Keith S.
Taber
Faculty of Education, University of Cambridge, UK. E-mail: kst24@cam.ac.uk
My concern here, however, is not with such stylistic details, but the more fundamental problem for those authors who may not be fluent in English, of preparing their manuscripts for publication. Researchers may lack such fluency on the quite reasonable grounds that they live and work in a context where English is not a dominant language. There are two sets of issues here: one set is pertinent to all authors in this situation, and the other arises in the kinds of research where it is appropriate, and indeed expected, to quote in research reports textual material where the sources are in a language other than English. It is this latter issue I particularly wish to highlight here.
Much advanced learning is based on communication through language – reading, and especially (as it is more inherently and directly interactive) talk. The development of scientific concepts is widely considered to be grounded in our direct experience of the world (Taber, 2014a) – that is, we make sense of taught scientific concepts in terms of our experience of the natural world, or in terms of other acquired concepts that themselves (perhaps through several such indirect stages) are ground in such experience (Lakoff and Johnson, 1980). Despite this need for ultimate grounding to make sense of abstract ideas, the learning process in formal educational contexts (e.g., school, university) is usually heavily dependent on symbolic communication. This may often include diagrams and graphs, but primarily occurs through verbal language. Textbooks often include various graphics to complement the text, but the choice of term (viz., textbook) is not usually a misnomer. So far, so good – language supports our abstract thinking and helps us understand each other.
It may be considered that each of us has learnt and developed a unique personal language including associations that are idiosyncratic, so that each speaker of (for example) English to some extent speaks their own slightly distinct version of English. The distinctions may be subtle, so any resulting breakdowns of communication can usually be overcome without moving outside the largely shared lexicon of the ‘common’ language (Bruner, 1987). Before I moved to another part of the country to attend University I had only ever heard the term ‘baps’ used as a slightly vulgar anatomical term, but the initial confusion at being offered baps in a university café was very quickly overcome by realising that these were what I would have called rolls. The language context of the use of the word (e.g., having a cheese salad bap with a cup of tea) allowed ready sense-making.
There are clearly much more substantial differences between the personal languages of speakers of different national languages such as Spanish, Chinese, Russian, and so forth. The well-known Sapir-Whorf hypothesis of linguistic relativity (Hussein, 2012) suggests that the language a person has available to some extent shapes how that person can perceive and understand the world. A ‘strong’ version of this conjecture framed as language determining (rather than influencing) thought is contentious, but the less extreme claim that the language a person has available effects their thinking to some degree is generally accepted.
Language provides affordances for thinking, and also puts a limit on how we can think in verbal terms. It has been suggested that bilingualism (or multilingualism) can offer alternative sets of thinking resources, such that switching languages can sometimes facilitate thinking not otherwise readily available in a first language (Pavlenko, 2006). It is clearly the case that phrases sometimes become adopted into use in other language contexts rather than translated. So, French phrases such as ‘c’est la vie’ and ‘vive la différence’ are widely used by English speakers who are not fluent French speakers – as if it is sensed that the French terms can offer emphasis and tone that would be missing in translation. Within research methods the French term bricolage has come to be used to describe an approach to research design for which there is no convenient English alternative. It is claimed that “bricoleurs move beyond the blinds of particular disciplines and peer through a conceptual window to a new world of research and knowledge production” (Kincheloe, 2005, p. 323). Arguably, the sense of bricolage needs a translation that encompasses notions of bespoke construction, of curation, and of ‘making-do’.
I found a website that offered some examples of this [https://www.mimicmethod.com/24-ridiculous-foreign-language-idioms-pronunciations/ accessed 20th May 2018]. Without having the idioms explained it was not clear what was supposed to be implied by foreign phrases that are literally translated as “to have mustard going up their nose”, “there is no standing on one leg”, “drawing a snake with feet”, “to treat you with fish in your face” or “take your little horse away from the rain”. There seems potential for an entertaining parlour game here: players offer their best guess explanations of how such translated terms are used in their native language context. Most of these allusions made sense once I saw the non-literal explanations (although not all: “you are very elbow”, for example, remained obscure), but the direct translations may not be transparent to the English reader. This works both ways of course: non-native readers of English may be nonplussed to find reports that it was raining cats and dogs, and would be barking up the wrong tree if they assumed that it was impact with the ground, rather than curiosity, which killed the cat.
