Corpus-based translation studies hypothesise that translated language differs systematically from originally produced language in the same language, showing explicitation, simplification, normalisation, and levelling out (reduced register variation). Kruger and Van Rooy (2012) tested operationalisations of these features simultaneously across six registers in a comparable corpus of translated and original South African English, finding significant translation effects for only two measures (optional that-complementiser use and lexical variety) and no support for levelling out; they themselves identified the register-feature interaction, uncontrolled translator expertise and source-language influence, unbalanced corpora, and generalisability beyond South African English as unresolved. The problem is to establish the relationship between register variation and the features of translated language: whether translation effects are uniform across registers or interact systematically with register, and whether translated language shows reduced register variation, in register-balanced comparable corpora with the expertise and source-language confounds controlled, and whether the pattern generalises beyond translated South African English. Follow-up work from 2015 to 2026 has shown that translationese properties differ by register in several language pairs and has produced a second negative synchronic test of levelling out alongside one positive diachronic test, but no register-balanced multivariate adjudication exists; the question remains open.

Background

"Translation universals", more neutrally called features of translated language, are hypothesised recurrent properties of translated texts, proposed by Baker (1996) to hold irrespective of the source and target languages involved; the cumulative linguistic profile that translated texts are claimed to share is often referred to as translationese. The four classical features are: explicitation (a tendency to spell out in translation what original texts leave implicit), simplification (translated language is claimed to be less lexically and structurally varied; Laviosa (1998), comparing translated and original English narrative prose in a comparable corpus — a corpus designed so that translated texts and originally written texts in the same language can be directly compared — found lower lexical variety in the translations), normalisation or conventionalisation (a tendency to conform to typical, conservative target-language patterns, e.g. fewer contracted forms, coinages and loanwords), and levelling out (a tendency for translated texts to gravitate towards the middle of the register continuum, that is, to show less register variation than original language). A register is a variety of language associated with a situational context and communicative purpose, such as academic prose, news reportage, or popular writing. Explicitation has frequently been operationalised as the rate of the optional complementiser "that" in finite complement clauses (as in "she said (that) it was late"), following Olohan and Baker (2000), who found that translated English uses the optional that more often than original English.

Kruger and Van Rooy (2012) investigated these features together — rather than one at a time, as most earlier studies had — in a comparable corpus of translated and original South African English: the Corpus of Translated English (CTE, about 271,000 words, translated predominantly from Afrikaans) and the South African component of the International Corpus of English (ICE-SA), each spanning six registers. For each feature operationalisation they fitted a factorial analysis of variance with corpus (translated vs original) and register as factors and examined the corpus-by-register interaction. Only two operationalisations showed significant corpus effects (translated texts omitted the optional that less often and had lower lexical variety), and, crucially, there was no interaction pattern consistent with levelling out: register variation was not reduced in the translated corpus. The study's own stated limitations define the open question: the registers were unequally represented, the source language was overwhelmingly Afrikaans (whose complementiser system differs from English), translator expertise and individual style were not controlled, the operationalisations may need refinement, and it is unknown whether the effects are specific to translated South African English or generalisable to other varieties and language pairs. What is at stake scientifically is whether "translated language" is a coherent unified variety at all, or whether the features attributed to it are artefacts of register composition that disappear or change once register is systematically factored in.

Problem Statement

Determine the relationship between register variation and the hypothesised features of translated language, as raised by Kruger and Van Rooy (2012). In register-balanced comparable corpora of translated and non-translated language in which translator expertise, individual translator style, and source-language influence are controlled or explicitly modelled, answer three jointly constitutive questions: (1) whether the linguistic operationalisations of the features of translated language — explicitation (e.g. optional that-complementiser use), simplification (e.g. lexical variety, word length), and normalisation (e.g. contracted forms, coinages and loanwords, lexical bundles) — show a uniform translation effect across registers, or whether the effects interact systematically with register; (2) whether translated language exhibits levelling out, i.e. significantly reduced register variation relative to comparable original language, as the translation-universals hypothesis predicts, or whether register variation is preserved in translation; and (3) whether the pattern established for translated South African English generalises to other varieties of English and other language pairs. These three questions are one scientific objective — characterising how the translation process and register jointly shape translated language — and an answer must address them together: the corpus-by-register interaction tests (the feature-by-register profile of translation effects) and the register-variance comparison (levelling out) are two faces of the same relationship, and the generalisation test establishes its scope. An answer that addresses a single feature, a single register, or a single language pair alone does not answer the question as stated.

