Syntactic Differences between European Parliament Interpreted and Translated English: How Interpreters Package Information for Listeners (Wu, 2024)

Conference English #4: KISS = Keep It Short and “Stupid"? Rethinking Spoken Clarity

In short: In this open access article, published in T&I Review 14(2) (December 2024), I compare the syntax of simultaneously interpreted English and translated English produced from the same European Parliament (EP) plenary speeches. I examined five constructions that move information to the beginning or end of a clause: the passive, it-extraposition, existential there, clefts, and inversion. The passive and it-extraposition are significantly more frequent in translation. Existential there and clefts are more frequent in interpretation, although the difference is not statistically significant. The examples show what lies behind the numbers: interpreters break heavy, written-style sentences into smaller units that move from Given to New information. This makes the output easier for interpreters to produce and easier for listeners to process.

Key facts at a glance

• Article: Wu, Y. (2024). Syntactic differences between European Parliament interpreted and translated English. T&I Review, 14(2), 7–30.
• DOI: https://doi.org/10.22962/tnirvw.2024.14.2.001 (open access)
• Journal: T&I Review, published by the Graduate School of Translation and Interpretation (GSIT), Ewha Womans University, Seoul; indexed in the Korea Citation Index (KCI).
• Research question: Do the syntactic differences between speech and writing found in non-mediated discourse also appear in mediated discourse, that is, in interpretation and translation?
• Interpreted English corpus: 35 simultaneous interpretations into English (18 from Spanish, 17 from Italian), 18,480 words, taken from the European Parliament Interpretation Corpus (EPIC), plenary sessions of February 2004.
• Translated English corpus: the English translations of the same 35 speeches from the EP's final verbatim reports of proceedings, 19,881 words.
• Source speeches behind the interpretations: read-out (16 speeches, 57% of words), impromptu (11 speeches, 22%), and mixed (8 speeches, 21%).
• Method: manual identification of five constructions in both corpora, Chi-square goodness-of-fit tests (SPSS 29), and qualitative comparison of aligned translation–interpretation pairs.
• Framework: Theme/Rheme and Given/New information structure (Systemic Functional Linguistics), building on Carter-Thomas and Rowley-Jolivet (2001) and Rowley-Jolivet and Carter-Thomas (2005).

Why compare interpreting and translation at the level of syntax?

Corpus studies comparing interpreting and translation (often called intermodal studies) have mostly looked at vocabulary: lexical variety, lexical density, core vocabulary coverage, and parts of speech. They generally find that interpreted language is more simplified than both spoken originals and translation. Syntax has received much less attention. In those studies the only syntactic measure was mean sentence length, and interpreted sentences turned out to be shorter.

Shorter sentences, however, do not tell us how information is organised. My starting point was a pair of studies on scientific discourse. Carter-Thomas and Rowley-Jolivet (2001) compared nine oral presentations by native English-speaking physicists at an international conference with the proceedings articles based on them. They found that five constructions are strongly mode-dependent. The passive and it-extraposition belong to writing, while existential there, clefts, and inversion belong to speech. In a follow-up study, Rowley-Jolivet and Carter-Thomas (2005) showed that non-native speaker scientists, although proficient in English, kept using written-style structures in their talks. They overused the passive and it-extraposition and underused clefts and inversion, the structures that break information into chunks listeners can process in real time.

My study asks whether the same mode differences appear when the speaker and the writer are not presenting their own ideas but mediating someone else's: interpreters speaking, and translators writing, from the same EP speeches.

Key concepts: Theme, Rheme, Given, New, end-focus, end-weight

In English, the Theme usually comes at the start of the clause and serves as its point of departure. It tends to carry Given information, which the listener can recover from context. The Rheme, the rest of the clause, tends to carry New information, the newsworthy part. Placing the most important information at the end is called end-focus. Placing grammatically heavy material at the end is called end-weight. Both principles make a message easier to plan for the speaker and easier to decode for the listener. The five constructions in this study are syntactic resources English uses to shift elements towards the beginning or the end of the clause.

