2. Literary translation

I will first briefly discuss my general ideas about translating literature. These have derived from my thirty-year career as a translator of English language prose and drama into Finnish and may not be universally applicable. My methods may be idiosyncratic and also, when asked to translate other kinds of texts, I often find that very different skills from the ones I am used to are needed. Some observations are reinforced by my experience of teaching literary translation to university students of languages and translation studies.

As a very crude division, it could be said that unlike other kinds of texts, which tend to be explicit and convey information, literary texts rely on the implicit (Riffaterre 1985: 217). Typically literature is written in the vernacular; and employs everyday vocabulary, which is the fuzziest element in any language, and can thus easily be charged with meaning, although not with any meaning, it must be added. Another important aspect is the fact that literature, particularly prose, more often than not, takes the form of a narrative. There are characters, there is a story, the characters speak, the story is told – so one could say that roughly speaking a narrative consists of dialogue and narrative proper .

I will attempt to draft a model for narrative translation and go on to suggest that non-narrative elements need a different approach.

2.1. Theory: What is narrative translation?

I propose a definition:

Translating narratives is a combined activity of highly disciplined reading and writing open to the (near)bilingual.

The command of the two languages involved is a prerequisite for this kind of translation. It need not be total bilingualism, but anything less than a near-native competence might create problems. Fawcett (1997) gives a daunting list of linguistic components that a translator needs to be able to recognize in both languages. On the word level he/she should be familiar with word sets and collocations, semantic fields, synonymy, polysemy, homonymy, antonymy, and connotations, and on higher levels he/she should be able to relate words to their context, recognize registers, and notice the way themes and rhemes are organized. He/she should recognize rhetoric devices, such as repetition, ellipsis, and referential relationships, and pay attention to links and cohesive devices. My experience as a translator and a teacher would confirm this.

 If one language is allowed to be weaker, I would agree with Dryden (1697: 30) that it must be the source language (SL), because as a rule, one should not attempt to translate literary texts into anything but one’s mother tongue. Bilingual status need by no means be achieved in childhood, the other language can be acquired later in life, and there is space for improvement in both all through one’s career (cf. Toury 1995: 250). 

There are two skills involved, reading and writing; of these reading will be looked at first. Self-evidently, it is impossible to translate a text one does not understand. Toury (1980) stresses that a text is not only ‘a sequence of words, phrases and/or sentences definable as an additive sum of these components’ but ‘a system where every constituent stands in interrelationships of various kinds with other elements of the same text and with the text as a whole.’ (93). Leppihalme (1997: 20) calls for ‘a holistic approach combined with intelligence and reading experience’ in order to find also implicit clues in the source text (ST). Paz (1971: 159) claims that the first stage is ‘no different from that of the reader or critic’. As the text in question is literature, a certain understanding of literary devices and conventions is necessary. Bonnefoy (1976: 186) goes even further, he calls for ‘empathy, shared existence’ in the reading experience and adds ‘If a work does not compel us, it is untranslatable’. It seems that for a literary text an intellectual reading is not enough, emotions must also be employed.

If one takes this view, reading is not a passive experience. Paz (1971: 157) claims that ‘reading is translation within the same language’ and Eco (2001: 14) seems to refer to something similar when he argues that ‘[t]ranslation does not concern comparison between two languages but the interpretation of two texts in two different languages’. This notion could be expanded further: having spent so much time translating texts I often feel my reading experience of an English book is incomplete without a translation into Finnish, it is as if I cannot quite grasp the book any other way.

It is possible to apply the same notion to writing, as Valéry (1953) does:

Writing anything at all, as soon as the writing requires a certain amount of thought and is not a mechanical and unbroken inscribing of spontaneous inner speech, is a work of translation exactly comparable to that of transmuting a text from one language into another. (116)

As Chesterman (1997: 113-14) points out, postmodernist ideas of intertextuality lead to a claim that all writing is translating. If there are no originals, all texts are translations.

