Sei sulla pagina 1di 26

Across Languages and Cultures 7 (2), pp.

145–170 (2006)
DOI: 10.1556/Acr.7.2006.2.1



School of Languages, Linguistics and Cultures
The University of Manchester
Oxford Road, Manchester M13 9PL, United Kingdom
Phone: +44 (0) 161 2753237

Abstract: In the paper I bring together two sets of theories from Narrative Theory and
from Retranslation Theory. Links and similarities between the theories are examined under
the headings of Essence, Social Conditioning, and Interpretation. A post-structuralist narra-
tive theory is presented, and I extrapolate from this to propose a post-structuralist retransla-
tion theory. After the theoretical discussion I report on the study of a corpus comprising
Zola’s novel Nana and its five major British (re)translations. The aim is to evaluate how
well the theories regarding narrative versions and retranslations hold up with respect to a
study of data. A conclusion is reached as to which theories best explain the data. The paper
concludes too that bringing together sets of theories from different but related disciplines
can be productive in conceptualizing translational phenomena, in this case the phenomenon
of retranslation.

Keywords: Retranslation Theory, Narrative Theory, Zola’s Nana


The purpose of this paper is to bring together two sets of theories, one from Lit-
erary Studies (Narrative Theory) and the other from Translation Studies (Re-
translation Theory1), in order to examine their links, similarities, and in the end
to promote cross-fertilization of the theories as a means of better understanding
translation practice. This follows in the line of the long and fruitful tradition in
Translation Studies of finding ideas and methods in other disciplines (linguis-
tics, psychology, sociology, etc.).
Before bringing the two sets of theories together in a deliberate manner, it
can be observed that there are obvious links between Narrative Theory and Re-
translation Theory. There is a topic in Narrative Theory which explicitly raises

1585-1923/$ 20.00 © 2006 Akadémiai Kiadó, Budapest


the issue of translation, the topic of narrative versions. Starting with a novel, an
abridged version can be produced, a film adaptation can be made based on the
novel, and a translation into another language can be produced. These are a few
of the types of narrative version; translation constitutes one type. When consid-
ering retranslation, we are also talking about many different versions, in this
case translational versions of one source text into the same language,2 usually
over a certain time period. There is a further link between retranslation theory
and narrative theory relating to the genre of texts under consideration. All kinds
of texts are retranslated. However, the genres which have been massively re-
translated are sacred texts, and canonical literary works. Much of the writing on
retranslation has discussed literary texts. Literary theory is relevant to literary
translation in a range of ways (Brownlie forthcoming), and thus there is a link
between retranslation theory and narrative theory as part of literary theory.
The discussion that follows is organized under sub-headings which express
the similarities between the theories. I group together theories from Narrative
Theory and from Retranslation Theory under the following sub-headings: Es-
sence, Social Conditioning, Interpretation, and Post-structuralism. The discus-
sion progresses in several ways. Firstly, the theories are discussed in rough
chronological order: the earliest narrative theory, for example, had its heyday in
the 1960s, and the most recent text on narrative was written in 2000. Secondly,
the two sets of theories become more and more closely related: although I dis-
cern a basic similarity between them, the first two theories discussed are very
different since they come from quite different intellectual traditions (Structural-
ism and Romantic Idealism), whereas later theories are very similar and even
intertwined. Finally, I start with a series of existing theories, before ending with
a theory which I have created by extrapolation (post-structuralist retranslation
After the theoretical discussion I shall report briefly on a case study in
which I investigated the relevance of the theoretical ideas to studying a corpus
consisting of an original text and its retranslations. The corpus comprises the
French novel, Nana, by Emile Zola, published in 1880, and its five major Brit-
ish translations (published in 1884, 1895, 1956, 1972, and 1992).


2.1. Essence

2.2.1. Dualism in Narratives

The dualist approach in narratology has a number of well-known proponents, of

which one of the more recent is Seymour Chatman. For Chatman each narrative

Across Languages and Cultures 7 (2) (2006)


has two parts: story and discourse. Story is the content, the deep structure con-
sisting of existents (characters and settings) and events in a chronological order.
Discourse is the surface manifestation, the means by which the content is com-
municated, the reworking and telling of the story. Chronological time may be
reordered in discourse to incorporate, for example, flashbacks and flashfor-
wards. Rather than being a series of traits, character is expressed in different
ways in discourse: what characters say, what they do, and what others say about
them. The discourse level also involves narrative techniques (first person, third
person narration) and focalization (telling from a certain point of view)
(Chatman 1978). For Chatman narrative versions can be explained by the fact
that story is able to be transposed and thus form the deep structure of different
versions, such as a novel and its film adaptation. Chatman argues that the trans-
posability of the story supports the claim that narratives are structures inde-
pendent of any medium; that words are not the ultimate components of narra-
tive, rather the ultimate components are elements of the deep structure
(Chatman 1981). With regard to narrative versions, Chatman does not mention
translation explicitly, but another narratologist, Schlomith Rimmon-Kenan,
does in her reference to languages:

Starting from a story, rather than with the text from which it is abstracted,
the former may be grasped as transferable from medium to medium, from
language to language, and within the same language. (Rimmon-Kenan

Thus the possibility of transfer of deep structure is linked to the possibility of

translation and retranslation.
Chatman later produced a more complex version of his notion of transfer in
saying that it is not only the deep structure that is transferred, but elements of
the surface structure: “different versions depend on transfer with minimal varia-
tion of discourse and story properties” (Chatman 1981:260). This statement was
produced in response to Herrnstein Smith’s (1980) criticism of his theory, and it
is the earlier and simpler version of his theory that is the object of Smith’s criti-
cism, as we shall see later.
I have categorized Chatman’s theory under the heading ‘Essence’ since the
story or deep structure constitutes a fixed essence which is said to be transfer-
able without change.