If the intended meanings of idioms may not be clear in translation, it is equally possible that a literal translation of a mundane phrase may actually have associations in the target language that were not present in the source. The US automobile manufacturer General Motors found that a model it called the Nova initially had very disappointing sales in Latin America, until they changed the name of the car. Although Nova suggests ‘new’ to most English speakers, ‘no va’ in Spanish refers to something that ‘won’t go’, which is hardly ideal for a motor vehicle (Vernon and Miller, 2001). Another car company, Ford, had a similar issue with their Pinto brand in Brazil where the term literally meant chick in Portuguese, but was also slang for having small male genitals (Groves et al., 2003). I have recently seen television advertisements in England for a car model (from a Romanian manufacturer) that has been called the Duster – a word that means a piece of cloth used to do household cleaning chores in my lexicon. It seems that just as intended associations may be lost in translation, so unintended associations may be found inadvertently.
They may also be found by design. In translating artistic works, translators avoid direct translations, and instead seek renditions that better catch the original meaning (which is, after all, what translation is about). This may happen with scholarly works as well. The translator deliberately finds a relevant phrase or term in the target language which does not exist in the source language, so arguably sometimes the translator may improve on a text by taking advantage of a resource provided by the target language that was not available in the original language. An example might be one of the English translations of Husserl's introductory exposition of phenomenology, ‘Ideen zu Einer Reinen Phänomenologie und Phänomenologischen Philosophie Erstes Buch: Allgemeine Einführung in die Reine Phänomenologie’ (Husserl, 1932/2012, commonly known in English as ‘Ideas’), where a German term that literally translated might be the conception of the ‘naive’ man was rendered as the view of the man in the street (Moran, 1932/2012), an idiom that means the thinking of a typical person one might come across going about everyday business.
Classic works may over time attract multiple translations, where the different translators make their own decisions about how best to represent (their understanding of) the author's intended meaning into, for instance, English. A well-known example is the translation of the only full length book that educationalist/psychologist Lev Vygotsky published in his lifetime. This book is normally known in English as ‘Thought and Language’ (Vygotsky, 1934/1986), but an alternative translation of the original Russian was published as ‘Thinking and Speech’ (Vygotsky, 1934/1962). Speech refers to something much more specific than language.
Translating poetry or song lyrics introduces further complications – as the metre (or, if you are reading this in the US, the meter) is an essential feature of the work, and should be retained in a translation: which simply may not be possible when using the most direct equivalent terms in the target language. It is therefore understandable that translators of such works are often given prominent credit, sometimes almost at the level of being seen as a co-author of the translated work (Zeller, 2000). In passing, it is worth noting that in English the metre is the same as the foot when discussing poetry, but not when measuring length and distance. Words (metre, bap, pants) take their meanings in specific contexts, so translators have to consider each word or phrase within its textual context.
These days people might prefer to make use of a tool such as ‘Google Translate’, and more often than not, this will provide a good deal of the original sense of a technical text. Google Translate is a powerful tool, that anyone with internet access can employ for quick translations. It is also possible to demonstrate the difficulty of using an algorithm to make translations by entering a phrase, and then translating, sequentially, through a number of languages. If perfect translation were possible the outcome should translate back into English unchanged in meaning from the original text. Of course, this does not usually occur. In translating a sentence from the text above through a sequence of European languages, finishing back in English, I discovered that
“Indeed, it works in two ways: readers who do not read in English may not be incomplete to say they are dogs and cats and measuring the tree inaccurate if they are The idea that it affects a blanket, irrespective of the cat specification killed.”
Clearly, very little of the original sense is retained here, when vernacular text is subjected to a sequence of algorithmic translations.
This may just suggest another amusing parlour game, perhaps suitable for keeping children occupied during long journeys: what is the minimum number of sequential translations before original meaning is lost; or, alternatively, how many sequential translations through different languages can you make before losing the sense of a given phrase? However, it also has something in common with one useful, if hardly foolproof, technique used to check a translation. This is to have the translated text translated back to the original language to check the meaning has not changed. I tested this by seeing what Google Translate made of my text in Chinese (see Table 1), and found that my suggestion that English idioms could perplex non-native readers was converted to an accusation of animal cruelty.
In the UK, the English system (but not the Scottish system) divides the compulsory school years into ‘KS1-4’ – terminology that does not mean anything to most overseas readers (the K has nothing to do with the K in ‘K-16’ used in the US context). ‘Public schools’ in England are prestigious private (independent of the state, fee-paying) schools, when in most countries ‘public schools’ are state schools. Within the English state sector, in recent years, there have been comprehensive schools, grammar schools, community schools/colleges, village colleges, city technology colleges, subject specialist schools, academies, free schools, and university technical colleges – whilst many 16–19 year olds attend further education colleges or sixth form colleges instead of secondary schools (many of which are now called colleges anyway). Arguably, even people living in England do not fully appreciate the subtle distinctions here.
