SOURCE:

BOOK TITLE: READINGS IN TRANSLATION THEORIES

11. Text types, translation types and translation assessment

- Katharina Reiss

Reiss' work on text types has been a major influence in contemporary trans­lation theory. Her book on the subject dates from 1976; the present article sum­marizing some of her ideas appeared in 1977. Her approach relates translation closely to text linguistics and communication studies.

With respect to the classification of text types, Reiss starts by sticking to the traditional three based on Bühler's functions of the linguistic sign, but adds an audio-medial type to cover the increasing use of language (and translation) which is linked simultaneously to other media. The special requirements of this text type can be very restricting indeed - such as the number of letters permitted on the TV-screen for a given subtitle - and it makes sense to consider this kind of translation separately.

In her book (1976) Reiss illustrates the relation between the traditional three text types and various text varieties in the form of diagrams. The main points of these can perhaps be summarized as follows. The diagram shows how examples of different text varieties can be approximately placed with respect to the three functions: no text variety represents only one function; each has its own characte­ristic mixture.

INFORMATIVE

reference book report

lecture

operating instructions

tourist brochure

biography sermon

official speech

play electoral speech

poem satire advertisement

EXPRESSIVE OPERATIVE

These placings are of course only rough indications. The primary function of a translated text clearly affects how the translator will operate. Reiss suggests (e.g. 1976: 20ff) that primarily informative texts

105

should be translated in plain prose, with expansions and explanations where necessary. A primarily expressive text needs an "identifying" translation method, where the translator aims at empathy with the original writer. Primarily operative texts require an "adaptive"translation, determined by the way the intended TL receivers are assumed to react to the text. Audio-medial texts should be translated in a"suppletory" way, supplementing what is expressed by the pictures, music, etc.

For Reiss's later work, see e.g. Reiss and Vermeer (1984, 1986). Translation assessment is taken up further in chapter 14, below.

The phenomenon of linguistic translation is probably not much younger than mankind itself, although of course this cannot be established with any certainty: mankind's collective memory, surviving in mythology, mentions the Tower of Babel and its disastrous consequences. We must be content to state that there has always been translation; there has always been criticism of translations; and there have always been clever heads to ponder the problems of translating, while streams of ink have flowed as both homogeneous and heterogeneous views on this theme have been passed down to future genera­tions.

This state of affairs might well lead one resignedly to the following con­clusion: enough of the gruesome game - it's all been said already, for in the

old days they were certainly no more stupid than we are today. How can one find anything new worth saying that has not already been said ages ago?

And yet - the old, eternally young problem of translation exerts an enor­mous fascination on each new generation. This remains true today, when the phenomenon of translation has even more relevance than earlier: people of the most varied tongues have surely never borne witness so clearly to the urgent need of permanent interlingual communication on all levels and in all spheres of life. Hence too the present-day trend to bring all kinds of transla­tion material into translation theory research: not only so-called "literary texts", but also "pragmatic texts", concerning which even Schleiermacher (1813/1963: 42) maintained that translating them was "almost a purely mechanical business which anyone can do who has a reasonable knowledge of the two languages".

Even if perhaps nothing Absolutely new can be said, it may well still be possible to discover hitherto unnoticed cross-relationships, or clarify hitherto overlooked associations. This would contribute towards increasingly removing the translation activity from the sphere of pure intuition, of subjective criteria;

106

it would thus serve to systematize translation problems, make translation itself teachable to some extent, and also objectivize the assessment of translations.

This aim fits with the development of modern linguistics, which aspires to precise verifiable or falsifiable results on the model of the natural sciences. Earlier, the problems of translation were primarily discussed, with very diffe­rent motives, by language philosophers, poets and writers, and also practical translators; in the second half of the present century, however - especially owing to the ambitious aim of constructing a "translation machine" (cf. Wilss 1973) - on the one hand linguistics has turned its attention towards translation, and on the other translation research itself has begun to make use of strictly linguistic methods. The significance and necessity of this cooperation between linguistics and translation research is well illustrated by a formulation of George Mounin (1967: 61): "Translation is never a uniquely and exclusively linguistic operation, but it is first of all and always a linguistic operation. "Of course, like all such apodictic sentences, this statement needs to be qualified: it all depends on how one interprets the word "linguistic". Without going any further into this problem area, I would just like to point out here that the statement is the more readily accepted the more modern linguistics extends its goals beyond the sentence boundary; in other words, the more it is prepared to adopt a textlinguistic approach and see genuine translation as an act of verbal textual communication. (This amounts to a semiotic definition of trans­lation. Semiotically, the sign systems of natural languages are only one means whereby texts can be realized; other possible sign systems include mime, gesture, colour, pictures, etc.)

