Linguistic Models of the Translation Process I 1

Linguistic Models of the Translation Process I 1/ Analysis and synthesis An exciting research topic: modeling a translator’s activity also means modeling text comprehension (the path from linguistic form to the mind) and text production (the path from the mind to linguistic form). Transfer from one sign system to the other: divided into two phases: at the first stage the translator analyses the SL text (analysis), whereas he constructs the TL text in the second (synthesis).

Linguistic Models of the Translation Process II The transfer phase Nida’s ford metaphor raises several questions: Where is the ford? How does analysis take place? In what language? Can it be divided into sub-phases? How does the phase of synthesis take place? In what language? What temporal relationship can be detected between the two? At which point does the second replace the first? What is there between the two?

Linguistic Models of the Translation Process III Phase of transfer • Between analysis and synthesis: the interference of some transmitting/mediator language (interlanguage) is hypothesized. • Opposite: no rigid dividing line between them can be drawn as the TL can play a significant role in the comprehension of the SL text, and, conversely, the SL is of pivotal importance to creating the TL text (reading a text in a foreign language, one instinctively translates it into one’s native language; a huge number of translations that reflect the strong influence of the SL. • Segments of analysis: morphemes, words, sentences, paragraphs? • Segments of synthesis: are they longer than those of analysis? • Do analysis and synthesis occur simultaneously?

Linguistic Models of the Translation Process IV Modelling the translation process Two basic types of translation models: denotative and transformational (Rezvin and Rosentzveig 1964) Denotative (situational): way from SL to TL leads through reality (the translator uses his/her knowledge of the world at the point of transfer interpretation) Transformational: direct path from SL to TL (translation proper) Two more models added by Komissarov (1972, 1973): Semantic model The model of equivalence-levels + other attempts at the systematisation of models

Linguistic Models of the Translation Process V The denotative (situational) model I Starts from the principle that, apart from some minor differences, the real world surrounding us is common to us all: which means that the objects signified (denotata) are the same, it is only the linguistic signs denoting them that are different. In the phase of analysis of the translation process the translator traces back the source language signs to the world of objects denoted by them (that is why the model is called denotative). As this world is common to everybody, by doing so the translator makes it clear what particular situation of objective reality is described in the SL text (that is where the term “situational model” derives from). In the phase of synthesis, the translator describes the same denotata and situations by using the tools of the TL. What the translator should avoid, however, is the setting up of a direct reference between the SL and the TL.

Linguistic Models of the Translation Process VI The denotative (situational) model II: advantages and disadvantages Correct reflection of the translation process inasmuch as in several cases equivalents in the TL can only be found if the real situation behind SL text is taken into account. If an object or phenomenon does not have a generally accepted name in the TL, the translator has to (re-)create it on the basis of what s/he knows about the object or phenomenon in question: realia (culture-bound terms, culture-specific terms). Devices translators can use: taking over the foreign word, finding some similar realia in the TL, inventing a new word for the TL (csűrdöngölő ’frolicsome Hungarian dance, barn-stamper, barn dance, Transylvanian dance’). However: „No one can understand the word ‘cheese’ unless he has a nonlinguistic acquaintance with cheese” (Bertrand Russell)

Linguistic Models of the Translation Process VII The denotative (situational) model III: advantages and disadvantages Used when there is only one solution in the TL: Keep off the grass! tracing back to reality Fűre lépni tilos! recreating the same denotatum in the TL Considering only what the TL text says; What happens or does anything happen between the two phases?

Linguistic Models of the Translation Process VIII The denotative (situational) model IV: advantages and disadvantages cannot be used to explain cases in which the target language has several linguistic tools for the description of an object, phenomenon or situation and it is up to the translator to choose the best of them. In such cases it is not just the piece of reality described with the SL signs that is taken into account by the translator. What is important is not just what is said but also how it is said.

Linguistic Models of the Translation Process IX The transformational model I The translator identifies the SL signs not with reality but with the TL signs, as the TL equivalents of the SL signs are given in his mind beforehand, which means that translation consists in the simple replacement of SL signs with TL signs. However: such a model would be practicable only if the SL and the TL were totally identical regarding the number, arrangements and functional rules of elements in them, which means that each SL sign would correspond to only one TL equivalent, and this single TL equivalent would be connected to all the other TL signs under the same rules and conditions as the SL sign is connected to all the other SL elements → obviously impossible in natural languages.