2.1.2. Progress in Retranslation

Antoine Berman’s theory of retranslation (1990) as a process of improvement

from one retranslation to the next also implies an essence. This time it is not a
deep structure, but something essential about the source text as a whole. In

Across Languages and Cultures 7 (2) (2006)


Berman’s view, the improvement in retranslations is realized as the successive

translations come closer to conveying the essence of the source text, to reveal-
ing the truth of the being of the source text. In order to arrive at a great or ca-
nonical translation, a series of stages is played out by retranslations:

First there is a courageous ‘introduction’ without literary pretension (usu-

ally for those studying the work); then comes the time of the first transla-
tions with literary ambition – they are generally not complete translations,
and as is well-known, full of flaws; then come the (many) retranslations…
Eventually a canonical translation may be produced which will stop the cy-
cle of retranslations for a long time. (Berman 1995:57, my translation)

Berman’s ideas are inspired by German Romanticism, notably Goethe who

speaks of a retranslation cycle: the first rendering or ‘introduction’ is word-for-
word, then come fairly free target-oriented translations, then translations which
are closer to the source text (Berman 1990:4). Since Berman is highly in favour
of source-oriented translation which respects meticulously the patterns and
specificities of the source text in its material and non-material aspects, the pro-
gression towards more source-oriented translation represents improvement for
him. Although we may not espouse Berman’s theory of global progress and im-
provement, it is certain that at a local level i.e. at the level of individual words
and phrases, errors may be corrected in a retranslation in much the same way as
in the second edition of a book.
Another theorist, Paul Bensimon, explains why earlier translations tend to
be target-oriented and later translations source-oriented.3 This is because ini-
tially a culture is often reluctant to embrace a text which is very foreign to it, so
in order for the foreign text to be accepted into the new cultural sphere, it has to
be adapted to the target culture. Later on, since the text has already been intro-
duced, it is really no longer foreign, and translations can return to the original
and be more source-oriented (Bensimon 1990:IX).
The idea that first translations are target-oriented, and later retranslations
are source-oriented has been taken up by more recent theorists (notably Ches-
terman 2000; Paloposki and Koskinen 2004). Most interestingly, these theorists
call the idea the ‘retranslation hypothesis’. This is a case of transfer of an idea
from one intellectual tradition or paradigm (Romantic Idealism) into another
(Natural Sciences): whereas Berman talks of ‘the truth of being’, Chesterman et
al. talk of ‘a hypothesis to be tested’. Because the idea is taken out of its origi-
nal context, it no longer carries with it the implication of improvement over
time and of essence, which is the main point of our present discussion. I shall
nevertheless report briefly on the findings of Paloposki and Koskinen,4 who set
out to test the retranslation hypothesis using a corpus of translations into Fin-
nish. Paloposki and Koskinen study the time period of 1809 to 1850, which

Across Languages and Cultures 7 (2) (2006)


represents the beginning of literary translation into Finnish. They do indeed find
that early translations are domesticating, but they propose that this may be a
feature of a phase in the development of a literature (in this case an initial
phase), rather than a feature of first translations in general. In examining trans-
lations from other historical periods, Paloposki and Koskinen conclude that
their data is not well described by the retranslation hypothesis, since there are
cases of the opposite scenario, and also cases where other factors are needed to
explain developments (Paloposki and Koskinen 2004).

2.2. Social Conditioning

2.2.1. Narrative as Social Activity

As mentioned above, Barbara Herrnstein Smith (1980) has criticized the dualis-
tic theory of narrative, and hers is a trenchant criticism. Dualism seems interest-
ing in theory, and it is easy to see why it has been convincing, but when in prac-
tice you ask what the deep structure (or ‘basic story’ as Smith terms it) of a nar-
rative or set of narrative versions actually is, there is immediately a problem.
This is because, contra Chatman’s assertion, it is impossible to get beneath or
out of language. This impossibility also implies that any statement is always
situated in a particular context, is uttered by a particular person who has a spe-
cific background and interests. A statement about basic structure will thus not
reveal any inherent structure of the narrative, but is an interpretation condi-
tioned by specific circumstances.
Smith discusses a study made of 345 variants of the Cinderella narrative
from around the world. It was found that the variation of the narratives was such
that it was difficult to establish a common core beyond a very general level.
Furthermore, even if there was some consensus on a common core or basic
story, this would be because the people producing this basic story had a similar
background, and similar purpose. Another situation is the case of a film which
is very complex in its presentation of time, with multiple flashbacks, flashfor-
wards, dreams, etc. While watching the film the audience may reconstruct the
narrative in a chronological order in order to help them comprehend the film,
since in Western culture we are used to perceiving chronological time as the
natural order of time. In both of these cases, establishing a basic Cinderella
story and establishing the chronological order of a film narrative, the people in-
volved have constructed a narrative for a particular purpose in particular cir-
cumstances, and this does not provide evidence for a deep narrative structure.
Smith concludes that for any narrative there is no single basic story subsisting
beneath it, but an unlimited number of narratives that can be constructed in rela-
tion to it.

