Theory of translation

Theory of Translation

    1.    TRANSLATION THEORY: OBJECT AND OBJECTIVES  

     Translation is a means of interlingual communication. The translator makes possible an exchange of information between the users of different languages by producing in the target language (TL or the translating language) a text which has an identical communicative value with the source (or original) text (ST). This target text (TT, that is the translation) is not fully identical with ST as to its form or content due to the limitations imposed by the formal and semantic differences between the source language (SL) and TL. Nevertheless the users of TT identify it, to all intents and purposes, with ST - functionally, structurally and semantically.

     The functional status of a translation is supported by its structural and semantic similarity with the original. The translator is expected to refrain from any remarks or intrusions in his text which may betray his authorship thereof. He is expected to efface himself as fully as he can to avoid interference with the process of communication between S and TR.

     The structure of the translation should follow that of the original text: there should be no change in the sequence of narration or in the arrangement of the segments of the text.

     The aim is maximum parallelism of structure which would make it possible to relate each segment of the translation to the respective part of the original. It is presumed that any breach of parallelism is not arbitrary but dictated by the need for precision in conveying the meaning of the original. The translator is allowed to resort to a description or interpretation, only in case “direct translation” is impossible. Similarity in structure is preserved in respect to the smallest segments of the text.

     Of major importance is the semantic identification of the translation with ST. It is presumed that the translation has the same meaning as the original text. No exchange of information is possible. The presumption of semantic identity between ST and TT is based on the various degrees of equivalence of their meanings. The translator usually tries to produce in TL the closest possible equivalent to ST.

     As a kind of practical activities translation (or the practice of translation) is a set of actions performed by the translator while rendering ST into another language. At its best translation is an art, a creation of a talented, high-skilled professional.

     Translation can be the object of scientific study aimed at understanding its nature, its components and their interaction as well as various factors influencing it or linked with it in a meaningful way. The science of translation or translatology is concerned both with theoretical and applied aspects of translation studies. A theoretical description of the translation phenomenon is the task of the theory of translation.

     The linguistic theory of translation is concerned with translation as a form of speech communication establishing contact between communicants who speak different languages. The basis of this theory is linguistics in the broadest sense of the word, that is, macrolinguistics with all its new branches, such as psycholinguistics, sociolinguistics, text linguistics, communicative linguistics, etc.

     The core of the translation theory is the general theory of translation which is concerned with the fundamental aspects of translation inherent in the nature of bilingual communication and therefore common to all translation events, irrespective of what languages are involved or what kind of text and under what circumstances was translated.

     An important part of the general theory of translation is the theory of equivalence aimed at studying semantic relationships between ST and TT. It has been noted that there is a presumption of semantic identity between the translation and its source text.  Let us take an elementary example. Suppose we have an English sentence ‘The student is reading a book”. Its Russian translation will be «Студентчитает книгу». This translation is a good equivalent of the English sentence, but it is not identical in meaning. It can be pointed out, for example, that the Russian sentence leaves out the meaning of the articles as well as the specific meaning of the Continuous Tense. In Russian we do not get explicit information that it is some definite student but not some particular book or that the reading is in progress at the moment of speech. On the other hand, the Russian sentence conveys some additional information which is absent in the source text. We learn from it that the student is a male, while in ST it may just as well be a female. Then the translation implies that the student in the case is a college undergraduate, while in ST he may be a high school student or even a scholar, to say nothing of the additional grammatical meaning conveyed by the grammatical aspect of «читает», the gender of «книга» and so on. Part of this information, lost or added in the translating process, may be irrelevant for communication, another part is supplemented or neutralized by the contextual situation, but it is obvious that translation equivalence does not imply an absolute semantic identity of the two texts. The theory of equivalence is concerned with factors which prevent such an identity, it strives to discover how close ST and TT can be and how close they are in each particular case.

     The general theory of translation describes the basic principles which bold good for each and every translation event. In each particular case, however, the translating process is influenced both by the common basic factors and by a number of specific variables which stem from the actual conditions and modes of the translator’s work: the type of original texts, in which ST is presented to him and the form in which he is supposed to submit his translation, the specific requirements he may be called upon to meet in his work, etc.