Quantitative results

• Passive: 274 in interpretation vs 444 in translation. Significantly more frequent in translation, χ²(1, N = 718) = 40.25, p < .001, Cramér's V = .24.
• It-extraposition: 36 in interpretation vs 83 in translation. Significantly more frequent in translation, χ²(1, N = 119) = 18.56, p < .001, Cramér's V = .39.
• Existential there: 66 in interpretation vs 51 in translation. More frequent in interpretation, but not significant, χ²(1, N = 117) = 1.92, p = .166.
• Clefts: 32 in interpretation vs 23 in translation. More frequent in interpretation, but not significant, χ²(1, N = 55) = 1.47, p = .225.
• Inversion: 7 in interpretation vs 13 in translation. Rare overall and not significant, χ²(1, N = 20) = 1.8, p = .180. This is the only construction that runs opposite to Carter-Thomas and Rowley-Jolivet (2001), where inversion was typical of speech.

The passive and it-extraposition are thus strongly associated with the written mode, in line with the earlier findings on scientific discourse. The trends for existential there and clefts also match the earlier findings, although they did not reach statistical significance in this corpus.

How interpreters repackage information: qualitative findings

In the examples below, T is the translation and I is the interpretation of the same passage.

1. The passive: from heavy noun subjects to "we" + active verb

Of the passives in translation, 39% have the same or a similar passive in interpretation, 30% are omitted, and 31% become active voice. In translation, sentences often open with long, precise noun groups as subjects. This condenses a lot of information into the Theme and keeps the text impersonal. In speech, such subjects are hard to produce under time pressure and hard to process by ear. Interpreters instead tend to use a personal pronoun subject followed by an active verb.

T: The material adopted on 21 January already contained the main political messages and key conclusions. A series of positive results was mentioned, notably concerning job creation. Six million jobs have been created despite the climate of weak growth.

I: What we adopted on the twenty-first of January already includes the main political messages we want to send out and also the main conclusions. We praise positive results, particularly the creation of six million jobs, despite the weak rate of growth.

The interpreter opens with a wh-cleft ("What we adopted…") instead of a post-modified noun phrase, then frames the message with "we praise". The subject and verb sit close together, and the parts with the highest information value move to the end of the clause, respecting end-weight. Passives do occur in interpretation too: 63% of them have the same or a similar passive in translation. But in interpretation they are not preceded by heavy, complex noun groups as subjects.

2. It-extraposition: from subtle evaluation to direct stance

Every case of it-extraposition in the translation corpus carries an evaluative element, such as likelihood, importance, or necessity. It lets the writer comment on a proposition in a subtle, impersonal way. In interpretation, 42% of these cases become personal pronoun constructions. Some typical correspondences:

• it is possible to → I'm sure we can / you can
• it is essential to → we must / we need to / we have to
• it is important to → we must / we need to / you've got to
• it is necessary to → we must / we need to
• it should be recalled that → we should also bear in mind that
• it needs to be borne in mind that → we need to remember that / don't forget
• it is not possible to → you can't say

None of the extraposition cases in interpretation uses a passive, while 16% of those in translation do (for example, "it must be said that"). Combined with the passive, extraposition can also soften criticism:

T: It must be said that the Commission's conduct, marked by inaction and complicity, has been deplorable.

I: And moreover, the Commission has failed to take action on this.

When extraposition does appear in interpretation, it serves end-weight: the empty "it" takes the Theme position and the heavy clause is pushed to the end.

3. Existential there: delaying New information

Existential there places a semantically empty word in the Theme position, so the New information can come later in the clause. In interpretation, it often replaces a translated sentence that opens with a specific noun:

T: The results are encouraging. / I: There have been very good results.

T: A day should be set aside to honour the victims of terrorism. / I: There should be an international day for the victims of terrorism.

T: The saying that the road to hell is paved with good intentions is common to all our languages.
I: In all our languages, there's a saying that the road to hell is paved with good intentions.

In the last example, the interpreter turns one sentence with a long clausal subject into three short units. Existential there may also work as a stalling device, giving the interpreter time to process more of the source speech. When the translation itself uses there, the interpretation is often even more direct, for example "There is no need to keep on reinventing the wheel" (T) vs "We shouldn't reinvent the wheel every other day" (I).

4. Clefts: springboards and signposts

Clefts split one clause into two parts: a background part with Given information and a focused part with New information. The corpus contains 12 wh-clefts in interpretation vs 7 in translation, 14 reverse wh-clefts vs 8, and 6 it-clefts vs 8. Wh-clauses such as "what I'm saying is" or "what we need is" act as springboards for starting an utterance, and they can replace complex noun phrases:

T: …and the margin for manoeuvre is indeed limited.

I: …and what we can do is limited.

Reverse wh-clefts such as "that is why" link back to the previous idea and give it focus. In one passage the interpreter uses "that is why" twice, where the translation uses "hence" and "for the same reason".