1. For if we ask why texts are normally translated, the answer must - in a general sense - be: a translation is a communicative service, and normally a service for a target language receiver or receivers. The normal function of a translation service is to include a new (target language) readership in a com­municative act which was originally restricted to the source language com­munity. This holds even for texts which in their source language form might not be thought to be truly communicative, such as diaries, personal memos, notes, etc. If such texts are translated, there must nevertheless be some (secondary) will to communicate. Communication theory will thus be prima­rily concerned to establish some basic systematic order in the enormous multi­plicity of actual translation material. In order to set up a text typology that would be relevant to translation, it thus makes sense to begin with the basic communicative situations in which texts fulfil quite specific and fundamentally distinct communicative functions.

107 -

Since natural languages form the basic material of verbal texts, we first take a look at language itself.

Language has long been classified intuitively, according to the pre­dominant mode of expression, as functional language, literary language or address. In the 1930's the psychologist Karl Bühler (1934/1965) distinguished three functions of a linguistic sign: informative (Darstellung), expressive (Ausdruck ) and vocative (Appell ). The semanticist Ulrich Stiehler (1970: 32) associated these three language functions with the realization of three types-of human cognition: thinking (or perceiving), feeling and willing. The Tübingen linguist Eugenio Coseriu (1970: 27) sees the three functions in terms of their relative dominance in linguistic utterances, and thus distinguishes three lan­guage forms: "a descriptive, declarative or informative language form, the main object of which is providing information about a given topic; an expres­sive or affective or emotive form, mainly expressing the speaker's state of mind or feeling; and a vocative or imperative form which primarily seeks to bring out certain behaviour in the hearer." This classification thus basically re­lates the main objective of a language form to one of the three main elements in the communicative process: sender (= speaker, writer); receiver (= hear­er, reader); and topic (= information).

This tripartite aspect of language itself suggests a similar tripartite divi­sion of basic verbal communicative situations; moreover, the many verbal constituents of the secondary system of language (i.e. its written form) can also be seen in terms of three rough types.

According to their communicative intention, verbal texts thus display three possible communicative functions, correlating with the dominance of one of the three elements of a communicative act as mentioned above. In this way we can distinguish the following three basic types of communicative situa­tion.

(a) Plain communication of facts (news, knowledge, information, argu­ments, opinions, feelings, judgements, intentions, etc.; this is also taken to in­clude purely phatic communication, which thus does not constitute a separate type: the actual information value is zero, and the message is the com­munication process itself: see Vermeer 1976). Here the topic itself is in the foreground of the communicative intention and determines the choice of ver­balization. In the interest of merely transmitting information, the dominant form of language here is functional language. The text is structured primarily on the semantic-syntactic level (cf. Lotmann 1972). If an author of such a text borrows aspects of a literary style, this "expressive" feature is neverthless only a secondary one - as e.g. in book and concert reviews, football reports and the

108

like. The text type corresponding to this basic communicative situation is the "informative" type.

(b) Creative composition, an artistic shaping of the content. Here the sender is in the foreground. The author of the text creates his topics himself; he alone, following only his own creative will, decides on the means of verbali­zation. He consciously exploits the expressive and associative possibilities of the language in order to communicate his thoughts in an artistic, creative way. The text is doubly structured: first on the syntactic-semantic level, and second on the level of artistic organization (Lotmann 1972). The text type correspon­ding to this communicative situation can be referred to as "expressive".