Linguistic Models of the Translation Process X The transformational model II But: this does not at all mean that none of their elements is arranged in the same way or functions under the same rules. Basis of translation from one language into another: the common domain, which the translator reaches after carrying out a series of intralingual, i. e. , language-internal transformations. Antecedents to the transformational model Chomsky’s Aspects of theory of Syntax (1965): concept of „deep structure” Nida’s Science of Translating (1964) and Theory and Practice of Translation (1969): kernels – „the minimal number of structures from which the rest can be … derived; ” basic structures – remarkable similarities between them in different languages

Linguistic Models of the Translation Process XI The transformational model II The translation process according to the transformational model SL surface structure intralingual transformation (analysis) TL surface structure intralingual transformation (synthesis) SL deep structure interlingual transformation (SL deep structure replaced with the equivalent deep structure of the TL )

Linguistic Models of the Translation Process XII The transformational model III Nida-Taber (1969) analysis phase: “back transformation” phase of synthesis: “restructuring” Between these two phases: “transfer” from language A to language B Result of analysis: “kernel” or “core” structures consisting of four classes: objects (e. g. man, dog, tree, sun, river, oil, etc. ), events (walk, see, run, jump, do, etc. ), abstracts (e. g. red, big, quickly, much, etc. ) and relation indicators (e. g. because, and, not, etc. ); Not identical with parts of speech: e. g. promise is a noun in the surface structure of the SL, is an event name in the deep kernel and at the surface level of the TL it can be realized even as a verb. Deep structures are characterized by the relation of their elements being expressed in the clearest and simplest possible way and being prepared for transfer.

Linguistic Models of the Translation Process XIII The transformational model IV: advantages and disadvantages Advantages 1/ By introducing the concept of intralingual transformations and describing their functions, it gives a true picture of different moments of the translation process: the constant consideration of various TL options, which is an organic part of the daily routine of translators, can be regarded as transformation. 2/ Intralingual transformations are also carried out in the process of understanding the SL text. E. g. a complex SL NP can be mentally broken down to several simple sentences. Back-and forth transformations can be useful in translator training

Linguistic Models of the Translation Process XIV The transformational model V: advantages and disadvantages 3/ Intralingual transformations: estimation of the degree of equivalence. A SL sign may not be directly related to a TL one not only because their respective positions are different in the systems of the two languages but also because their “transformational histories” are different. E. g. E development is not always equivalent to H fejlődés, since the former is transformed from develop vt and develop vi, while the H noun is derived from one verb only: fejlődik. 4/ Important role in machine translation: if there is a domain that two languages share (equivalent lexical and grammatical structures, transfer can be done by a machine (pre- and post-editing required).

Linguistic Models of the Translation Process XVI The semantic model I In the transformational model it is tacitly posited that meaning (sense), remains unchanged. Antecedents of the semantic model 1/ Research in semantics (beginning: the 60 s) has influenced TT as well. 2/ Relevant for translation: the ability to decide whether • a sentence has only one or more than one meaning, (He might have told us he would be late) • two (or more) sentences have a common meaning, ie. , are synonymic (Peter likes/cares for/takes to/is keen on/is attracted to/has affection for Mary)

Linguistic Models of the Translation Process XV The transformational model VI: advantages and disadvantages Disadvantages: 1/ Separates the SL and the TL and neglects the fact that the way the SL original is formed plays a significant role in shaping the TL surface. 2/SL and TL structures are related to each other in a way that cannot be traced back to the common domain that the two languages share. These equivalencies can be interpreted under the denotative-situative model, in which the linguistic form plays hardly any role, so there is a way of putting something in one language and a different way in the other: Keep off the grass – Fűre lépni tilos, Staff only – Idegeneknek tilos a bemenet.

Linguistic Models of the Translation Process XVII The semantic model II 3/ the decomposability of meaning, e. g. bachelor: (Noun), (Human), (Male), (one who has never married), etc. Words of the SL broken down to such semantic components → in the phase of SL analysis it is not basic lexical units or grammatical structures that the original text is traced back to but to elementary semantic units, called semes. They form a certain system depending on the context and speech situation: some of them will be relevant (or central), others will be secondary (or peripheral).