Across Languages and Cultures 7 (2) (2006)


Smith therefore conceives of narrative as always situated, as a social activ-

ity, and proposes that it is futile and misguided to seek for deep structures or to
consider that a deep structure constitutes the basis for a set of narrative versions.
Rather, Smith proposes that “similarities and differences among sets of narra-
tives can be explored and explained on the basis of similarities and differences
in the specific conditions that elicit and constrain them” (Smith 1980:234). The
form and function of a narrative version will depend on the circumstances and
motives that elicited it. With respect to translations as a particular type of narra-
tive version, Smith says that “[s]ome versions such as translations and transcrip-
tions may be constructed in order to preserve and transmit a culturally valued
verbal structure” (Smith 1980:221). For Smith, the eliciting conditions of narra-
tives include broad social factors, as well as local circumstantial conditions in-
cluding the motives and interests of narrators and audiences.

2.2.2. Ideologies and Norms Impacting on Translation

In retranslation theory it is the broad social factors that have been taken up as
major explanatory sources. Translations and retranslations are said to vary ac-
cording to different ideologies and norms prevailing in different time periods in
the culture which initiates the translation. The distinction between ideologies
and norms is not always clear. Roughly, ideologies can be defined as sets of be-
liefs, whereas norms are related to practices. There is an ambiguity at the heart
of the term ‘norm’, in that it can apply to what is normal, and also to what is
approved behaviour. This ambiguity no doubt arises because a prescriptive
norm evolves from a descriptive norm: a particular behaviour is initially a re-
peated practice, then comes to be considered as an acceptable and approved
practice in society (see Toury 1995; Hermans 1996). Granted this ambiguity, we
can consider that the main types of norms which affect translation are linguistic,
literary, and translational. Language, poetics, and notions of approved transla-
tional behaviour evolve over time. Retranslations are undertaken because there
has been a change in ideologies and/or norms in the initiating culture (usually
the target culture), and the translation is thought to have aged or is unacceptable
because it no longer conforms to the current ways of thinking or behaving. The
study of retranslations can thus reveal changing norms and ideologies in soci-
In order to illustrate the above, I shall discuss one study which shows that
retranslations have been influenced by changing norms (Du-Nour 1995), and
another which shows the influence of different ideologies (Kujamäki 2001).
Du-Nour studies the translation of children’s literature into Hebrew from
the early 20th century up to the early 1990s. Her corpus consists of nine original
books and several (two to five) translations of each book. At the start of the pe-

Across Languages and Cultures 7 (2) (2006)


riod under study Hebrew was being revived as a modern living language. At an
early stage in the revival, ‘classical’ language and style in the written language
were promoted. This norm was particularly strong in children’s literature. Thus
the early children’s translations into Hebrew were in a quasi-biblical elevated
style. Gradually the norms evolved and by the 1970s modern Hebrew usage was
being employed consistently in retranslations, and consequently a less elevated
style was used. The (re)translations conformed to linguistic and stylistic norms
of children’s literature at the particular time in question. There were also spe-
cifically translational norms: prior to about 1980 deletions were permissible in
the translations, and adaptation of cultural specificities was undertaken, for ex-
ample, ‘Dorothy’ of The Wizard of Oz was replaced by a Hebrew equivalent,
Dorit. Post-1980 retranslations show that such operations were no longer ac-
ceptable. Despite the adoption of current usage in translations, there was a reti-
cence about using colloquial language to reflect colloquial source text dia-
logues. This reticence only recently started to be overcome. Across the century,
then, retranslations were conditioned in various ways by changing norms.
Kujamäki studies eight retranslations of one Finnish work, Aleksis Kivi’s
Seitsemän veljestä, into German undertaken between 1901 and 1997. The title
of the German translations is Die Sieben Brüder. I shall discuss as examples
two translations from specific time periods. The first translation was undertaken
in 1921 when Finland had recently gained independence from Russia. The Fin-
nish government was keen to promote the Finnish identity and its separateness
from the Slavs. One of the means of promotion was translation, and the gov-
ernment promoted and financed translation of Finnish literature. The ideology
of promotion of a national identity is evidenced in the translation of Seitsemän
veljestä, particularly in the footnotes where careful explanations of elements of
Finnish culture are given. The ideology of Finland distancing itself from Slavic
ties is evidenced in the footnotes too, where emphasis is put on Finland’s links
with Western Europe. The 1935 translation of the same work had quite a differ-
ent commissioner, the Nordische Gesellschaft, an association incorporated
within the German National Socialist organization, and responsible for cultural
relations between the 3rd Reich and the Nordic area. The ideology of the time in
Germany considered the Nordic race and culture to be superior, and to be the
forefathers of Germany. The influence of this ideology of a quasi-mythical su-
perior people is evidenced in the translation, in that any negative comment
about Finland or the Finns in the original text is omitted or toned down: this
concerns quarrelling and swearing, high consumption of alcohol, an attempted
suicide, and famines.
Literary norms and traditions also affect the choice to (re)translate certain
types of text. Venuti (2004:34) points out, for example, that the dominance of a
neoclassical aesthetic in 18th century Britain was instrumental in the repeated
translation of classical epics.

