Indirect Translation

This article concerns indirect translation (ITr), understood broadly as translation of translation, and has the aim of facilitating systematic research on this long-standing, widespread yet underexplored phenomenon. The article thus provides an overview of some of the main patterns in ITr practice and research and explores suggestions for related future studies. The overview follows the ‘Five W’s and One H’ approach. The what question concerns terminological and conceptual issues related to ITr and explores the relevance of systematic studies on ITr. The who question considers the profile of agents involved in ITr processesas well as the profile of ITr researchers. The where question relates to the spatial dimension of ITr as well as to the geographic spread of ITr research. The when question concerns the time coordinates of ITr practice as well as the diachronic evolution of ITr studies. The why questions looks into the motivations for ITr and into the historical neglect in the Translation Studies discipline. Finally, the how question considers selected details of ITr processes as well as the methods used in identifying most probable mediating texts and languages. The article ends with a brief consideration of prospects for research on ITr training. The what question concerns terminological and conceptual issues related to ITr and explores the relevance of systematic studies on ITr. The who question considers the profile of agents involved in ITr processes as well as the profile of ITr researchers. The where question relates to the spatial dimension of ITr as well as to the geographic spread of ITr research. The when question concerns the time coordinates of ITr practice as well as the diachronic evolution of ITr studies. The why questions looks into the motivations for ITr and into the historical neglect in the Translation Studies discipline. Finally, the how question considers selected details of ITr processes as well as the methods used in identifying most probable mediating texts and languages. The article ends with a brief consideration of prospects for research on ITr training.


INTRODUCTION
When translation scholars talk about indirect translations they are usually referring to translations of translations (Gambier 1994).The English language terminology used to label this practice varies immensely.Designations in current use include compilative, double, eclectic, intermediate, mediated, pivot, relay, second-hand, secondary translation.Some scholars understand indirect translation in a narrow sense: as a translation via a third, typically more well-known, language.In these narrow terms, an indirect translation is understood as being done in two steps: a first translation from language A into language B, and then a second translation by a different translator from language B into language C. Others, however, define indirect translation more broadly.They recognize that the term is broad enough to include a great diversity of interlinguistic* and intermodal translation* processes.Possible examples of indirect translation subsets include: • retranslation*: the translation of a text which has been translated into the same language at least once before (A-B-B) • back-translation: a translated document that is translated back into the source language (A-B-A) • support translation: when only isolated fragments of previous translations are used as a part of a research conducted by a translator • the novelization of a film or video-game that is itself based on a book: when a written text is translated into image and sound and from there again into a written text.Indirect translation has been common throughout history (e.g., translations of the Bible, I Ching, The Arabian Nights) and continues to exist as a phenomenon worldwide, in various settings, platforms, modes and media.For example, streamed media not originally in English is generally first translated into an English template which is used to produce subtitles in other languages.In international news production, media outlets often rewrite wire dispatches that already include translations.With the increase in the number of language combinations, multilingual institutions (e.g. the European Parliament) are using a system of relay languages, where documents are first translated into a major language (e.g.English, French, German) and then from that language into many others.The same happens with neural machine translation* engines (such as Google Translate), which rely on machine learning and hence the availability of large parallel data for training.Since such data are not available for most language pairs, present-day technologies often need to use pivoting approaches as a work-around to translate between languages with scarce resources for the development of language technology.
There is no evidence of reduced relevance of indirect translation in the foreseeable future; rather the opposite.For example, globalization often prompts situations where there is a sudden need to translate between language pairs for which there are not enough competent direct translators.This is when translation often goes indirect: translators translating specifically with a further translation in mind or from an already translated text.
This entry provides an overview of indirect translation research.For an overview of indirect translation practice, see the entry on relay translation*.The focus here will be on written texts.Spoken texts are dealt with in the entry on relay interpreting*.

CHALLENGES AND DEVELOPMENTS
Indirect translation research has been characterised by terminological and conceptual confusion, as well as by methodological constraints.Lately, efforts have been made to organize the metalanguage and to lift indirect translation from a theoretical basis.Assis Rosa et al (2017, 117) identify a number of patterns in English-language terminology.They show that 'indirect translation' has gained ground over competing terms, although 'pivot' is still preferred in publications on audiovisual and machine translation, whereas 'relay translation' is privileged in publications on Chinese translation traditions.
Relatively systematic typologies of indirect translation have been proposed (Assis Rosa et al 2017; Washbourne 2013).While increasingly used, these typologies have also been criticised for adopting too complex a labelling system and for being inadequate for certain text types (e.g.journalistic texts).
Some hypotheses about indirect literary translation have been developed and tested.Ringmar (2007, 11) has suggested that translators working indirectly take more liberties with the mediating text than they would with the ultimate source text.Špirk (2014) posits that more changes are introduced during the transition from the ultimate source text to the mediating text than in the passage from the mediating text to the ultimate target text.In turn, Hadley's (2017) "concatenation effect hypothesis" makes the claims that indirect translations tend to omit elements of the text that are identifiable with the source culture, and also describe themselves as something other than translations.Several studies have put these hypotheses to the test without reaching a unanimous conclusion.It also remains to be seen to what extent these claims hold true for non-literary translations.
The identification of the most plausible mediating texts and languages is one of the trickiest elements of studying indirect translation.Challenges hindering easy identification typically derive from the fact that there is a plethora of indirect translation situations and that indirectness is often unacknowledged or explicitly hidden (due to the non-standard, often controversial status of this practice).As a result, information on the mediating languages and texts is typically unavailable or unreliable.Diachronic reconstructions of the origin of specific translations often prove to be merely probabilistic.Research on journalistic translation and translation history also shows that from a certain point onwards the identification of an ultimate source text may be unfeasible.supplementary material (archival documents, bibliographies, databases, historiographies).The first approach has recently yielded particularly promising insights, tapping into methodological advances from Forensic Linguistics and Genetic Criticism (e.g., using technologies developed in order to detect plagiarism, (cf.Marin-Lacarta 2017); putting an indirect translation through computational source language detection, ( cf.Ivaska 2019)).One noteworthy methodological conclusion seems to be that, irrespective of the type of text, product-oriented methods (e.g.frame analysis, discourse analysis, corpus-based studies, recourse to stylometry) often prove insufficient and need to be combined with process-and/or participant-oriented approaches (e.g.eye-tracking, keystroke logging, interviews, non-participant observations).