     The translator has to deal with works of the great authors of the past and of the leading authors of today, with intricacies of science fiction and the accepted stereotypes of detective stories, must be able to cope with the elegancy of expression of the best masters of literary style and with the tricks and formalistic experiments of modern avant-gardists.

     His duty is to translate diplomatic representations and policy statements, scientific dissertations and brilliant satires, maintenance instructions and after-dinner speeches, etc.

     Translating a play the translator must bear in mind the requirements of theatrical presentation, and dubbing a film he must see to it that his translation fits the movement of the speakers’ lips.

     In simultaneous interpretation the translator is expected to keep pace with the fastest speakers, to understand all kinds of foreign accents and defective pronunciation, to guess what the speaker meant to say but failed to express due to his inadequate proficiency in the language he speaks.

     In consecutive interpretation he is expected to listen to long speeches, taking the necessary notes, and then to produce his translation in full or compressed form, giving all the details or only the main ideas.

     In some cases the users will be satisfied even with the most general idea of the meaning of the original, in other cases the translator may be taken to task for the slightest omission or minor error.

     Each type of translation has its own combination of factors influencing the translating process. The general theory of translation should be supplemented by a number of special translation theories identifying major types of translation activities and describing the predominant features of each type.

     Another important branch of the theory of translation is concerned with the study of ST and TT units which can replace each other in the translating process. A bilingual theory of translation should study two separate sets of equivalents, with either language considered, in turn, as SL and the other as TL.

     Of particular interest is that branch of the theory of translation which is concerned with the translating process itself, that is, with the operations required for passing over from ST to TT. It is a great challenge to the translation theory to discover how the translator does the trick, what are his mental processes which ensure production in TL of a text of identical communicative value with the given ST.

     The study of the translating process reveals both the translator’s general strategy and specific techniques used to solve typical translation problems.

     In conclusion, mention should be made of one more branch of the theory of translation which deals with the pragmatic aspects of the translating process. The communicants involved in interlingual communication speak different languages but they also belong to different cultures, have different general knowledge, different social and historical background.

     The translator has to assess the possible communicative effect of TT and take pains to ensure an adequate understanding of its message by TR. This may necessitate expanding or modifying the original message to make it more meaningful to the members of a different language community.

     In some cases the pragmatic value of translation is the major factor in assessing the quality of the translator’s performance.  All branches of the theory of translation are concerned with important aspects of the translator’s work and constitute a body of theoretical thought of indisputable practical value. 

    2.    EQUIVALENCE IN TRANSLATION  

     Translation equivalence is defined as a measure of semantic similarity between ST and TT.

     If we compare a number of TTs with their STs we shall discover that the degree of semantic similarity between the two texts involved in the translating process may vary. In other words the equivalence between ST and TT may be based on the reproduction of different parts of the ST contents. Accordingly, several types of translation equivalence can be distinguished.

     Let us first of all single out translations in which the degree of semantic similarity with ST seems to be the lowest. This type of equivalence can be illustrated by the following examples (cited from the published translations):

     (1) Maybe there is some chemistry between us that doesn’t mix.Бывает, что люди не сходятся характерами. – Este ceva la mijloc/ce nu merge.

     (2) A rolling stone gathers no moss. - Кому дома не сидится, тот добра не наживет./ Piatra ce se rostogoleşte nimic (nu) dobîndeşte.

     (3) That’s a pretty thing to say. - Постыдился бы!/ Îi uşir de spus/ Ruşine să-ţi fie !

     Неге we cannot discover any common seems or invariant structures in the original and its translation. An absolute dissimilarity of language units is accompanied by the absence of obvious logical link between the two messages which could lead to the conclusion that they arc “about the same thing”, Le. that they describe one and the same situation. Yet, it is evident that the two sentences have something in common as to their meaning. This common part of their contents is obviously of great importance, since it is enough to ensure an adequate communication.

