What it is
Componential analysis splits a lexical unit into its sense components, compares them with those of the nearest target item, and renders the source term by that item plus one to three of the components it does not carry. The output is a term with a modifier or a short qualifying phrase, and it is longer than the source. Newmark accepts that cost: the procedure trades economy for accuracy, and is chosen where accuracy is what the passage requires.
Newmark sets it out at page 90 of his Textbook and devotes a chapter to it. He separates it from synonymy, which he reserves for items not important enough to justify analysis, typically adjectives and adverbs of quality in uncritical positions. The translator's componential analysis is also not the linguist's: a semanticist decomposes a lexeme exhaustively, aiming at a full feature description, whereas the translator compares two items across two languages and supplies only the components the text puts in play. The analysis is local, and the same term may be analyzed differently in another document.
The nearest neighbours are description and explicitation. A descriptive equivalent abandons the target term for an account of the item's form or function; componential analysis keeps the term and augments it, preserving a hook for the reader and for later occurrences. Explicitation makes overt what the source left recoverable, whereas this procedure supplies what the source term carried and its target counterpart does not.
When it applies
- A system-bound term has no functional counterpart, and the text turns on precisely the features that make it incomparable.
- The nearest target term is a near-synonym whose divergence matters here, because the text goes on to exploit the distinction.
- The term is central and recurs, so a footnote would be consulted too often to be usable.
- Kinship systems, ranks, offices and legal statuses cut the semantic field differently in the two languages.
Examples
TR>EN
- Source: yenge
- Target: his brother's wife
- Comment: Turkish distinguishes the wife of a brother or uncle from the wife's sister and from the husband's sister, each with its own lexeme. English gathers all three under "sister-in-law", keeping the components female and affinal and losing the one that fixes which relative is meant. In an inheritance document that component carries the operative information.
DE>EN
- Source: Beamter
- Target: civil servant with tenured public-law status
- Comment: German employment law separates the Beamter from the salaried public employee, and the difference governs appointment, pension and dismissal. English "civil servant" covers both. Where the passage concerns those consequences the distinguishing components must be carried; where it merely reports that someone works for the state, they need not.
NL>EN
- Source: notaris
- Target: civil-law notary
- Comment: English has "notary", but in common-law jurisdictions it names a far narrower office with no drafting or registration function. One component separates the two without giving up the target term, and Newmark's ceiling of three is a limit rather than a target.
Watch out for
The procedure is disciplined by relevance. Every component added is a claim that the text needs it, and components supplied out of thoroughness are over-translation of the kind that reads as particularization: the target becomes more specific than the source, and a reviser is entitled to ask what licensed the extra precision.
- Analyze at first occurrence, then let the term run. Repeating the full analysis each time turns a document into a glossary.
- Derive the components from the source system, not from the target's categories. Describing a foreign office in terms of the reader's own institutions produces a functional equivalent, a different technique with different risks.
- In certified and sworn work the augmented rendering is an addition to the text. Most jurisdictions require such additions to be marked as the translator's, in brackets or in a note, rather than folded silently into the term.
References
- Newmark, P. (1988). A Textbook of Translation. London: Prentice Hall, p. 90 and Ch. 11.