PAST AND CURRENT TOPICS AND CONTRIBUTIONS
For many years indirect translation has been marginalized by translation trainers, practitioners and researchers.This is partly because translation reflection and education have long been anchored in the standard Western model of translation, which prioritises translating from the original.Moreover, mainstream Translation Studies do not typically deal with situations where translation occurs between the so-called low-diffusion, which are the very situations where indirect translation is most common (see minority languages and translation*).This may also explain why systematic studies on indirect translation are a recent development.
For quite some time now, most studies have adopted a historical approach that rarely looks beyond the 20th century (Pięta 2017).Traditionally, the emphasis has been on a small range of language pairs and geographic areas (mostly in Asia, Europe and South America), one medium (the printed book) and one text type (literature).Much has been said and written about the negative consequences of indirect translation, adding to the common perception that the practice is a threat to: • translation quality; (the reasoning is that mistakes in the pivot versions are necessarily replicated in further translations) • the jobs and ethics of translators (as direct translation between less-known languages tends to be more costly than indirect translation via a more popular language, for which there are more translators available; translating from a major pivot language may therefore mean taking translation work away from people who are already marginalized because of the minor language they use) • the balance between languages and cultures (with English acting as a main pivot language worldwide there is a danger of economic, cultural and epistemological homogenization, whereby consumer preferences are anglicized and English mediating is preferred to direct translation from more peripheral languages).
However, in recent years indirect translation research has experienced a wave of changes.
Indirect translation is going digital and global: there is an increasing number of studies focusing on digital domains and using digital technologies to identify and study indirect translation (Ivaska 2019;Oziemblewska and Szarkowska submitted).The range of examined text types, media, regions and language combinations is clearly expanding (see, e.g., van Rooyen 2018 on news translation in South Africa).Most studies deal with translation situations where English is the main pivot language but increasing attention is being paid to regional pivot languages (Japanese in Asia, Portuguese in Lusophone Africa, Russian in the former Soviet Republic, Swedish in Nordic Europe, etc.).
Dedicated studies are starting to shift the focus to the advantages of this practice, highlighting its potential as an empowering aid that allows speakers of low-diffusion languages Some methodological guidelines have been developed with regard to the indirect translation of literature (Assis Rosa et al 2017; Marin-Lacarta 2017; Ringmar 2007).It remains to be seen to what extent these can be adapted to the translation of other text types.These recommendations boil down to a combination of two approaches: analysing and comparing the texts involved (ultimate sourcemediatingtarget text) and extracting information from to access important information and to be heard on the global stage or as a pedagogical resource that can help teachers make the most of linguistic diversity in translation classrooms(Torres Simón et al. forthcoming;Van Rooyen 2018).It thus seems that high-quality indirect translation may play a crucial role in realizing global aspirations of sustainable development, where there is a true equality and the same access to opportunities exists, independent of the language spoken.Although still peripheral to Translation Studies, research on indirect translation has already produced insights that are relevant to the discipline as a whole.For instance, while foregrounding the intricate threefold nature of many translation processes (source-mediatingtarget text), research on indirect translation make us rethink some of the foundational concepts, such as equivalence*, original* and translation*.Since indirect translation studies ask questions about the ethical, technological and practical implications of power struggles among languages and cultures, these studies have opened up useful entry points for interdisciplinary dialogues with other fields of inquiry interested in these issues.These studies can also add new perspectives to ongoing debates about inaccessibility, inequality, and other global concerns that hinder sustainable growth and are thus of keen interest to scientific communities, governments, companies, and society at large.At the same time, the need to train future translators to translate from translation and with a particular translation in mind has been recognized; different teaching and training approaches are beginning to be explored and tools and aids for translators are starting to be developed (Torres Simón et al. forthcoming; Oziemblewska and Szarkowska submitted).Researchers are no longer only describing and analyzing how people in the past translated indirectly; they are also developing research-informed recommendations on how to produce high-quality indirect translation in the years to come.All these different aspects point to the diminishing marginality of indirect translation practice.They also indicate that the issue of indirect translation is a burning one for translation research and for training and language technology providers.