     It comprises the information which must be preserved by all means even though the greater part of the contents of the original is lost in the translation.

     In plain English, the translation does not convey either “what the original text is about”, or what is said in it” or “how it is said”, but only “what it is said for”, i.e. what the Source meant, what the aim of the message is.

     This part of the contents which contains information about the general intent of the message, its orientation towards a certain communicative effect can be called “the purport of communication”.Thus we can deduce that in the first type of equivalence it is only the purport of communication that is retained in translation.

     The second group of translations can be illustrated by the following examples:

    -         He answered the telephone. -Он снял трубку./ El a luat receptorul.

    -         You see one bear, you have seen them all.-Все медведипохожи друг на друга./ noaptea toate pisicile sunt negre.

    -         It was late in the day. - Близился вечер./ Se întunecă spre sfîrşitul zilei.

     This group of examples is similar to the first one, as the equivalence of translations here does not involve any parallelism of lexical or structural units. Most of the words or syntactical structures of the original have no direct correspondences in the translation. At the same time it is obvious that there is a greater proximity of contents than in the preceding group.

    Consider, for instance, the translations:

    -(1) Maybe there is some chemistry between us that doesn’t mix. Бывает, что люди не сходятся характерами.

    - (2) Не answered the telephone. Он снял трубку.

     In (I) the things referred to are different, so that there is hardly any logical connection between the two statements. We can draw identical conclusions about the speaker’s sentiments: there is no love lost between him and another person.

     In (2) the incomparable language units in the original and in the translation describe, in fact, the same action, refer to identical reality, as a telephone call cannot be answered unless one picks up the receiver. Both texts give different information about the same, or, as one sometimes says, they express the same idea “using different words”. It is the type of equivalence that can be well explained in terms of the situational theory. We may presume that such phrases describe identical situations but each is presented in a different way. Since in each of the two texts the situation is described in a different way, the common feature is not the method of description but the reference to the situation, the possibility of identifying the situation, no matter how it is described in the text. The information which characterized the second type of equivalence can, therefore, be designated as “identification of the situation“.

     In the next group of translations the part of the contents which is to be retained is still larger. This type of equivalence can be exemplified as follows:

     Scrubbing makes me bad-tempered. Оть мытья полов у меня настроение портится.

     - London saw a cold winter last year.-В прошлом году зима в Лондоне была холодной./ A fost  o iarnă grea pentru Londra în anul trecut.

     - You are not serious? – Вы  шутите?/ Glumeşti ?? Faci glume?

     The translation contains the same general notions as the original. This means that the translation is a semantic paraphrase of the original, preserving its basic semes and allowing their free reshuffle in the sentence. The common semes are easily discovered in the comparative analysis of the translations of this group. Consider the first of the examples cited. Both in the translation and in the original the situation is described as a “cause-effect” event with a different pattern of identical semes. In the original: A (scrubbing) causes В (I) to have С (temper) characterized by the property D (bad). In the translation: С (temper) belonging to В (I) acquires the property D (bad) because of A (scrubbing).

     The use of the identical notions in the two texts means that the basic structure of the messages they convey remains intact. Here it indicates “what is said in the original”, i.e. what aspect of the described situation is mentioned in the communication.

     We can now say that the third type of equivalence exemplified by the translations of the third group, implies retention in the translation of the three parts of the original contents which we have conventionally designated as the purport of communication, the identification of the situation and the method of its description.

     The fourth group of translations can be illustrated by the following samples:

      -  He was never tired of old songs.Старые песни ему никогда не надоедали./

      -  I don’t see that I need to convince you. He вижу надобности доказывать это вам./ Nu e nevoie să te conving

      - Не was standing with his arms crossed and his bare head bent.Он стоял, сложив руки на груди и опустив непокрытую голову./ El stătea cu mîinele încruucişate şi cu capul plecat.

     In such translations the syntactic structures can be regarded as derived from those in the original through direct or backward transformations. This includes cases when the translation makes use of similar or parallel structures.

