Text analysis in translation

Checklist

The following questions may help to elicit the relevant information

about the content of the text:

How are the extratextual factors verbalized in the text?

Which are the information units in the text?

Is there a difference between the external and the internal situation?

Are there any gaps of cohesion and/or coherence in the text? Can they be filled without using additional information or material?

What conclusions can be drawn from the analysis of content with reference to other intratextual factors, such as presuppositions, composition, and the stylistic features?

Presuppositions

What is a presupposition

The notion of presupposition is rather complex because What we mean here is the "pragmatic presupposition". These presuppositions are implicitly assumed by the speaker, who takes it for granted that this will also be the case with the listener. Communication can therefore only be successful if speaker and listener both implicitly assume the same presuppositions in sufficient quantity.

For example, the answer Twelfth Night or What You Will presupposes the knowledge on the part of the receiver that this is the title of a play, and this presupposition forms the basis on which the joke "works".

In everyday communication it is usually the factors of the communicative situation which are presupposed to be known to the participants and which are therefore not mentioned explicitly. Nevertheless, they have to be taken into consideration when the utterance is made. If, for example, the referent of the information is a person present in the room, the speaker may lower or raise his/her voice or choose simple or complicated or even coded formulations, etc. Of course, it is usually superfluous to mention the things and persons one can point at.

Presuppositions comprise all the information that the sender expects (= presupposes) to be part of the receiver's horizon. Since the sender wants the utterance to be understood, it seems logical that s/he will only presuppose information which the receiver can be expected to be able to "reconstruct".

Presuppositions may refer not only to the factors and conditions of the situation and to the realities of the source culture, but can also imply facts from the author's biography, aesthetic theories, common text types and their characteristics, metric dispositions, details of subject matter, motives, the topoi and iconography of a certain literary period, ideology, religion, philosophy and mythical concepts, cultural and political conditions of the time, media and forms of representation, the educational situation, or the way a text has been handed down.

Since it is one of the social conventions of communication that an utterance must be neither trivial nor incomprehensible, the sender has to judge the situation, the general background knowledge of the addressee, and the relevance of the information that will be transmitted in the text, in order to decide which presuppositions can be made and which cannot. This convention applies not only to the relationship between the ST sender and the ST receiver, but also to that between the TT producer, i.e. the translator, and the TT receiver. The translator has to take account of the fact that a piece of information that might be "trivial" to the ST receivers because of their source-cultural background knowledge (and therefore is not mentioned in the source text) may be unknown to the TT audience because of their target-cultural background knowledge (and therefore has to be mentioned in the target text) - or vice versa.

How to identify presuppositions in the text

Since a presupposition is by definition a piece of information that is not verbalized, it cannot be "spotted" in the text. In their role as ST receivers, translators are familiar with the source culture and - ideally -understand the presupposed information in the same way as a source-culture receiver. This makes it rather difficult to discover the presuppositions which are contained in the text.

In order to identify the presuppositions, the translator has first of all to ascertain which culture or "world" the text refers to (which may have already been established in the content analysis). Here, an important distinction must be made between factual and fictional texts. Factual texts claim to make a proposition about reality (as generally accepted in the culture in question) whereas fictional texts make no such claim - or at least not in the same way as factual texts. The difference lies in the relationship between the text and the (assumed) reality. Fictional texts are, of course, as real as factual texts, and fictitious information can be contained both in fictional and factual texts.

The categorization of a text as factual or fictional does not primarily depend on the structure of the text itself. It is the author and, above all, the reader who classifies the text according to the concept of reality prevailing in their culture - a concept which is, of course, determined by philosophical and sociological conventions. A text intended to be factual by the ST sender can therefore be "understood" as fictional (and vice versa) by a TT receiver who has a different, culture-specific view of what is "real".

If the ST is "anchored" in the world of the source culture, some information on this world will usually be presupposed in the text because of the maxim of relevance, to put it in Gricean terms. If, on the other hand, the ST refers to the world of the TT receiver, which cannot be assumed to be familiar to the ST receiver, it would seem logical for the ST producer to verbalize a certain amount of information for the ST receiver which then would seem irrelevant to the TT receiver. In either case, the translator will normally adjust the level of explicitness to the (assumed) general background knowledge of the intended TT audience using, for example, expansion or reduction procedures.

If the ST refers to a world that is equally "distant" to both the ST and the TT receivers, it is less probable that translation problems will arise from the contrast of ST and TT presuppositions. In these cases the subject matter dealt with in the ST can be regarded as "generally communicable" or, at least, as "transculturally communicable", i.e. between the two cultures involved in the translation process.

