What it is
Literal translation carries the source across structure by structure and item by item, altering only what the target grammar compels. The result is ordinary target language, not an interlinear gloss: segmentation, constituent order and lexical selection are preserved wherever they survive the crossing intact.
Vinay and Darbelnet listed it third among the direct procedures, after borrowing and calque, and made it the pivot of the entire model. Their claim was normative: between languages of the same family and the same cultural sphere, the literal rendering is the expected solution, and the translator leaves it only when one of five conditions holds. The literal version gives a different meaning; it gives no meaning at all; it is structurally impossible; it has no corresponding expression within the metalinguistic experience of the target language; or it corresponds to something at another level of language, that is, at a different register (V&D 1995: 33-35). Everything oblique, from transposition to adaptation, is on their account a response to one of those five failures rather than a free stylistic preference.
Later frameworks kept the procedure and dropped the primacy. Chesterman files it as G1 and characterizes it as the default from which translators depart as target norms require: a starting position, not a merit. Molina and Hurtado Albir treat it as one technique among eighteen, operating on individual units, and separate it explicitly from the older usage in which "literal translation" named a whole-text method or a school of thought.
The nearest boundary is calque. Literal translation reuses target items that exist; calque manufactures a new target item by translating the parts of a source expression.
When it applies
- Source and target are structurally and culturally close, and none of the five abandonment conditions is triggered.
- The passage is terminology-dense and any reformulation would cost definitional precision: standards, specifications, patent claims, technical annexes.
- The text is binding, and the reader is entitled to the drafter's own formulation rather than to a smoother version of it.
- The unit consists of names, figures or citations, where the wording itself is the object of interest.
Examples
FR>EN
- Source: J'ai laissé mes lunettes sur la table en bas.
- Target: I left my glasses on the table downstairs.
- Comment: The standard illustration, given by Vinay and Darbelnet in the opposite direction. Every element finds a counterpart in the same position and the same class. Nothing is gained by reformulating, and a reviser has no case against it.
TR>EN
- Source: Dilekçe süresi içinde verilmediği için reddedilmiştir.
- Target: The petition was rejected because it was not filed within the time limit.
- Comment: Turkish places the subordinate reason before the main predicate and marks it with a suffix rather than a conjunction. Word-for-word transfer is structurally impossible, which is the third abandonment condition, yet the rendering above is still literal in the operative sense: nothing has moved except what the target grammar forced.
NL>EN
- Source: Hij heeft er geen kaas van gegeten.
- Target: He knows nothing about it.
- Comment: Word-for-word gives "he has eaten no cheese of it", which is the second condition: the literal version has no meaning in the target language. The solution is an established equivalent, reached because the literal route failed a test, not because it read badly.
Watch out for
The label is doing two jobs at once, and revisers exploit the ambiguity. As a procedure it names a defensible unit-level choice; as an accusation it names a rendering that is grammatical, traceable to the source and unusable. The distinction is testable: apply the five conditions in order, and record which one sent you elsewhere.
- Literal is not a defence for a rendering no target reader would produce. Failing the fourth condition, no corresponding metalinguistic experience, is the failure translators most often miss, because the output looks correct.
- Neither is fluency a defence for leaving it. In binding and certified work the burden falls on the departure, and an unlogged departure is a finding.
- Direction matters. Between typologically distant pairs the literal option is available far less often than the procedure's name suggests.
References
- Vinay, J.-P. & Darbelnet, J. (1958). Stylistique comparée du français et de l'anglais: méthode de traduction. Paris: Didier. English translation by J. C. Sager & M.-J. Hamel (1995), Comparative Stylistics of French and English: A Methodology for Translation, Benjamins Translation Library 11. Amsterdam: John Benjamins, pp. 33–35.
- Newmark, P. (1988). A Textbook of Translation. London: Prentice Hall, pp. 68–69.
- Chesterman, A. (1997). Memes of Translation: The Spread of Ideas in Translation Theory. Amsterdam: John Benjamins. Revised edition 2016, Benjamins Translation Library 123, Ch. 4.
- Molina, L. & Hurtado Albir, A. (2002). "Translation Techniques Revisited: A Dynamic and Functionalist Approach." Meta 47(4), 498–512.