Translation is, however, a special case, distinguished from other kinds of writing by the presence of constraints. There are norms a translator has to follow in his/her writing, as Toury (1980) has so perceptively demonstrated. Chesterman (1997) takes this concept further and lists three main norms: (1) the accountability norm or the ethical norm, which means that a translator should be loyal to all parties involved, the writer, his/her readers, him/herself etc; (2) the communication norm or the social norm, which refers to the need to optimise communication; and (3) the relation norm or the linguistic norm, which calls for an appropriate relationship between source text and target text (69). Adherence to these norms leads to translation being a highly disciplined activity. There are other norms operating in translation which relate to reader’s expectations and patronage (ibid), but unlike the above three, they could apply to all writing. Chesterman’s summing up is well thought out and is quite likely to be applicable to actual translation situations.

Nord (1997) uses the concept function or skopos as the axis round which her theory revolves. She stresses the role of the recipient of the text: ‘[i]t is the recipient who ‘completes’ the communicative action by receiving (i.e. using) the text in a certain function’ (47) and adds that the recipient of a translated text cannot be the same as that of the ST as they belong to different cultures (52). This approach to translation has served well in shifting focus to the function of a text and leaving behind abstract speculation on texts without context or readers.

Gutt’s (1990) analysis of translation is based on communication theory and the role of relevance. He claims that when producing his/her TT a translator will keep in mind what is relevant for his/her readers and make sure they can understand it without too much effort (140). Gutt’s arguments are convincing but his theory is rather general and a bit of a dead end.

Paz (1971), who like Valéry is a poet and not a scholar, has an interesting way of looking at translation. He says that translating poetry resembles writing a poem but ‘unfolds in the opposite direction [...]The poet, immersed in the movement of language, in constant verbal preoccupation, chooses a few words – or is chosen by them’ but instead the translator is ‘dismantling the elements of the text, freeing the signs into circulation, then returning them to language’ (159). Surprisingly, I have not found this idea of translating as reversed writing anywhere else, although it seems to capture something rather fundamental about the process. It addresses the crucial question: how and when do words enter the process?

2.2. Practice: How does narrative translation happen?

2.2.1. Models of the translation process

In the literature of translation studies that I have read I have not encountered a description of the practical translation process which I could wholeheartedly apply to my work. Literary translation is a special case, and scholars such as Gutt, Chesterman and Nord are mainly interested in a more general approach, but there is another serious obstacle. It may not be true of other translation processes, but some crucial aspects of translating narratives seem not to be conscious and intuitive actions do not lend themselves to study very easily.

Nord (1997) gives what she calls a ‘looping model’. First the translator establishes the target situation, then he/she analyses the source text and decides which elements are particularly relevant in the production of the target text, and the final step is structuring the target text (33). While there is nothing particularly inappropriate in this model, when applied to translating narratives it overemphasizes the analytical element and fragments the process. The same problems can be found in Leppihalme’s (1997) three stage model which consists of analysis of ST and task, problem-solving and reverbalisation (19). She goes on to add, though, that strategies can be either conscious or intuitive (25).

Chesterman (1997) suggests that an expert (translator or other) uses mainly intuition, and his/her work is beyond rational analysis (149). Rather than seeing the use of intuition as a sign of expertise in translation, I would like to claim that it is an attitude. When translating narratives it pays to deliberately avoid analysis and conscious thinking to keep all options open, otherwise the fuzzy qualities of words can freeze before they are fully explored. In the last few years thinking-aloud-protocol (TAP) –studies have tried to shed light on the translation process and, according to Chesterman, (1997) indicate that

translators of all kinds tend to proceed in jerks: there are smooth, ‘automatic’ patches of activity, interrupted by pauses, problem points where the translator appears to have to think in a non-routine manner. (89)

These different methods have normally been interpreted to reflect routine and non-routine tasks (see Jääskeläinen 1999: 50-51). Anyone observing a translator at work should be able to notice this variation, but if there are two kinds of states in translation, activity and pauses, the question arises, which of these becomes the focus of attention for the scholar. Usually he/she seems to concentrate on the problems, whereas I would like to claim that it is the smooth periods, misleadingly called routine tasks, that form the main translation process. It looks like the TAP-method does not produce what it claims: thinking aloud interferes with the object of study, the translation process. There are indeed interruptions in a translator’s work, but these are an exception, not the rule.

