Multiword Constructions in European Parliament Conference English: A Corpus-Driven Pedagogical List for Simultaneous Interpreting (Wu, 2024, Open Access)

Conference English #3: Open Grammar Devices in Simultaneous Interpreting

Start free, stay free.

In short: In this article, published in The Interpreter and Translator Trainer 18(3) (2024), I compiled the European Parliament Conference English Multiword Construction Pedagogical List (EP-CE MWC List). It contains 232 multiword constructions (MWCs) drawn from a 664,732-word corpus of English spoken in European Parliament plenary sessions, covering both interpretations into English and original English speeches. The list is designed for interpreters working into English as a B language (retour interpreting). Its most practical part is 118 "Less Specific" MWCs, which I group into Heads, Hinges, and Highlighters. These work as open grammar devices, padding, and tone-softeners in simultaneous interpreting (SI). The full list (Appendix A) and an unannotated interpreted speech for classroom use (Appendix B) are freely available with the article.

Key facts at a glance

• Article: Wu, Y. (2024). Multiword constructions in European Parliament conference English: a corpus-driven pedagogical list. The Interpreter and Translator Trainer, 18(3), 423–441.
• DOI: https://doi.org/10.1080/1750399X.2024.2334582 (open access, CC BY-NC-ND 4.0)
• Corpus: the EP Conference English Corpus (EP-CEC), 664,732 words in 1,770 speeches. It was built from the TC-STAR transcriptions of European Parliament plenary sessions held from May 2004 to January 2005.
• Interpreted English: 1,410 speeches (517,872 words, 78% of the corpus), interpreted from 19 source languages.
• Original English: 360 speeches (146,860 words, 22%), mostly delivered by native speakers.
• Extraction criteria: four-word lexical bundles occurring at least 13 times (20 per million words) in at least five different speeches, identified with WordSmith Tools 7.
• Result: 645 four-word bundles, combined and collapsed into 232 MWCs (about one third of the original number).
• Specific MWCs (114): 12 meaning-based groups, e.g. EU Institutions, People, Procedures, Rules and Regulations, Time, Action.
• Less Specific MWCs (118): 53 Heads, 28 Hinges, and 37 Highlighters.

Why ready-made chunks matter for interpreters working into English

With English as a lingua franca, many conferences now use only two languages: the host language and English. Interpreters are therefore often expected to work into English even when it is not their native language. Interpreter trainers such as Gile, Jones, and Setton and Dawrant advise trainees to memorise fixed and semi-fixed expressions in their B language. These expressions are available immediately, so they reduce production effort, and they make the output more fluent and idiomatic.

Research on formulaic language points the same way. Prefabricated chunks are stored and retrieved whole from memory. They make speech more fluent for speakers and easier to process for listeners, because a familiar frame tells the listener where the sentence is going. They also mark the speaker as a member of a professional community. EU interpreters, for example, pick up phrases from senior boothmates and build a shared repertoire.

What was missing was an empirically derived list of such chunks for conference interpreting. The EP-CE MWC List supplements the stock phrases trainers offer from experience with phrases that professional interpreters and Members of the European Parliament actually use again and again.

From lexical bundles to multiword constructions

Four-word lexical bundles are often structurally incomplete (it comes to the). Different bundles also overlap (when it comes to, it comes to the). To make the list teachable, I combined overlapping bundles into larger MWCs, each with a root structure and variable slots. I distinguish three types of overlap:

• Complete overlap: two bundles come from one five-word sequence, e.g. those in favour please + in favour please show → those in favour please show.
• Complete subsumption: a higher-frequency bundle absorbs a lower-frequency one, e.g. when it comes to (141) + it comes to the (29) → when it comes to the.
• Partial subsumption: a measure created in this study for bundles that only partially overlap, e.g. the work of the + of the Irish presidency → the work of the Irish presidency.

Variable slots are shown with options separated by slashes. One example is when/if/here we are ['re] talk[ing] about a/the, which merges seven bundles. Frequent context words are added in parentheses, as in (this) is the best way (to).

Heads, Hinges, and Highlighters: categories built for simultaneous interpreting

Earlier bundle studies used functional categories such as "stance bundles" and "referential bundles" (Biber, Conrad, and Cortes 2004). Those labels describe bundles for researchers. The EP-CE MWC List instead groups the Less Specific MWCs by what they let an interpreter do in real time:

• Heads (53) open a sentence. They have a subject (mostly I or we) and a verb, and usually end in that or to. Subgroups are Request (25), Stance (21), and Negative Stance (7). Examples: we need to make sure that…, I'd like to take this opportunity to…, we have to bear in mind that…, I don't think we should…
• Hinges (28) bridge two clauses or verb phrases. Subgroups are Neutral (13) and Cause/Result (15). Examples: at a time when…, this means that…, with a view to…, in such a way that…, be in a position to…
• Highlighters (37) put a spotlight on the noun phrase that follows. They usually take the form "preposition + the + (blank) + of the". Subgroups are Topic (24), People/Institutions (5), Cause/Result (4), and Connector (4). Examples: when it comes to the…, in the light of the…, on the question of…, on the part of…

What is open grammar?