Across Languages and Cultures 7 (2) (2006)


Although it is usually the case that retranslations conform with reigning

norms and ideologies, this is not always so. On occasion retranslations may
challenge such values (as we shall see in the discussion of Tyndale’s Bible in
the next section).

2.3. Interpretation

2.3.1. Multiple Readings

Literary criticism is called by Barbara Johnson (1980) the “art of rereading”, of

producing new interpretations. As such, the question of interpretation must be
taken into account in narrative theory. The term ‘rereading’ implies that it is the
reader who is responsible for producing a new interpretation, and this is the
case. But in Johnson’s deconstructive outlook it is also the text which allows
different interpretations: “A text’s difference is not its uniqueness, its special
identity. It is the text’s way of differing from itself. And this difference is per-
ceived only in the act of rereading” (Johnson 1980:4). Difference subverts the
idea of identity, infinitely deferring the possibility of adding up the sum of a
text’s parts or meanings, and reaching a totalized integrated whole. The decon-
structive approach to reading proceeds by teasing out warring forces of signifi-
cation within the text itself. Another reason for multiple interpretations is that
each time the text is re-read, it is re-read in a new context.

2.3.2. Reinterpretation and Translation

Certain genres of text, notably literary and religious texts, are open-ended, and
thus readily lend themselves to multiple interpretations. At a fundamental level
every translation can be considered to be an interpretation, and every translator
an interpreter. Thus reinterpretation has been given as an important reason for a
retranslation, particularly of religious and literary works. From a survey of arti-
cle abstracts5 which mention reinterpretation and translation, it was deduced
that entire reinterpretations of long texts such as a novel are probably less com-
mon than new interpretations and corresponding translations of a passage, a
chapter, some verses, or a short text. Reinterpretations may occur on an even
smaller scale within translations: odd phrases and sentences may be interpreted
differently from one translation to the next. Reinterpretation thus occurs at all
textual levels.
In the abstracts surveyed various motivations are given for reinterpretation
and thus retranslation. A reinterpretation may be based on allusions, ambiguity
or obscurity of the text or passage in question. Ancient texts pose a particular
challenge: reinterpretations/retranslations of such texts involve new understand-

Across Languages and Cultures 7 (2) (2006)


ings of an ancient language and culture. A reinterpretation may be undertaken

within a specific framework or approach, for example a psychoanalytic reading
or a Christian reinterpretation. Different interpretations immediately imply in-
tertextuality, for a new interpretation/translation demarcates itself and justifies
itself in comparison with previous versions. The intertextual and intratextual
network is dense, since reinterpretations not only position themselves in con-
trast to former interpretations/translations, but they draw on support from other
parts of the text in question, and from other texts such as scholars’ work.
Vanderschelden uses the metaphor ‘hot’ and ‘cold’ translations to distinguish a
first translation (hot) undertaken soon after publication of the source text, and
retranslations (cold) which are undertaken with the distance afforded by passed
time, and which can make use of knowledge of earlier translations, evaluations
of those translations, and of the critical reception of the work. All of these fac-
tors may impact on interpretation (Vanderschelden 2000:8).
Venuti’s (2004) appreciation of retranslation combines considerations of
reinterpretation, intertextuality, ideology and institutional pressures. With re-
gard to ideology and institutions, retranslations can maintain and strengthen the
institutionalized interpretations of a canonical text. Alternatively, retranslations
can challenge that interpretation. A good example is found in the history of Bi-
ble translation. The King James Bible consolidated the authority of the Angli-
can Church in the 17th century by drawing on previous Protestant versions such
as Tyndale’s. And yet before the Reformation in England, Tyndale’s translation
was considered heretical because it ran counter to the Vulgate of the Catholic
Church. In the case of non-canonical texts, retranslation may be undertaken in a
bid to achieve canonicity through inscription of a different interpretation. This
has occurred in the feminist rediscovery of neglected women writers (Venuti
In general, different interpretations have a diachronic basis, but it is also
possible that differing interpretations and manners of translating exist synchron-
ically in two published translations, although for economic reasons the exis-
tence of such twin texts is rare (Vanderschelden 2000:12).
It could be said that a text itself contains the possibilities for a new inter-
pretation. At the same time each of the motivations for reinterpretation is de-
pendent on a new context, whether that involves a new interpreter, a different
time period, a new conceptual framework, a changed institutional goal, a new
interest group, and/or a new intertextual set, thus corroborating the notion that it
is a new context which gives birth to a reinterpretation informing a retransla-
tion. Reiteration in a new context is a key concept in post-structuralist theory.

