Translation & Interpretation

Translation (written) and interpretation (spoken) are the oldest and most practically important language skills. They require not merely bilingual competence but a distinct set of meta-skills: source text analysis, equivalence judgment, register matching, cultural mediation, and the discipline to produce a target text that serves the same purpose as the source. This skill treats translation and interpretation as learnable meta-skills rather than natural by-products of bilingualism.

Agent affinity: saussure (signifier-signified, structural equivalence), lado (contrastive analysis, cross-linguistic transfer)

Concept IDs: lang-language-culture-link, lang-linguistic-relativity, lang-formality-register, lang-high-frequency-words

Equivalence Theory

Nida's Distinction

Eugene Nida (1964) defined two types of translation equivalence that remain the foundational framework:

Formal equivalence. Translates form for form: word for word where possible, phrase for phrase, preserving the structure of the source text. Prioritizes fidelity to the source.

Dynamic equivalence (functional equivalence). Translates the effect of the source on its original audience into an equivalent effect on the target audience. Prioritizes the reader's experience.

Example. The French expression "il pleut des cordes" (literally: "it rains ropes"):

• Formal equivalence: "it's raining ropes" -- preserves the metaphor but produces a nonsensical English sentence
• Dynamic equivalence: "it's raining cats and dogs" -- different metaphor, same communicative effect

Neither approach is universally correct. The choice depends on purpose: literary translation often favors dynamic equivalence; legal and religious translation often favors formal equivalence for accountability.

Venuti's Domestication vs. Foreignization

Lawrence Venuti (1995) reframed the debate:

• Domestication. Makes the text read fluently in the target language, minimizing foreignness. The reader barely notices the text is a translation.
• Foreignization. Preserves markers of the source culture, producing a reading experience that reminds the audience they are encountering a foreign text.

Domestication is the dominant commercial strategy (especially for English-language publishing). Foreignization is advocated by translation theorists who argue that the "invisible translator" erases cultural difference and reinforces linguistic imperialism.

Translation Strategies

Vinay & Darbelnet (1958) cataloged seven fundamental translation strategies, arranged from most literal to most free:

#	Strategy	Description	Example (French -> English)
1	Borrowing	Import the source word unchanged	"croissant" -> "croissant"
2	Calque	Translate morpheme by morpheme	"gratte-ciel" -> "skyscraper" (scrape-sky)
3	Literal translation	Word-for-word when grammatically possible	"Le livre est sur la table" -> "The book is on the table"
4	Transposition	Change the grammatical category	"avant son depart" (before his departure) -> "before he left" (noun -> verb)
5	Modulation	Change the point of view	"ce n'est pas difficile" -> "it's easy" (negative -> positive)
6	Equivalence	Replace with a functionally equivalent expression	"comme un chien dans un jeu de quilles" -> "like a bull in a china shop"
7	Adaptation	Replace with a culturally equivalent situation	A French character eating a croissant for breakfast -> an English character eating toast (full cultural transposition)

Professional translators move fluidly between these strategies at the sentence and sub-sentence level, selecting the approach that best serves the text's purpose.

Source Text Analysis

Before translating, a competent translator analyzes the source text along these dimensions:

Text Type (Reiss, 1971)

Type	Function	Translation Priority
Informative	Convey facts, data, knowledge	Accuracy of content
Expressive	Convey the author's aesthetic choices	Style and form
Operative	Persuade or move the reader to action	Effect on target reader

A technical manual is informative (translate for accuracy). A poem is expressive (translate for aesthetic impact). An advertisement is operative (translate for persuasive effect). Many texts are mixed.

Register Analysis

The translator must identify and preserve the source text's register:

• Field. The subject matter (medical, legal, literary, casual).
• Tenor. The relationship between writer and reader (formal/informal, expert/layperson).
• Mode. Written or spoken, planned or spontaneous.

Register mismatches in translation are common and damaging: translating a casual email in formal academic register, or a legal contract in conversational register, undermines the text's purpose even if every word is correctly translated.

