What it is
Most fixed wordings are written down nowhere as rules. They sit in earlier texts of the same kind, and the work is to find the earlier text rather than to compose the sentence.
Two different things answer to that description. One is the parallel text, which Schäffner defines as texts in the two languages of equal informativity produced in comparable communicative situations, consulted so that the target text can be brought into line with the text-typological conventions of the target culture. The other is a body's own back catalogue of what it has already issued, which is not a comparison at all but a record of decisions the house has taken and expects to go on taking.
The terminology record on sources of term evidence keeps the authority behind a term pair. This record keeps the authority behind a passage, and the two scales do not convert: a designation can be standardized while the sentence around it rests on nothing firmer than the last time somebody wrote it.
Where the fixity comes from
A body's own requirement, stated as a quality parameter rather than as a rule of language.
Biel lists among the parameters of institutional translation quality internal terminological consistency and consistency with reference materials, and consistency or continuity with preceding and higher-ranking texts. Hanzl and Beaven, on the Council's translation service, put it more strongly: consistency of terminology and phraseology inside one document and across documents is paramount, and the texts carry a large amount of repetition and interconnection, explicit and implicit. Continuity is not a translator's habit here. It is asked for, and it changes when the body revises what it asks for.
The tension is the record. What continuity points at is an archive, and the archive is by construction translated text. Biel, Koźbiał and Wasilewska, measuring lexical bundles across four institutional genres in Polish, found that translations develop formulaic profiles of their own, levelled out against the corresponding source-language corpora and overlapping only minimally with domestically drafted genres. The requirement to follow the archive and the finding that the archive is its own variety are both sound, and they do not cancel out. No figure is given here for how much prior wording is reused, because no study establishing one was found.
Consequences for translation
The archive is at once the fastest route to a defensible wording and the least reliable evidence about the language, and both have to be carried at the same time.
- A parallel text and an archive answer different questions. The first shows how the target culture writes this kind of text; the second shows how one body has written it before, and it is the second that institutional work asks for.
- An objection that a rendering departs from precedent names the precedent. Without it, a preference is being reported in the vocabulary of a finding.
- The warrant reaches as far as the house. Inside the body that keeps the archive a departure is a finding; outside it the same departure is a suggestion, and a client's archive is not the institutions'.
- A wording taken from the archive is not thereby the target community's usage. It is the usage of a text population, and where the readers sit outside that population the choice is worth stating rather than assuming.
- The empty approximation tier is a claim. This record covers the case where an earlier text holds the wording and the translator retrieves it. Where none holds it there is no model to approximate, and composing the substitute falls to the records that keep the unit itself.
Examples
EN>NL
- Source: Member States shall encourage the initial and further training of mediators
- Target: De lidstaten bevorderen de basis- en vervolgopleiding van bemiddelaars/mediators
- Comment: Directive 2008/52/EC, Article 4(2). The English offers no reason to prefer bevorderen over its neighbours, and none to keep it apart from the verb the same act uses at Article 6(1), where shall ensure becomes zorgen ervoor dat. The act holds the two apart because the acts before it did. A translator composing from the English would have one Dutch verb for both.
EN>PL
- Source: Member States shall encourage the initial and further training of mediators
- Target: Państwa członkowskie wspierają wstępne szkolenia mediatorów i kontynuację tych szkoleń
- Comment: Same provision. Polish keeps the same split, wspierają here against zapewniają at Article 6(1). Two target communities inheriting one distinction did not inherit it from the source, which marks it with two ordinary verbs; each took it from its own stock of earlier acts.
References
- Schäffner, C. (1998). "Parallel texts in translation." In L. Bowker, M. Cronin, D. Kenny and J. Pearson (eds), Unity in Diversity: Recent Trends in Translation Studies, 83 to 90. Manchester: St Jerome Publishing.
- Biel, Ł. (2017). "Quality in institutional EU translation: parameters, policies and practices." In T. Svoboda, Ł. Biel and K. Łoboda (eds), Quality Aspects in Institutional Translation, 31 to 57. Translation and Multilingual Natural Language Processing 8. Berlin: Language Science Press.
- Hanzl, J. and Beaven, J. (2017). "Quality assurance at the Council of the EU's Translation Service." In the same volume, 139 to 153.
- McEnery, A. and Xiao, Z. (2007). "Parallel and comparable corpora: What is happening?" Chapter 2 of Anderman, G. and Rogers, M. (eds), Incorporating Corpora: The Linguist and the Translator. Clevedon: Multilingual Matters. Pagination unconfirmed, and one library catalogue gives the year as 2008 against the publisher's own 2007; no statement on this page rests on either.
- Biel, Ł., Koźbiał, D. and Wasilewska, K. (2019). "The formulaicity of translations across EU institutional genres: a corpus-driven analysis of lexical bundles in translated and non-translated language." Translation Spaces 8(1), 67 to 92.
- Directive 2008/52/EC of the European Parliament and of the Council of 21 May 2008 on certain aspects of mediation in civil and commercial matters, Articles 4(2) and 6(1), English, Dutch and Polish language versions.