Across Languages and Cultures 7 (2) (2006)


2.4. Post-structuralism
2.4.1. Post-structuralist Narrative Theory

Andrew Gibson (1996) criticizes traditional narratology (cf. Chatman) as being

geometric, static, hierarchized and universalizing, in contrast with postmodern
energetics, fragmentation, movement and multiple spaces. For the purposes of a
link to translation, Martin Mcquillan’s (2000) post-structuralist narrative theory
is highly appropriate. It is a rich theory drawing on a range of theorists. I will
report on that part of Mcquillan’s theory inspired primarily by Derrida and Fou-
cault, and will outline some basic aspects of the theory. In a departure from tra-
ditional units of narrative text, Mcquillan redefines narrative as any minimal
linguistic written or verbal act and calls it a ‘narrative mark’: “the representation
of an instance, no matter how small, of inter-subjective experience” (Mcquillan
2000:8). The reference to inter-subjectivity evokes context, in that language is
not simply operating on its own but is inextricably linked to contexts of utter-
ances, incorporating both intersubjective and broad historical and social ele-
ments. Text and context are inseparable: context provides the text/utterance
with meaning, and the utterance gives rise to the context. In order for meaning
to be possible, texts and language must be iterable: it is only through repetition
that language accrues meaning, but at the same time repetition in a different
context entails the introduction of difference.6 Insofar as iterability is a neces-
sary condition of meaning, the narrative mark is inevitably iterable. This is one
way in which the narrative mark is limitless. Another way has to do with iden-
tity. A narrative mark gains its identity through being distinguished from what it
is not, other narrative marks. At the same time, because of the narrative mark’s
dependence on others, those others structurally constitute its identity. Thus lim-
its are set up but are permeable; one narrative is haunted by others.
Mcquillan calls the field of narratives a ‘narrative matrix’. The matrix is
boundless. All narrative marks are related to the others, and, as we have seen,
the possibility of limits is denied to the narrative mark. Every mark initiates fu-
ture responses in the syntagmatic construction of the matrix, and it is the ex-
change of narratives within the narrative matrix that builds up inter-subjectivity.
The desire to produce narrative can never be satisfied by the reception of the
narrative mark, hence the interminability of exchange. Iterability ensures that
reception by the socially positioned other is always incomplete. There are thus
contextual and textual workings which together ensure the impossibility of clo-
sure. This impossibility and the necessity of exchange lead to the constant pro-
duction of narrative and counter-narrative. With regard to power, again textual
system and context are interlinked. The functioning of the narrative matrix is
such that counter-narratives are always produced, but it is only by the necessary
anchoring in the real world that those counter-narratives are attributed specific

Across Languages and Cultures 7 (2) (2006)


political force. Immanent narrative force relations are both independent of the
politico-ethical determination of power struggles of real others, and always al-
ready implicated within them.
With regard to the issue of narrative versions, we may link this to Mcquil-
lan’s notion that any understanding of an object is constituted in the form of
narrative. Thus any narrativisation is necessarily a renarrativisation of another
narrative. Each new narrative both carries on and cancels out narrative material
from the previous item in the narrative chain (Mcquillan 2000).

2.4.2. Proposal of a Post-structuralist Retranslation Theory

Drawing on the above discussion, I shall propose elements of a post-

structuralist retranslation theory. A source text is reiterated in a new context,
thus instituting stability through repetition, instability through the new situation,
and contextual specificity, all of which enter into the translation. This contex-
tual specificity includes elements of specific subjectivity, as well as broader
situational and historical elements. The reiteration of the source text calls forth a
translation in the ongoing process of exchange and production of narratives, and
further reiterations call forth retranslations. A (re)translation both carries on and
cancels out aspects of the source text and of previous translations. A set of re-
translations of the source text constitutes a subset of the narrative matrix. Each
translation is related to all other past and future narrative marks, thus constitut-
ing and challenging its own identity; in particular, each translation is haunted by
the other translations in the set of retranslations. The open set of retranslations
has its own power dynamics, as well as expressing power relations in society.

2.4.3. Revisiting the Norms/Ideology Approach to Retranslation

The norms/ideology approach to retranslation is a powerful approach, but due to

the emphasis on dominant broad social patterns, it could tend to neglect com-
plexities. Some fundamental notions from post-structuralist theory can serve to
remind us of those complexities.
The norms/ideology approach implies conditioning by large social forces,
and thus a hierarchical relationship. In some guises a hierarchical relationship
may be depicted as deterministic. Of course there are always governing rela-
tions of unequal power at play in language and society, but there are also rela-
tions of more equal power. In post-structuralist theory there is emphasis on mul-
tiple relations of many kinds, which are rhizomatic (like a network of plant
roots) rather than hierarchical. In this connection Robert Stam has suggested the
useful trope of dialogue, specifically intertextual dialogism:

Across Languages and Cultures 7 (2) (2006)


The concept of intertextual dialogism suggests that every text forms an in-
tersection of textual surfaces… In the broadest sense, intertextual dialo-
gism refers to the infinite and open-ended possibilities generated by all the
discursive practices of a culture, the entire matrix of communicative utter-
ances within which the text is situated. (Stam 2000:64)

