Décalage fascinates and frustrates in equal measure. It is the most obvious feature of simultaneous interpreting, it’s the first thing student interpreters ask about – and yet there are very few actionable explanations of it out there. In this article, I will argue that we should not see décalage as a goal or technique to be achieved but instead as the quirky by-product of techniques that make simultaneous possible and which can be clearly described, consistently applied and learnt.
Table of Contents
- Executive summary
- What is décalage?
- Why décalage is a poor metric
- Décalage is not an actionable technique
- The right end of the stick
- Techniques
- Conclusion
- Terminology note
- Footnotes
Executive summary
A variety of techniques for setting simultaneous décalage have been described in the literature with overlapping and inconsistently used terminology.
This article seeks to review the various techniques, clarify the terminology associated with them and recommend what the author sees as the best two techniques within a general approach, namely… when interpreting simultaneously interpreters should think in terms of when-to-start-speaking rather than in terms of length-of-décalage. And interpreters should wait until they have heard enough information to create either prosodic units or an entire sentence before starting to speak. Either or both techniques can be used in the same speech according to the interpreter’s choice at any given moment.
What is décalage?
Décalage is the French term meaning “lag” or “difference”, which describes the time between an interpreter hearing something and their re-expressing it in the target language. It has also come to be called “ear-voice-span” and “(time) lag” in English, but interpreters usually use the French term in English.
Décalage is what makes interpreting possible. It is the time required for the interpreter to understand, analyse, and reformulate the meaning of the original speech (before speaking it). Unfortunately, because we talk about “simultaneous” interpreting, many non-interpreters don’t immediately recognise the existence and importance of this crucial period of time.
Research tells us that lag – usually measured as the time in seconds between two words or information items – can vary from a negative value (when an interpreter anticipates) through zero and up and beyond 10 seconds[1]. Research also tells us that at least minimum and maximum values above zero are defined by cognitive capacity. The minimum by how quickly you can process the source into output and the maximum by your ability to manage (not exceed) working memory capacity. Research also tells us[2] that experienced interpreters have an advantage at both ends of the spectrum. They can process information more quickly and they can work with larger chunks of information, so better manage working memory capacity.
Why décalage is a poor metric
Décalage has fascinated researchers and student interpreters alike, but it is a flawed metric for a number of reasons.
1) Décalage is inevitably measured across words, but there are at least two problems with this. If one word in the source is translated by three words in the target, where do we take our measurement of décalage? Also, our mantra as conference interpreters has always been that we interpret meaning, we do not transfer words. So how do we pinpoint the moment that meaning is transferred between languages?
2) Décalage is measured in seconds. Even if we accept that words are a usable reference for measuring décalage, the time/word measurement becomes unreliable. Taking a very hypothetical example below, in which the number and order of words in source language and target lagnuage are almost identical, décalage – as represented by the coloured lines – can still vary within a single sentence. When (in this hypothetical example) the interpreter speaks at the same speed as the speaker, then décalage will be the same across the whole sentence…

… but if the interpreter just speaks more quickly than the speaker, then their décalage will get progressively shorter with each word they speak. You can see that the blue line below is shorter than the green line and the blue line in the first example.

Different speaking speed, different décalages at any and every measurement point, all other things (like processing) being equal. Therefore this variation tells us little of use about the interpreting process. And in the real world of course, the other things are not equal!
3) Décalage also varies depending on the type of information presented; the syntactical similarities between different passages of the two languages in question; and the interpreter’s familiarity with the topic. Since familiarity with the topic is a very personal affair, different interpreters will have different décalages for a given passage.
4) Décalage can vary between interpreters depending on the experience of the interpreter and their familiarity with the topic.
5) Décalage is often measured as an average. This may be of scientific interest, but an average (for example of the décalage above in 2) is no good to the student who asks “how long should my décalage be?”. Let me use just one example to show how décalage varies and how an average therefore doesn’t paint a reliable picture…

I would argue therefore that décalage doesn’t give us a useful picture of what we are doing in simultaneous, because the data are inherently inconsistent and unreliable. Worse still, décalage in this form is not a technique we can teach or practise.
