What it is
Most of the equivalence literature describes. This part of it decides. Šarčević's framework is an acceptability test: it does not merely report how close two concepts are, it says whether a translator is permitted to use a target-system term for a source-system concept at all.
The procedure runs in three stages. First, a conceptual analysis that identifies the essential and the accidental characteristics of the source concept, and then of the candidate target concept. Second, a comparison that classifies the pair as near equivalence, partial equivalence or non-equivalence. Third, an acceptability judgement, and this is where the test bites. Near equivalence is always acceptable. Non-equivalence is never acceptable. Partial equivalence, which is where most real pairs land, is acceptable only where the two terms correspond in structure and classification, in scope of application, and in legal effects.
That last triple is what distinguishes the test from the scale it sits on. A pair can share most of its characteristics and still fail, because legal effects are not a characteristic among others: they are the thing the term is for.
Where the authorities stand
The framework is cited constantly and the book itself is not open, which has produced a disagreement in the secondary literature worth setting out, because a reviser who meets both versions needs to know what happened.
Wilson reports Šarčević as identifying seven degrees of functional equivalence, ranging from full equivalence through partial equivalence to none. Doczekalska, Kusik, Way and Vezzani all report three, the near, partial and non-equivalence triad above. The two counts are not a contradiction. Wilson's own footnote anchors her seven to pp. 238 and 239 of the book, where Šarčević is expanding a classification of Lane's, while the sources reporting three cite pp. 236 and 237, where Šarčević sets out her own. Duběda supplies the reconciling observation: Šarčević classifies by type and by degree together, which is exactly how one scheme gets recounted as another.
The honest thing to say about all of this is that nobody in the chain has read the book to check. Every page number in circulation, including the ones above, comes from a source citing a source. The substance is stable across all of them; the anchoring is not.
Two things the test is often asked to do and does not. It does not tell a translator what to write when the answer is that no target-system term is admissible, which is a separate problem with its own record. And it does not settle by itself whether the text is one in which admissibility is even the right question, because the criteria assume a text where legal effects can be compared.
Consequences for translation
The test is worth running because its output is a permission or a refusal, which is something a translator can put in front of a reviser or a client.
- Compare essential characteristics before comparing anything else. A candidate that fails on an essential characteristic is out, and no weight of usage or dictionary support rescues it.
- The three partial-equivalence criteria are a checklist, not a mood. Structure and classification, scope of application, legal effects: each can be answered from the two legal systems, and each answer can be recorded.
- A rejection is a deliverable. Recording why a plausible candidate was refused is worth more to the next translator than recording the one that was chosen, because the plausible candidate is what they will reach for.
- Where the test refuses everything, that is a finding about the concept pair, not about the translator, and it should be reported as such rather than absorbed silently into a compromise rendering.
Examples
PL>EN
- Source: spółka cywilna
- Target: civil law partnership
- Comment: Kusik runs the test in public. Civil partnership is rejected because in English law it names a union of two people as an alternative to marriage. General partnership is rejected on an essential characteristic: a general partnership exists to do business for profit, and the Polish company need not. Non-trading partnership is rejected as well, since the Polish company is very often a business vehicle. What he recommends is civil law partnership or Civil Code partnership, on the ground that they carry the distinctiveness of the concept and point to where it is regulated.
EN>PL
- Source: going equipped
- Target: no admissible target-system term
- Comment: Non-equivalence, and therefore an automatic refusal rather than a close call. Wilson notes that the offence under section 25 of the Theft Act 1968 has no counterpart in the Polish criminal code. The test returns the same answer as the scale here, which is the uninteresting case; its value lies in the pairs where the two answers differ.
References
- Šarčević, S. (1997). New Approach to Legal Translation. The Hague: Kluwer Law International, pp. 235 to 262.
- Doczekalska, A. (2013). "Comparative Law and Legal Translation in the Search for Functional Equivalents: Intertwined or Separate Domains?" Comparative Legilinguistics 16, 63 to 75.
- Duběda, T. (2021). "Direction-Asymmetric Equivalence in Legal Translation." Comparative Legilinguistics 47, 57 to 72, at p. 59.
- Kusik, P. (2022). "English Translation Equivalents of Selected Polish Partnership Types Revisited from the Perspective of Comparative Law." Lingua Legis 30, 7 to 21, at pp. 9 and 16 to 18.
- Way, C. (2023). "Legal translator terminology training: Unravelling the mysteries." In Ł. Biel and H. J. Kockaert (eds), Handbook of Terminology, Volume 3: Legal Terminology, 594 to 618. Amsterdam: John Benjamins.
- Wilson, P. E. (2023). "Comparative law outside the ivory tower: an interdisciplinary perspective." Legal Studies 43(4), 641 to 657, at pp. 646 and 648.