In the same way, how many readers outside the US fully appreciate what a charter school is; or how the district, state, and federal, authorities are involved in determining school curriculum? Readers of articles in the US may not realise that entrance to a chemistry degree programme in the UK will usually be most heavily determined by student grades in specialist examinations in chemistry and related subjects, whilst readers in the UK might not appreciate that in the US much greater weight is normally given to standardised tests that assess (largely by multiple choice) core scholarly aptitudes and attainment: reasoning, language skills, numeracy, etc. Some other countries put more weight on internal assessments made within and by the candidates’ own schools. ‘University entrance examinations’ would mean something substantially different in these different contexts.
The content of the chemistry curriculum will vary from place to place, as will the age at which chemistry may appear as a discrete subject in a student's career. Even references to ‘science lessons’ may mean something rather different in different places as in some countries ‘science’ may be used as a covering term including other STEM (science, technology, engineering, and mathematics) subjects. Norms about moving between year groups (or being held back), the sorting of students into groups (mixed-ability, banded/streamed, set, or with gifted learners creamed-off), whether chemistry is taught by specialist chemistry teachers (rather than more generalist science teachers) at different grade levels or in different kinds of schools, and so forth, may need to be explained for readers around the world to make sense of a particular research report. The extent to which a degree student specialising in chemistry will focus on chemistry courses or will study and be assessed in courses in other subjects as part of their programme varies considerably.
Assumptions about teacher education may not be shared in different national contexts: chemistry teachers take multi-year teaching degree courses in some systems, but are trained as post-graduates (already holding chemistry or cognate degrees) elsewhere. The amount of time spent in classrooms during training, when this usually occurs, the role of the trainee in the class, whether there is a research-training component, who arranges the placement, and the role of host teachers in planning and assessing the training experience, can all vary considerably from place to place. The alert reader will have noticed that these are all matters of local culture and custom, rather than language, but it is not just that language is not always directly translatable when there are no entirely equivalent terms, but also because in different places the things and situations that invite labelling with words (the components of the local ontology) themselves diverge.
This adds complications in reporting chemistry education research that are not found in most areas of chemistry. Aluminum may always be assumed to be entirely equivalent to aluminium, but who is to say if a chemistry degree classed 2(i) from an English university can be equated to a chemistry degree from a US University with a GPA of 3.5? Phosphorus is phosphorus (and has the same properties) wherever it is found, even if it is called elsewhere, and its physical and chemical properties will not change if in time it suffers the fate already met by sulphur (i.e., sulfur) and is renamed fosforus (more akin to its existing name in several other languages). By contrast, elementary school might stop at various ages around the world. Distillation works the same anywhere – but progression to senior high school courses in chemistry may require passing examinations in some countries but not others. These issues arise even when researchers are working in English (e.g., comparing Australia, the United States, England, Ireland, etc.) but if authors are translating their work into English from their own language they may not be sure whether or not the object they are referring to is best labelled as a ‘middle school’ (or ‘college course’, or ‘trainee teacher placement’, or ‘national standard’, etc.), or something else, in English.
Data has to be analysed, and again it is usually appropriate for data to be analysed in the language in which it is collected, by someone who is (or people who are) fluent in that language. Someone not fluent in a language is very likely to miss nuance, if not misinterpret responses. Indeed, in transcribing interviews or classroom talk it can sometimes be difficult to clearly hear precisely what is being said even when one is working in a first language.
Moreover, transcripts are constructions that necessarily compromise between providing a precise transcription of the sounds made and interpreting the communication being offered. When two people have a conversation, such as a research interview, clear communication depends upon interpretation of what is being said and not just hearing the sounds. Unlike in written communication, where drafting and editing is possible, conversation usually occurs in ‘real time’ and speech often includes mumbling, incomplete words, shifts between tense mid-sentence, slips of the tongue, and sentences left incomplete (or completed by emphasis or facial expression or gesture rather than actual utterance of the words implied).
The process of completing a research transcript involves decisions about how much tidying up needs to be done to best represent this. An accurate transcription of the sounds an interviewee (or even an interviewer) actually made may be difficult to read. However, too much tidying up risks over-interpreting what was intended – and producing a transcription where what was a transcriber's best guess is presented on the page as definitive data. Ideally, the interview occurs in the first language of the interviewer and interviewee; the transcription is carried out by the interviewer (who brings some knowledge of the context of utterances and may recall non-verbal cues); and analysis occurs in the same language by the interviewer/transcriber, with access to the recordings to check unclear points.