Décalage is not an actionable technique
In addition to not being a reliable metric, I would argue that décalage is not an actionable technique that students, trainers and interpreters can work from. Décalage is no more than a description, quantification, of what we can see and hear when an interpreter interprets. It doesn’t give us a tool that we can use to improve our interpreting, because if students are occasionally told to “shorten” or “lengthen” their décalage we are not giving them any useful instruction as to HOW to operate that change. Indeed I would argue that décalage is not a technique or a skill- it is the by-product of the techniques interpreters use to interpret simultaneous. And just as measuring the amount of yeast thrown away by a brewery will not teach you how to make beer, worrying about décalage is the wrong angle to approach simultaneous interpreting from.
From the trainer’s point of view, I see some excellent professional interpreters and trainers who are unable to explain how they manage their own décalage, even though they clearly do successfully manage it. And the literature (with the exceptions cited below) has not always been much help. Seleskovitch & Lederer[3] set the tone by describing lag as a pendulum, “students must learn to get alternately closer to and further away from the [source] speech”.
My impression is that when students and trainers talk about setting décalage, they are neither talking about an average décalage over a sentence, nor about all the different décalages over the course of any sentence, but in fact only about that first moment when the interpreter speaks. And that brings us to what I believe is a much more practical way of approaching the problem.
The right end of the stick
So what are we to teach if not décalage? Well, Roderick Jones[4] was the first to flip the viewpoint when he focussed on “When to start speaking” in simultaneous. Setton and Dawrant[5] have followed that up more recently in talking of the “entry points” for the interpreter. Both these approaches start then not from the length of the lag, but from how much information the interpreter needs before they can start speaking, how to recognise that amount, and what to do with that information when they have it. That wait is not measured in seconds but in (different) blocks of meaning. And the process is repeated every time the interpreter starts to speak, not just at the beginning of the speech. This allows interpreters to consistently apply a single, but flexible, technique to help them interpret. The fact that the resulting lag is extremely varied is irrelevant.
Gile explained that interpreting was a series of efforts undertaken simultaneously – the Listening and Analysis Effort; the Memory Effort: the Production Effort and crucially the Coordination Effort. But that we had finite mental capacity to make these efforts. Entry point techniques are very much about the tension between the waiting for information and the limits of mental capacity. The interpreter needs enough information to compose something meaningful to say, but needs to say it as soon as possible to avoid exceeding the limit of their mental capacity and to free up that capacity to listen to and analyse the incoming language. By starting from an entry point, not the lag, each interpreter can decide their own entry point at any moment in the speech according to their own abilities, knowledge, experience etc. Their (varying) lag after the entry point will be far less important.
Techniques
So let’s have a look at the entry point techniques for simultaneous interpreting. I warn you in advance that I will recommend only two from this list at the end of the article.
- 1. Word for word
- 2. Units of meaning
- 2.4 Prosodic units
- 3. The idea
- 4. Jones’ when-to-start technique
- 5. The salami technique
- 6. Stalling & fillers
1. Word for word
The interpreter speaks immediately after almost every word or expression
I mention this because, unfortunately, some interpreters work like this. But it is really the example of what the interpreter should NOT be doing. In languages that are structurally similar enough (FR-EN, FR-ES, or NL-DE), the interpreter can chase after the speaker, just a word or two behind, and they will produce something that may superficially sound acceptable, but which won’t bear up to closer scrutiny. In languages with more structural differences, you would just end up with nonsense – word-for-word would be quite impossible between DE and EN, or JP and EN, for example, where subjects, objects and verbs come in different orders..
Even for similar languages, lexical and syntactical calque will abound, if the output makes sense at all. Let’s take a simple example from French:
FR – l’assurance garantit, sauf déclaration spéciale, la totalité des bagages de l’Assuré
EN – Unless otherwise stated, your insurance covers all your luggage…[6]
The part of the sentence in blue would naturally appear in a different position in French and English so if the interpreter starts to speak on a word-for-word basis, it will be impossible to arrive at the stylistically better syntax in English. And since French does a lot of this sort of parenthesis, it’s going to stick out.
More seriously, word-for-word also doesn’t allow the interpreter the time needed to understand the point being made and to adapt their output accordingly (so the listener can understand it too).
