Audience Dialogue

Working with interpreters

The first time I worked with an interpreter was in early 1997, in Hanoi, Vietnam. I was running a course on audience research methods, and at first it seemed to work smoothly enough. I'd say a sentence, then pause, and the interpreter would translate the sentence for the trainees. This was happening in a Vietnamese government conference room: quite a formal setting, with a dais, on which I and the interpreter (a charming woman called Ngoc Anh) stood. The dozen trainees sat in one corner of the room, which had seats for about 200 people. After an hour or so, it became clear that Ngoc Anh preferred to translate longer chunks of speech. "Keep going," she'd say, without interpreting, till I'd spoken for maybe half a minute.

Mostly, she was spending more time speaking Vietnamese than I was spending speaking in English. It takes less time to say something in English than in most other languages, but sometimes she was taking much, much longer - a minute of my talking became five minutes of hers. Though I understood no Vietnamese, I could tell from her tone of voice that sometimes she was dramatizing what I'd said. She put on different voices, as if two people were arguing. It sounded quite entertaining, so I began to use that style myself.

At one point I spoke for longer than usual, explaining some point in detail, but Ngoc Anh's translation took only a few seconds.

"Did you tell them everything I said?" I asked. "That seemed very short."

"It's OK," she replied. "I told them that other stuff before."

I was puzzled. How could she possibly have told them, before I'd said it? Then I remembered that I'd given her a copy of our training manual - a forerunner of Know Your Audience. She'd probably read the first chapter, in preparation for the first day of the course. I began to wonder how necessary my presence was.

Since then, I've run about 20 courses through an interpreter, mainly in Asia. Gradually I worked out better ways of passing knowledge and skills on to the trainees. This page summarizes those ways, for the benefit of people working with interpreters for the first time. Though this page doesn't quite belong on the Audience Dialogue website, I've added it because I couldn't find a comprehensive web page (or a widely available article or book) on this topic. To make the scope of this page crystal clear: it's about running short courses in developing countries, by English-speaking people from developed countries.

In a broader context, this page is not only about working with interpreters; it's about the most effective ways of communicating a message to a group of people who don't understand English. So these suggestions go beyond tips for working with interpreters. Effective communication involves not only lecturing and listening, but also reading, writing, observation, and questioning.

Preparing the course

To make communication as complete as possible, preparation for working with an interpreter is best begun when planning the course. You need to work out the ideal duration of the course, learn something about the culture, and if possible prepare some written materials for translation.

How long will the course need to be?

One point that should be obvious - but took me a while to realize - is that when a course uses an interpreter, it takes much, much longer. In an English-speaking environment, the trainer speaks and the trainees understand, instantly. No time is wasted.

But with an interpreter, the trainer speaks, then pauses while the interpreter translates. In most other languages, to convey a point takes longer than in English because of their "inefficient" grammars. Japanese is specially slow, taking about twice as long as English. On average, one minute in English corresponds to about one and a half minutes in another language. But there are other slownesses, too. For a start, we modern English speakers are so task-oriented; few other cultures are in such a rush. In Cambodia, for example, the art of conversation is prized: people give little speeches - not the snappy repartee that's prized in Western business settings. The amusing dramatizations that Ngoc Anh improvised also take longer. And when the interpreter asks for clarification on some specialized term before the interpretation can begin, that too slows down the process.

Also, interpreting is exhausting work. The interpreter needs a break every hour, for five minutes or so. And of course a break that's planned to be five minutes generally turns into a ten-minute one, as trainees - unable to communicate with the trainer while the interpreter has a break - go off for a coffee or a smoke or to check their phone messages, and have to be rounded up again.

Thus what would take an hour in English often takes about three hours when you are working with an interpreter. Therefore, if a course runs for one week back home, it will take about three weeks in another language. If you haven't planned for this, you'll end up either skimping on the details, or drastically cutting the coverage of the course. And as any experienced trainer knows, it's not a viable solution to run the course for more hours per day: 6 hours a day, in two 3-hour blocks, is about the longest that trainees can concentrate, specially if they have to attend to their normal job as well.

Action learning

So if training funds are limited (as they usually are), and trainees' time is also limited (because they have to do their regular jobs as well) what can you do? Audience Dialogue's solution is to make courses highly interactive. We think of the students as participants rather than trainees. We use action learning and problem-based learning, in which participants work on issues that are relevant to them, rather than merely sitting and listening to the interpreter.