Another key concept, the unit of translation, calls for attention. Malmkjaer claims that the clause is generally accepted as the main unit of translation because even if the text as a whole can be seen as such in theory, ‘it is not possible, in the process of creating a target text, to consider an entire source text at once and to render it as a target text in one fell swoop’ (in Baker 1998: 287). I would very much like to claim that this is what actually happens, albeit during a process of weeks and months. Toury (1980) suggests there could be ‘a constant back-and-forth movement between the single units and the entire text’ (110), but I would like to go further than that and claim that it is the entire text that is indeed the only relevant unit of translation. I see it as an organic whole with all its elements in constant motion where the smaller units are always strictly subordinated to the whole. This must, of course, be seen as a matter of emphasis or principle, rather than an absolute statement, as smaller units certainly exist and can be discussed.

Modern translation scholars do not spend much time discussing the rather useless concept of translatability. There seems to be a consensus on the futility of the hunt for formal equivalence and the ever swinging pendulum between free and literal translation seems to have been buried for awhile. It has been established that translation does not belong to the domain of langue but that of parole. Catford’s shifts (Catford 1969: 73-82) nevertheless, are alive and kicking. A lot of energy is spent trying to establish exactly what kind of shifts a translator makes in relation to the ST to reach a TT. Again this approach fragments the process. My main argument against the way shifts are discussed, however, is more fundamental. Contrasting and comparing source language (SL) elements with target language (TL) elements during the translation act denies the independence of the TL and leads to SL interference in the TT. I believe it is of utmost importance to keep the two languages separate when bridging the gap. Any link between SL and TL will bind the signs that should be freed into circulation, to use Paz’s words (1971: 159). The emergence of actual TL words into the process too early can be an obstacle to free association and may fix the wording before the translator has properly absorbed and grasped the meaning. (This is why bilingual dictionaries can be misleading.) I would like to suggest that shifts do not occur between the ST and the TT, they occur within the TT. Once there is a draft in the TT, shifts can become an issue. It must be stressed that this claim refers to the translation process, translation in action: obviously it is possible, if perhaps not fruitful, to point out shifts between the original ST and the finished TT.

All these models leave something to be desired. Even if translation studies have moved on from prescriptive and purely speculative areas, there is still a certain negativity in the air. Toury (1995) has pointed out that there is a tendency to concentrate on what translation succeeds or fails to be instead of trying to describe in positive terms what translation is (84-5). If translation is seen in terms of shifts from ST to TT, as decision-making and problem-solving, the creative, dynamic nature of the process is undermined, if not completely misunderstood. To call the core of the translation process routine tasks is not very perceptive.

2.2.2. A personal account of narrative translation

 I will now try and describe my experience of how translation of narratives takes place, what happens when I produce a Finnish text based on an English one.

My overall method is almost always the same. After having read the text (normally a novel) once, I launch on producing a draft. I have come to the conclusion that this is the crucial stage of my work and the hardest to explain and describe in words, as at this point I work almost on auto pilot, minimise conscious thinking and tap into my subconscious. During this stage my state of mind is  similar to simultaneous interpreting. It is as if I read the text in English and write it in Finnish. When I translate narrative proper my mental images tend to be visual, I can imagine the setting, the movements, the characters. When translating dialogue, my imagination becomes auditive, it is as if I hear the characters speak.

During the first phase – while, or through, writing – I also hope to absorb the ST, own it, live it in a way that has echoes of Method acting [1] . I hope to relate the text to my own experience and express what I read through my own personality, in my own words. Just like an actor I am constrained by an existing text, there is no space for invention, only expression. Other translators seem to have experienced something similar. Valéry (1953) notices ‘an subconscious identification with the imagined state of mind of [Virgil]’ and adds defiantly ‘Why not?’ (119), Rossetti (1861: 65) talks about ‘self-denial’, and Dryden (1697) calls for men

 capacious of the soul and genius of their authors, without which all their labour will be of no use but to disgrace themselves, and injure the author that falls into their slaughter-house. (31)

Even if this process can take months, I would like to claim that this first stage is ‘one fell swoop’. The process is broken by other activities, by leisure and rest, but continues in the background. By the time I have finished my first draft, the transition from English into Finnish has taken place and from now on I work in Finnish. The draft is nothing like a finished text, but it is Finnish, however crude and stilted.

