When a text is being translated it is usually segmented into smaller parts which are easy and convenient to translate. Such segments of a text are known as units of translation. V. N. Komissarov uses the term “переводема” to refer to this notion. One of the foreign pioneers of the theory of translation J. Catford introduced the term ‘rank of translation’ in his book “A Linguistic Theory of Translation” (London 1965) which was used in a similar meaning.
According to S. B. Tyulenev, the unit of translation should be termed translateme (транслатема) and defined as a combination of a linguistic unit of the source language expressing a certain contextual meaning and the minimal corresponding linguistic unit of the target language with the same meaning .
The term ‘unit of translation’ was first used by J-P. Vinay, J. Darbelnet who claimed that its size may be variable as it serves only practical purposes. In this connection P.
Torop who investigates translation in close connection with language and culture points out that for a translator/interpreter it is necessary to operationally distinguish elements of culture and language both on the level of a text and the level of language units.
It reasonable to keep apart three aspects of the unit of translation that should be discussed separately and not confused as they arouse their own particular problems: (1) theoretical: its understanding and definition which should take into account the most important features of a unit of translation; (2) size-of-the-unit: relation of a init of translation to language levels; (3) operational: criteria applicable in the process of their identification in a SLT, i.
e. segmenting a SLT into textual elements that are convenient to translate.
The definition of the unit of translation may be given from three points of view: 1) with reference to a SLT, 2) with reference to a TLT, 3) with reference to both a SLT and a TLT. Within the above three approaches scholars may take into account various criteria: (a) content, or (b) form, or (c) both content and form. The first aspect of the unit of translation is connected with its definition. The most well-known definition of the unit of translation based on the criterion of content with reference to a SLT was suggested by J. -P. Vinay and J. Darbelnet.
According to them, a unit of translation is a unit of sense. The criterion of form is very important in form-oriented translation when formal peculiarities serve as an additional channel of conveying some information, e. g. English favours alliteration as a stylistic device which it is possible to render in translation often at the expense of its content (cf. Papa, potatoes, poultry, prunes and prism are all very good words for the lips (Ch. Dickens. Little Dorrit). – Папа, пряники, палисандр, персики и призмы – прекрасные слова для губ (Е. Конашева).
The definition which takes into account both content and form in Russian translation studies was suggested by prof. Komissarov V. N. who defined a unit of translation as a minimum language unit in SLT which is chosen as an independent object of translation process. With reference to a TLT, definitions of the unit of translation first appeared in the works of scholars who tackled the problems of machine translation. A unit of translation was understood as a combination of certain lexemes and grammemes which corresponds to a certain lexical or grammatical category in TLT (I.I. Revzin, V. Yu. Rozenzweig).
Definitions of the unit of translation which are formulated with reference both to a SLT and a TLT are also numerous. The most wide-spread definition based on these criteria was given by prof. Barkhudarov L. S. It runs as follows: It is a minimum language unit in a SLT which taken as a whole has a certain correspondence in TLT, but the constituent parts of which taken separately do not have a correspondence with a similar meaning in a TL [Бархударов 1975: 175].
To sum up the discussion of the understanding of this category of LTT it is necessary to stress its main features which are as follows: (a) a unit of translation should always be found in a SL text, not in a TL text, (b) it is established as a special category relevant from a translator’s point of view and therefore it has no correlation with the existing language levels, (c) it is a minimum unit in a sense that it cannot be segmented in translation into smaller parts without detriment to its sense.
The second aspect of the unit of translation has to do with its relationship with language levels. At present there are various approaches to the solution of this problem. 1. Some scholars believe that the unit of translation is always larger than a separate word. It may be a sentence or a clause, a group of sentences, a paragraph or even a whole text. This point of view is supported by Ya. I. Retsker, V. N. Komissarov, etc. They claim that a word cannot be chosen as a unit of translation for several reasons: 1) its boundaries are not always clear-cut;
2) words blend together in speech forming various semantic and structural complexes, the elements of which must be viewed together by a translator in order to retain in translation the meaning of a whole unit; 3) the outer (formal) aspect of a speech event to be rendered in translation may be often more important than the semantic one. E. g. in actual speech stylistic effect is often due to special devices that are based on the formal aspect of speech units recurring in a succession of words in a speech string (repetition of some sounds, sound symbolism – e.g. the effect of the sounds [d], [i], [ing] recurrent within one or even more than one utterances.
2. V. N. Komissarov points out that a sentence is usually chosen as an operational unit of translation since it provides a microcontext necessary and sufficient to comprehend its meaning properly [Комиссаров 1999a]. True, he admits the possibility of choosing other units of translation within a sentence which are of a lower rank, and associated with a word-level or a word-group level.