This extends the interpreter's role, in a way that some interpreters are not comfortable with. In many Asian cultures, a Confucian model of education is assumed - that the teacher is a revered expert, who passes on his (sic) knowledge to the humble students by lecturing. The students take copious notes, and would not dare to ask the teacher a question when they don't understand something. (Instead, they have confused whisperings.) The Confucian concept is that learning is transmitted purely by listening.

Do you recognize this model? It happens in Western cultures too, at most universities - despite the fact that decades of educational research have found that lecturing is about the least effective way of transmitting knowledge. The main difference is that Western students don't revere the teacher, and they dare to ask questions.

We've found that the Confucian model is maligned. In practice, it's not much different from Western education, and Asian students quickly welcome an action-learning approach. But the philosophy needs to be explained, in detail, right at the beginning of the course. It's also important to persuade the trainees' supervisors, who might regard an action learning course as too much fun, and not enough learning.

The method that we've evolved is based on a one-hour cycle - an idea we borrowed from short-wave radio programming. Each cycle focuses on one particular topic. First, the trainer talks for about 30 minutes, explaining the topic. (Actually the trainer talks for about 12 minutes, and the interpreter for about 18.) In the next 20 minutes, groups of two or three trainees are given a practical exercise, trying out some aspect of what they've just learned. The final 10 minutes of the hourly cycle is a discussion: a question and answer session about the exercise and the topic.

The timetable is flexible- sometimes the cycle takes more than an hour, sometimes less. But we make a point of never speaking for more than half an hour at a time.  The one-hour cycle is popular both with interpreters (who aren't overworked) and students (who spend a lot of time trying out their new knowledge, instead of just listening all day). And this is a much more time-efficient method than straight lecturing, taking only about 30% longer than the same course in an English-speaking country.

Choosing an interpreter

For a course to be successful in imparting knowledge and skills, it's vital to choose a suitable interpreter. Action-learning interpreting is a situation that many interpreters are not used to, and not comfortable with.

Varieties of interpreters

Normally the host organization in the other country will choose the interpreter, but that organization needs to understand how the course will be run. I was surprised to find that some interpreters have skills only in one direction: for example, they might be excellent at interpreting from English to Indonesian, but unskilled at interpreting in the other direction. I don't quite understand how this is possible, but have encountered it several times. For the traditional Confucian-mode lecture, it's not a problem if the interpreter isn't good at turning the local language into English, because the students don't dare to ask questions, but for a more interactive course, this can be a big problem.

Another aspect of interpreting is written translation. Again, you might expect that an interpreter would also be able to translate brief course materials on the spot. It often helps if trainees can read what they've also been told - using a whiteboard, for example. However, some interpreters I've worked with have declined to write anything in English, saying they haven't been trained in this, or it's not part of their job. (Maybe they were embarrassed about making spelling mistakes in front of me.) If you insist on one who can interpret in both directions and also translate on the fly, this is a wider skill-set and it will therefore cost more.

Because of this specialization among interpreters, one method I've adopted is to use two interpreters at once: one in each language direction. When a course is highly interactive, using two interpreters can save a lot of time. Though some agencies I've worked with have been horrified at the idea of two interpreters for one trainer ("a wasteful luxury"), others have understood the benefits. In developing countries, interpreters are paid a tiny fraction of what a visiting expert gets, so if the expert could convert a three-week course into a two-week course by using two interpreters, the savings on the expert's fee would be far larger than the cost of the extra interpreter.

During a course, who is paying the most attention? The interpreter, of course - who has to attend to every word you utter, in order to translate it. Therefore, after the course has been running for several days, the interpreter (when not actually interpreting - e.g. when the students are working on projects in small groups) can assist the trainer in advising the small groups. When you have several interpreters, the course will progress much more quickly if two people can answer trainees' questions at the same time.

Preliminary review of technical terms

If you will be using technical terms in the course - words that the interpreter may not know in English (or even in their own language) it's best to provide a glossary of these in advance, so that the interpretation won't be interrupted by the interpreter having to ask you to explain the terms. Having to ask the trainer to explain makes the interpreter lose face, so interpreters generally welcome the chance to sort out these terms in advance.

Sometimes the terms just don't exist in the target language, and it will be necessary to create neologisms - or to just use the English term, and explain it. Thus interpreters should read the course notes at least several days before the course begins, so they can ask you about any difficult points. If the course notes are going to be translated, it's best if the interpreter can do the translation, because this means the interpreter will already know a lot about the subject when the course begins - as I found in Hanoi.