Setton and Dawrant (2016) describe open grammar as the ability to start a sentence in a way that keeps options open, and then either continue it or wind it up without being trapped by the speaker's syntax. In SI, a long source sentence can be sliced into short, self-contained units. The interpreter then avoids both backtracking and falling far behind the speaker. Heads, Hinges, and Highlighters are ready-made tools for this. They can also serve as padding, buying time when a speaker is slow or unclear, and as tone-softeners for suggestions and disagreement.

Example 1: a Highlighter sets up the topic first

Here is a real European Parliament interpretation. List items are in brackets:

[We also need to] recognise that [when it comes to the] most important issues, we do have an overlap of interests. Be it [in the case of] terrorism, be it [the fight against] nuclear proliferation, be it [in the case of] our climate situation.

A more succinct version would be: Our interests overlap on important issues like the fight against terrorism and nuclear proliferation, and our climate situation. The Head we also need to buys time. When it comes to the turns the topic into a stand-alone unit before the comment. This is similar to the SI strategy of "framing and filling", and it is especially handy when the source language is topic-prominent, as Chinese is. Each list item is rendered as it arrives, without the interpreter needing to know how long the list will be.

Example 2: parallel Heads and a Hinge

This example comes from the English interpretation of a speech delivered in German by Helmut Scholz on 5 July 2021:

[We need to have] communication between peoples. [We need to make sure that we have] participation in elections, in a representative democracy. That's one thing. [But what we need even more perhaps is] daily involvement in debates — political, economic, social. [That is what] can actually contribute to us all shaping the future and the kind of society that we want.

One long idea becomes a string of short, complete sentences. Three parallel Heads set a rhythm that listeners can follow. The Hinge that is what lets the interpreter cut and restart, tying the new sentence back to the previous point.

A later insight: information structure

A later study of mine compared interpreted and translated English in the European Parliament (Wu 2024, T&I Review 14(2)). It helped me see what else these chunks do. They put something light and given at the start of the sentence (we need to, that is what) and push the heavy, new information to the end. This pattern follows the principles of given-before-new and end-weight. It makes the sentence easier for the interpreter to produce and easier for the audience to process.

A caveat for B-language interpreters

Less Specific MWCs are padding, and overusing them makes the output wordy. A native speaker says we need to make sure that in one light breath. Said slowly and heavily, or mumbled, the same chunk makes the interpreter lag behind the speaker and can crowd out the real message. Into a B language, these frames pay off only once they come out whole, fast, and clear. Until then, it is better to go straight to the point.

A three-step classroom task

The article proposes speech analysis as a complement to concordance-based activities:

1. Listen to the prosodic features of MWCs in a speech: intonation, pauses, and pace. Shadow the interpreter.
2. Highlight MWCs on the transcript.
3. Analyse how the MWCs frame ideas and develop the discourse.

The demonstration uses the five-minute English interpretation of Helmut Scholz's speech on citizens' participation in EU decision-making (European Parliament plenary, 5 July 2021). In this 743-word interpretation, 35 MWCs from the list account for 151 words, so at least one fifth of the speech is formulaic. Of the 35, six are Specific MWCs, 16 are Heads, three are Hinges, and 10 are Highlighters. Appendix B gives the unannotated transcript so that trainees can mark the chunks themselves first.

Frequently asked questions

What are multiword constructions (MWCs)?

MWCs are recurrent word sequences with a fixed root structure and variable slots. In this study, they were built by combining overlapping four-word lexical bundles into units that are more complete in structure and meaning.

Who is the EP-CE MWC List for?

It is for conference interpreters and trainees working into English (especially English as a B language), interpreter trainers, and teachers of English for specific or academic purposes.

Is a corpus from 2004 still relevant?

Some Specific MWCs are tied to their time; the fight against terrorism might today be the fight against Covid-19. Most Less Specific MWCs, however, are sentence frames rather than topical terms. The 2021 speech shows that they are still in active use.

How does this study relate to the author's other work?

It expands Wu (2021), a study of four-word lexical bundles across interpreted, translated, and spoken registers of EU parliamentary English, by using a much larger corpus and interpreter-oriented categories. It sits alongside Wu (2023) on phrasal verbs in European Parliament conference English, and Wu (2024, T&I Review) on syntax and information structure.

Where can I get the full list?

Appendix A of the open access article contains all 232 MWCs. The full list of 645 four-word bundles is available as supplemental material.

Citation

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

Open access under CC BY-NC-ND 4.0. Please cite the source when quoting.

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. (2024). Syntactic differences between European Parliament interpreted and translated English. T&I Review, 14(2), 7–30. https://www.kci.go.kr/kciportal/landing/article.kci?arti_id=ART003153959
• 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  Read my blog post on this study

About the author

Yinyin Wu teaches at the Graduate Program in Translation and Interpretation, National Taiwan University, Taiwan. Her research uses corpus methods to study conference English, interpreting, and B-language enhancement for interpreters. ORCID: 0000-0003-4877-7719 

This is the third post in the Conference English series, which moves from two-word phrasal verbs to longer word sequences and the rhetorical structure of whole speeches.