The next phases are more relaxed, the urgency is gone. The rest of my task tends to take as long if not longer than the first stage. I edit and polish my text meticulously to make sure it is coherent and makes sense. If something does not feel right I change it. If asked, I might be able to explain why, but as there is normally no one asking, I do not look for explicit justification. The story is taking shape in Finnish and I am the writer. Every now and then I am forced to go back to the original to check that I have understood correctly, but I do not feel comfortable with this. I worry that the English will interfere with my Finnish. There is a conscious effort to keep problems of comprehension separate from problems of expression. I should have it all in my head by now, know it by heart like an actor knows his/her lines, my Finnish TT providing the cues as I go along. It never works hundred percent, but I try to look at the original as seldom as I can and in big chunks, preferably chapters. When there is no more need to refer to the original and the text reads as text written in Finnish, my work is done. As with any piece of writing there is, nevertheless, always space for improvement, as is well demonstrated by the many changes I have made to my published translation (TSF/P) for the revised version of the first volume (TSF/R) which is included as an appendix. This modification cannot probably go on ad infinitum, as at the end of the road there looms a danger of overworking a text so that it loses the spontaneity and inspiration of the first draft and becomes what in translators’ slang is pejoratively called ‘academic’.

2.2.3. Interruptions in translating narratives

I have explained above how in my view translating, when all is going well, is a dynamic highly concentrated activity similar to performing arts, where most of the processes elude conscious analysis. My claim is that the more there are internal interruptions, the more difficult and painful it will be for the translator to produce a coherent and relevant translation, which can be called literature in the target culture. External interruptions do not seem to matter that much, in fact they can be quite welcome, as the processes can go on subconsciously. By internal interruptions I mean elements that defy the comprehension-interpretation act so violently that they shake the translator out of his/her concentration. These elements force the translator to change his/her working method: give up what I called an attitude to deliberately avoid analysis and conscious thinking in order to keep all options open. He/she will have to solve these problems consciously, either when he/she encounters them in the text, or later. No doubt different translators have different ways of dealing with these interruptions, or bumps. If I encounter them in the middle of writing my first draft, and a quick look into a CD dictionary or encyclopaedia does not help, I tend to ignore them so that I do not lose momentum, and deal with them later, either during my editing process or sometimes completely separately.

 It is not surprising that translation studies often focus on aspects of translation that can be seen as problems and require conscious processing. It is very difficult to imagine a method for the study of such elusive acts of translating that I have tried to describe above. Who knows what happens when a musician plays his/her instrument? Who can read the heart of an actor holding Yorick’s skull? This does not mean that the results cannot be studied empirically, only that the act itself seems unfathomable. The ‘black box’ remains inaccessible. A translator can discuss his/her performance afterwards and a scholar can examine various aspects of the end product, compare it to the original, speculate and analyse. In fact, I would like to suggest that if a translator comes back to his work, his/her approach will not be very dissimilar to that of an outsider, as there are seldom actual memories of what he/she has done and why, and most of these are likely to relate to an interruption, some problem of comprehension or expression, not the nuts and bolts of reading a text in one language and writing it out in another. Admittedly I have very little proof for my claim, mainly only my own experience.

This is why, after having explained the rule, I will concentrate on the exception in my study on translating Tristram Shandy below. I have decided to focus on the elements in Tristram Shandy which required conscious problem solving. They are the bumps in the ride of reading, and indeed translating Sterne’s masterpiece. I cannot claim to discuss all the problems I had, as Tristram Shandy is a highly complex novel. Even so, most of my work did happen as I described earlier, the great bulk of the text was conveyed into Finnish during days and months of frenzied typing -- interrupted by very long pauses indeed. Analysis came afterwards.

These problematic elements are called non-narrative in this study, and will be discussed in detail later on. Problems which arose from simple lack of comprehension or a block in writing are not included: I will concentrate on elements which are objectively problematic because they do not consist of straightforward narrative proper and dialogue. For practical reasons they have been classified in three groups: insertions, ambiguity and visuals. The analysis will focus on the first volume of Tristram Shandy, but to illustrate a point, material from later volumes has occasionally been used as well. Before starting this discussion, however, the book, Laurence Sterne’s Tristram Shandy, will be introduced.