(c) The inducing of behavioural responses. Texts can ,be conceived as stimuli to action or reaction on the part of the reader. Here the form of verba­lization is mainly determined by the (addressed) receiver of the text, by virtue of his being addressable, open to verbal influence on his behaviour. The text is doubly, or even triply structured: on the semantic-syntactic level, (in some circumstances, but not necessarily, on the level of artistic organization,) and on the level of persuasion. The corresponding text type may be called the "operative" one.

(One consequence of this threefold division is of course that in addition to these linguistic functions, an expressive text must also fulfil an artistic func­tion in translation, and an operative text a psychological one.)

2. We now have three basic types which are relevant to translation. If we now apply this classification to the assessment of translations, we can state that a translation is succesful if:

- in an informative text it guarantees direct and full access to the concep­tuäFcontent of the SL text;

- in an expressive text it transmits a direct impression of the artistic form of the conceptual content; and

- in an operative text it produces a text-form which will directly elicit the desired response.

In other words:

(a) If a text was written in the priginal SL communicative situation in order to transmit news, facts, knowledge, etc. (in brief: information in the eve­ryday sense, including the "empty" information of phatic communion), then the translation should should transmit the original information in full, but also without unnecessary redundancy (i.e. aim in the first place at invariance of con­tent). (This relates to the controversy about target text additions or omissions vis-à-vis the source text - see e.g. Savory 1957: 49.)

109 -

_ An example, from Ortega y Gasset (1937/1965: 18-19): - "... es usted una

especie de último abencerraje, último superviviente de una fauna desapare­

cida ..." -> "you are a kind of last 'Abencerraje', a last survivor of an ex­tinct fauna..." This translation is inadequate, because the English reader lacks the Spanish reader's understanding of what the name Abencerraje signifies (a once famous Moorish family in Granada).

(b) If the SL text was written because the author wished to transmit an artistically shaped creative content, then the translation should transmit this content artistically shaped in a similar way in the TL (i.e. aim in the first place at an analogy of the artistic form).

An example: two translations of a line from Rilke's first Duineser Elegie:

"Ein jeder Engel ist schrecklich".

-> (i) "_IRound every angel is terror" (trans. by Wydenbruck)

-> (ii) "Each single angel is terrible" (trans. by Leishman and Spender). This second version mirrors the form of the original. (Cf. Reiss 1975: 57f.)

(c) If the SL text was written in order to bring about certain behaviour in the reader, then the translation should have this same effect on the behaviour of the TL reader (i.e. aim in the first place at the production of identical be­havioural reactions).

An example: an advertisement "Füchse fahren Firestone-Phoenix". If this slogan is only translated "informatively", as "Foxes drive (use) Firestone", the psychologically persuasive ("operative") alliterative element is lost and false associations are evoked: metaphorically, Fuchs is not equivalent to "fox". Suggested version, preserving alliteration: "Profs prefer Firestone­Phoenix".

If a given translation fulfils these postulates, which derive from the com­municative function of a text, then the translator has succeeded in his overall communicative task. ¡

Of course, the full achievement of this goal entails not only a considerati­on of the text type in question - this only indicates the general translation method - but also the specific conventions of a given text variety (Textsorte). Text varieties have been defined by Christa Gniffke-Hubrig (1972) as "fixed forms of public and private communication", which develop historically in lan­guage communities in response to frequently recurring constellations of lin­guistic performance (e.g. letter, recipe, sonnet, fairy-tale, etc). Text varieties can also realize different text types; e.g. letter: private letter about a personal matter -> informative type; epistolary novel -> expressive type; begging-let­

110

ter -> operative type. Limitations of space prohibit a further discussion of this in the present context, but see Reiss (1974) on the problem of text clas­sification from an applied linguistic viewpoint.

The three text types mentioned cover in principle all forms of written texts. However, one must not overlook the fact that there are also compound types, where the three communicative functions (transmission of information, of creatively shaped content, and of impulses to action) are all present, either in alternate stages or simultaneously. Examples might be a didactic poem (in­formation transmitted via an artistic form), or a satirical novel (behavioural responses aroused via an artistic form).