Finally, the interpreter’s sentence intonation will be all wrong too. The interpreting word-for-word is monotone and as unpleasant to listen to as the following is unpleasant to read.
your
insurance
covers,
unless
otherwise
stated,
all
your
luggage…
As a general rule, interpreters avoid this and interpreter trainers do not teach it. Word-for-word is often the mark of the interpreter who has not had formal training. I leave you with Lynn Visson[7]‘s verdict: “Literal mot-à-mot [word-for-word] translation requires considerably less mental effort but produces far worse results”.
2. Units of what?!
2.1 Unit of meaning was not a décalage technique
The unité de sens may be the most famous unit in interpreting. It was not, however, intended as a method for describing or setting décalage. The unité de sens / “unit of sense” was devised by Marianne Lederer as a way to explain how interpreters arrived at meaning-based target language versions. She defined it as “the smallest element leading to the establishment of equivalence in translation”, (where equivalence takes place at segment level, not word level). It has been translated as both “unit of sense” and “unit of meaning” in interpreting literature.
Lederer describes creating a “unit of meaning” in terms of “button tufting”, “capitons” in French. The interpreter waiting for a unit of meaning is like “the upholsterer who pins his needle [multiple times] into the back of an armchair”. With each needle thread the “interpreter is … forced to translate words according to their immediate linguistic meaning”. Only when they get to a certain stage, does the upholsterer “make divisions [buttons] appear at a pull of his thread”. This is “the point that they start to glimpse what the speaker is getting at and can create a unit of meaning.”[8]

It is clear from this that Lederer sees the interpreter speaking BEFORE a unit of meaning has been identified, so it can’t be the unit of meaning that determines décalage or the entry point.
That interpreters have found the description of this method difficult to apply can be seen in the fact that this rather nice analogy has not survived the test of time in the same way that Seleskovitch’s “raisins in the cake”[9] analogy has.
Seleskovitch & Lederer’s approach then, was that “The interpreter is obliged to work with sub-segments of meaning”[10]. This distinction between unit of meaning and sub-segment of meaning has been lost over the years and may be the reason why the unit of meaning’s meaning and function are now understood differently.
2.2 The unit of meaning’s new meaning & function
Having begun life as a description of a phenomenon encompassing fairly large chunks of source speech the “unit of meaning” has metamorphosed in two fundamental ways over the last 30 years:
2.2.1 to mean something much smaller than a sentence, (probably what Seleskovitch & Lederer called “sub-segment of meaning”[11]. Have a look at Setton & Dawrant’s examples below:

Setton & Dawrant[12]
2.2.2 The function of the unit of meaning has also changed. It is now seen as that unit which defines the point at which the interpreter should start speaking in simultaneous, or “entry point”. When the interpreter has heard a unit of meaning, they start speaking.
Evidence of both these changes can be found in Setton & Dawrant when they say, “the simultaneous interpreter… must work with sub-sentence chunks of input… [with] pauses inserted between the chunks, each a relatively self-contained meaningful segment, or ‘unit of sense'”[13]. Kader and Seubert say something very similar when they describe “individual units of meaning contained within a source text sentence.”[14]
As we can see, the new meaning of “units of meaning” is synonymous with chunks and this technique is also called “chunking”.
The interpreter should not start speaking until they have heard a unit of meaning in the source language.
To my mind, this shift in the meaning and function of the unit of meaning and the fact that two English translations are in circulation – unit of meaning and unit of sense – mean that this term is no longer very clear, and therefore no longer very helpful.
2.3 Unit of meaning’s new name
So what do we call this unit of meaning that used to be something else and do something else? Should we distinguish between 20th and 21st century units of meaning? And which English translation to use? And whatever we choose, will what we mean be clear to other interpreters and students? Should we use “chunks”? It’s simple, if not terribly high-brow. Perhaps we need a new term. But what?
One answer presented itself one day when I wasn’t looking for it! I was talking to a group of trainers about using the “entry point” as a technique in simultaneous teaching and I explained Setton & Dawrant’s understanding of the unit of a meaning and a more practical description of it that I use in classes, which is “at least the minimum amount of information that can be surrounded by pauses in normal conversation without it sounding weird.”