Linguistic Models of the Translation Process XVIII The semantic model III The translation process as described by the semantic model SL surface text TL surface text intralingual transformation (analysis) (synthesis) SL semantic deep structure interlingual transformation (SL semantic deep structure replaced with the equivalent deep structure of the TL )

Linguistic Models of the Translation Process XIX The semantic model IV Melchuk and Zholkovsky’s semantic model “Text-sense-text” semantic model: insight into the hypothesised semantic deep structure → a translator first understands the text to be translated then expresses what he has understood - the sense of the text - in the TL. Sense: the speaker’s ability to express the same thought in different ways and the listener’s ability to perceive the identity of sense between statements that look/sound different. The sense of a given text means the common elements that can be found in all the texts that are intuitively regarded as equivalent. This common semantic invariant is described in terms of a special semantic language, or basic language.

Linguistic Models of the Translation Process XX The semantic model V Process of translation (E→H): idiomatic English sense analysis E semantic invariant (basic language) idiomatic Hungarian sense synthesis switching from E basic language to the H one (translation proper)

Linguistic Models of the Translation Process XXI The semantic model VI What does the lexis of the basic language consist of? Three types of elements: 1/ predicates, 2/ nominals, 3/ adjuncts, subdivided into constant elements + changing elements (lexical functions/correlates of the given word) a/ substituting correlates, e. g. : synonyms: storm → Syn → tempest antonyms: good → Ant → bad b/ connecting correlates , e. g. : rain → Incep → begin mourning → Magn → deep sleep → Fin → wake up engine → Fin → shut off

Linguistic Models of the Translation Process XXII The semantic model VII Lexical functions • make the paraphrasing of the SL possible, i. e. explain, express, clarify semantic relations below the surface of the sentence; • contribute to the deconstruction of SL meaning and to the reconstruction of the TL meaning. Each language has tools to express meanings like: • beginning (Incep) • end (Fin) • operation (Oper) which vary from language • causing something to happen (Caus) to language. • annihilation or liquidation (Liqu), • degradation (Degrad), etc.

Linguistic Models of the Translation Process XXIII The semantic model VIII If phrases like a hal/hús megbüdösödik, a tej/bor megsavanyodik, a kenyér megszárad, a gép leáll, az óra megáll are to be translated from H into a foreign language (E), what has to be looked for is not the dictionary equivalents of these verbs but how the TL language expresses the semantic function (Degrad) after the equivalents of hal, hús, tej, kenyér, gép, óra, etc. : • • fish taints quickly meat becomes stinky bread becomes dry/stale milk tuns sour • wine turns acid • engine fails • the clock is out of order

Linguistic Models of the Translation Process XXIV The semantic model IX Such a limited number of semantic functions (cca. 60) – makes it possible to systematise core meanings as well as decompose and restructure them. Melchuk and Zholkovsky have added a few syntactic rules and filters to eliminate incorrect solutions on the way to the TL surface. Advantages and disadvantages I 1/ The translator traces the SL sentences back to sense relations rather than to basic lexical units or elementary grammatical structures. The equivalence between basic lexical units and elementary syntactic structures – no matter whether it is intralingual or interlingual – can be stated only on the basis of identity of sense. 2/ Can answer questions like why should the H verb csinál go into the translator’s mind on one side and the E verb tidy come out on the other, although a dictionary will not give them as equivalents in one entry? It is because under the semantic function (Caus) they should come together in the phrase rendet csinál (a szobában) – tidy (up) the room.

Linguistic Models of the Translation Process XXV The semantic model X Advantages and disadvantages II 3/ Even if it were more exhaustive, the semantic model would not be able to explain instances in which the SL form has no effect whatsoever on the TL surface phrase, which should then be found on the basis of the given situation or the given goal of communication.

Linguistic Models of the Translation Process XXVI Komissarov’s model: levels of equivalence Nida’s “ford” metaphor: • the models have put the ford where the river can be crossed either a long way away (denotative/situative: distancing from linguistic signs); • or very near (transformational: switching from one language to the other without having to leave the level of linguistic signs). Neither of them reflects what the translator actually does. In reality, the translator always wants to cover the route from the SL to the TL at the level of linguistic signs, which is the shortest, and the longer is chosen only if there are no other options. All this means that he applies different models and the level of switching from one language to the other is always determined by the specific task to be carried out.