In other words, the author introduces the idea of hierarchy of units of translation which may be chosen on different levels (higher and lower). 3. K. Reiss and H. Vermeer hold that a unit of translation should be chosen on the level of a text as it may help to overcome contextual ambiguity and ambivalence of words and even sentences that arise from differences in languages and cultures; a text realizes the speaker’s intention (‘scopos’) (K. Reiz, H. J. Vermeer and others).
Translation practice testifies that the unit of translation can be chosen on various language levels: (a) phonemes which is usually the case in dealing with proper and geographic names, sport, space, computer and other sets of terms, etc: Challenger – Челенджер, bobsleigh – бобслей, minority interests – миноритарные интересы, prime time – прайм-тайм, merchandising – мерчандайзинг. The method of reproducing SL phonemes by corresponding phonemes in TL is called transcription.
It is regarded as one of the types of borrowed translation and underlies borrowing of the so-called international words. Since words have not only a sound form, but also a graphic form, graphemes may also serve as units of translation. In practice it may be a combination of both phonemes and graphemes chosen as units of translation: off-shore – офшор, Waterloo – Ватерлоо, poster – постер. The method of reproducing graphemes of SL by similar graphemes of TL is known as transliteration and it is very popular today alongside transcription when translating neologisms, terms, etc.
(b) morphemes may be chosen as units of translation when translating bymorphemic and polymorphemic words: самореклама – self-promotion, вневедомственная охрана – non-departmental security guard, отказник — refusenik. The method of segmenting words into morphemes which are further translated into TL which reproduces the morphological structure of a word is called loan translation. It is also qualified as a kind of borrowed translation because the translator not only reproduces the morphemes, but also retains the morphological structure of a word and arrangement of morphemes.
(c) words are often chosen as units of translation. Their proper translation depends on two main factors: 1) semantic peculiarities, i. e. whether a word is monosemantic or polysemantic, the correlation of denotational and connotational components of meaning and some other features; 2) the degree of dependence on a context and a situation of discourse. From this point of view words are divided into context-free and context-bound words. In the former case the translation does not depend on a context, e. g. here refer units of precision lexis, proper and geographical names, inch – дюйм, New Zealand – Новая Зеландия.
In the latter case the context plays a great role, especially when translating polysemantic words, e. g. to make an assault upon a fortress – штурмовать крепость; some music is an assault on the ears – есть музыка, которая оскорбляет слух; assaults upon the prerogatives of Parliament – резкие выступления против прерогатив парламент (d) word-groups as units of translation are especially important when translating from English into Russian because in analytical (lexical) languages like English the dependence of a word on its environment is much greater than in synthetical (grammatical) languages like Russian.
E. g. Three hundred workers went on strike over a bonus claim. – Триста рабочих забастовали, требуя выплаты премиальных. e) sentences serve as units of translation in two main cases: 1. when translating utterances used to describe identical situations in the two languages in their own idiomatic ways, cf. Authorized Personnel only – Посторонним вход запрещен / Служебный вход 2. when translating proverbs and sayings which are based in different languages on different images, though they are related in sense, cf. Still waters run deep. – В тихом омуте черти водятся.
Apart from these two cases the unit of translation may be chosen on the level of a sentence to retain a proper stylistic effect. E. g. Their bodies were covered with fur. – Они были одеты в меха. (Их одежда была сшита из меха. – пер. Н. Волжина). Word-for-word translation of the sentence would sound misleading, cf. Их тела были покрыты мехом. f) a text. The term ”text” itself needs some clarification since linguists are not at one concerning its understanding and definition. O. I. Moskalskaya, for instance, qualifies any stretch of speech that expresses a complete thought as a text.
It may be a single word in a telegram message. According to I. R. Galperin a communicative speech unit only then turns into a text if it possesses a number of text-forming features such as the category of information, modality, cohesion, and some others. Without going into details on this complicated matter we confine our understanding of a text for the purposes of translation as a combination of several sentences which forms a connected stretch of speech possessing as a whole unit certain content and formal features.
Practitioners of translation claim that the unit of translation on the level of a text is often chosen when translating poetry. In dealing with prose it is hard to overestimate the importance of a text to achieve an adequate translation. Though a prosaic text as a whole cannot serve as a unit of translation it influences to a great extent correct translation of its parts. This influence can be easily observed in translations of titles of books. The title of the famous novel by J. Braine “Room at the top” was translated by several variants from word-for-word to more adequate, cf.«Мансарда» > «Место наверху» > «Путь наверх».