     Equivalence imply the retention of the linguistic meaning, i.e. the information fixed in the substantial or structural elements of language as their plane of content. The translation conveys something of the “how-it-is-said in the original”.

     The fourth type of equivalence presupposes retention in the translation of the four meaningful components of the original: the purport of communication, the identification of the situation, the method of its description, and the invariant meaning of the syntactic structures.

     The fifth group of translations can be discovered when we analyse their relationships with the respective originals. Here we find the maximum possible semantic similarity between texts in different languages. These translations try to retain the meaning of all the words used in the original text. There is considerable semantic proximity of the correlated words in the two sentences:

    -   I saw him at the theatre.-Я видел его в театре./ L-am văzut la teatru.

    -  The house was sold for 10 thousand dollars.- Дом был продан за десять тысяч долларов./ Casa s-a vîndut cu 10 000 de dolari.

    -  The Organisation is based on the principle of the sovereign equality of all its Members.-Организация основана напринципе суверенного равенства всех ее членов./Organizaţia se bazează pe egalitatea suzerană a tuturor membrilor săi.

     Here we can observe the equivalence of semes which make up the meaning of correlated words in the original text and the translation; parallelism of syntactic structures implying the maximum invariance of their meanings; the similarity of the notional categories which determine the method of describing the situation; the identity of the situations; the identical functional aim of the utterance or the purport of communication.

     We can sum up. We have discovered that there are five different types of semantic relationships between equivalent phrases (texts) in two languages, differing as to the volume and character of the information retained in each.

     Every translation can be regarded as belonging to a certain type of equivalence. Since each subsequent type implies a higher degree of semantic similarity we can say that every translation is made at a certain level of equivalence.

     Each level of equivalence is characterized by the part of information the retention of which distinguishes it from the previous level. The list of levels, therefore, includes: 1) the level of the purport of communication; 2) the level of (the identification of) the situation; 3) the level of the method of description (of the situation); 4) the level of syntactic meanings; 5) the level of word semantics.

     A translation can be good at any level of equivalence.  

    3.    TYPES OF EQUIVALENTS  

     The structural similarity of ST and TT implies that relationships of equivalence are established between correlated units in the two texts.

     Some of the SL units have permanent equivalents in TL, that is to say, there is a one-to-one correspondence between such units and their equivalents. Thus “London” in Russian is «Лондон», “a machine-gun” as «пулемет» and “hydrogen” is always rendered as «водород». As a rule this type of correspondence is found with words of specific character, such as scientific and technical terms, proper or geographical names and similar words whose meaning is more or less independent of the particular contextual situation.

     Other SL units may have several equivalents each. Such one-to-many correspondence between SL and TL units is characteristic of most regular equivalents. The existence of a number of non-permanent (or variable) equivalents to a SL units implies the necessity of selecting one of them in each particular case, taking into account the way the unit is used in ST and the points of difference between the semantics of its equivalents in TL.

     Depending on the type of the language units involved regular equivalents can be classified as lexical,phraseological or grammatical.

     The choice of the equivalent will depend on the relative importance of a particular semantic element in the act of communication.

     A variety of equivalents may also result from a more detailed description of the same object in TL. The English word “attitude”, for instance, is translated as «отношение, позиция, политика» depending on the variant the Russian language prefers in a particular situation. Here the choice between equivalents is determined by TL factors.

     Even if a SL unit has a regular equivalent in TL, this equivalent cannot be used in TT whenever the unit is found in ST. An equivalent is but a potential substitute, for the translator’s choice is, to a large extent, dependent on the context in which the SL unit is placed in ST. There are two types of context: linguistic and situational. The linguistic context is made up by the other SL units in ST while the situational context includes the temporal, spacial and other circumstances under which ST was produced as well as all facts which the receptor is expected to know so that he could adequately interpret the message.

     Thus in the following sentences the linguistic context will enable the translator to make a correct choice among the Russian equivalents to the English noun “attitude”:

     - (1) I don’t like your attitude to your work.