The level of explicitness varies according to text type and text function. It is interesting in this context to note that in fictional texts the situation is often made more explicit than in non-fictional texts. While the comprehension of factual texts is based on the fact that sender and receiver share one model of reality, the fictional text has to start building up a model of its own, either referring explicitly to a realistic model or creating a fictitious one in the text, which can then be related in some degree to an existing realistic model. It can even be contrary to the normal truth values of non-fictional utterances (e.g. in fairy tales). A fictional text must, however, also contain some reference or analogy to the receivers' reality because otherwise they would not be able to find access to the world of the text.

If the information on the internal situation is hidden in certain elements of a fictional text, such as in proper names, regional or social dialect (e.g. Shaw's Pygmalion) etc., it is often extremely difficult to transmit it to the target text, as for instance in the following example, because in a literary text it is often not appropriate to use substitutions, explanatory translations or footnotes.

Example

In Ana Maria Matute's short story Pecado de omision the characters are socially classified by their names. The main character, a simple village boy who in spite of his talents does not get the chance to train for a profession, is only called by his Christian name Lope, whereas his class mate, whose father can afford to let him study law, is introduced by Christian name and surname: ManuelEnriquez. Lope's uncle, the village mayor, has the rather pompous name Emeterio Ruiz Heredia; the school teacher is referred to
by the respectful combination of don together with his Christian name (don Lorenzo). The simple shepherd with whom Lope has to stay in the mountains cannot even boast an individual name: he is called Rogue el Mediano (i.e. "Roque the middle one").

These hidden clues cannot be explained to the TT receiver without running the risk of losing the literary charm of the text. Fortunately, most authors do not rely exclusively on implicit characterizations, but include some explicit hints, as does Ana Maria Matute in the above-mentioned text.

Presupposition indicators

The probability of presuppositions being present can be calculated from the distance of the ST and TT receiver to the cultural environment of the subject matter, as well as from the level of explicitness and the level of redundancy. Text contains certain "elements of crystallization" which may indicate presuppositions. These elements might be attached to certain syntactic or lexical structures, such as the gerund, infinitive, or passive constructions, modal auxiliary verbs or valences of lexemes, as in the following example.

Example

"John will be picked up at the station. Peter is always in time." Since the verb to pick up requires two actants, semantically specifiable as agent and patient, the reader will automatically know that Peter has to refer to the person who is going to pick up John at the station. If the two sentences are to constitute a text, the existence of the agent is presupposed in the first sentence .

Other signals pointing to presuppositions can be provided by the intra-textual dimensions of subject matter, content, sentence structure, and suprasegmental features. The negation left out in an utterance meant to be ironic can, for example, be signalled by a certain intonation: "How very, very clever of you!" Non-verbal elements, such as a photo showing the skyscraper environment of the "immaculate garden flat", can also illustrate presupposed situational conditions.

The analysis of the extratextual dimensions of sender, receiver, time, place, and motive of communication can also reveal presupposed information, as has been pointed out above. With their TC competence, translators will be able to check the comprehensibility of the verbalized information from the TT receiver's point of view. Thus, any possible information gap or surplus in the background knowledge of the intended TT receiver, as described by the translation brief, can be localized and, if necessary, compensated for.

Checklist

The following questions may help to discover the presuppositions made in the source text:

Which model of reality does the information refer to?

Is the reference to reality verbalized explicitly in the text?

Are there any implicit allusions to a certain model of reality?

Does the text contain redundancies which might be superfluous for a TT receiver?

What information presupposed to be known to the ST receiver has to be verbalized for the TT receiver?

Lecture 2. Text Composition

General considerations

The text has an informational macrostructure (i.e. composition and order of information units) consisting of a number of micro-structures. The text segments forming the macrostructure are marked or delimited primarily by the continuity or discontinuity of tenses.

There are several reasons why both the macro and microstruc-ture of the text are important aspects of a translation-oriented text analysis.

If a text is made up of different text segments with different situational conditions, the segments may require different translation strategies according to their different functions.

The special part that the beginning and end of a text play in its comprehension and interpretation means that these may have to be analysed in detail in order to find out how they guide the reception process and influence the effect of the whole text.

For certain genres, there are culture-specific conventions as to their macro and/or microstructure. The analysis of text composition can therefore yield valuable information about the text type (and, perhaps, the text function).

In very complex or incoherent texts, the analysis of informational microstructures may serve to find out the basic information or subject matter of the text.

Text ranks

A source text can be part of a unit of higher rank, which we may call a text combination or hyper-text. Thus, a short story or a scientific article might be included in an anthology or a collection, in which the other texts constitute a frame of reference, and a novel might be intended to form part of a trilogy or tetralogy. The different texts can be related and linked in various ways.