One of the trainers huffed and suddenly gave expression to a very sincere frustration (that I have often felt myself) about interpreting literature being so sloppy in its use of terminology. “That annoys me” she said, “because that is just a prosodic unit!” I duly went away and researched this a little and discovered two things:
1) that my colleague was absolutely right and
2) that despite some disagreement at least some linguistics experts had arrived at something comparable to what Setton & Dawrant described and I had tried to explain. Prosodic units, I read, are… “minimal stretches of speech marked by … by prosodic cues such as pausing … These units function as the basic building blocks of spoken discourse and provide insight into how speakers package information in real time”[15].
So what do prosodic units look like? Well, like those units described by Setton & Dawrant in the diagram above as units of meaning. Compare also acceptable and unacceptable pausing – where the pause is the line-break – in the examples below.
Example 1 of acceptable pausing (= prosodic units) where the line break represents the pause
I very much welcome
the plan arrived at
at the North Sea Summit
to dramatically ramp up renewable energy
in the North Sea[16]
Example of unacceptable pausing (not = prosodic units)
I very much welcome the
plan arrived
at at the North
Sea Summit to
dramatically ramp up renewable
energy in the North Sea
The good thing for the interpreter is that prosodic units can vary enormously in length and information content and that there are many correct ways to break down a sentence into prosodic units. It will be the interpreter’s choice to make, and will define when they start speaking, their “entry point”. That gives the interpreter the flexibility needed to react to unfamiliar topics, or draw on knowledge they do have before speaking.
Example 2 of acceptable pausing
I very much welcome the plan
arrived at at the North Sea Summit
to dramatically ramp up renewable energy in the North Sea
NB Lynn Visson[17] advocates exactly this approach for simultaneous production and presents very similar examples but does not explicitly mention the link to lag.
So now we have something we can explain and a word for it – a “prosodic unit”.
2.4 Prosodic units as a technique for simultaneous
The interpreter should not start speaking until they have heard enough information to compose a prosodic unit in the target language.
So we’ll call these chunks, segments, units of meaning “prosodic units”, and use that as the unit that defines when we start speaking in simultaneous, our entry point.
Note that over and above the change in terminology there is a subtle shift here from “hear a unit of meaning” to “hear enough to compose a prosodic unit”. The interpreter waits until they can produce something meaningful, not until they have heard something meaningful. It’s an important distinction that affords the interpreter some extra flexibility. The interpreter can start before the speaker’s meaning is clear AND there is no obligation to include all the information heard in the source text in the prosodic unit the interpreter speaks. Information can, if the interpreter wishes, be kept in memory, or jotted down on paper, and added to a later prosodic unit spoken.
But what does this all mean in practice? A speech is spoken, one word at a time. If we represent that chronology as shown below…
I
I very
I very much
I very much welcome
…then the interpreter applying the prosodic unit method will not start speaking until after “welcome” because that is the first place they will have enough information to start mentally composing a prosodic unit. Once they have done that they can start speaking it. And while they are speaking that prosodic unit the interpreter is listening to what follows…
the
the plan
the plan arrived
the plan arrived at
… they will not start speaking again until after “at”, because that is the next prosodic they have pieced together. Pauses between the and plan and arrived would be weird, so we wouldn’t have a prosodic unit. And so on.

But equally, the interpreter is free to wait longer to compose their first prosodic unit until “at”, if they are comfortable with that amount of information….
I
I very
I very much
I very much welcome
I very much welcome the
I very much welcome the plan
I very much welcome the plan arrived
I very much welcome the plan arrived at
Indeed there is no fixed length to a prosodic unit, only limitations as to where it can’t be (because a pause would sound odd in a given place). This gives the interpreter incredible flexibility to adapt their entry point to their own strengths/needs.
One key tactic in simultaneous which applies to prosodic units is that information items do not have to appear in the same prosodic unit, or the same order, in the target language as it did in the source language.
Taking the same example above we could “interpret” as follows, moving “the North Sea Summit” and later “renewable energy” between prosodic units.