     - (2) There is no sign of any change in the attitudes of the two sides.

     - (3) He stood there in a threatening attitude.

     It is obvious that in the first sentence it should be the Russian «отношение (к работе)», in the second sentence — «позиции(обеих сторон)», and in the third sentence - «поза (угрожающая)».

     The fact that a SL unit has a number of regular equivalents does not necessarily mean that one of them will be used in each particular translation. True, in many cases the translator’s skill is well demonstrated in his ability to make a good choice among such equivalents.

     Geographical names have such equivalents which are formed by imitation of the foreign name in TL. And the name of the American town of New Haven (Conn.) is invariably rendered into Russian as «Нью-Хейвен». But the sentence “I graduated from New Haven in 1915″ will be hardly translated in the regular way since the Russian reader may not know that New Haven is famous for its Yale university. The translator will rather opt for the occasional equivalent: «Я окончилЙельский университет в 1915 году».

     The same goes for phraseological equivalents. Phraseological units or idioms may also have permanent or variable equivalents. Such English idioms as “the game is not worth the candle” or “to pull chestnuts out of the fire for smb.” are usually translated by the Russian idioms «игра не стоит свеч» and «таскать каштаны из огня длякого-л.», respectively. These equivalents reproduce all the aspects of the English idioms semantics and can be used in most contexts. Other permanent equivalents, though identical in their figurative meaning, are based on different images, that is, they have different literal meaning. Cf. “to get up on the wrong side of the bed” —«встать с левойнога», “make hay while the sun shines” —«куй железо, покагорячо». Now an English idiom may have several Russian equivalents among which the translator has to make his choice in each particular case. For instance, the meaning of the English “Do in Rome as the Romans do” may be rendered in some contexts as «С волками жить- по-волчьи выть», and in other contexts as «В чужой монастырьсо своим уставом не ходят». But here, again, the translator may not infrequently prefer an occasional equivalent which can be formed by a word-for-word reproduction of the original unit: «В Риме поступайтак, как римляне».

     The choice of grammatical units in TT largely depends on the semantics and combinability of its lexical elements. Therefore there are practically no permanent grammatical equivalents. The variable equivalents in the field of grammar may be analogous forms in TL or different forms with a similar meaning. In the following English sentence “He was a guest of honour at a reception given by the Soviet government” both the Russian participle «устроенном» and the attributive clause «который был устроен» can be substituted for the English participle “given“. And the use of occasional equivalents is here more common than in the case of the lexical or phraseological units. We have seen that in the first three types of equivalence no equivalents to the grammatical units are deliberately selected in TL.

     Semantic dissimilarity of analogous structures in SL and TL also result in SL structures having several equivalents in TL. For instance, attributive groups are common both in English , Russian and Romanian: “a green tree“—«зеленое дерево/ copac verde». But the semantic relationships between the numbers of the group are broader in English, which often precludes a blue-print translation of the group into Russian. As often as not the English attributive group is used to convey various adverbial ideas of location, purpose, cause, etc. Consider such groups as “Madrid trial” (location), “profits drive” (purpose), “war suffering” (cause). Such groups may also express various action-object relationships. Cf. labour movement” (movement by the workers), “labour raids” (raids against the workers), and “labour spies” (spies among the workers).

     A word within an attributive group may sometimes alter its meaning. So, “war rehabilitation” is, in fact, rehabilitation of economy after the war, that is, “post-war rehabilitation”.

     As a result, many attributive groups are polysemantic and are translated in a different way in different contexts. “War prosperity” may mean “prosperity during the war” or “prosperity in the post-war period caused by the war”. ‘The Berlin proposals” may imply “proposals made in Berlin” (say, at an international conference), “proposals made by Berlin” (i.e. by the FRG), “proposal on Berlin” (of political, economic or other nature).