In the practice of professional translating, the parts of a text combination are sometimes translated by different translators, as is shown in the following example.

Example

The German version of the textbook on linguistics edited by Andr6 Martinet (Martinet 1973) was produced by two translators: Chapters 1 to 25 were translated by I. Rehbein, and Chapters 26 to 51 by S. Stelzer. Each of the chapters is an independent text and, at the same time, part of a larger unit, whose characteristics have to be taken into account by both translators.

The inclusion of a text in a unit of higher rank is usually signalled by the title and/or the title context, which can be regarded as a sort of "hyper-sentence" or "metacommunicative utterance".

On the highest rank this hyper-sentence is often replaced by the information about the communicative situation which the receiver infers from extratextual clues. If the extratextual analysis shows, however, that the situation of the TT will differ considerably from that of the ST and that the TT receiver cannot infer sufficient information about the ST situation, the translator may feel obliged to add some kind of hyper-sentence (e.g. in the form of an introductory lead) to the translation.

Example

In German newspapers, comments taken from other papers are usually introduced by hyper-sentences, such as "President Reagan's speech before the UN is commented on by The Times (London)" (cf. Suddeutsche Zeitung, Oct. 26/27, 1985; my translation). The form of these hyper-sentences is culture-specific, and they may even be rather elliptic. In the International Herald Tribune, for example, texts quoted from other papers are printed in a special column under the heading "Other Comments" and signed with the name and place of publication of the reference paper, e.g. "Asiaweek (Hong Kong)".

Macrostructure

Metacommunicative sentences of the type "A says (to B)" can also be signals for the beginning of an embedded text (cf. example 3.1.0./1), these signals separating the different levels of communication. This is particularly important in translation, because, as was pointed out earlier, each level of communication may require a situational analysis of its own. One of the crucial aspects in the analysis of macrostructure is therefore the question of whether there are any sub-texts or in-texts embedded in the ST.

Other forms of in-texts are quotations, footnotes, and examples (e.g. in scientific texts, such as the present study). The main task of the translator is to find out which function the in-text fulfils in the embedding text. Although other extratextual factors (e.g. audience, place, time, medium) may be the same for the embedding text and the in-text, the function must be analysed separately.

Example

Quotations, like other texts, can have an informative, expressive, appellative, and phatic function. The function of a quotation is basically independent of that of the embedding text, although there seems to be a certain correlation between genre and quotation types. For example: In scientific and technical texts we find more informative quotations, whose form is rather conventional (especially where bibliographical references are concerned) than in popularizing texts or (literary) essays, which more often contain expressive quotations stressing the author's own opinion, or quotations appealing to the reader's own experience or which are intended to impress the reader by citing a famous authority, such as Aristotle or Shakespeare.

Footnotes inserted into a target text in order to provide background information or give additional explanations, can also be regarded as in-texts. Since the effect that a text with footnotes has on the reader is different from that of a text without footnotes, the translator has to consider carefully whether other procedures, such as explanatory translations or substitutions, would be more appropriate to the genre and function of the target text than footnotes.

The relationship between the in-text and the embedding text can be compared with that between titles or heading(s) and the text they belong to. A title is a metatext which tells us something about the co-text in question and can equally fulfil various other communicative functions.

Example

The title of Chapter VII of Charles Dickens' The Pickwick Papers not only informs the reader about the contents of the chapter but also recommends the text to the reader: "How Mr. Winkle, instead of shooting at the pigeon and killing the crow, shot at the crow and wounded the pigeon; how the Dingley Dell cricket club played All-Muggleton, and how All-Muggleton dined at the Dingley Dell expense; with other interesting and instructive matters." The metacomrnunicative function of the title is in this case signalled by the form of an indirect question introduced by how. In the title of Chapter I of Jonathan Swift's A Voyage to Lilliput it is made even more explicit: "The author gives some account of himself and family (...)".

Inclusions commenting on the text itself (e.g. so to speak or as I pointed out earlier or to put it into a nutshell) can also be regarded as metacommunicative utterances. At the same time they have the (phatic) function of giving a signal to the receiver, thus representing the (extratextual) audience orientation by intratextual means.

Within the text itself, macrostructure is defined from a semantic point of view. Hierarchical delimitations of text sections (chapter, chunk, paragraph, complex sentence, non-complex sentence, etc.) can only provide a rather superficial orientation. Since the days of classical rhetoric, the beginning and the end of a text are considered to be of particular importance in the interpretation of the whole text. This is why they should be analysed separately.