The trope of dialogue allows for exchange and negotiation with forces of vari-
ous power differentials.
De-emphasizing large governing social forces allows for more complex
and detailed discussion. Explanation for what is going on in retranslations may
be found not (only) at the broad social level, but in specific contextual circum-
stances which give a significant role to the individual commissioner and transla-
tor. A return to Herrnstein Smith’s theory, which emphasizes local context, in-
dividuality, and interests, is warranted. With regard to interests, a retranslation
may be motivated by the particular purposes, and ideological and/or poetologi-
cal investments of commissioner or translator. A retranslation of a classic may
be commissioned for purely commercial reasons on the part of a publishing
house. Collombat (2004) argues that the influence of translators’ interests is a
characteristic of the spate of late twentieth century and early twenty-first cen-
tury retranslations, in which translators have deliberately adopted particular
postulates. A well-known example that she cites is the strategy adopted by a
group of translators/researchers at McGill University to translate Faulkner’s
Southern United States sociolect in The Hamlet by Quebec vernacular (Collom-
bat 2004:11). Although the broader environment provides the ideology and op-
tion of translating into Quebec vernacular, it is the translators who have chosen
that option in the particular case. It is not easy for translators to decide on a
general strategy, and make decisions on particular instances, since there are al-
ways multiple forces at play. In such a situation Frank expresses the negotia-
tional, agential, and context-sensitive role of the translator: “By adopting, adapt-
ing, and rejecting…norms, a translator brings together source side, work, and
target side in an act of interpretive, context-responsive translational transfer”
(Frank 1990:54).
Another area of potential over-simplification in the norms/ideology ap-
proach is the tendency to consider that there are different time periods each with
a different set of norms/ideologies, which explains the changing characteristics
of translations. Post-structuralist theory works to challenge boundaries. In the
context of retranslation this challenge occurs in two ways. Firstly, instead of
conceiving of a retranslation occurring in a different time period, and being mo-
tivated by and reflecting the change of norms/ideology, the study of retransla-
tion must be open to and conceptualize the possibility of more than one transla-
tion being undertaken during one time period. A case in point is studied by Seb-
nem Susam-Sarajeva (2003), when she considers the multiple translations into

Across Languages and Cultures 7 (2) (2006)


Turkish of works by the French theorist Barthes undertaken during a fairly short
time span (1975–1990). Susam-Sarajeva finds that retranslations are not neces-
sarily the consequence of ‘ageing’ translations or ‘changing times’. In the case
of Turkish Barthes, retranslations occurred as the result of an effervescent situa-
tion in the receiving system, which was struggling to create an indigenous liter-
ary critical discourse through competing terminological proposals in transla-
tions. We have also mentioned above the case of literary translations published
synchronically, which offer different readings of and different manners of ren-
dering the source text.
The second way in which time period boundaries can be challenged is that
in practice there is not always a neat and homologous relationship between time
period and norms/ideology. Norms typically associated with one time period
may appear occasionally in another time period. Within one translation there
may be evidence of heterogeneity of norms. Earlier and later translations may
haunt the present one. We can take these phenomena to be due to the operation
of unbounded textuality, and/or they may be explained by the translator’s role
of deliberation with regard to various options (as referred to above).


Space permits only a limited consideration of the case study corpus which com-
prises Emile Zola’s novel, Nana, and its five main British translations. I have
labeled the translations with letters from the alphabet to aid presentation: A is
the 1884 translation published by Henry Vizetelly (the translator is anony-
mous); B is the 1895 translation by Victor Plarr; C is the 1956 translation by
Charles Duff; D is the 1972 translation by George Holden; and E is the 1992
translation by Douglas Parmée. I shall study the corpus data in relation to the
ideas presented in the theoretical section, and thus evaluate how well the theo-
ries hold up with respect to the data.

3.1. Changing Ideologies

I shall start with what seems to be the most striking aspect of the set of transla-
tions, particularly with regard to the first translation (published in 1884) in con-
trast to the retranslations. Zola’s Nana is the story of a Parisian working-class
girl, Nana, who rises in society to become a famous courtesan. Given the sub-
ject matter as well as Zola’s naturalist writing style, there are many references
to sensuality, and the language is quite crude. Such features were not acceptable
to the British Victorian middle class ideology of moral uprightness and ‘deli-
cacy’. An important aspect of delicacy was linguistic prudishness: verbal refer-

Across Languages and Cultures 7 (2) (2006)


ences to sex, sensuality, bodily functions, and sensual parts of the body were
avoided, or euphemisms were used. Another aspect of delicacy was the avoid-
ance of coarse language and swearing. In particular, taking the Lord’s name in
vain was not acceptable. This is linked to the prevailing piousness and religious
conservatism of the time (Weeks 1981; Perrin 1969). The effect of the Victorian
middle class ideology is apparent in the way Translation A was undertaken. The
contextual background to Translation A is that the publisher, Henry Vizetelly,
wished to publish popular editions in English of Zola’s works. Given the aim of
publication for a broad readership, the powerful middle-class and its ideology,
and the threat of prosecution under the Obscene Publications Act, the pub-
lisher/translator undertook ‘self-censorship’ in the translation (Merkle 2000).
Later translations do not show this heavy degree of self-censorship. Let us ex-
amine some examples.
Minor omission is a common way of dealing with ‘offensive’ parts of the
source text in Translation A. In the first example the list of body parts of the
original is replaced by the single term, a superordinate, ‘body’, whereas the re-
translations do not operate this reduction:

Example 1
…avec ses rires, avec sa gorge et sa croupe, gonflés de vices (159)
[with her laughter, her bosom and her rump swollen with vices]
A: with her smiles and her body full of vice (124)
B: with her laughter, and her bosom, and her hips, which seemed swollen
with many vices (142)
C: with her laughter, her bosom and her rump swollen out with wickedness
D: with her laughter, her breasts and her crupper, which seemed swollen
with vice (155)
E: with her laugher, her breasts, the curves of her buttocks…vicious to the
core. (129)7