Interpretation

Interpretation Modes

Mode	Timing	Use Case	Cognitive Demand
Simultaneous	Interpreter speaks while the source speaker continues	Conferences, international organizations	Extremely high -- parallel processing of input and output
Consecutive	Interpreter speaks after the source speaker pauses	Business meetings, medical appointments	High -- requires note-taking and memory
Sight translation	Interpreter reads a written text aloud in the target language	Immigration proceedings, document review	Moderate -- visual input, oral output
Liaison/bilateral	Interpreter mediates between two parties, alternating	Community interpreting, police interviews	Moderate -- short segments, bidirectional
Whispered (chuchotage)	Simultaneous interpretation whispered to one person	Diplomatic settings, small group	High -- simultaneous but without equipment

Note-Taking for Consecutive Interpretation

Consecutive interpreters develop personal notation systems that record:

• Key ideas (not individual words)
• Logical connectors (cause, contrast, sequence)
• Names, numbers, and dates (verbatim)
• Emphasis and speaker attitude

The notes serve as memory triggers, not transcription. A skilled consecutive interpreter can render a 5-minute speech from a single page of symbols.

Untranslatability and Compensation

Linguistic Untranslatability

Some expressions resist translation because the target language lacks an equivalent structural resource:

• Grammatical categories. Japanese honorific levels have no English equivalent. The translator must use circumlocution, added words, or footnotes.
• Wordplay. Puns depend on sound-meaning relationships that rarely transfer. A pun on "sole" (fish / bottom of shoe) in English has no equivalent in languages where these are phonetically distinct.
• Culturally bound concepts. German "Schadenfreude," Japanese "wabi-sabi," Portuguese "saudade" -- single words encoding complex cultural concepts with no one-word equivalent.

Cultural Untranslatability

The referent itself may not exist in the target culture:

• A recipe calling for "kimchi" in a culture that has never encountered fermented cabbage
• A legal text referencing "habeas corpus" in a legal system without that tradition
• A joke that depends on knowledge of a specific political figure

Compensatory Strategies

When direct translation is impossible, translators compensate:

1. Footnotes/endnotes. Explain the source concept for the target reader.
2. Glossing. Add a brief explanation in the text: "saudade (a deep longing for something absent)."
3. Compensation elsewhere. If a pun is lost in one sentence, create a different pun in a nearby sentence to preserve the text's playful tone.
4. Cultural substitution. Replace the source referent with a target-culture equivalent.
5. Omission. As a last resort, drop the untranslatable element when it is not essential to meaning.

Machine Translation Literacy

Modern neural machine translation (NMT) systems produce fluent output that often masks serious errors:

Where NMT Excels

• High-frequency language pairs (English-French, English-Chinese)
• Formulaic text types (weather reports, product descriptions, standard business correspondence)
• Sentence-level translation where context is local

Where NMT Fails

• Low-resource language pairs (Basque-Vietnamese, Welsh-Swahili)
• Pragmatic meaning (sarcasm, implicature, politeness levels)
• Document-level coherence (pronoun resolution across paragraphs, consistent terminology)
• Culturally loaded text (humor, poetry, advertising)
• Ambiguity resolution requiring world knowledge

Back-Translation

Translating the NMT output back into the source language is a simple but effective quality check. If the back-translation diverges significantly from the original, the forward translation likely contains errors.

Cross-References

• saussure agent: The signifier-signified relationship -- translation as mapping between sign systems. The arbitrariness of the sign explains why translation is always approximate.
• lado agent: Contrastive analysis predicts which translation directions and language pairs produce systematic errors.
• baker agent: Community interpreting, bilingual mediation, and the politics of language access.
• crystal agent: Historical language change that creates and destroys cognate relationships. Diachronic translation challenges.
• grammar-syntax skill: Structural differences between languages that force transposition in translation.
• pragmatics-communication skill: Pragmatic equivalence -- the hardest dimension to translate.
• vocabulary-acquisition skill: Cognate networks that facilitate translation between related languages.

References

• Nida, E. A. (1964). Toward a Science of Translating. E. J. Brill.
• Venuti, L. (1995). The Translator's Invisibility. Routledge.
• Vinay, J.-P. & Darbelnet, J. (1958). Stylistique comparee du francais et de l'anglais. Didier. (English translation: Comparative Stylistics of French and English, 1995.)
• Reiss, K. (1971). Moglichkeiten und Grenzen der Ubersetzungskritik. Max Hueber. (English translation: Translation Criticism: The Potentials and Limitations, 2000.)
• Baker, M. (2018). In Other Words: A Coursebook on Translation. 3rd edition. Routledge.
• Pochhacker, F. (2016). Introducing Interpreting Studies. 2nd edition. Routledge.