The beginning and end of a text can be marked by certain verbal or non-verbal features, which in some genres will be even conventional, such as the moral at the end of a fable or the expression once upon a time at the beginning of a fairy tale. The end tends to be less frequently marked than the beginning (the words The End at the end of a film are probably a remnant from the time when the end of a text was conventionally marked by finis). The imminent end of a text can also be signalled by the shift to a higher level of communication, e.g. a metacommunicative recapitulation like "in conclusion, let me restate...". Thus, in the fable The Lover and his Lass, for example, the moral ("Laugh and the world laughs with you, love and you love alone.") establishes a direct communication between sender and receiver.

The example of the fable shows that certain features of text composition are genre specific. Certain text types are characterized by a particular macrostructure and particular structural markers, as well as particular means of conjunction between the text parts. A good example is the text type "letter" with the conventional text segments date, address, salutation, message, and complimentary closing. In an instrumental translation the translator should observe the target-cultural convention for the text type in question.

Microstructure

Both in macro and microstructure we have to distinguish formal and semantic or functional structures. If the highest rank is that of meta-communication and the second rank is constituted by macrostructural units such as chapters and paragraphs (formal structure) or beginning and end (functional structure), the third rank will be that of simple and complex sentences (formal structure). From the semantic or functional point of view we can distinguish information units, utterances, steps of the course of action or plot, or logical relations, such as causality, finality, specification, etc. The fourth rank will then be that of sentence-parts and their relation, such as the theme-rheme structure (TRS).

In written texts, a "sentence" is the unit between two full stops (or question marks, exclamation marks, etc.). In spoken texts it is delimited by intonatory devices, such as pitch or lengthy pauses. In either case, grammatical completeness is not taken into account as a criterion. In spite of all possible reservations regarding this definition, the division into sentences can provide a first approximation to the micro-structure of a text. Moreover, it will lead into the analysis of sentence structures. In a second step, the analyst has to prove whether the formal division into sentences corresponds to the semantic division into information units.

In narrative texts, the information units can coincide with the steps of the course of action. One of the intratextual features of text composition is, in this connection, the order of tenses used in the text.

A composition which follows the course of action represents a structure with an analogy to objects and situations in the real world ("ordo naturalis"), which is not language-specific and therefore does not raise unsolvable problems for the translator - at least where there is no great distance between SC and TC. This applies also to dialogues, which can be regarded as a (chronological) sequence of various monologues.

Composition structures which do not follow the "ordo naturalis" are determined - both on the macro and microstructural level - by culture-specific norms. They are marked by language-specific linking devices (such as renominalization, adversative conjunctions, etc.) or even by means of metre, rhyme, alliteration, and other sonorous figures, which may help to structure the text.

Thematic organization of sentences and clauses

The semantic and functional division of sentences or information units into theme and rheme (TRS, also topic and comment), which belongs to the microstructure of a text, is independent of the syntactic structures, although it is frequently combined with certain syntactical features. Linking the information units by the device of thematic progression the writer at the same time produces a certain macrostructure. Thus, TRS is a feature overlapping micro and macrostructural composition.

For translation-oriented text analysis, we can confine ourselves to the context-bound aspects of TRS. From this point of view, the theme refers to that part of the information presented in a sentence or clause which can be inferred from the (verbal or non-verbal) context (= given information) whereas the rheme is the non-inferrable part of the information (= new information). Irrespective of its grammatical function as subject or predicate or its position at the beginning or the end of the clause, the theme refers to the information stored in what Brown & Yule (1987) call the "presupposition pool" of the participants. This pool contains the information gained from general knowledge, from the situative context of the discourse, and from the completed part of the discourse itself. Each participant has a presupposition pool and this pool is added to as the discourse proceeds.

TRS has to be regarded as a semantic universal which is realized in different ways by different languages.

Markers of text composition

The macrostructure of a text is first and foremost signalled by formal devices used to mark the boundaries of segments of both written and spoken discourse which form large units, such as chapters or paragraphs in written texts and "paratones" in spoken texts. Chapters are marked by chapter headings or numerals, paragraphs by indentations, and paratones by intonation, pauses of more than a second, etc. These non-verbal markers are often combined with lexical markers, e.g. adverbial clauses in initial (first - then -finally) or focussed position (on the one hand - on the other hand). In text types with a conventional "ordo naturalis" (e.g. reports) the composition is marked according to subject matter and content.

Microstructures are marked by means of syntax structures (main/subordinate clauses, tenses, inclusions, etc.) or lexical devices (e.g. cataphora) and by suprasegmental features (focus structures, punctuation, etc.).

Checklist

The following questions may help to discover the main characteristics

of text composition:

Is the ST an independent text or is it embedded in a larger unit of higher rank?

Is the macrostructure of the text marked by optical or other signals?

Is there a conventional composition for this type of text?