For this reason translation of the titles of books very often undergoes varied transformations, «На всякого мудреца довольно простоты» (А. Н. Островский) – “Even a Wise Man Stumbles”, “Even the Wise Can Err”, “Scoundrel”. Thus, in conclusion it must be stressed that the unit of translation should not be tied in with any particular language level since as has been shown above it can be located on various levels on condition that it satisfies the criteria and requirements stated in the quoted definition.
The third important aspect of a unit of translation is connected with criteria on which a translator may rely in order to identify it in the process of segmenting a hSL text. The analysis of various approaches testifies to diverging solutions to this problem. Most translatologists contend that such criteria do not include the criterion of the size of a unit of translation as it is variable in this respect in various translation acts. Some linguists connect the size of the unit of translation not with a SLT, but with the capacity of the TL to express the same notion that is expressed in the original [Тюленев 2004].
According to L. S. Barhkudarov, we should depend in this matter on the existence of TL correspondences for certain segments of a text. In search of reliable criteria some linguists associate units of translation with translation problems, hence the only criterion of a unit of translation is claimed to be connected with its being sufficient for making proper translation solutions (cf. «Величина отрезка текста не может быть критерием единицы перевода. Таким критерием должна быть возможность принять решение на перевод» [Миньяр-Белоручев 1999]).
In other words, units of translation call for an individual translation solution. Such an approach to the criteria of a unit of translation was elaborated by Ye. V. Breus who defined units of translation in connection with a text which is viewed as a matrix («сетка») of translation problems. This matrix covers problem translation areas and problem-free translation areas. In the former case a translator deals with units of translation which he has to translate creatively and of his own accord, while in the latter case he transcodes a text using regular correspondences according to Ya.
I. Retsker’s theory. Thus apart from the theoretical aspects of understanding and defining a unit of translation there arise practical difficulties associated with segmenting a SLT into minimal translation units. It is obvious that the knowledge of SL and TL systems alone is not sufficient to cope with this task. In fact, the process of segmenting a SL text into units of translation should be viewed as a multi-stage procedure the success of which depends to a great extent on both linguistic and extra-linguistic factors.
As a rule, translatologists establish three stages in this process: 1) considering linguistic units of a certain level (usually words), 2) evaluating their contextual interdependence and 3) modifying it in various ways (enlarging initially chosen linguistic units to the size of word-combinations, sentences and even entire paragraphs and texts; or diminishing them and choosing on a lower level or combining several into one; or dropping in translationl.
We believe that it is not enough to take into account only linguistic factors (systemic and contextual), but it is also necessary to pay attention to such important aspects of translator’s activity as translation norm, situational and broader cultural parameters which determine the sense and implications of a SL text affecting the choice of proper units of translation on a given occasion.
Thus, in connection with these factors and parameters of a SLT it is possible to single out the following stages in the text-segmentation procedure: 1) segmenting a text, depending on our knowledge of the language, into minimal units of sense (in regard to meaningful language units) or minimal units of outer expression (phonemes or graphemes) which cannot be further segmented into smaller parts without detriment to either sense or form.
(cf. This is a new building – I’m fond of new potatoes, where new for translation purposes should be singled out on a word-level in the first sentence (новое) and a word-group – level in the second case (молодой картофель); 2) analyzing all contextual modifications of established semantic and formal features of those units (cf. … Не проси, мин херц, сейчас и не думай об Монсихе (А. Толстой, 102) – Don’t ask for that, myn Herz. You mustn’t even think of the Mons girl now…(A. Miller)
3) considering all situational and background factors which determine particular uses of the established units (cf. I made out your voucher (W. Saroyan. Selected short stories) – Я выписал тебе расчёт (пер. Л. Натан); 4) enlarging, diminishing or dropping selected units of translation under the influence of translation norms as a verification of their correct choice in a given context (cf. In a panic he shoveled hay over the puppy with his fingers. – В испуге он быстро зарыл щенка в сено (Хайруллин 1997). 5) applying the principle of interchangeability , e. g. Don’t turn over – Не кантовать.
6) taking into account the addressee factor (cf. recycling – рециклирование/вторичная переработка), 7) choosing a proper translation strategy, cf. Бог с ними! – God with them! (word-for-word translation) God is with them! (literal translation) Never mind about them! (adequate translation). Thus, it is apparent that segmenting a SLT into units of translation is a complicated procedure of its interpretation which is the decisive factor of adequate rendering in a TL and at various stages the SLT analysis and segmentation into parts depend upon both linguistic and non-linguistic factors.
We may conclude that the discussion of the unit of translation should bring to light its main aspects: theoretical (connected with the understanding and definition of this notion), the size-of-the-unit aspect (connected with the choice of units of translation on different language levels on the basis of identificatory criteria), operational aspect (connected with principles of segmenting a SLT into units of translation).