Another regularly used technique of self-censorship is substitution. In the fol-

lowing example Translation A substitutes a phrase with a quite different mean-
ing from the source text:

Example 2
– Tout de même on coucherait avec, déclara Fauchery. (100)
[‘All the same you’d sleep with her’, declared Fauchery.]
A: ‘All the same she’s a fine woman,’ declared Fauchery. (72)
B: ‘All the same, all right in bed,’ declared Fauchery. (85)

Across Languages and Cultures 7 (2) (2006)


C: ‘All the same, she’s nicely made,’ declared Fauchery. (67)

D: ‘All the same, she’d be all right in bed,’ declared Fauchery. (96)
E: ‘All the same, she’s eminently bedworthy’, said Fauchery. (74)

A less frequently found technique is to leave the offending word in the foreign
language without explanation, as in the following example:

Example 3
Nana était enceinte de trois mois (388)
[Nana was three months pregnant]
A: Nana was three months enceinte (324)
B: Nana had been in the family way for the past three months (364)
C: Nana was three months pregnant (313)
D: Nana had been pregnant for the last three months (385)
E: Nana was three months pregnant (344)

The examples examined so far display obvious cases of self-censorship in

Translation A. There are also many examples of more subtle self-censorship.
This occurs when the particular rendering of a word or phrase is ideologically
motivated. Translation A, for example, displays a more negative and moralizing
attitude towards illicit sexual relations than the original text. At one stage Nana
has an infatuated relationship with a young man, Georges. Since he is so young
and inexperienced, so in love with her, the relationship between Nana and
Georges resembles a first innocent and passionate love. The French text states
that love blossoms in Nana despite her being experienced and disgusted with
men, whereas in Translation A it is Nana’s knowledge of men (i.e. knowledge
gained through prostitution or prostitution itself) which disgusts her:

Example 4
C’était, sous la caresse de cette enfance, une fleur d’amour refleurissant
chez elle, dans l’habitude et le dégoût de l’homme. (194)
[Beneath the child’s caresses, a flower of love bloomed again in her, de-
spite her experience of and disgust for men.]
A: Beneath the child’s caresses, the flower of love bloomed again, in spite
of her knowledge of man, and the loathing it caused her. (155)
B: under the caressing influence of this renewed childhood, love’s white
flower once more blossomed forth in a nature which had grown hack-
neyed and disgusted in the service of the other sex. (176)

Across Languages and Cultures 7 (2) (2006)


C: Under the caress of this renewed childhood a flower of love bloomed

again in Nana, a woman habituated to men and disgusted with them.
D: under the caressing influence of this new childhood the flower of love
blossomed again in a nature jaded and disgusted by experience of men.
E: She was used to being with men, who disgusted her; but now love was
rekindled by the caresses of a child. (162)

The above examples relate to sex, sensuality, and the body. Religion is another
area in which self-censorship is undertaken in Translation A. In Example 5,
there is a case of expletives, where ‘God’ (Dieu) is mentioned in all versions
except Translation A:

Example 5
Nom de Dieu! Foutez-moi la paix (291)
A: Damnation! Go to the deuce. (240)
B: Good God! Why the hell can’t you shut up? (270)
C: God damn it, shut up! (229)
D: God almighty! Shut your trap, will you! (288)
E: Christ almighty! For God’s sake, why don’t you bugger off? (253)

In Zola’s novel Count Muffat, who is Nana’s principal benefactor, is a very re-
ligious man. Count Muffat’s experience of religion and religious ecstasy is
compared with his feelings and experience with Nana, the courtesan. Finding
similarity between religion and prostitution would certainly not have been ac-
ceptable to the Victorian church.8 In the following passage Nana in a stage role
is compared with God. Translation A makes impossible any reference to the
Christian God:

Example 6
Paris la verrait toujours comme ça, allumée au milieu du cristal, en l’air,
ainsi qu’un bon Dieu. (464)
[Paris would always see her like that, illuminated in the midst of the crys-
tal, in the air, like a good God.]
A: Paris would ever see her thus, beaming in the midst of the crystal,
poised in the air like a goddess. (389)
B: Paris would always picture her thus – would see her shining high up
among crystal glass like the good God Himself. (436)

Across Languages and Cultures 7 (2) (2006)


C: Paris would always see her like that: brightly illuminated in the centre
of the crystal, high in the air, just as a good God might be. (378)
D: Paris would always see her like that, shining high up in the midst of all
that glittering crystal, like the Blessed Sacrament. (460)
E: Paris would always see her like that, blazing with light in the middle of
all that crystal, floating in the air like an image of the good Lord. (415)

I have discussed Translation A so far under the category of ideology, but there
is a clear overlap between social ideologies and literary norms, in that what is
considered acceptable in literary texts is affected by current social mores. I shall
continue the discussion of the group of texts as a whole under the topic of liter-
ary norms.