Which form of thematic progression is realized in the text?

Non-verbal elements

General considerations

Signs taken from other, non-linguistic, codes, which are used to supplement, illustrate, disambiguate, or intensify the message of the text, are referred to by the functional concept of "non-verbal elements". The term, comprises the paralinguistic elements of face-to-face communication (e.g. facial expressions, gestures, voice quality, etc.) as well as the non-linguistic elements belonging to a written text (photos, illustrations, logos, special types of print, etc.). However, intonational features, pauses, etc. and the graphical devices that perform analogous functions in written communication (punctuation, capitalisation, itali-cisation, etc.) are classified as "suprasegmental features".

Example

If you're an American living abroad and you need to keep track of your calls, you really ought to get the AT&T Card. First of all, you get a monthly itemized bill. A new option even lets you bill your AT&T Card calls to your American Express® Card account. Or, you can choose to be billed to your VISA® or Master Card.

In addition to itemized billing, the AT&T Card makes it easy to reach family, friends and business associates in the States. And, you can take advantage of AT&T USADirect® service , which gets you through to an AT&T Operator in seconds.

For an AT&T Card application, call us collect at 816-6004 Ext. 60, or write to AT&T Card Operations, P.O. Box 419395, Kansas City, MO 64141-0434.

So if you want to know who you called, get the AT&T Card.

Non-verbal elements are particularly audience-oriented.

Forms and functions of non-verbal elements

We have to distinguish non-verbal elements accompanying the text (e.g. layout or gestures) from those supplementing the text (e.g. tables or graphs) or those constituting an independent text part (e.g. pictures of a comic strip) or replacing certain text elements (e.g. the * that replaces a taboo word).

In face-to-face communication we tend to use gestures of the face and the body (such as winking or shrugging). We distinguishe between gestures used more or less involuntarily by speakers to express their feelings and those used intentionally with a specific meaning. While involuntary gestures constitute a universal phenomenon, which, apart from differences in temperament and certain culture-specific conventions, are common to all the peoples of the world, intentional gestures are signs belonging to a culture-specific code. In an interpreting situation it may therefore be necessary for the interpreter to verbalize certain gestures made by the speaker, if there is any risk of misinterpretation. The receivers only see the gestures of the ST speaker and do not usually notice the interpreter in the booth "translating" the gestures into a TC code.

The interplay of verbal and non-verbal text elements is particularly important on the stage. Plays in which the word is subordinate to the gestures are less problematic in translation than plays in which there is a carefully balanced tension between words and gestures. This tension should be regarded as an intentional feature of the text, which the translator may have to reproduce in the TT.

In spoken discourse there are situations where the hearer would not perceive any mimical expressions or gestures of the speaker because of the spatial distance between them (e.g. in an electoral speech on a market square). And there are text types or functions where the use of non-verbal signals is conventionally forbidden. In these cases, non-verbal elements are more and more replaced by suprasegmental linguistic signs, such as stress, intonation, slowing down, etc., which can even develop into genre-specific features (e.g. sermon).

In written communication, mimical expressions or gestures cannot be used; but the reduced pragmatic contextuality of written texts must, of course, be compensated for. This is done partly by the selection of particular verbal elements, especially those representing supra-segmental features in writing (e.g. punctuation, dash, bold type), and partly by additional non-verbal means, such as pictures (a photo of the author, a cartoon illustrating the subject, a drawing showing how to hold the handle of a machine). It may happen that the non-verbal elements convey a piece of information that is even more relevant to the reader than the message transmitted by the text. A number in small print on the label of a wine bottle may in itself be of little interest, but it tells the "connoisseur" more about the quality of the wine than the name.

The range of non-verbal elements used in literature extends from the ancient acrostics to the typographical means which are found in the poems of Klopstock or Stefan George, Apollinaire or E. E. Cummings.

Non-verbal elements can belong to the conventional form of certain text types, such as the shorter lines of traditional poetic texts or the "small print" in contracts.

Of course, it is not always the author or sender with their specific communicative intention who is responsible for the layout and format of a text. But no matter who makes the final decision on text organization - the effect that these elements produce on the receiver remains the same. If the translation skopos requires "equivalence of effect", the translator must, therefore, take account of all types of nonverbal elements.

Illustrations, diagrams, drawings of certain operations, etc. are conventional supplements or even form an integral part of operating instructions or manuals. In some cases it may even be convenient for the translator to try and carry out the instructions him or herself in order to check the coherence of verbal and non-verbal elements and the functionality of the text.