Linguistic Models of the Translation Process XXVII Levels of transfer in Komissarov’s theory I (1973) 1/ the level of linguistic signs, 2/ the level of utterance, 3/ the level of message, 4/ the level of situation, text can 5/ the level of the communicative goal content levels (to which both the SL and the TL be broken down)

Linguistic Models of the Translation Process XXVIII Levels of transfer in Komissarov’s theory II 5/ The level of communicative goal: The speaker, say, wants to inform the reader, the listener, i. e. , the receiver about something, would like to raise some emotion in them, or would like to make them carry out an act (e. g. , a tenant would like the landlord to mend the roof). 4/ The level of situation: To achieve this goal, the speaker informs the recipient about real objects, persons, abstract phenomena or the relationship between them (miserable living conditions: the roof is leaking). 3/ The level of message: The situation, however, cannot be completely covered at one single time, and its different characteristics are described in different ways: The roof needs mending. When the autumn rains set in you have to wear a macintosh even indoors. Or, as Petőfi put it in Pató Pál úr: the sky peeps in through the attic “Pallásról néz be az ég. . . ”

Linguistic Models of the Translation Process XXIX Levels of transfer in Komissarov’s theory III 2/ The level of utterance: When the message is being worded, the linguistic signs available for it have to be arranged in a linear order, in which the elements can be organized into different structures. The message of the roof having to be mended can be conveyed by various utterances: The roof should be repaired; It is to be mended; The roof wants mending; Why not mend the roof? 1/ The level of linguistic signs: In utterances of similar structures different words can be used: The roof should be mended. The roof ought to be fixed.

Linguistic Models of the Translation Process XXX Levels of transfer in Komissarov’s theory IV • The production of a text is the result of numerous decisions. The sender can choose one of the situations available for the desired goal of communication. To describe the selected situation he chooses one of the messages. To formulate the selected message one of the several utterances is chosen and a selection of the linguistic signs available is made. • The comprehension of a text: the receiver covers the same way in the opposite direction, i. e. , starts from the linguistic signs and reaches the sender’s communicative goal. • The translator, both a receiver and a sender, covers both ways. In the analysis phase he proceeds from the linguistic signs to the goal of communication, whereas in the synthesis phase he proceeds from the goal of communication to the linguistic signs.

Linguistic Models of the Translation Process XXXI Levels of transfer in Komissarov’s theory V • Thus, the selection of the TL equivalent is determined by the goal of communication, irrespective of what situation, message, utterance or linguistic signs it is expressed by in the SL, even if one’s linguistic behaviour is described, for example, if a character has a foreign accent. The fact that someone pronounces the consonant /r/ in a French manner can be translated only with words containing this consonant. The following example is a frequently quoted translation error: “Ó, de boldog vagyok, hogy eljött – mondta a grófnő franciásan ropogtatva az r-eket. ” (I’m so happy you’ve come, said the countess, crunching the rs like the French do) • The level of situation: can also exclude the equivalence of all the other lower levels. Vigyázat, mázolva. I in E: Wet paint! irrespective of how it is structured in H.

Linguistic Models of the Translation Process XXXII Levels of transfer in Komissarov’s theory VI • The level of message: excludes lower level equivalents, e. g. , if you are to comfort someone by saying Cheer up! Chin up! Don’t lose heart!, it can only be translated into H Ne lógasd az orrod! • The level of utterance determines the formulation of the TL equivalent of an E passive in H: The doctors were puzzled by the fact. . . Az orvosok rejtélyesnek találták a tényt. . . , his meals were brought to him in bed. . . a kosztot ágyba hordták neki. . . (Heller: Catch 22, A 22 -es csapdája). • The level of linguistic signs allows no choice in the translation of international organisations which have a constant equivalent in the TL (ENSZ-UNO).

Linguistic Models of the Translation Process XXXIII Levels of transfer in Komissarov’s theory VII Advantages and disadvantages of the equivalence level model 1/ Advantages: gives a more nuanced picture of the work of a translator, who pursues various strategies at the same time. 2/ On the other hand, he idealizes the translator’s work saying that in the phase of analysis the translator always covers all the steps from the level of linguistic signs to that of the goal of communication, and some of this trouble can be spared only if the selection of the TL equivalent is determined by the goal of communication or the situation, which are “higher levels”. 3/ In reality, translators translate elements, as a rule, at the level of linguistic signs and step a stage higher only if setting up equivalents at this level meets with some unexpected difficulty.

Linguistic Models of the Translation Process XXXIV Summary • The translator is constantly moving to and fro between the two languages, is constantly relating the SL and TL signs to each other and reality. Once the TL equvalent has been found, the understanding of the SL text may be made easier, however, it can also lead to misunderstanding. • The deeper the translation grasps the SL text, the more TL versions that have already been found may be rejected by him. The translation process can, perhaps, be best described as the simultaneous functioning of the two languages. The linguistic models of translation contribute to a better understanding of this complicated process.