Similarly, the prosodic unit technique gives us a way to rearrange syntax, which is not always the same between languages. Because prosodic units function as single chunks for the purposes of working memory, they can also be moved around in sentence order as single chunks (for little extra cognitive effort). That means flexibility and a way to produce more natural syntax. The text in blue could be considered as prosodic units and have both been moved.

Here’s an authentic example of the same thing…[6]

Advantages of the technique
- the prosodic unit can be very short, often a single word, like “Hello”, or conjunction like “next”. Speaking a short prosodic unit frees up the working memory that would have been holding on to it for other tasks, like listening to what follows;
- prosodic units function as single chunks for the purposes of working memory. This means that they can also be moved around in sentence order as single chunks (for little extra strain on working memory). That means flexibility and a way to produce more natural syntax;
- because it is based on conversational units, the intonation of the interpretation is relatively natural and easy on the listener’s ear;
- the technique is sustainable; it can be applied from start to finish in any speech.
Disadvantages of this technique
- can lead to long pauses between very short utterances
- a string of prosodic units will not make complete sense alone and will often need repair, compensation, and/or “stitching together”.[18]
Example: an interpreter saying the following…
If the government ignored its election manifesto promises and decided to raise income tax for middle income workers then they would most likely pay the electoral price for that in municipal elections in May.
in prosodic units would have to remember having used the “if… ignored” while dealing with quite a lot of new information before getting to the second verb, the tense of which will depend on the “if… ignored“. That’s not easy. And it is harder into a B language than into an A.
The shorter the prosodic units, the more repair will be required later and the narrower the options available to the interpreter for the next unit will become.
3. The idea as a technique
Start speaking when you have heard and understood an idea
You may have heard trainers say, “wait for the idea before interpreting”. For example Seleskovitch, “Il s’agit donc d’attendre d’avoir conscience d’une idée pour la réexprimer spontanément” (The aim being to wait until you become aware of an idea so that you can express it spontaneously).[19] Unfortunately, you’ll struggle to find a definition of “idea” anywhere. Indeed, there are at least 3 definitions current in interpreting literature. As a trainer, I have a problem with that. You can’t teach, or expect a student to apply, a technique, or the use of a concept, that isn’t clearly defined. The vagueness of the term idea probably also means this technique is actually just one of the other techniques listed here with a different name.
4. Jones’ when-to-start technique
Start speaking when you can create and say a complete sentence
In his seminal work, Conference Interpreting Explained, Roderick Jones presents a technique for setting lag in simultaneous interpreting. And Jones was the first (in the literature at least) to flip the point of view from passive lag, a phenomenon inflicted upon us, to active, to be decided upon by the interpreter. Jones talked about “when the interpreter should start speaking”[20]. From a teaching and training point of view, this is a much more actionable approach.
So what does Jones propose?
- Don’t start speaking until you know you can complete a grammatical sentence, any sentence (that is faithful to the original).
- But you don’t have to complete the sentence you originally had in mind…
- …nor the same sentence the speaker finishes.
Let’s have a look at how that works in practice.
A speech is spoken, one word at a time. If we represent that as shown below…
Once again we are facing rhetoric from the extreme right that would have us believe that all the car industry’s problems come from the 2035 rule.[21]
once again
once again we
once again we are
once again we are facing
once again we are facing rhetoric
once again we are facing rhetoric from the
… then the interpreter applying Jones method will not start speaking again until after “rhetoric”, because that is the first moment they have enough information to create a full sentence. The sentence they have created in their mind is not the same as the speaker’s, it is shorter. For example, “Once again we are facing this rhetoric.” But as the interpreter is speaking they are also hearing more information. As new information arrives, they have a choice between 2 approaches. Create another short sentence and repair so that the two sentences go together, or adjust the first sentence they had started with and incorporate then new information into that, always with a complete sentence in mind. Those options might look like this:
1.
Once again we are facing this rhetoric.
It’s the rhetoric of the extreme right.
or
2.
Once again we are facing rhetoric from the extreme right.
The second and third points of Jones’ technique are crucial here to distinguish Jones’ method from a simple chopping up of the original into short sentences (the salami technique below). Jones’ nuanced approach means that the interpreter 1) always has a sentence in mind as they speak. But 2) that short sentence can be lengthened as new information arrives (for example “from the extreme right” in the example above). Equally, the interpreter may turn two short sentences from the speaker into a longer sentence if it sounds better, or perhaps to compensate for short sentences elsewhere.
