What it is
A unit shift occurs when the translation equivalent of a source unit belongs to a different rank from that unit. One word is answered by a group, a group by a clause, a clause by a word. The content is unchanged; what changes is the size of the grammatical package that carries it.
Catford introduced the category in his chapter on translation shifts, defining it against formal correspondence: the ranks of his description are those of the scale morpheme, word, group, clause, sentence, and a shift is registered whenever the equivalent departs from the rank of its source. He treats such departures as ordinary rather than exceptional, since equivalence is established by what the two texts do in a situation, not by any requirement that they use units of the same size. Chesterman adopts the category unaltered as his syntactic strategy G4, keeping Catford's rank scale.
A terminological warning belongs here. Much secondary literature prints "rank shift" as a synonym, and Catford's own gloss of "changes of rank" invites it. In systemic functional grammar, however, rankshift is Halliday's established term for embedding: a unit operating in the structure of a unit of equal or lower rank, as when a clause functions as a qualifier inside a nominal group. That is a fact about the internal structure of one language, not a relation between two texts, and it can occur in a source sentence and its translation alike. Using the two terms interchangeably collapses a description of monolingual structure into a description of translational departure. Catford's own heading is "unit shift", and it is the term to use.
The boundary against transposition is the cleaner one to hold: transposition changes word class at constant rank, unit shift changes rank. The two co-occur often enough that they are worth separating in a revision note.
When it applies
- The source language packs into one word what the target must spread across a group, as in agglutinative and heavily compounding languages.
- The target has a single lexicalized item where the source uses a fixed group.
- Subordination is expressed by a nominalized or non-finite form that the target can only render as a full clause, or the reverse.
- A grammatical relation marked by an affix in the source requires a preposition or an auxiliary in the target.
Examples
TR>EN
- Source: evlerimizden
- Target: from our houses
- Comment: One Turkish word, four English words, morpheme for morpheme: house plus plural plus first person plural possessive plus ablative. The rank moves from word to group with no change of class on either side, which is what distinguishes this from transposition.
DE>EN
- Source: Rechtsschutzversicherung
- Target: legal expenses insurance
- Comment: German compounding produces single words where English produces nominal groups. The shift is systematic in the pair and accounts for much of the length difference between German and English texts.
FR>EN
- Source: pomme de terre
- Target: potato
- Comment: The same shift in the opposite direction: a fixed French group answered by a single English word. Nothing has been condensed; the two languages simply lexicalize the concept at different ranks.
Watch out for
The practical consequence of unit shifts is that source and target word counts are not comparable across structurally distant pairs, and length ratios make a poor completeness check. A Turkish source will look short beside its English target while carrying the same content. Volume estimates, per-word arrangements and length-based quality flags all need pair-specific baselines.
- Where a nominalization is expanded into a clause, the target must supply a subject, a tense and often a modality that the source left unspecified. Check whether the addition is recoverable from context or is new information.
- Where a clause is compressed into a word or a group, subordinating relations can disappear. What was stated as a condition can end up reading as a description.
- Do not confuse a unit shift with distribution change. The unit shift is a claim about rank; distribution change is a claim about how many surface items carry the same semantic components.
References
- Catford, J. C. (1965). A Linguistic Theory of Translation: An Essay in Applied Linguistics. London: Oxford University Press, p. 79 (Ch. 12).
- Halliday, M. A. K. (1961). "Categories of the Theory of Grammar." Word 17(3), 241–292, for rankshift in the systemic sense.
- Chesterman, A. (1997). Memes of Translation: The Spread of Ideas in Translation Theory. Amsterdam: John Benjamins. Revised edition 2016, Ch. 4.