On Idioms, Intertextuality, Puddings, and Quantum Physics

(all of them in simultaneous, please)

hese notes were triggered by an anecdote recently reported by trainee interpreters at the European Court of Justice (ECJ). The anecdote has to do with the way interpreters handle idiomatic expressions and more generally what is known as "culture-bound" items, or realia. I will suggest some insights from Translation Studies and draw attention to the practice of translation at the European Court of Justice, in an attempt to question received ideas as to how to translate idioms in the booth. Finally, I shall touch on the role of language work in multilingual organizations.

Here is the anecdote: at a hearing of a case at the ECJ, spring 2004: novice colleagues, after studying the usual pile of case documents, are practicing in the dummy Polish booth, coping with a case referred by the German jurisdiction for labour issues, the Bundesarbeitsgericht. The language of the case is German, with one party intervening in English. Suddenly one of the parties refers in English to the need to wait and see the results of a particular development, because, he argues, "the proof of the pudding is in the eating." The Polish colleague renders the idiomatic expression in plain language, neutralizing it, because, as she reports later, "we don't say that in Polish."

So far so good, inasmuch as she gets the main point across to her listeners. Yet, as colleagues will have witnessed, speakers like to re-cycle imagery: later on, another participant refers to what the previous speaker has said, takes up the "pudding" image again and develops it, half-jokingly, for what must have seemed an endless time to the other colleague on the mike at the moment, who cannot refer back to anything similar in the output of her booth-mate.

In scholarly terms, this is a textbook example of intertextuality, i.e., roughly, of a text referring to or re-using an item taken from another text. And of course this characterizes oral as much as written communication. The notion may seem trivial but has consequences for those who study and teach interpreting, since it implies that, strictly speaking, you should not evaluate an interpreting performance only on the basis of a single stretch of "text," e.g. an individual intervention at a conference, a plenary speech, a plea in court and so on. By doing so you would loose sight of all intertextual references to other texts in the same event, whereas for those listening to the original those references were visible and part of the overall meaning associated with the event. In other words, the unit of measure when studying interpreting should be the whole event. The point was aptly made among others by Franz Pöchhacker in his seminal Simultandolmetschen als komplexes Handeln (1994), where it is suggested that we should see the whole interpreted event as a hypertext, a network of cross-references between individual texts.

Thus armed with a solid notion of intertextuality, we are perhaps ready for a change of perspective on the anecdote. But first let me note that such incidents are by no means rare. In fact, a similar one (referred to in the admittedly obscure title of these notes) occurred at the same hearing, this time involving two languages and a slight misunderstanding. One of the parties suggested that interpreting the existing case-law as requested by the counterpart would amount to a "quantum leap" in EU legal practice, meaning, I suppose, a very radical change. The image itself displays an interesting overlapping of meanings: it refers both to a complex scientific notion from quantum physics and to a popular sci-fi TV series, where the main character travels in time thanks to a machine allowing "quantum leaps." Predictably, the image did not go unnoticed and was taken up again by a German-speaking participant in a lengthy and nebulous digression. The comment started with the inevitable "and since quantum physics has been mentioned...," and developed on the original scientific meaning of "quantum leap." Whether the reference to physics was accurate, I am in no position to judge. What is certain is that this expression caused the same problems in the Italian booth (we had not used the original "quantum leap" image the first time) as the "pudding" one did to the Polish colleagues.

The anecdote puzzling the Polish colleagues is thus by no means isolated. Going back to the possible solutions, the point here is not whether in Polish you do or do not say "the proof of the pudding is in the eating" or whether the literal "salto quantico" makes any sense to Italian listeners (in the algid setting of the ECJ, on top of that). Asking only these questions leads us to translate in a way that is certainly satisfactory, but fails to take into account the whole event as a hypertext. The Polish colleague omitting the reference to puddings did in fact provide a fairly adequate translation of that individual utterance. The same goes for the solution adopted by the Italian booth in the "quantum leap" case. The problem is that these solutions deprive listeners, and colleagues taking the next turn in the booth, of references that will later be used intertextually within the event. In plain words, solutions that omit the original image in idiomatic expressions may create difficulties some time later, when references to that very image surface up again, in the same or another language.

Before going on, let me briefly deal with a legitimate objection: of course we are talking about a minor problem compared, for example, to the interpreter not knowing what the Bundesarbeitsgericht or a preliminary ruling by the ECJ are. The fact that listeners of the Polish booth would not have had the opportunity to smile at the second reference to puddings hardly has any impact on the successful conduct of an ECJ hearing. Point taken. Yet, the fact that the anecdote was discussed at some length by colleagues after the hearing shows that minor things like these do challenge our skills, so we had better be equipped with conceptual tools for that challenge. Moreover, apprentice interpreters can easily learn what a preliminary ruling is, whereas coming up with a solution for the hearing's quantum leaps and puddings involves some more reflection. In turn, this can cast some light on the wider issue of our role as interpreters and translators.