Jones’ technique really comes into its own when the interpreter gets prior warning of long sentences in the form of particular language at the beginning of a sentence. For example a conjunction (link) like although below.
Although the Empire kept building bigger and louder, it was outflown by rebels in smaller X-wings who were simply smarter and better.[22]
An interpreter following this speech could apply the prosodic unit approach and start with something like “although the Empire…“. But this is to create a hostage to fortune. The interpreter is committed by the link “although” to 1) a two part sentence that the speaker may or not produce themselves AND 2) to the grammatical and syntactical agreements that the “although” imposes. That’s a gamble AND extra cognitive load. Jones’ technique would be to create a sentence as soon as possible, and then repair.
The Empire kept building bigger and louder.
Despite that, it was outflown by rebels in smaller X-wings.
The interpreter puts the “although” on hold. Creates the short sentence “The Empire…” and then repairs by turning the “although” into “despite that“, where “that” refers to the content of the previous sentence. The interpreter could equally have said “however” or “but” instead of “despite that”. The meaning over the whole extract is maintained.
The use of Jones method for subordinate clauses sentences is reprised by Setton[23] in his ‘Taking the plunge’ (or ‘starting differently’).
Advantages
- sentences always completed;
- full meanings;
- it makes for perfectly natural sentence intonation, because the interpreter is speaking in complete sentences;
- the technique is sustainable, it can be applied from start to finish in a speech.
Disadvantages
- it requires consistently a little more cognitive effort, and
- arguably is not always necessary (because the prosodic unit technique will work);
- my experience is that there are very technical or very fast speeches where it is not actually possible and/or the additional cognitive capacity required is not available;
- repair often takes the form of repetition of the subject/theme of sentences. If overused, this becomes noticeable to listeners. See example 1. above.
5. The salami technique
What many interpreters[24] have retained from Jones’ book and the training methodology coming out of the EU institutions is the famous “salami” technique. Chop the sentence up into shorter sentences. This works for a lot of what we need to interpret and is a great pedagogical tool. But it lacks the nuance of Jones’ method which leaves the option of creating long sentences open (because it allows for adaptation of the sentence originally thought up as it is spoken and as new information arrives).
Also, in some languages long sentences are stylistically desirable. A technique that rules out long sentences does not therefore seem to be a technique that interpreters can use all the time.
Example: salami
Once again we are facing rhetoric from the extreme right that would have us believe that all the car industry’s problems come from the 2035 rule.
Once again we are facing rhetoric
That rhetoric from the extreme right.
Their rhetoric would have us believe something.
Namely that the car industry has problems.
And that all those problems come from the 2035 rule.
Advantages
- very little working memory burden;
- full sentences guaranteed.
Disadvantages
- repair often takes the form of repetition of the subject/theme of sentences. If overused this becomes noticeable to listeners.
- some languages require long sentences for good style. Salami precludes long sentences.
6. Stalling & fillers
Stalling and fillers are the same thing with a different name. Fillers, as the name suggests, fill the silence. They serve to reassure the listener that the interpreter is working, neither add nor subtract meaning, while requiring only a minimum of cognitive effort, thus allowing the interpreter time to listen and process more of the incoming speech. Equally, we could say that the interpreter is stalling, playing for time, while waiting for more information to build what they will say on.
I don’t consider the use of fillers to be a fully-fledged entry point technique, but rather either:
1) they are prosodic units of a sort and therefore part of that technique or
2) a tactic to facilitate either the prosodic unit technique or Jones’s technique by giving the interpreter more time to think about what they will say (for minimal extra cognitive load).
Examples of fillers:
| Ladies and Gentlemen, Madam Chair, Well then, Now, | I think, If I may, Let me just say that Moving on… |
Conclusion
Now, of course, setting the entry point is not the only thing we do when we interpret simultaneously. We must also learn to understand conceptually the points speakers are trying to make and the topic areas they are discussing. We must also incorporate repair and compensation – stitching and button tufting – to successive units we speak. But this article was only about décalage and entry point techniques, not all of simultaneous.