The analysis of non-verbal text elements usually yields some information about the aspects of text composition (e.g. paragraph markers), presuppositions (e.g. marks of omission), lexis (e.g. facial expressions which suggest an ironic meaning), and suprasegmental features (e.g. shortened lines in a poem). Of the extratextual factors it is mainly the intention of the sender and the function of the text which may be characterized by non-verbal elements.

The importance of non-verbal elements in translation Non-verbal text elements are, like verbal elements, culture-specific. Within the framework of a translation-relevant ST analysis the translator has to find out which of the non-verbal elements of the ST can be preserved in translation and which have to be adapted to the norms and conventions of the target culture. A particular logo or name which is intended to have a positive connotation in the source culture may be associated with a negative value in the target culture; the TC conventions may not allow the graphic representation of a certain piece of information; the TC genre norms may require non-verbal instead of verbal representation, etc. What is taken for granted as regards linguistic text elements (that they have to be "translated"), is not always accepted for non-verbal elements, because initiators are often unwilling to commit themselves to the extra expense involved in adapting nonverbal material.

It is not difficult to identify the non-verbal elements of the source text, as they are usually fairly obvious and often predictable in certain media or text types. But it is important in each case to analyze the function of these elements. Quotation marks, for example, can point to an ironical meaning (in which case they represent a suprasegmental feature, i.e. a certain intonation) or to a neologism introduced ad hoc and explained in the text or to a reference to somebody else's utterances (in which case the text producer may want to express a mental reservation, which would have been marked by a wink of the eye in spoken discourse).

Checklist

The following questions may lead to a functional interpretation of

non-verbal elements:

Which non-verbal elements are included in the text?

Which function do they perform with regard to the verbal text parts?

Are they conventionally bound to the text type?

Are they determined by the medium?

Are they specifically linked to the source culture?

Lecture 3. Lexis and Sentence

Lexis

The choice of lexis is determined by both extra and intratexrual factors. Therefore, the characteristics of the lexical items used in a text often yield information not only about the extratextual factors, but also about other intratextual aspects. For example, the semantic and stylistic characteristics of lexis (e.g. connotations, semantic fields, register) may point to the dimensions of content, subject matter, and presuppositions, whereas the formal and grammatical characteristics (e.g. parts of speech, word function, morphology) refer the analyst to predictable syntactic structures and suprasegmental features.

Intratextual determinants of lexis

The selection of lexical items is largely determined by the dimensions of subject matter and content. Depending on the subject matter, certain semantic fields will be represented by more items than others, and the textual connection of key words will constitute isotopic chains throughout the text.

In this context, morphological aspects (suffixes, prefixes, compositions, acronyms, etc.), collocations, idioms, figurative use (metonymy, metaphor), etc. have to be analysed from the point of view of textual semantics. Componential analysis, etymological investigations, and comparative lexicological studies can also be helpful when the meaning of certain words, especially of neologisms, is not clear.

Extratextual determinants of lexis

The field of lexis, on the other hand, illustrates particularly well the interdependence of extratextual and intratextual factors. The extratextual factors not only set the frame of reference for the selection of words, but they are themselves often -directly or indirectly - mentioned in the text. I will therefore deal with the extratextual factors one by one in order to explain the impact these factors can have on the choice of lexical items.

The first question is whether or not the expectations deriving from the external information and clues as to the general character of the sender (time, geographical and social origin, education, status, etc.) or his/her particular position regarding the analysed text (e.g. communicative role) are verified by the text. This also applies to any internal sender who may be mentioned or presupposed in the text, e.g. in the case of quotations or in fictional texts. If the analysis confirms the expectations, such characteristics can be assumed to be non-intentional; if not, it seems likely that by disappointing the receiver's expectations the sender wanted to produce a certain effect. If there is little or no external information on the sender, the analysis of the pragmatic aspects of lexis may provide some clues to the person of the sender.

The second question is whether the author is mentioned in the text as sender. In such a case, the use of the first person, of expressions like in my view in contrast with other persons' opinions, etc. gives the readers the impression that the sender is addressing them directly. In non-fictional texts we can assume that the first person really does refer to the author. For some text types, there are even conventions as to how authors should refer to themselves, e.g. the use of the first person plural or the third person singular

As far as the impact of the sender's intention on lexis is concerned, we have to ask whether and how the intention is reflected by the selection of words or, if there is no external information, what intention can be inferred from the use of words in the text. It is the pragmatic aspect of intentionality in the sense of "concrete interest" underlying the text production which is being analysed in this context.

This intentionality is reflected by those characteristics of lexis which are not due to the specific situational conditions or to norms and conventions, as well as by those features which appear to signal an intentional "violation" of any norms and conventions valid both for the genre in question and for the conditions of medium, place, time, and motive of communication characterizing the situation of the text. This means that a feature of lexis can be assumed to be intentional if the translator has to analyse the interest and the purpose which induced the author to use precisely this expression, this figure, this word.