So, assuming that this is a problem, what is the solution, then? The first thing to note is that idioms and cultural references like the ones we are dealing with can be accommodated in a broader definition of realia, i.e. lexical items designating elements specific to a particular culture. See for example the following definition, originally by the Bulgarian scholars Vlahov and Florin, and quoted by B. Osimo in an online course on translation theory (http://www.logos.it/lang/transl_en.html):

A couple of examples from Vlahov and Florin's list of "political and social" realia can clarify the kind of translation problems we are talking about: think of how you would "translate" into your own language county, canton, princedom, bidonville, arrondissement, suk, promenade, corso, prospekt, agora, storting, kneset, duma, czar, doge, vizier, alcalde, ayatollah, satrap, Bürgermeister, Union Jack, fleur de lis, and so on.

Set against this definition of realia, both our examples appear to designate objects or concepts typical of a given culture: traditional British culture—British cuisine?—in the case of the pudding, American sci-fi in the case of the quantum leap. Neither of them has "exact matches" (whatever this means) in our target languages. Both phrases carry some "local colour," as is witnessed by the fact that other speakers felt compelled to take them up again and develop on them, perhaps in order to enrich their own speech with some humour. That's why I argue that phrases like these can be considered as a specific type of realia.

As is often the case in legal reasoning, and occasionally in scholarly argument on translation too, once you agree on how to categorize something, you already have a clue as to how to handle it. So how do you go about translating realia? The standard answer is that you choose from the weaponry of translation techniques ranging from transcription, through loan and calque, to various ways of explaining what the item is. One end of the range is more controversial: the search for a cultural or functional equivalent to substitute for the original. Basically, techniques mentioned in scholarly discussion of realia are still the ones exemplified decades ago by Vinay and Darbelnet in their epoch-making Stilistique comparée du français et de l'anglais of 1958 (but see Molina & Hurtado Albir 2002 for an update on translation techniques).

Even more than with single-word realia, when dealing with set phrases like the ones in our examples, language professionals are keen to search for a cultural equivalent, as is witnessed for example by the many multilingual lists of idioms circulating in interpreter-training institutions. One important point to be made is that translators will to some extent mix the available techniques, partly depending on factors such as text type and function. A good example of this are bilingual in-flight magazines: here, substitution of the realia by a (supposed) cultural equivalent is frequent, perhaps because the two versions are meant to be read independently (that is, unless the readers are frequently traveling interpreters suffering from a serious professional deformation). This would imply that articles are to be entertaining in each language separately. The translation consequently focuses on naturalness and ease of reading, and is highly idiomatic, sometimes even more so than the original. On occasions, an entirely different realia or idiom is deemed equivalent to the original, on other occasions the realia is neutralized and "de-localized" into something less culturally marked; all these techniques can be seen in the following extracts from a recent issue of a German-English in-flight magazine. The examples are from an interview with a German actress who discusses her humble origins:

Once we realize that there are in fact several ways in which a translator can handle idioms and cultural items, even in one and the same text, we can perhaps go back (hopefully not back to square one, itself a nice idiom) to our interpreting practice at the ECJ. When participants talk of puddings and quantum leaps, automatically neutralizing the cultural references and replacing them with something that sounds more natural in our TL is not the best option in terms of intertextual cohesion; plainly, we leave our colleague helpless against subsequent occurrences of the same phrase.

We need alternatives, then. If adaptation or substitution do not work, why not try importing idioms in our own translation? At first sight this may seem impracticable. The objection is obvious: listeners would simply not understand if they suddenly heard their interpreter talk about puddings and quantum leaps that happen to be unfamiliar things in their language. Point taken. But then, and here is where written translation hopefully comes to rescue my argument, how is it that the very case documents we study for hours before ECJ hearings are full of examples of source language realia imported in the target text without any attempt at adaptation? And how does it work?

Let us have a look at it in more detail. ECJ translators, and legal translators in general, seem to rely heavily on non-translation when it comes to a specific category of cultural items, the ones designating institutions in the source language.