5. Inversion: rare in this setting

Only 20 inversions were found, 5 subject–verb and 15 subject–operator (for example, "Only then can…"). In the physics conference talks studied by Carter-Thomas and Rowley-Jolivet, subject–verb inversion often referred to slides ("Here's a picture of…", "On the left are…"). EP plenary speakers rarely use visual aids, which may explain why the construction is uncommon here.

What this means for interpreter training and public speaking

Interpreter training often presents the "salami technique", slicing long sentences into shorter pieces, as a way to manage the interpreter's own cognitive load. This study supports a broader view. Repackaging information into smaller, Given-to-New units helps the interpreter produce speech in real time, and it also helps the audience process it. The same principle applies to anyone speaking English in public, especially speakers of English as an additional language, who may assume that more advanced English means longer and more nested sentences.

The aligned examples can be used directly in class. Students can practise replacing heavy noun subjects with "we" + active verb, turning "it is essential to" into "we must", using "there is/are" to delay new information, and using wh-clefts ("what we need is…") or "that is why" to frame and link ideas.

Limitations

The two corpora are small and come from parliamentary debate, which is political in nature, so the statistical results cannot be generalised to all interpreting and translation. Some differences may also reflect the source texts: the interpretations were based on read-out, impromptu, or mixed speeches, while the translations were based on revised written versions of those speeches. The study covers only five constructions.

How this study connects to my other work

My earlier studies of European Parliament English focused on vocabulary: lexical bundles (Wu, 2021), phrasal verbs (Wu, 2023), and a pedagogical list of multiword constructions (Wu, 2024, The Interpreter and Translator Trainer). This article moves from words to syntax and discourse. Looking back, the information structure perspective also offers a way to explain why multiword constructions help interpreters: they may help speakers follow the Given-to-New principle and place heavy material at the end of the clause, which reduces their own processing load and makes speech easier to follow. This connection is my later reflection and is not argued in the original article.

Frequently asked questions

What is the main finding of this study?
Interpreted and translated English from the same EP speeches package information differently. The passive and it-extraposition are significantly more frequent in translation. Existential there and clefts are more frequent in interpretation, though not significantly. This mirrors the difference between spoken and written scientific discourse.

Do interpreters simply use shorter sentences?
Not only. They reorganise information: they replace heavy noun subjects with pronouns and active verbs, express evaluation as direct stance ("we must" instead of "it is essential to"), and use structures that present one unit of information at a time.

Do interpreters avoid the passive?
No. The interpretation corpus contains 274 passives, and 63% of them match a passive in translation. The difference is that passives in interpretation are not preceded by long, complex subjects.

Which corpus was used?
The interpreted English data come from the European Parliament Interpretation Corpus (EPIC), distributed by ELRA. The translated English data come from the EP's final translated verbatim reports of the same February 2004 plenary sessions.

Who might find this article useful?
Interpreter trainers, interpreting students working into English as a B language, teachers of English for specific purposes and public speaking, and researchers in corpus-based translation and interpreting studies.

Is the article free to read?
Yes. It is open access via its DOI: https://doi.org/10.22962/tnirvw.2024.14.2.001

Citation

Wu, Y. (2024). Syntactic differences between European Parliament interpreted and translated English. T&I Review, 14(2), 7–30. https://doi.org/10.22962/tnirvw.2024.14.2.001

Related publications

• Wu, Y. (2021). Lexical bundles in English EU parliamentary discourse: Variation across interpreted, translated, and spoken registers. Compilation and Translation Review, 14(2), 37–86. https://doi.org/10.29912/CTR.202109_14(2).0002
• Wu, Y. (2023). Phrasal verbs in European Parliament conference English: A corpus-based pedagogical list. The Interpreter and Translator Trainer, 17(2), 301–318. https://doi.org/10.1080/1750399X.2023.2183452
• Wu, Y. (2024). Multiword constructions in European Parliament conference English: A corpus-driven pedagogical list. The Interpreter and Translator Trainer, 18(3), 423–441. https://doi.org/10.1080/1750399X.2024.2334582

About the author

Yinyin Wu teaches at the Graduate Program in Translation and Interpretation, National Taiwan University, Taipei. Her research interests include conference interpreting, retour interpreting, B language enhancement, interpreting pedagogy, and corpus linguistics, with a focus on the linguistic features of European Parliament conference English. ORCID: 0000-0003-4877-7719