My conclusion and recommendation – and also what I believe the best interpreters do – is to apply the prosodic unit technique as a default and apply Jones’ when-to-start technique when the sentence begins with language that clearly indicates a more complex sentence is to follow.
The real deal-closer of this recommendation for me is that both of these techniques can be applied consistently by the interpreter (and trainers demonstrating them) and their application can be assessed by the trainer – “you have/haven’t applied the prosodic unit technique there”. The same applies for an interpreter listening to their own performances. That clarity is a huge advantage for anyone trying to improve, be it in a classroom or as part of their own professional development.
Terminology note
To give an overview of how interpreter trainers have been less than consistent with their use of terminology, here is a table of the entry point techniques described in this article and what they are called by various authors in their publications. (NB no one recommends word-for-word, so I haven’t included it!)
| prosodic unit | idea | Jones | salami (full, short sentences) | stalling | |
|---|---|---|---|---|---|
| Gillies | prosodic unit, chunking | – | Jones | salami | stalling |
| Jones | – | – | When-to-start | – | – |
| Seleskovitch & Lederer | sub-segment, idea | sub-segment, idea | – | – | – |
| Setton & Dawrant | chunk, segment, salami, unit-of-meaning | – | chunking, segmentation | – | stalling, padding, fillers |
| Tiselius | chunking, segmentation, salami | – | salami | – | – |
| Llewellyn Smith | salami, unit of meaning | – | salami | salami | – |
Footnotes
- Encyclopedia of Interpreting Studies, Routledge, p419 ↩
- Encyclopedia of Interpreting Studies, Routledge, p419. ↩
- Seleskovitch & Lederer, Pédagogie raisonnée de l’interprétation, p160. ↩
- Jones, Roderick, Conference Interpreting Explained, 2002. ↩
- Setton & Dawrant, Conference Interpreting – a complete course, 2016, p288. ↩
- Guillemin-Flescher, Syntaxe comparée du français et de l’anglais – problèmes de traduction, 1981. ↩
- Visson, L. From Russian into English – An Introduction to Simultaneous Interpretation, 2nd edn, 1999, Focus Publishing, p13. ↩
- Lederer, M., Unité de traduction ou unité de sens ? Du linguistique au cognitif, 2018, p37. ↩
- Seleskovitch & Lederer, Pédagogie raisonnée de l’interprétation, p264. ↩
- Seleskovitch & Lederer, Pédagogie raisonnée de l’interprétation, 1989, p158. ↩
- Seleskovitch & Lederer, Pédagogie raisonnée de l’interprétation, 1989, p158. ↩
- Setton & Dawrant, Conference Interpreting – a complete course, 2016. ↩
- Setton & Dawrant, Conference Interpreting – a complete course, 2016, p473. ↩
- Kader S. & Seubert S., ‘Anticipation, segmentation … Stalling? How to teach interpreting strategies’, in D. Andres & M. Behr (eds) To Know How to Suggest …: Approaches to Teaching Conference Interpreting, Berlin: Frank & Timme, 2015, pp. 125–44. ↩
- Du Bois, J.W., Transcription Design Principles for Spoken Discourse Research ↩
- Speech by Barry Andrews, Energy: europarl.europa.eu. ↩
- Visson, L. From Russian into English – An Introduction to Simultaneous Interpretation, 2nd edn, 1999, Focus Publishing, p123. ↩
- Setton & Dawrant, Conference Interpreting – a complete course, 2016. ↩
- Seleskovitch & Lederer, Pédagogie raisonnée de l’interprétation, 1989, p159. ↩
- Jones, R., Conference Interpreting Explained, 2nd ed. 2002. ↩
- Speech by Pascal Canfin: europarl.europa.eu. ↩
- Speech by Reinier Van Lanschot: europarl.europa.eu. ↩
- Setton, Progression in SI Training, 2008, FORUM Revue internationale d’interprétation et de traduction / International Journal of Interpretation and Translation 6(2):173-193. ↩
- Holcomb: prezi.com; Lavrenchuk: scribd.com. ↩
top image: generated by Gemini AI