In the language professions we may be tempted to believe that languages do lead separate lives and never shall meet except in our talented multilingual brains or nomadic intercultural lives; yet the obvious consequence of the above approach is that in ECJ texts languages happily mix with each other. I have in front of me the cover page of a document that uses French, abbreviations and even pictograms to tell me that it is an English translation: a little black square appears near the word traduction and then EN, for English, follows; the title is in French and German, with no italicizing of the foreign words: Demande de décision préjudicielle introduite par le Bundesarbeitsgericht (for reference, it is case C-284/02, but any ECJ document will do). The cover page also tells us in French that the language of the whole proceedings is German. The cover of another order for referral (C-245/02) claims, in French, to be an Italian translation, and the language landscape is even more diverse: it bears the French title Demande de décision préjudicielle introduite par le "Korkein Oikeus" (this time in inverted commas, but why?) and the parties' names are in English and Czech. The French cover also tells me that the language of the case is Finnish, thereby telling me at least where to look for a Korkein Oikeus if I happen not to know what it is.

This is of course the result of in-house standards with regard to document formats, for ease of archiving and retrieval, Yet the interesting thing is that the very appearance of ECJ translated texts betrays their nature as translations. Interpreters, who are often told that their output should sound like an original speech, should perhaps start wondering to what extent this is possible and advisable, especially in a setting where written texts by no means pretend to be originals, unless they really are.

A closer look can help clarify the way ECJ translators handle institutional references, and possibly give us some guidance as to how to treat instances of the "pudding" type. Of course this implies agreeing that from the point of view of translation there is not much difference between an item like Bundesarbeitsgericht and "the proof of the pudding is in the eating." Looking back at the definition of realia mentioned above, both items can be said to be "peculiar to one nation," to "carry some local colour" and not to have "direct matches" in another language. The main difference I can see is that the name of the German highest labour court designates an institution materially existing in the social life of Germany, whereas the idiom is no less an institution, but one that only leads a de-materialized existence within the system of the English language and British culture.

So how do ECJ translating colleagues render institutional names? Pending a more structured research, the prevailing technique consists in keeping the original denomination, especially in headings and operative parts of the text, and, very often, inserting a translation or explanation near the original item, at least on one of its first occurrences in the text.

Other aspects of the ECJ's translators' approach to institutional realia are worth noting: for example, the policy prevailing for institutional names also applies to less material institutions such as the title and indeed the acronyms designating national laws: the following excerpts show instances of this in the English and Italian translations of the German text:

All this may seem far from our own practice in the booth. Yet I am sure that a similar approach to institutional names is used by interpreters. In the hearing of this case they will probably have used, in their respective languages, the German word Bundesarbeitsgericht, possibly with an approximate pronunciation and perhaps followed by a translation or explanation of what this is. Some interpreters may even have used the German title of the law, Mutterschutzgesetz (I, for one, did)—although few will have dared to utter the acronym MuSchG, pace our translating colleagues.

Now, if you have followed my argument to this point you will probably expect the next step: what is the difference between uttering a German word like Bundesarbeitsgericht, or Mutterschutzgesetz, in your Polish or Italian output in the booth, and then explaining to what social or legal institution it refers, and uttering the English word pudding when used in an idiomatic expression—if time allows, of course? Or even, and this takes my argument a bit further, what prevents us from uttering the whole idiomatic expression, either in the original language, or in a very literal translation, and then explaining what it means? This would give you or your colleague enough material to develop on, should the idiom be used again in the course of the proceedings. Besides, it would give your listeners the benefit of a compact lesson in cross-cultural awareness, to the extent that they would learn, from you, that the British only trust their puddings when they eat them.

So why are we so reluctant, when interpreting, to say things that sound slightly odd—slightly foreign—to our listeners? Of course it is difficult to have the meaning of an idiom on the tip of your tongue, and find the time to explain it; yet this must not be more difficult than decoding, say, a Tarifvertrag zur Anpassung des Tarifsrechts. My impression is that there must be other reasons why interpreters are reluctant to approach idioms in the same way as translators approach legal realia.

The first reason has to do with the interpreter's role as both they themselves and society perceive it. If you approach the idiom as a realia, and you explain it instead of replacing with an equivalent one (but who decides that two idioms are equivalent? Dictionaries? Your interpreting teacher? Do you really trust them?) in your target language, you may seem to be getting across to your readers two messages: a message from the speaker, i.e. the meaning of the idiomatic expression, and a message about the speaker and his cultural background: in this case, this would be the fact that he chose to put his point in idiomatic terms, trying to strike a lighter note perhaps, and that his culture uses that particular image. More plainly, one could say you are adding something that was not there in the original. The question is, of course, if as an interpreter you are entitled to get the latter message across and to add this little something. In my opinion you certainly are. You are entitled to add the explanation of the idiom because your listeners need it to have a full picture of who is talking to them. And you are entitled to get the second message across for the very simple reason that the message about the speaker and his language use was already there in the original, for listeners who share the same cultural background. In sum, by treating the proof of the pudding like you would treat the Bundesarbeitsgericht you are not really adding anything to the original message as perceived by source language listeners—and you are safeguarding your ability to reestablish intertextual references in your output.

