RECOUP/Transcription and translation/handout on issues of translation/handout: Difference between revisions

From OER in Education
m (1 revision: Import and update from manual.recoup.educ.cam.ac.uk, CC BY-NC-SA - RECOUP)
m (1 revision: Import and update from manual.recoup.educ.cam.ac.uk, CC BY-NC-SA - RECOUP)
 
(2 intermediate revisions by the same user not shown)

Latest revision as of 12:20, 5 February 2015

What language should be used? When and how should translation into the language of analysis take place?

(For the sake of simplicity, in what follows we assume that the language of analysis will be English. If you will be working in some other language, please change to suit your local circumstances.)

In many qualitative social research projects in the South, researchers who are not native speakers of English carry out data collection in their native language, or in another local language, and record field-notes (or responses to interview questions) in either English or in another language. If the project generates transcripts from recorded interviews or focus groups, they will usually not be in English. Although best practice is for research results will eventually be translated back into a local language, much of the analysis of texts or interview responses will be carried out in English. Sometimes researchers may be working in a third language. Even when researchers work in their mother tongue with participants, everyday uses of that language can differ, (perhaps because of differences in region, class, education), and this needs to be taken into account during translation processes.

Firstly, does translation matter?

Yes it does: 'translation-related decisions have a direct impact on the validity of the research and its report' - but social researchers rarely talk about how they organise translation of materials. Why not? Perhaps because they are somewhat embarrassed by the ad hoc nature of how and when translation takes place. Some say that translation (and how it is dealt with) raises issues of how we represent what people say to us, and this should be of concern to all researchers. Others say that it is important to reflect on the potential impact of translation and interpretation on the research process, especially when disseminating research.

  1. Is it just a matter of the linguistic competence of the researcher and/or the translators?

No, there's more to it: whoever is doing the translation needs a thorough understanding of both languages and the cultures or sub-cultures within which they are being used. There is sometimes a trade-off between knowing the local 'language' (dialectic, slang, jargon etc), on one hand, and understanding the purposes of the research and the academic language (of sociology, or whatever) on the other. Depending on the research context, one person rarely has both sets of skills.

  1. Is there a 'perfect translation' which can be achieved by using appropriate methods?

No. 'Gaining conceptual equivalence or comparability of meaning' is, according to some authors, 'in absolute terms an unsolvable problem'. This is because any phrase or sentence in one language carries with it a set of assumptions, feelings and values that cannot be directly translated into another language without setting off resonances in the second language that are almost never going to be exactly equivalent. Direct lexical equivalences are sometimes misleading or not as obvious as people might immediately think. For example, to translate the Hindi word chacha ('father's younger brother') as 'uncle' is too simplistic: the English word hides the fact that in South Asia different kinds of 'uncle' have different roles—a mother's brother has very different rights and responsibilities from a father's brother, a father's sister's husband, or a mother's sister's husband—all 'uncles', in English. Where there is no exact and uncomplicated word for translation, one should aim for 'conceptual equivalence without concern for lexical comparability'. But this will inevitably mean either the 'introduction of pseudo-information' or the 'loss of information' - or both.

  1. Does it matter who does the translation?

Yes. If the researcher and the translator are the same person, the researcher will be able to draw on additional implicit information (body language, tone of voice etc) that may help them to choose the appropriate translation (e.g. if it involves irony). On the other hand, researchers may be tempted to translate in ways that support their arguments. If the researcher and the translator are different—for example, a research assistant [RA], a friend or colleague, or a professional translator—then the 'material circumstances' of the translator may make a major difference. We know that what data is collected may depend on the social position of the RA: similar issues of 'impression management' are also involved when people decide how to translate what they regard as 'sensitive' material. The translator's biography (and their sympathy or otherwise with the views being expressed, their willingness to use words that reflect slang, regional- or class-specific language, or swear-words, for example) and their relationship to the researcher may all affect how conversations or responses are translated.

  1. When is it legitimate to use participants' words, so that they 'speak for themselves'?

Some would argue, only from tape-recorded interviews (or written texts) where the researcher can be sure of exactly what was said, when the original text can be given as well as the translation, and any reader with competence in both languages can check the translation for themselves. That is an ideal case scenario: there is a continuum involving verbatim note-taking, short-term recollections of what was said, and longer-term memories. Those who use field-notes rather than tape-recorded interviews have rarely specified exactly how short extracts from fieldnotes were translated into the form they are published in a book or article.

5.How might we make the best possible job in an imperfect world?

    1. 'Back translation' (into English and then, without access to the original, back into the foreign language, comparing this with the original to see how much difference there is). But this can give a misleading sense of confidence.
    2. Consultation with several native speakers or experienced translators (but this is only possible for really significant elements of a project, such as the wording of a questionnaire—it is too expensive in time and money to be used for the translation of extracts from interviews or fieldnotes).
    3. For questionnaire items, pre-testing and asking participants what they think a question means can be helpful.

What are basic working principles for a good qualitative project?

  1. RAs must have a full understanding of the key issues in the research, so that they know why the questions have been selected and how they should be explained to people in interview situations, or followed up with prompts in focus groups or semi-structured interviews. RAs must try out a variety of ways of putting the questions (or topics) in a local language. These exercises should yield better questions and enrich our cultural understandings of the conceptual issues involved. Debriefing RAs as well as possible and keeping notes of these discussions will provide insights into the topic.
  2. RAs should keep fieldnotes (or comments from the questionnaires) in the most convenient, accurate and rapid form possible: but each option (in their mother tongue or in English) has costs as well as benefits. If they keep their fieldnotes in their mother tongue, they should be able to write faster, and retain turns of phrase used by respondents. Translation can then be done later, but both sets of material must be kept, preferably in electronic format. If they keep their fieldnotes in English, they must include phrases, or whole sentences, in the original language as well, whenever something particularly significant is said.
  3. Material used for publication must be shown to native speakers (competent in whatever local languages data has been collected) who are also social scientists. All 'direct quotations' in published work should specify its source and how (and by whom) it was translated. Aim to retain the conceptual equivalence of the original to the best of your (combined) ability, and give key phrases or words in the original language so that bilingual readers can check your translations.

Cc-by-nc-sa-narrow.png Singal, N., and Jeffery, R. (2008). Qualitative Research Skills Workshop: A Facilitator's Reference Manual, http://oer.educ.cam.ac.uk/wiki/RECOUP, Cambridge: RECOUP (Research Consortium on Educational Outcomes and Poverty, http://recoup.educ.cam.ac.uk/). CC BY-NC-SA 4.0. (original page)