When the local language doesn't have a precise enough term, the interpreter needs to be encouraged to be brave enough to warn you of the problem, and the trainer and interpreter then need to work together to find a suitable translation. With new concepts such as Causal Layered Analysis, most languages will not yet have terms. If the language does not use our alphabet, specially if it is read from right to left, terms written in English may not be recognizable - just as a word written in Arabic would be hard for you to remember in English text.

One problem with using approximations in the local language is that the meaning may not be as precise as it should. For example, many languages seem not to distinguish between "random sampling" in the mathematical sense and "randomness" in the sense of "just ask anybody, it doesn't matter who." Surveys done using the latter principle tend to produce biased results (favouring the entity organizing the course), so this is an important distinction.

If your course notes are comprehensive and self-explanatory, it's generous to donate copies of the translation to local libraries, for reference by others after the course has finished. After all, the reason why this country has hired you - an expert who doesn't even speak its language - is because it lacks the expertise.

During the course

This section covers practical issues that arise during a course.

Who are you talking to?

A common mistake is for the trainer to address the interpreter - as if talking to the interpreter, not the class. Instead, the trainer should address the class, and the interpreter should be as unobtrusive as possible - a voice in the background, not getting between the speaker and the audience.

That's a fine principle - but some interpreters want to be noticed, and this can be a big problem. You don't know what they are telling the audience (otherwise you wouldn't need an interpreter) but maybe you suspect it's not quite accurate. And when a participant asks a question, the interpreter may answer it directly, without referring it on to you. Even if the answer is obvious, these tactics impede communication between the speaker and the audience. To prevent such behaviour from the interpreter, the principles need to be agreed before the course begins.

Some interpreters, after a few days, believe they are as expert as the trainer; I've known some of them to give misleading advice - and have only found out later from participants who know some English. If you suspect your interpreter is translating rather loosely, it's a good idea to warn them when important points arise, that need to be translated very precisely.

If you can't trust your interpreter to provide an accurate rendition of what you are saying, you may even need to find another interpreter, at short notice. If you suspect this is happening, you need to confirm it. This can be done by getting participants to provide feedback in some way that the interpreter can't easily change, such as a daily multiple-choice test, from which students have to choose one of several similar responses. Though I've never found an interpreter who wilfully changed what I've said, I suspect that several of them have not cared much about conveying the precise meaning.

In most courses that I've run in developing countries, it turns out that many of the participants understand some English. A common motivations for going on an interpreted course is to improve their English. So even if the interpreter's English is perfect, participants will be more involved if you make some attempt to communicate directly to them, by speaking slowly, and not using obscure words or slang. Our page on Global English gives some advice on this.

The duration of speech turns

At one extreme, you can say a few words at a time, then the interpreter can translate them. But even if the sentence structure of the other language is the same as English (i.e. the subject - verb - object "SVO" sequence) and it therefore makes sense to utter one phrase at a time, interpreters hate this. They can't work well if all they have to go on is a few words, without knowing the rest of the sentence. So one sentence at a time (averaging maybe 10 to 15 words) is the absolute minimum. The resulting interpretation will be precise, but often stilted.

At the other extreme, the trainer can talk for several minutes at a time. This makes it much easier for the interpreter, but if you speak for too long, the interpreter tends to forget some of the early points.

The ideal length of a speech turn varies with the target language, but is generally several sentences - around 50 words, or several sentences. With this duration, there's no danger of the interpreter forgetting the beginning, but it's a long enough statement that it can be clearly shaped in the other language.

When the other language differs greatly from English in its structure, the longer a speech turn should be. Though many languages share with English the SVO sequence, other languages have different structures. For example, in German, where the verb can come last, and sentences are often long, translation into English can't begin until the sentence has finished in German. Other languages where sentences have verbs at the end are Japanese, Turkish, many Indian languages, and Farsi (Persian). Arabic has the verb at the beginning, and Latin and Finnish can have any sequence. (See the Wikipedia article on Word Order for more on this.)

Sometimes it's not just words that need to be translated, but whole concepts. In Vietnam, I tried to illustrate the concept of randomness by mentioning dice. I thought every country had dice, but the interpreter had no idea what I meant. So I switched to coins (heads and tails) - but there are no coins in Vietnam, only banknotes. Luckily, the banknotes had serial numbers, so I could use the last digits of those as an example.