The second reason why this is not done systematically has to do with what is expected of language work in multilingual organizations. To understand this, we need another theoretical notion, that of documentary versus instrumental translation. The distinction, which builds upon ages of scholarly debate on for instance free vs. literal translation (or formal vs. dynamic equivalence, or overt vs. covert translation), is originally by Christiane Nord (see 1999); it is also reported clearly in an online article by Andrew Chesterman (see 2000):

The distinction is important because it makes clear the contradictory requirements imposed on language work in multilingual organizations. Most of the training you receive as an interpreter, and of the deontological discourse within the profession, seem to point at an instrumental translation. Your output as an interpreter has to be as "deverbalized" as possible from the original wording, as convincing as possible as a speech in the target language in its own right. Ask seasoned interpreters which remark flatters them most, and many will probably answer "when my customers tell me that I did not even sound like I was interpreting." This is certainly justified when it comes to the usability of your output (fluency, register, intonation, pauses). However, the need for a usable, instrumental output must not be misconstrued: it cannot imply that an interpreter's output must not contain anything that sounds odd, or foreign, such as an unexpected image taken from an idiomatic expression in the source language. Because in this way we would be stuck in good old untranslatability, a ghost often haunting scholars, every time an idiom like the pudding one is used, with, as we have seen, a loss of intertextual references and ultimately a failure in communication (although admittedly a minor one).

This looks all the more strange if we go back to ECJ case documents. As we have seen, they announce their status as a translation on the first page, they are full of word and phrases in different languages with translations in brackets; they are interspersed with bracketed dots (...) indicating that items from the original have been left out (by the translator) as irrelevant, and at times they even carry the original page numbering at the margin, for ease of reference. In short, the translation of ECJ case documents are certainly more on the documentary side. So why should language work in a single institution like the ECJ follow two contradictory norms, a documentary one in writing and an instrumental one orally? The paradox is even more obvious if you think that the readership of translated documents and the audience of interpreted pleas at the ECJ is often the same group of legal experts, who are used to functioning in a multilingual environment. So why should they be deceived into believing that what they are getting in their headphone is an original speech in the target language?

With reference to his experience in bilingual Canada, Brian Mossop (see 1990) has aptly argued that having translations look completely idiomatic in the target language, neutralizing all foreign elements, implies masking important aspects of the social and political reality in which the translation came into being. Going back to our anecdotes, is there any reason to mask puddings and quantum leaps? Is there any reason why our listeners should not perceive, through our translation, that the original speaker draw on resources from his traditional British cultural background?

Possibly. But this French word is a handy way to introduce a conclusion. If they have followed me until now, readers will forgive the loan. In these notes I hope to have shown

• that neutralizing phraseology, or replacing idioms with natural-sounding equivalents in the target language deprives our listeners and colleagues of meaningful cultural information, and that this information was there in the original;
• that in terms of translation technique there is little difference between the dreaded idioms and all other cultural and institutional references—realia—found in our texts;
• that, broadly speaking, both types of items can be handled in the same documentary way by the interpreter in the booth, similarly to how translators handle institutional realia in ECJ documents.

In doing so, I have referred to sources as disparate as in-flight magazines, legal translations, scholarly work done in the 1970's in Bulgaria and the 1990s in Germany and Canada. Whether I have convinced you remains to be seen. After all, the proof of the pudding...

Chesterman, A. 2000. "Translation typology." In A. Veisbergs and I. Zauberga (eds.), The Second Riga Symposium on Pragmatic Aspects of Translation. Riga: University of Latvia, 49-62. Online with other papers by Chesterman at http://www.helsinki.fi/~chesterm/2000bTypes.html.

Molina, L., and Hurtado Albir, A. 2002. "Translation Techniques Revisited: A Dynamic and Functionalist Approach" in META XLVII, 4, pp. 498-512.

Mossop, B. 1990. "Translating Institutions and 'Idiomatic' Translation," in META XXXV,2, pp. 342-355. Recent issues of META are online at http://www.erudit.org/revue/meta.

Nord, C. 1999. Translating as Purposeful Activity. Manchester: St. Jerome.

Pöchhacker, Franz. 1994. Simultandolmetschen als komplexes Handel. Tübingen: Narr.

Vinay, J.-P., and Darbelnet J. 1958. Stylistique comparée du français et de l'anglais: méthode de traduction. Paris: Didier; Montréal: Beauchemin.