Interdisciplinary · Language · Translation
What must a translation preserve to remain faithful?
A translation can preserve the main meaning while changing tone, implication, politeness, cultural reference or the effect on its reader.
The question is therefore not whether two texts are identical, but what should be preserved when words, tone, presuppositions and context do not line up exactly across languages.
Purpose and scope
Compare without confusing equivalence with identity.
Translation studies distinguish lexical meaning, register, textual function, relations between speakers, presuppositions, implicatures and other pragmatic effects. This dossier uses those distinctions without proposing a general theory of language, claiming that every translation entails one global loss, or treating strong linguistic relativity as established.
Plurality of solutions ≠ absence of criteria. Difficult translation ≠ absolute untranslatability. Context ≠ total subjectivity.
Equivalence
The word is not always the right unit.
Mona Baker distinguishes analysis at word level, above the word, in grammar, text and pragmatics. A lexically close solution can fail on register, cohesion, reference or speech act, while a less literal solution can preserve a more important function.
Types of translation
Translation is not only movement between languages.
Jakobson distinguishes intralingual, interlingual and intersemiotic translation. Rewording within a language, translating between languages and transposing into another sign system do not involve exactly the same constraints.
Form and audience
Audience response is not an objective measure of “the same effect.”
Nida made the contrast between formal orientation and dynamic or functional equivalence influential. Here it is used as a historical and methodological reference, without assuming that identical reception across audiences can be measured simply.
Situation and function
Who speaks, to whom, where and for what purpose also matters.
Juliane House compares dimensions such as field, tenor, mode, genre and function. The model can reveal shifts in register or social relation even when propositional content looks similar. It is analytical and partly interpretive, not a universal metric.
Lexical case
“Speaker”: translatable, but not reducible to one word.
Baker uses the Speaker of the House of Commons as an example of an institutional concept that does not map neatly term for term. A nearby title may fail to carry all features of the British role. The concept is not absolutely untranslatable; it may require explanation, paraphrase or retention of a specialist term.
Pragmatics
What is said does not exhaust what is communicated.
Presupposition, implicature, deixis, speech acts, politeness and context shape interpretation. Explicitation can help the target reader while fixing an ambiguity too rigidly; omission can streamline a sentence while removing important pragmatic information.
Pragmatic case
Japanese → Cantonese: adding can sometimes compensate.
In Pei Chun Shih’s corpus of Cantonese dubbing of Japanese anime, forms of address absent from the source can be added to reconstruct some speech levels and politeness functions. This is partial pragmatic compensation, not perfect reproduction of the original social and linguistic systems.
Additional case
A public directive also carries tone and relationship.
Christina Lees’s Thessaloniki case shows that a translation can preserve an instruction or prohibition while altering mitigation, social distance or the appearance of authority. It remains a supplementary case, not a universal model.
Cultural context
Shared knowledge does not travel automatically with the text.
Institutions, holidays, historical references, religious practices, academic titles or administrative systems may be familiar to the source audience and absent for the target audience. Keeping, explaining, glossing, adapting or accepting partial loss serves different priorities depending on genre and function.
Fluency and foreignness
A very fluent translation can also erase differences.
Venuti’s concepts of domestication, foreignization and translator visibility are useful critical tools. This dossier does not turn them into automatic moral rules: preserving foreignness or adapting must be judged against function, audience and accepted losses.
Institutional case
Multilingual EU law requires more than “one word = one word.”
The Court of Justice treats the authentic language versions of EU law as equally authentic. One version cannot alone serve as the interpretive basis or automatically outrank the others. EU legal concepts may also diverge from national concepts that use similar words, so context and regulatory purpose matter.
Comparative terminology
VJM and IATE make types of legal equivalence visible.
The Comparative Multilingual Legal Vocabulary brings together material from several legal systems. Entries may include definitions, comparative notes, functional equivalents, paraphrases, false friends and warnings about likely confusion. VJM is available through IATE.
Explicitation
Helping the reader can also close an ambiguity.
Making a strongly constrained inference explicit can be useful; turning a plausible interpretation into a definite assertion is different. The dossier therefore distinguishes restored information, added information and overinterpretation.
Philosophical safeguard
Quine challenges the ideal of one uniquely correct translation, not the possibility of communication.
Indeterminacy of translation is used here to show that more than one translation scheme may fit the available evidence. It does not show that all translations are equal, that dictionaries are useless or that communication is impossible.
Errors and trade-offs
Not every difference is a legitimate choice.
The dossier separates referential, logical, grammatical, terminological and major pragmatic errors from explicit trade-offs in which real dimensions of fidelity conflict.
Levels of evidence
Structuring distinctions, case-by-case claims, and claims not to generalize.
Structuring
Multiple levels of equivalence; role of register and function; observable pragmatic compensation; multilingual authenticity in EU law; need to compare legal concepts.
Case by case
Whether explicitation improves understanding; adaptation better preserves an effect; compensation sufficiently repairs a local loss; foreignness is editorially preferable.
Do not generalize
“Every translation betrays”; “everything can be translated without loss”; “word-for-word = fidelity”; “fluency = quality”; “language fully traps thought.”
Laboratory
A comparison protocol without one global fidelity score.
The candidate tool examines separately propositional content, referents, terminology, logical relations, ambiguities, presuppositions, implicatures, register, speech act, social relation, function, cultural knowledge, readability, losses, additions, compensations and uncertainties.
PEDAGOGICAL PROPOSAL / CANDIDATE TOOL — NOT A VALIDATED INSTRUMENT.
Noosophical confrontation
A representation can be transformed without being merely copied or destroyed.
A candidate map follows source expression → linguistic cues + shared context → translator interpretation → priority hierarchy → target expression → recipient interpretation → comparison of shifts → possible revision.
The map identifies transformation points but establishes no Noosophical theory of language and does not show that a constructed or multimodal language would systematically reduce loss.
STRUCTURED ANALOGY / MAPPING TOOL — NOT A LINGUISTIC THEORY AND NOT EMPIRICAL VALIDATION OF NOOSOPHY.
Limits
Genre, audience and function change the priority of criteria.
There is no universal criterion for “the same effect.” Audiences are heterogeneous, editorial norms differ and legal translation has specific constraints. The examples in this dossier support no broad conclusion about Sapir–Whorf or causal relations between language and cognition.
Bibliography
Structuring references.
Baker (2018), House (1997, 2015), Jakobson (1959), Nida (1964), Quine (1960), Venuti (1995), Shih (2020), Lees (2024), CILFIT / Consorzio Italian Management case law, Court of Justice multilingual resources, VJM and IATE.
This bibliography structures the dossier; it does not exhaust translation studies.