Adjusting the pace

I've noticed that some interpreters are reluctant to admit they don't understand what you are saying, or are having trouble keeping up with you, or remembering what you just said. It therefore helps to offer, fairly often, to change the pace. In fact, there are two paces involved: the pace at which you speak to the interpreter, and the pace at which the interpreter speaks to the students. Each pace is limited by the listeners' understanding. The better your interpreter's English, the faster you can speak. But this may not flow through to the participants' understanding of what the interpreter says. If a few participants have a much lower initial knowlede level than most others, those slower people may need special attention later, so that everybody else isn't slowed down to their pace.

Visual aids

Whenever somebody speaks for a long time, it's easy to lose the drift of what they are saying. I've therefore found it helps, even when I'm talking to English-speakers, to display a set of headings and sub-headings, and to point to each new sub-heading as I begin discussing it. A co-operative interpreter will write the equivalent subheadings in their own language, next to what you have written.

This can be done using various media - a whiteboard, a computer presentation, or large sheets of paper. Presentation software, such as Powerpoint and Keynote, look impressives, but generate time-wasting problems in developing countries - odd keyboard layouts, frequent power cuts, and the like. Whiteboards are more reliable than presentation software, but the best alternative is to use large sheets of paper - "butcher paper" as it's often called. There are lots of reasons for this:

  1. The sheets can be kept, for later reference by students.
  2. Poorly maintained whiteboards are often hard to erase - but paper is always new.
  3. Sheets of paper can be easily moved to another location.
  4. Usually, only one whiteboard can be seen at a time, but there's no practical limit to the number of sheets of paper used - so summaries from days earlier can be quickly referred to, when the sheets of paper are stuck on the walls of the training room or stacked on an easel.
Paper wins!

A "split screen" layout works well - even if the "screen" is paper. I write the points in English on the right, there's a dividing line vertically down the centre, and the interpreter writes in their own language on the left side of the line. Participants who know some English appreciate this type of layout, because it helps them improve their English, and better understand what the trainer is saying when the terms are next mentioned. Here's a photo of a whiteboard laid out like this, with a course evaluation questionnaire in English on the right and Lao script on the left:

split-screen bilingual presentation

If the course involves using diagrams or pictures, captioning them in both languages (by the trainer and the interpreter) is also helpful.

What if the interpreter doesn't turn up?

This has happened to me several times: while running a course in a country where I don't speak the local language, depending on an interpreter for everything (outside the class, as well as in it) - and one day the interpreter doesn't turn up. No matter how good the reason - such as being involved in a motorcycle accident - if there's no alternative interpreter instantly available, what happens with the course and the trainees? This happened to me in Indonesia, and it wasn't disastrous. By then, I'd spent a lot of time in Indonesia, so knew something of the language, but several participants knew a lot more English than I knew Indonesian. Between us, we managed. But in other situations it would have been more difficult. This is a question worth raising with the host organization: how quickly a standby interpreter could be arranged, if the main interpreter fails to turn up. (This is another reason for having two interpreters.)

Evaluating student learning

In Asian cultures, trainees can lose face if they admit to others that they don't understand - particularly if the others on the course are their workmates, bosses, or underlings. By giving a quiz (with multiple responses, for ease of marking) at the end of each day, marking it overnight, and returning the papers the next morning, students' progress can be assessed, and slow ones identified. (Evaluating courses and other events is discussed on another page on this site.)

If most students don't get a high mark on these overnight tests... Well, this is not a school. It doesn't mean the students have failed, but that the training has failed. It means that you and the interpreter have failed to convey the content, and therefore that you need to change your approach - as soon as possible. A good way to begin this change is by discussing it with the trainees.


The Web doesn't seem to have much on this topic - hence this page. But these links could be useful...

The International Law Enforcement Academy of Bangkok: General guidelines on working with interpreters. The most detailed page I found about how to work with interpreters in training courses.

AAFP (American Association of Family Practice - Its journal Family Practice Management, for June 2004, has an article with some useful tips, though not all of them are relevant here. It' focuses on how to use interpreters on a one-to-one basis in a medical context. Don't miss the ingenious picture by Christine Schneider.

If the host organization can't suggest a suitable interpreter, there's AIIC - the International Association of Conference Interpreters, at, with 2771 interpreters on tap, covering 46 languages, in a wide range of countries, specially in Europe.