     No small number of SL units have no regular equivalents in TL. Equivalent-lacking words are often found among SL names of specific national phenomena, such as the English words “coroner, condominium, impeachment, baby-sitter” and the like. However, there are quite a number of “ordinary” words for which TL may have no equivalent lexical units: “fluid, bidder, qualifier, conservationist”, etc. Some grammar forms and categories may also be equivalent-lacking. (Cf. the English gerund, article or absolute participle construction which have no counterparts in Russian.)

     The absence of regular equivalents does not imply that the meaning of an equivalent-lacking SL unit cannot be rendered in translation or that its translation must be less accurate. We have seen that words with regular equivalents are not infrequently translated with the help of contextual substitutes. Similarly, the translator, coming across an equivalent-lacking word, resorts to occasional equivalents which can be created in one of the following ways:

     1. Using loan-words imitating in TL the form of the SL word or word combination, e.g. tribalism — трайбализм, impeachment — импичмент, backbencher — заднескамеечник, brain-drain — утечка мозгов.

     2. Using approximate substitutes, that is TL words with similar meaning which is extended to convey additional information (if necessary, with the help of foot-notes), e.g. drugstore — аптека, witchhunter — мракобес, afternoon — вечер. The Russian «аптека» is not exactly a drugstore where they also sell such items as magazines, soft drinks, ice-cream, etc., but in some cases this approximate equivalent can well be used.

     3. Using all kinds of lexical (semantic) transformations modifying the meaning of the SL word, e.g. “He died of exposure” may be rendered into Russian as «Он умер от простуды» or «Онпогиб от солнечного удара».

     4. Using an explanation to convey the meaning of the SL unit, e.g. landslide-победа на выборах подавляющим большинством голосов, brinkmanship — искусство проведения политики на грани войны, etc.

     This method is sometimes used in conjunction with the first one when the introduction of a loan-word is followed by a foot-note explaining the meaning of the equivalent-lacking word in ST.

     There are also quite a number of equivalent-lacking idioms. Such English phraseological units as “You cannot eat your cake and have it”, “to dine with Duke Humphrey”, “to send smb. to Coventry” and many others have no regular equivalents in Russian. They are translated either by reproducing their form in TL through a word-for-word translation or by explaining the figurative meaning of the idiom, e.g.:People who live in glass should not throw stones. — Люди,живущие в стеклянных домах, не должны бросать камни; to see eye-to-eye with srnb. - придерживаться одних взглядов.

     Equivalent-lacking grammatical forms give less trouble to the translator. Here occasional substitutes can be classified under three main headings, namely:

     1. Zero translations when the meaning of the grammatical unit is not rendered in the translation since it is practically identical to the meaning of some other unit and can be safely left out. In the sentence “By that time he had already left Britain” — К тому времени онуже уехал из Англии the idea of priority expressed by the Past Perfect Tense needn’t be separately reproduced in TT as it is made superfluous by the presence of “by that time” and “already”.

     2. Approximate translations when the translator makes use of a TL form partially equivalent to the equivalent-lacking SL unit, e.g.: I saw him enter the room — Я видел, как он вошел в комнату.  The Russian/ Romanian languages have no comple6x objects of this type but the meaning of the object clause is a sufficient approximation.

     3. Transformational translation when the translator resorts to one of the grammatical transformations e.g.: Your presence at the meeting is not obligatory. Nor is it desirable — Ваше присутствие на собрании необязательно и даже нежелательно (the syntactical integration).

     As has been emphasized, equivalents are not mechanical substitutes for SL units but they may come handy as a starting point in search of adequate translation. The translator will much profit if he knows many permanent equivalents, is good at selecting among variable equivalents and resourceful at creating occasional equivalents, taking into account all contextual factors.    
 
 

     4. ASPECTS OF TRANSLATING PROCESS  

     Description of the translating process is one of the major tasks of the translation theory. Here we deal with the dynamic aspects of translation trying to understand how the translator performs the transfer operation from ST to TT.

