli (Source: Wikimedia Commons)

In poetry, as in other genres, translations form a minority of overall work published, though the actual percentage varies enormously between target cultures. Nevertheless, translated poetry can exert an important influence on target-culture poetry. Davis (2002: 82-83), for instance, claims that the “apparently secondary and cautious activity [of translation] has initiated virtually every major rhetorical innovation in English poetry”. And for Chinese, Mao (1922: 204) noted that “the translation of foreign poetry can be a means of revitalizing our own poetry”.

Despite accounting for just a small proportion of translating activity, poetry translation has also strongly influenced translation theory world-wide, regardless of how well poetry-based theories apply to other genres. Venuti’s argument that translators should use foreignising rather than domesticating strategies (1995: cf. Debates in the re-creative approach above), for instance, arguably derives – at least in part – from his experiences as a poetry translator: for most poetry translators, recreating source-poem style is a high priority. Poetry translation’s theoretical clout has arisen partly because of poetry’s cultural prestige, but largely because, as noted above, poetry translators have, throughout history, written prefaces explaining their practice (Boase-Beier 2020: 411) – unlike translators in many other genres (one notable exception being translators of scripture).

Seminal poetry translations can also shape the target language, especially if it is renewing itself or in flux. Thus Robert Garioch’s translations of Giuseppe Gioachino Belli’s 19th-century Roman-dialect sonnets into Scots were instrumental not just in the 20th-century revival of literary Scots, but also in the language’s gradual standardisation (Whyte 2004: 195-196).

Societies, cultures, politics

Teams and fields, like those of poetry translation and publishing, operate in an even wider social context. In Jones’s (2011a: 27-28,45-50) theoretical model of poetry translation, these are the looser and wider “third-order” networks of people, discourses and ideas such as “US society”, “Spanish culture” or “the Japanese people”.

Because poetry, as mentioned, is a highly-prized form of cultural expression, source-culture individuals (e.g. Serbian nationalist bloggers – Jones 2010: 238-242) or institutions may promote poetry translation, or extol translated poetry, as a means of promoting their culture or nation. The purpose of such promotion may be simply to spread knowledge of the source poetry internationally, though motivation is often also political: soft power or even propaganda. The translator’s own sense of cultural ambassadorship is therefore often supported by wider source-culture networks. Such networks may be sub-groups within the overarching source culture, as noted in Loyalty and identity above. They may even act in opposition to other sub-groups, such as those in and beyond Bosnia opposing a model of Bosnian society based on inter-ethnic difference (Jones 2011a: 81). Networks may also be transnational, as this example shows.

Wider networks that commission or support poetry translation may also be largely or wholly target-culture-based. These, too, may be influenced by underlying social trends or ideologies. A target country’s community of poets may feel a need “to seek poetry outside their own tradition”, as with US poets “during the political upheaval of the 1960s” (Rosenthal 2011: 221). Or translation may be encouraged by images and ideologies, political or otherwise, current in the target culture. Thus the admiration for Eastern European poetry in the West during the Cold War was arguably fuelled in part by the image of the source poet as heroic opposer of Nazism and then Soviet Communism (Sampson 2001: 83) – or, less starkly, by the idea that a more valid poetics had been forged in the crucible of harsh life experiences and the need to circumvent literary censorship.

Once a translation is published, translators have no control over which social groups use it, or for what ends. For instance, I was once appalled to see a translation of mine (of Vyacheslav Kupriyanov’s late-Soviet Russian poem about the futility of “the paper lie”) quoted in a Holocaust denier’s web post.

Research potential

Some areas of poetry translation have been thoroughly researched – how translators translate certain iconic poets, for instance. Other areas have been initially mapped, but in few studies: poetry translators’ real-time working processes, for instance. Here there is a need for studies that look at how translators deal with a bigger range of poem-types and language pairs. Similarly, of all the collaborative translation modes in common use, only the poettrio has been intensively researched: hence we lack data on how other groupings work, such as poet-poet pairs, or translation workshops. There are also recent debates in translator practice, such as whether source poets and translators should share certain aspects of personal identity – and if so, which aspects, and under which social and personal circumstances. These too could do with exploring further.

Methodologically, we need to go beyond translators’ after-the-event reports and case studies of single source poems or poets, valuable though these have been for mapping the poetry-translation field. Comparative surveys of multiple cases (be they poets, translators or poems) are more likely to generate generalisable findings, especially if underpinned by relevant theory. Concordance and corpus work is strikingly underused for poetry translation research, and is therefore worth exploring. Real-time data on solo and collaborative translating can be gathered by recordings of think-alouds and collaborative co-translators’ discussions respectively. Finally, readers’ views are often neglected in poetry translation: these can be gathered by questionnaires, interviews, think-alouds, and reader notes (Low 2002).