I recently saw a beautiful book about Mercedes. It’s all pure, glossy prestige — 400 richly illustrated pages in German, English and French. But there’s a catch — a translation catch.

The text begins with a story: Es ist der 19. August 1997, so um 9 Uhr vormittags herum. Die Verbindung funktioniert, auf der anderen Seite der Leitung Herr S.

There’s a translation into English: It’s 19 August 1997, around 9 am. The link is OK, and Mr S. is on the other end of the line.

And then there’s one into French: Nous sommes le 19 août 1997, vers neuf heures du matin. Le téléphone sonne, c’est un certain monsieur S. qui m’appelle.

From these translations it appears that the English and French translators don’t agree on the meaning of the German phrase “Die Verbindung funktioniert” (literally “the connection functions”). The English translator has figured that it means that the call has successfully gone through (i.e. it didn’t fail) and that the telephone connection between the storyteller and Mr S. is working.

The French translator, however, has decided that it simply means that the phone rang (“Le téléphone sonne”).

Has an error of interpretation been made? Who is right? How did this happen? What can translators learn from this example? I posed these questions to veteran translator John Jamieson.

Paul: What’s going on here, John? Why did the English and the French translators come up with differing interpretations of the original German?

John: The French translator got it right. The phone rang. The English translator didn’t recognize that the German word for “the connection” (die Verbinding) is a derivative, an oblique expression which refers back to the more specific concept of “telephone” or “telephone connection” (die Telefonverbindung). Similarly, “functioned” was a derivative expression from the more informative “rang”. So in other words, what happened was simply that the phone rang.

Paul: You’ll need to explain what you mean by derivative.

John: This sort of translation problem can be very nicely described in terms of rudimentary calculus. If you have the coordinates of the location of a moving body (such as a train or an automobile), over a period of time, you can calculate its velocity using a simple mathematical rule. If you apply the rule again you can then calculate its acceleration. From position, we can derive velocity and from velocity, we can derive acceleration.

In the example from the Mercedes book, the expression (“the connection functions”) is derived from and refers back to a specific concrete event — the phone rings.

Paul: So a derivative expression is more ambiguous?

John: Yes. Just like in calculus, every time you apply the rule to form a derivative, there is some information loss. In this case, “phone” is lost and we are left with a more ambiguous “connection.” “Rings” is lost and is replaced by “functions.”

Paul: Can you give me some more examples of such derivatives in language?

John: Take for example a newspaper headline “Tremor shakes capital.” Tremors come in many guises, but according to the convention for newspapers in our part of the world, it is taken to mean “earthquake.” And the more general term “capital” is readily understood as more specificly Wellington, New Zealand’s capital city, but only for New Zealand readers.

So the writers of headlines such as these expect you to know the specific information from which they have been derived. We know that “tremor” means earthquake because of journalistic speech habits in English, and we know that “capital” refers to Wellington if we are in New Zealand. In other words, the use of derivative expressions presupposes the existence of some sort of in-group.

Paul: So they often present particular difficulties when they need to be translated.

John: Yes, because foreign readers are unlikely to be in that same in-group. If the differentiated phrase is translated directly they may well miss the intended meaning.

Paul: So how can the translator avoid such errors?

John: Firstly, the translator has got to spot that the phrase is actually a derivative. Once spotted, the translator may have to modify the expression to ensure it is meaningful to readers in the target linguistic or cultural context. This is often best done by anti-differentiating the phrase.

In mathematics, the reverse process of differentiation, anti-differentiation, is called integration. When you integrate acceleration you get velocity and when you integrate velocity you get position. But integration is not as easy as differentiation because you have to restore the lost information.

This was the approach taken by the French translator in the Mercedes book. He recognized that “connection” was a derivative of “phone,” and “functions” was a derivative of “rings” and so he integrated the phrase, restoring the original missing information.

The English translator didn’t recognize that “the connection functions” was a derivative and he rendered the literal meaning in his translation. There is no way that “the link is OK” can be interpreted in English as “the phone rings.”

Paul: So how can translators spot what is derivative and what should be taken literally?

John: The translator can often disambiguate tricky text by establishing where the primary and second stresses lie. Some material in a text is new information, which tends to carry the primary stress, whereas other information is secondary and derivative, referring back to that primary information, and generally carries a lesser degree of stress. But with a sentence like “die Verbindung funktioniert” both terms are derivative, so you try putting the stress on the first main word, then on the second, and see what happens.

If I read “Die Verbindung funktioniert” (“the connection functions“) where “functions” carries the primary stress, then I will assume that some sort of connection is working as opposed to not working. But if I read it as “die Verbindung funktioniert,” I assume that the “connection” is doing something or other. I then realise “it must mean Telefonverbindung,” interpret it as “telephone,” and the meaning of “funktioniert” falls into place as a matter of course.

Paul: Are you suggesting that all derivatives need to be integrated when translating?

John: Not necessarily. But we have to remember that these derivatives are often based on writer and reader being in the same in-group, and the reader of the translation may well not be in that in-group. So we must be careful to use expressions that allow the reader to recognise the antecedent it refers back to.

We need to be aware of the differing conventions in different languages. In French, for example “échéances” (deadlines) can be derivative shorthand for “échéances électorales” (electoral deadlines or just elections). We can’t force the literal English equivalent “deadlines” into that straitjacket, but we could use “polls” as the English conventional derivative from the primary concept of “elections.” So in that case we would be translating a conventional derivative in French with a differential conventional derivative in English.

As translators, we have to distinguish between primary and secondary beats in the text, to detect what is new and what is derivative and then reflect the different weightings of the two categories in the target language. This may involve some manipulation of the text and the restoration of some missing information — something that would not be approved by some of my colleagues in the profession, and indeed by some translation examiners!

Even the most boring, everyday texts are imbued with their own form of prosody and music — strong and weak beats dictated by the semantic flow of the communication. Reading difficult texts aloud to find these beats, as musicians do when they read a score, can enrich the reading experience, particularly in the case of good or great writing.