Example

Language can be used, for example, to camouflage the real significance of an event, as is shown in the following paragraph from an article on "doublespeak": "Attentive observers of the English language also learned recently that the multi-billion-dollar stock market crash of 1987 was simply a fourth-quarter equity retreat; that aircraft don't crash, they have uncontrolled contact *with the ground; that janitors are environmental technicians; that it was a diagnostic misadventure of a high magnitude which caused the death of a patient in a Philadelphia hospital, not malpractice; and that Ronald Reagan wasn't really unconscious while he underwent minor surgery, just in a non-decision-making form." (THE SUNDAY TIMES, 7 January 1990)

In order to elicit the sender's intention it seems advisable to analyse the "degree of originality" of the lexis used in the text. This is common practice with similes and metaphors. But it can also be applied to other figures of speech, such as the adoption of words from other areas of lexis (e.g. language for special purposes in a general text), other registers (e.g. slang words in a formal text), or from regional or social dialects, and to the metonymic use of words (e.g. the Pentagon for the US Ministry of Defense). In all these cases the translator has to examine whether the choice of words is common or at least standardised for certain text types or whether it can be regarded as original or even extravagant.

The analysis of various lexical items in a text can often show that a particular stylistic feature is characteristic of the whole text. If the translation skopos requires the preservation of such features, individual translation decisions (in the field of lexis as well as content, composition, sentence structure, etc.) have to be subordinated to this purpose. The translator has to plan the translation strategies with this overall purpose in mind, looking for the stylistic means which serve to achieve this purpose in the target language and culture instead of translating metaphor by metaphor or simile by simile.

Similarly, the translator should also assess the stylistic implications of the author's "semantic intentionality". Semantic intentionality refers to the reasons which have induced the author to select one particular piece of information for his or her text from the wide range of all possible information, and to the effect that this choice has on the audience. This can be of particular importance in fictional texts since it may be assumed that the number of informational details which the author may choose from is limited only by the situational conditions. The decision to take one specific detail rather than another constitutes an important clue to the author's (stylistic, literary) intention.

A text may not only contain implicit clues to the sender's intention, but also explicit expressions or (often conventional) cliches by which the sender's intention is announced.

Example

"Our aim is therefore to replace a sporadic approach with a systematic one; to minimise - we can never remove - the intuitive element in criteria of analysis." (From the Introduction to Crystal & Davy 1969: 14).

The medium mainly influences the level of style of the lexical elements (colloquial, formal), word formation (e.g. abbreviated words or acronyms as used in mobile phone messages) and deictic expressions (e.g. operating instructions, which come to the receiver together with the machine).

Example

Just a few examples of typical newspaper abbreviations and compounds, collected from one page of THE SUNDAY TIMES (7 January 1990, p. El); Ј215m fraud, pre-tax profits, RAF, ISC, CSF, GEC, GrandMet, Bond Corp, a pubs-for-breweries swap, the UK dairy-produce company, cash-rich institutions, PR group.

The aspect of time is also reflected in deictic elements, in internal time references, and in temporal markings of certain lexical items. This last aspect is particularly relevant both to the translation of old texts and to that of texts whose language is marked as "modern". In old texts we would not expect "modernisms" (and vice versa).

However, the translator has to decide whether the translation skopos requires a "synchronous" or an "actualizing" translation. As it might be difficult for a 21^st century translator to render a text in the language of the 18^th century, s/he should at least take care not to use typically 21^st century lexis (e.g. fashion words).

In Jonathan Swift's A Voyage to Lilliput, archaic forms like giveth, mathema-ticks, physick, Old Jury instead of Old Jewry, my self and words like hosier (in the 1735 edition, reprinted in Gulliver s Travels, Everyman's Library, London 1940) mark the text as "old" without being an obstacle to comprehension. The German translation (Swift 1983), however, is written in unmarked modern German.

Certain text types, such as legal documents, are characterized by archaic lexis.

The motive or occasion for communication may influence the choice of lexis by requiring a particular level of style (e.g. in a funeral address) or certain formulas or cliches. This can be an important aspect when the target text is intended to be used on a different occasion from that of the source text.

Text function (in correlation with the text type) is also frequently reflected in the choice of lexical items. For example, some examples of typical lexical features of the language of newspaper reporting: complex pre and postmodification, typical adjective compounds such as more and faster-arriving, sequences of adjectives; emphatic and colloquial lexis, etc. Language for special purposes and metalanguage are other function-specific fields of word use. Genre conventions point to the fact that the sender is interested in subordinating form to content, thus setting guidelines for a particular effect of the text. If the function changes within the text, the use of text-type conventions or of functional style can signal a particular stylistic interest on the part of the author.