     Psychologically viewed, the translating process must need to include two mental processes - understanding and verbalization. First, the translator understands the contents of ST, that is, reduces the information it contains to his own mental program, and then he develops this program into TT. The problem is that these mental processes are not directly observable and we do not know much of what that program is and how the reduction and development operations are performed. That is why the translating process has to be described in some indirect way. The translation theory achieves this aim by postulating a number of translation models.

     model may describe the translating process either in a general form or by listing a number of specific operations (or transformations) through which the process can, in part, be realized. Translation models can be oriented either toward the situation reflected in the ST contents or toward the meaningful components of the ST contents.

     The existing models of the translating process are, in fact, based on the situational (or referential) model and the semantic-transformational model. These models are supposed to explain the dynamic aspects of translation. In other words, it is presumed that the translator actually makes a mental travel from the original to some interlingual level of equivalence and then further on to the text of translation.

     In the situational model this intermediate level is extralinguistic. It is the described reality, the facts of life that are represented by the verbal description. The process of translating presumably consists in the translator getting beyond the original text to the actual situation described in it. This is the first step of the process, i.e. the break-through to the situation. The second step is for the translator to describe this situation in the target language. Thus the process goes from the text in one language through the extralinguistic situation to the text in another language. The translator first understands what the original is about and then says “the same things” in TL.

     For instance, the translator reads in A. Cronin’s “Citadel”  the description of the main character coming by train to a new place of work: “Manson walked quickly down the platform, searching eagerly for some signs of welcome“. He tries to understand what reality lies behind the words “searching eagerly for some signs of welcome”. The man was alone in a strange place and couldn’t expect any welcome committee or deputation. Obviously, he just wanted to see whether anyone was there to meet him. So, the translator describes the situation in Russian in the following way: «Мэнсон быстро прошел поперрону, оглядываясь, не встречает ли его кто-нибудь».

     A different approach was used by E. Nida who suggested that the translating process may be described as a series of transformations. The transformational model postulates that in any two languages there is a number of nuclear structures which are fully equivalent to each other. Each language has an area of equivalence in respect to the other language. It is presumed that the translator does the translating in three transformational stages.

     First — the stage of analysis — he transforms the original structures into the nuclear structures, i.e. he performs transformation within SL.

     Second —the stage of translation proper —he replaces the SL nuclear structures with the equivalent nuclear structures in TL. And third —the stage of synthesis — he develops the latter into the terminal structures in the text of translation.

     Thus if the English sentence “It is very strange this domination of our intellect by our digestive organs” (J.K. Jerome) is translated into Russian as «Странно, до какой степени пищеварительныеорганы властвуют над нашим рассудком» we presume that the structures “domination of our intellect” and “domination by our digestive organs” were first reduced to the nuclear structures “organs dominate” and “they dominate intellect”, respectively. Then they were replaced by the equivalent Russian structures «органы властвуют/ organele domină» and «они властвуют над рассудком/ ele domină asupra raţiunii», after which the nuclear structures were transformed into the final Russian/Romanian variant.

     A similar approach can be used to describe the translation of semantic units. The semantic model postulates the existence of the “deep”  semantic categories common to SL and TL. It is presumed that the translator first reduces the semantic units of the original to these basic semantic categories and then expresses the appropriate notions by the semantic units of TL.

     Thus if he comes across the sentence “John is the proud owner of a new car“, he is first to realize that it actually means that “John has a new car” and that “he is proud because of that’. After transferring these basic ideas to Russian/Romanian and converting them to the semantically acceptable phrases he will get the translation «У Джона(есть) новая машина, которой он очень гордится / John are o maşină cu care el se mîndreşte».

     Training translators we may teach them to use these models as practical tools. Coming across a specific problem in ST the translator should classify it as situational, structural or semantic and try to solve it by resorting to the appropriate procedure. If, for instance, in the sentence “He is a poor sleeper” the translator sees that the attributive group cannot be directly transferred into Russian/ Romanian, he can find that the transformational model will do the trick for him here and transform the attributive group into a verb-adverb phrase: «Он плохоспит / El doarme rău/ El are insomnie».

     Another approach to the description of the process of translating consists in the identification of different types of operations performed by the translator. The type of operation is identified by comparing the initial and the final texts.