3. However, there is one circumstance which still needs special attention. Written texts often occur in communicative acts together with "texts" of other signs, where the texts in the different sign systems have been produced to re­late to each other in a constant way. The written language is supplemented and accompanied by "texts" in the 'language" of music or of pictures. Examples are: songs, comic strips, advertisements, medieval morality ballads etc. Translation must also take account of these mutual references within the text, lest the interrelation be lost in the TL text. (In the translation of songs, for instance, the target language intonation and prosody must be made to fit the rhythm and melody of the accompanying music, which of course remains constant and, as it were, "sets the tone".) Furthermore, it should be recalled that not all written texts to be translated are ultimately intended to be read; some are better seen as written substrata for an oral communicative act (see Chiu 1973). Examples include songs again, and also plays, speeches, texts for radio and television, etc. The translation of these texts too is based on certain principles, which derive from the special characteristics of the spoken lan­guage and oral communication. These factors do not in any way diminish the validity of the three basic communicative situations and corresponding text types outlined above. Rather, they represent a kind of superstructure; all texts exhibiting these additional factors can be included, as regards their transla­tion, within a single audio-medial text type. From the point of view of transla­tion method, the special requirements of this type take precedence over what­ever basic text type a given text otherwise belongs to, thus:

Figure 1. audio-medial text type

informative expressive operative

text type text type text type

111

There is not the space here to examine in more detail how translation methods and optimal choices of translation procedures are affected by the above-mentioned requirement that translation should normally bring about an integral communicative performance. (For detailed discussion of text types and translation methods, see Reiss 1976).

4. The assessment of translations does not only have to take into account the ideal case of integral communicative performance, in which the aim in the TL is equivalence as regards the conceptual content, linguistic form and com­municative function of a SL text. The practice of translation is subject to a great many conditions which determine that such an integral communicative performance cannot, or even should not, be achieved. Theodore Savory (1957: 49) listed ten translation principles gleaned from the literature, some of which are directly contradictory while others are mutually complementary; they provide an impressive picture of the abundance of opinions about what a cor­rect translation should be like.

1. A translation must give the words of the original.

2. A translation must give the ideas of the original. 3. A translation should read like an original work. 4. A translation should read like a translation.

5. A translation should reflect the style of the original.

6. A translation should possess the style of the translation. 7. A translation should read as a contempory of the original.

8. A translation should read as a contemporary of the translation. 9. A translation may add to or omit from the original. 10. A translation may never add to or omit from the original.

These heterogeneous views cannot simply be dismissed as more or less scur­rilous notions of individual theoreticians or practising translators, for they recur far too frequently in the translation literature, and are often advocated by a broad stream of adherents.

It is easy to see that none of these principles, taken alone, can be valid for all text translations. On the other hand, they could never have arisen and been defended without some support from translation practice. What speaks against them is above all their undifferentiated claim to be defining the best way of translating. If we examine the causes of such contradictory translation principles - which at the same time should also reflect valid principles for

112

translation assessment - then we find at least three ways of explaining them, which can be appealed to either separately or in combination.

(a) The principles apply to only one kind text at a given time - usually this means the translation of the Bible or major works of world literature (espe­cially the Greek or Latin classics) - and their validity is then disputed or erroneously applied to all texts, i.e. made absolute. This is approximately the case with the controversies on which Jerome and Luther took issue (the word­for-word principle against the sense-for-sense principle), and which thereafter reappeared in a great many theoretical discussions.

(b)^-T^o some extent, the principles are closely bound up with a given con­ception of a text. Roughly speaking, if the attainment of equivalence between source and target text is the aim of all translation, it follows that this equivalence will be sought and achieved to different degrees according to the conception of the text, and of equivalence, that one happens to hold in any given case (cf. Vermeer 1973).

Thus, if a text is regarded as being built up of a sum of words, equiva­lence on the word-level will be the ideal of all translation - word-for-word translation (= an interlinear version) is then a "good translation". This view carries the day when the word is taken not only as a formal linguistic unit but also as an independent semantic unit, - or even, as used to be the case with many Bible translators, as something "holy"; its position in the sentence, the next linguistic unit above, is therefore seen as fixed.