If data includes samples of student writing or talk, this may comprise discussion of the chemistry in a manner quite distinct from an expert's technical use of the language of the subject, and indeed student conceptions (whether the focus of the research or not) are likely to be different from the canonical concepts of chemistry (Taber, 2014b). Thus, this material may not be as open to ready translation into another language as formal technical reports in chemistry that have been honed by professional chemists according to the conventions of the subject. Translating this material may therefore have additional challenges. In the case of interview transcripts (or transcripts of classroom talk and the like) a written transcription of speech that may be far from complete and grammatical in the source language will need to be translated into a different language in a way that best represents the original communication.
It is not being suggested here that what has to be done necessarily offers insurmountable barriers – but rather that there are additional challenges in preparing research reports in another language, and that these challenges may be more extreme when this includes material such as students’ written work and talk. These points may seem self-evident, but experience at Chemistry Education Research and Practice suggests this is not always so.
This is certainly not impossible. There are networks of international schools that teach and assess in English in many countries where English is not the local language. There are universities in some countries which teach in English rather than a national language as it is considered that this will offer greater opportunities for their graduates. There are schools in contexts where students speak a diversity of local languages where English is used as a lingua franca (or perhaps a lingua anglais), and there are schools in parts of Asia where science is commonly taught in English as an immersive language (i.e., giving a real-life context for students to learn a second language as they need to use it to study their science). Collecting data in English may well be sensible in these circumstances, but this needs to be explained, so that editors and reviewers – and, if peer review goes well, readers of the published paper – can appreciate this.
More commonly, however, what is being reported is not the text that was located or constructed in the research context, but a translation. Authors have reasonably assumed that as they are writing-up their study in English, they should also translate any materials they wish to present to illustrate their claims. Whilst this is sensible, what is not appropriate is to present translated material in English without acknowledging and discussing this. Translation is part of the methodology of undertaking and reporting research, and adds a further stage of processing where quality assurance is needed. Yet submissions often make no mention of the language of instruction in which students were taught, or the language in which data was collected, but simply present everything in English assuming that is what is needed. This would not matter if translation was a straightforward, algorithmic, process. Yet, as discussed above, translation introduces an additional step between the raw data and its interpretation – whether that interpretation is by an analyst working with translated material, or by reader of a research paper being presented with translation of the material that was analysed in the study.
(1) If a research paper only includes material in English, and there is no discussion of translation having occurred, a reader should be able to safely assume that the material was collected in English and no translation has taken place. Therefore, if the work reports from a context where English is not the local language or is unlikely to be considered by readers as the usual language of instruction, this should be clarified for a reader (e.g., explaining that the work reports from an international school where the language of instruction is English).
(2) Therefore, whenever material included in a paper is a translation, this should be reported. It should be explicit which material has been translated, and which language(s) it has been translated from. It should be pointed out that translation may not always be a perfect process and nuances of the original may be lost in translation.
(3) Ideally the original text that has been translated should be included (i.e., in the original language), juxtaposed against the translation, or reported in an appendix (Taber, 2016) as seems most appropriate. Where both original and translated material is available, then readers of the original language will be able to consider the data in its original form, and bilingual readers will be able to evaluate the translation offered. Sharing the original text provides a means of quality assurance of translations made.
(4) The authors should report who translated the material, and their qualifications for doing so. Normally a translation should be made by someone with a high level of fluency in both languages, as well as close knowledge of the research context and the specific study.
(5) There should be some checks on the translation (in the way that data analysis usually has built in consistency checks or is tested for reliability), and these should be reported. For example, back translation can be used, and any inconsistencies subject to scrutiny; or there may be multiple translators producing material that is triangulated; or – at the very least – translations should be checked by another person fluent in both languages. By reporting on how translations were checked, as well as produced, the authors give those readers who cannot read the original language (which is often most of them) good grounds for having confidence in the translation.
When these steps are taken, editors, reviewers, and readers of the journal, are able to see that authors appreciate that translation requires care, and that the researchers have indeed taken due care in reporting translated material. An academic article makes knowledge claims that are supported by arguments based on the collection and interpretation of evidence, and readers have to be persuaded that all stages in the process have been carried out to a high standard. The claims in an academic article are only as convincing as the weakest link in the chain of argument: whether that be framing of the project, research design, sampling, data collection procedures, analytical framework applied, analysis of data, interpretation of findings, implications drawn, or the processes of composing the report. So, when the authors are initially working in another language, this includes, inter alia, the processes of translating the argument of a research report into English, and the translation of any evidence being presented to support the argument made. Insufficient care or skills here can undermine the research paper as much as substandard work at any other point. Good practice in translation, and in reporting it, is part of good research.
This journal is © The Royal Society of Chemistry 2018 |