Checklist

The following questions may be helpful in analysing the lexis used in a text.

1. How are the extratextual factors reflected in the use of lexis (regional and social dialects, historical language varieties, choice of register, medium-specific lexis, conventional formulas determined by occasion or function, etc.)?

2. Which features of the lexis used in the text indicate the attitude of the sender and his/her "stylistic interest" (e.g. stylistic markers, connotations, rhetorical figures of speech, such as metaphors and similes, individual word coinages, puns)?

3. Which fields of lexis (terminologies, metalanguage) are represented in the text?

4. Are there any parts of speech (nouns, adjectives) or patterns of word formation (compounds, prefixed words, apocopes) which occur more frequently in the text than would normally be the case?

5. Which level of style can the text be assigned to?

Sentence Structure

General considerations

The formal, functional and stylistic aspects of sentence structure are mentioned as an important factor in almost all approaches to translation-relevant text analysis, although they are not dealt with in any systematic way.

In spite of the transcultural repertoire of syntactic figures of speech, such as parallelisms, chiasms, rhetorical questions, etc., the effect of these figures may vary slightly according to the different language structures. Complex hypotactic sentences are generally regarded as an appropriate means to describe complex facts. However, in German, hypotactic sentences are much more likely to look complicated and intricate (partly because the verb has to be put at the end of subordinate clauses) than, for instance, in Spanish, where the syntax has a principally linear character and where isolated non-finite constructions (gerund, participles, infinitives) are often preferred to subor-i clinate clauses.

The analysis of sentence structure yields information about the characteristics of the subject matter (e.g. simple vs. complex), the text composition ("mise en relief, order of informational details), and the suprasegmental features (stress, speed, tension), and some syntactic figures, such as aposiopesis, may indicate presuppositions. Among the extratextual factors it is primarily the aspects of intention, medium and text function that are characterized by particular sentence structures.

How to find out about sentence structure

The translator gets a first impression of the typical sentence structure of a text by analysing the (average) length and type of the sentences (statements, questions, exclamations, ellipses) and the other constructions which replace sentences (infinitives, past and present participles, gerunds), the distribution of main clauses and subordinate clauses -and inclusions - in the text (paratactical vs. hypotactical sentence structures), and the connection of sentences by connectives, such as conjunctions, temporal adverbs, substitutions, etc.. On the basis of such an analysis, s/he is able to find out how the information given in the text is structured. I wish to stress the point, however, that the analysis of sentence structure is not an aim in itself but must lead to a functional interpretation.

Below the level of sentences and clauses it is the order of the constituents (such as Subject-Predicator-Complement/SPC) or words (e.g. the position of adverbials) that may lead to a further structuring. Depending on their respective norms of word order, intonation, pitch patterns, etc., different languages use different means of focussing certain sentence parts or of giving a "relief to the text. By analysing the different aspects of syntax (e.g. distribution of main and subordinate clauses and non-finite constructions, "mise en relief by tense and aspect) the translator may achieve a solid basis for text interpretation.

In addition to the classical figures of speech it is (mainly, but not only, in literary texts) the deviation from syntactic norms and conventions which is used in order to produce a particular stylistic effect. In these cases, the translator has first to find out what kind of deviation is used and how it works before s/he can decide, whether or how to "translate" it (in the widest sense of the word) in the light of the translation brief.

Example

In his short story Los cachorros ("The Little Dogs"), the Peruvian author Mario Vargas Llosa plays with syntactic structures, boldly mixing narration, direct speech and stream-of-consciousness technique: "Y un dia, toma, su mama, corazon, le regalaba ese pic-up, ipara el solito?, si...". By a syntactic analysis, we can separate the narrative sentence, which conforms to the syntactical norms ("Y un dia su mama le regalaba ese pic-up"), from the inserted elements of direct speech {toma, corazon, si) and interior monologue Qpara el solito?). Reversing these steps of analysis, the translation is easy: "And one day, here you are, his mummy, darling, gave him that record-player, just for him?, yes..."

The syntactic features, too, depend on various other intratextual features, especially content and composition (e.g. distribution of informational details both in the text and in the sentences), lexis (e.g. verbal or nominal constructions), and suprasegmental features (especially focus, intonation). Among the extratextual factors it is mainly the aspects of intention, audience, medium (e.g. speech vs. writing), and function (e.g. conventional structures), which affect the syntactic features.

Checklist

The following questions may be helpful in analysing sentence structure:

Are the sentences long or short, coordinated or subordinated? How are they linked?