If a text is considered to be a collection of individual sentences, equiva­lence on the sentence-level will be taken as the ideal for all translation. With­in the sentence the syntax should be adapted to that of the TL. "Literal trans­lation" is "good translation". (For the difference between word-for-word and literal translation, see e.g. Wilss 1975.) This kind of translation is still practised today as so-called "grammar translation", in order to point out and teach differing linguistic structures in SL and TL to foreign language learners.

If the text is seen as the basic linguistic sign, as in modern text linguistics, equivalence on the text level will be the ideal of all translation. Yet account is taken here only of linguistic equivalence, regarding content and style. The source language sets the criteria. Pragmatic or cultural differences, which come up not only in text variety rules but also in clichés (turns of speech, pro­verbs, etc), are ignored or at best tacitly flattened out. This kind of translation leads to a kind of marked translation (Verfremdung) of the original. The method has been defended inter alia by Ortega y Gasset (1937), following Schleiermacher's work on the different methods of translation (1813); Ortega

113

Gasset argued that this method would lead to a better understanding of the linguistic and cognitive structures of the foreign (i.e. source) language com­munity. Fritz Güttinger (1963) calls this "learned translation".

Only when the text is seen as a verbal component of a total communica­tive event, i.e. as a text- with- a- function, can all the factors of the communica­tion situation be brought into account. The aim is then a communicatively ef­fective translation, of an appropriate text type and text variety, with equivalence not only of content and meaning but also of effect. It is this type of translation that constitutes the "integral communicative performance" referred to above.

(c) A third and final explanation of translation types that appear to deverge from this integral communicaton has to do with the intended function of the SL text and the original communication situation. This concerns all cases where a translated text undergoes a change of function, either because the original function can no longer be identified (e.g. in particular texts of earlier periods and/or past cultures); or because the original situation can no longer be reproduced for a modem reader, so that it needs to be clarified by means of notes, explanations, commentaries etc, which supplement the actual text (as in scholarly Bible translations and other "learned translations"). If this is not done, the function of the translation necessarily alters: a satirical novel such as Gulliver's Travels - i.e. an operative text - turns into ordinary enter­taining fiction, an expressive text. Finally, a change of function may be de­liberately preferred because the translated text has different communicative aims from the original. It is an open question whether such translation per­formances should ultimately still be called translations (Übersetzungen), or whether they should rather be designated as transfers (Übertragungen) because of the change of function.

5. It goes without saying that all the types of translation mentioned may be justified in particular circumstances. An interlinear version can be extremely useful in comparative linguistic research. Grammar translation is a good aid to foreign language learning. Learned translation is appropriate if one wishes to focus on the different means whereby given meanings are verbally expressed in different languages. And the changing of function of a text, as a verbal com­ponent within a total communicative process, may also be a justified solution. However, when the translation is an end in itself, in the sense of simply seeking to extend an originally monolingual communicative process to include receivers in another language, then it must be conceived as an integral

114

communicative performance, which without any extra-textual additions (no­tes, explanations etc) provides an insight into the cognitive meaning, linguistic form and communicative function of the SL text.

The main points discussed above are summarized in the following diag­

ram.

Figure2.

Text concept Translation type Translation aim

Text = sum of word-for-word comparative linguistic

words translation research (interlinear)

Text = sum of literal trans. foreign language

sentences (grammar trans.) learning

Text = basic learned trans. study of culture­

linguistic sign (deliberately bound language

marked + differences

commentary)

Text = verbal communicative

component of translation a) integral comm.

a comm. process a) normal case performance

(text-with-a- b) special subtype b) all kinds of

function) changes of function Î

6. In conclusion: the assessment of a translation, therefore, requires that in the first place one must determine the kind of text the original represents (in terms of text type and text variety); the translator's conception of the transla­tion (to be inferred from his manner of translating, and perhaps also explicitly stated in a translator's preface); and the aim of the translated text. Only when these factors have become established is one in a position to judge a transla­tion "fairly", in accordance with the appropriate criteria. In this way one can avoid the risk of taking one of Savory's provocatively juxtaposed principles as an absolute and biased criterion for the evaluation of translations of all kinds.

---