References

  1. Kruger, Haidee, and Bertus van Rooy (2012). Register and the features of translated language. Across Languages and Cultures 13(1): 33–65. DOI 10.1556/acr.13.2012.1.3, https://doi.org/10.1556/acr.13.2012.1.3
  2. Baker, Mona (1996). Corpus-based translation studies. In: Terminology, LSP and Translation (Benjamins Translation Library 18), pp. 175–186. DOI 10.1075/btl.18.17bak, https://doi.org/10.1075/btl.18.17bak
  3. Olohan, Maeve, and Mona Baker (2000). Reporting that in translated English. Evidence for subconscious processes of explicitation? Across Languages and Cultures 1(2): 141–158. DOI 10.1556/acr.1.2000.2.1, https://doi.org/10.1556/acr.1.2000.2.1
  4. Laviosa, Sara (1998). Core patterns of lexical use in a comparable corpus of English narrative prose. Meta 43(4): 557–570. DOI 10.7202/003425ar, https://doi.org/10.7202/003425ar
  5. Redelinghuys, Karien, and Haidee Kruger (2015). Using the features of translated language to investigate translation expertise. International Journal of Corpus Linguistics 20(3): 293–325. DOI 10.1075/ijcl.20.3.02red, https://doi.org/10.1075/ijcl.20.3.02red
  6. Redelinghuys, Karien (2016). Levelling-out and register variation in the translations of experienced and inexperienced translators: a corpus-based study. Stellenbosch Papers in Linguistics 45. DOI 10.5774/45-0-198, https://doi.org/10.5774/45-0-198
  7. Kruger, Haidee, and Bertus van Rooy (2016). Constrained language: A multidimensional analysis of translated English and a non-native indigenised variety of English. English World-Wide 37(1): 26–57. DOI 10.1075/eww.37.1.02kru, https://doi.org/10.1075/eww.37.1.02kru
  8. Prieels, Lynn, Isabelle Delaere, Koen Plevoets, and Gert De Sutter (2015). A corpus-based multivariate analysis of linguistic norm-adherence in audiovisual and written translation. Across Languages and Cultures 16(2): 209–231. DOI 10.1556/084.2015.16.2.4, https://doi.org/10.1556/084.2015.16.2.4
  9. Lapshinova-Koltunski, Ekaterina, and Mihaela Vela (2015). Measuring 'Registerness' in Human and Machine Translation: A Text Classification Approach. In: Proceedings of the Second Workshop on Discourse in Machine Translation (DiscoMT), pp. 122–131. DOI 10.18653/v1/W15-2517, https://doi.org/10.18653/v1/W15-2517
  10. Kruger, Haidee, and Bertus van Rooy (2018). Register variation in written contact varieties of English. English World-Wide 39(2): 214–242. DOI 10.1075/eww.00011.kru, https://doi.org/10.1075/eww.00011.kru
  11. Kruger, Haidee (2019). That Again: A Multivariate Analysis of the Factors Conditioning Syntactic Explicitness in Translated English. Across Languages and Cultures 20(1): 1–33. DOI 10.1556/084.001, https://doi.org/10.1556/084.001
  12. Vandevoorde, Lore (2018). Register, Source Language, and Cognateness Effects on Lexical Choice in Translated Dutch. Meta 63(3): 627–648. DOI 10.7202/1060166ar, https://doi.org/10.7202/1060166ar
  13. De Sutter, Gert, and Marie-Aude Lefer (2020). On the need for a new research agenda for corpus-based translation studies: a multi-methodological, multifactorial and interdisciplinary approach. Perspectives 28(1): 1–23. DOI 10.1080/0907676X.2019.1611891, https://doi.org/10.1080/0907676X.2019.1611891
  14. Pontrandolfo, Gianluca (2020). Testing Out Translation Universals in Legal Translation: Quantitative Insights From A Parallel Corpus of Spanish Constitutional Court's Judgments Translated Into English. Comparative Legilinguistics 43: 17–55. DOI 10.2478/cl-2020-0007, https://doi.org/10.2478/cl-2020-0007
  15. Kunilovskaya, Maria, and Gloria Corpas Pastor (2021). Translationese and Register Variation in English-To-Russian Professional Translation. In: New Frontiers in Translation Studies, pp. 133–180. DOI 10.1007/978-981-16-4918-9_6, https://doi.org/10.1007/978-981-16-4918-9_6
  16. Pang, Shuangzi, and Kefei Wang (2025). What transcends "translation universals" across time? Babel 71(4): 449–477. DOI 10.1075/babel.24133.pan, https://doi.org/10.1075/babel.24133.pan
  17. Afzaal, Muhammad, Baorong Huang, and Dina Abdel Salam El-Dakhs (2025). Decoding the digital: a corpus-based study of simplifications and other translation universals in translated texts. Frontiers in Psychology 16. DOI 10.3389/fpsyg.2025.1517107, https://doi.org/10.3389/fpsyg.2025.1517107
  18. Jiang, Zeyuan (2025). Does LLM translation align with translation universals? A cross-genre simplification study on English-Chinese translation based on dependency grammar. PLOS ONE 20(6). DOI 10.1371/journal.pone.0324830, https://doi.org/10.1371/journal.pone.0324830
  19. Liu, Yanmeng (2026). Translation as language contact: a multidimensional perspective of syntactic variation. Frontiers in Language Sciences 4. DOI 10.3389/flang.2025.1712405, https://doi.org/10.3389/flang.2025.1712405