3.2. Changing Literary, Linguistic and Translational norms

In Victorian times the dominant norm in the writing of novels was a certain
‘delicacy’ of expression, of which we have seen evidence in Translation A ren-
derings in the previous section. Novels had to be written with young innocent
girls in mind as prospective readers (Perrin 1969). This is not the norm in con-
temporary English literature: coarse language, including (religious) swear words
(see Example 5), and explicit reference to sex and sensuality are acceptable. The
following example is wonderfully ‘tidy’ in that it could illustrate how the norm
with regard to the expression of sensuality in literature changed gradually over
time. Each rendering is more explicit than the previous one:

Example 7
ce fleuve d’or dont le flot lui coulait entre les membres (416)
[this river of gold whose stream flowed between her limbs]
A: this everflowing river of gold (347)
B: a river of gold, the tide of which almost enveloped her (390)
C: this great river of gold, the flood of which ran between her legs (336)
D: this river of gold which flowed between her legs (412)
E: this stream of money flowing through her thighs (369)

Of course the progression is not so easily displayed in other passages. One in-
teresting effect of the norm of sensual explicitness in contemporary literature is
that on occasion where there does not seem to be a sensual connotation in the
French, the most recent translations (D and E) add a sensual reference, as in the

Across Languages and Cultures 7 (2) (2006)


Example 8
La grande chaise avait une mine chiffonnée, un renversement de dossier
qui l’amusaient, maintenant. (88)
A: The big easy-chair had a tumbled look, and a curve in the back which
now rather amused him. (63)
B: The big chair had a rumpled look – its nether cushions had been rum-
bled, a fact which now amused him. (74)
C: The big chair had a rumpled expression with its back cushion reversed;
and now it amused him. (57)
D: The big chair had a rumpled look, its back a suggestive slant which now
amused him. (84)
E: That large armchair had a saucy look, its back was tipped up in a way,
which, on reflection, was amusingly suggestive. (63)

I shall now turn to linguistic norms. The justification often given for retransla-
tions is to produce a translation in contemporary target language, which is there-
fore acceptable to a contemporary readership. A striking example of a change of
linguistic norms reflected in the translations is the following. At issue is the
rendering of the French fille which means loose woman/prostitute. The term is
translated ‘gay women’ in Translation A: in late 19th century English, this meant
a loose woman, whereas in contemporary English the term means a lesbian.
Clearly the later translators were obliged to use contemporary language in order
to avoid miscomprehension. Here is the passage concerned:

Example 9
éprouvant cette sorte d’obsession qu’exercent les filles sur les bourgeoises
les plus dignes. (199)
A: experiencing that kind of witchery exercised by gay women over the
most respectable ladies. (159)
B: gave evidence of the absorbing curiosity with which notorious courte-
sans are able to inspire even the worthiest old ladies. (180)
C: feeling the sort of obsession which strumpets arouse in the most worthy
middle-class matrons. (150)
D: revealing that obsessive fascination which courtesans exert on the wor-
thiest of ladies. (195)
E: with the sort of obsessive fascination that the most respectable women
feel towards ladies of easy virtue. (166)

Notice the rendering in Translation E, ‘ladies of easy virtue’. This phrase is

typical of Victorian expressions. Translators may not conform with the modern-

Across Languages and Cultures 7 (2) (2006)


izing trend, and may deliberately archaize in translating a source text from an
earlier period, in order to evoke a flavour of the period, or feel for the character
being described.
As far as translational norms are concerned, there has evidently been a
change with respect to the completeness and ‘faithfulness’ of a translation. In
Translation A there are throughout a large number of minor changes with re-
spect to the source text, in particular substitutions and omissions. And yet it is
stated on the title page that this translation has been undertaken “without
abridgment”. Today’s translational norms would not accept such a translation as

3.3. Reinterpretation

In the corpus there are examples of differing interpretations among the transla-
tors for particular phrases. This occurs where the French text allows a double
meaning. In the following example the French mal allows more than one inter-
pretation (‘evil’, ‘disease’):

Example 10
en face de la mort, avec la peur sourde du mal. (473)
A: …with the secret dread of evil. (396)
B: …a dull dread of coming ill possessed them. (445)
C: …feeling the mute fear of evil to themselves. (386)
D: …and filled with a vague fear of disease. (468)
E: …remembering their secret fear of the disease. (423)

Different interpretations also occur when the French text is vague. In Example
11 it is difficult to know exactly what the French means: literally it is ‘she
wanted to go there’, but this is metaphorical in the context. Among the different
renderings, there is one which is not supported by the context: translation B,
which implies a desire to commit suicide:

Example 11
Elle s’ennuyait trop, elle voulait y passer… (371)
A: She felt so dull, she would try the change. (310)
B: She was too much bored by existence, she said; she wanted to get out of
it. (348)
C: She was too bored, she wanted to get away from that place. (299)
D: She was bored, she said, and wanted a different sort of life. (369)
E: She was so bored, she said she wanted to have a go. (329)

Across Languages and Cultures 7 (2) (2006)


Interpretation may be influenced by extra-textual factors. In Example 12 the ex-

pression comme s’il avait voulu entrer en elle translates literally ‘as if he had
wanted to enter into her’. This expression could be taken metaphorically or
spiritually, ‘becoming a part of somebody’. Predictably in the Victorian Trans-
lation A it is the metaphorical interpretation which is privileged, whereas in the
two most recent translations D and E, the interpretation is clearly physical. The
direction of interpretation is thus influenced by the two very different social
contexts, Victorian and contemporary:

Example 12
la face entre ses genoux, qu’il s’enfonçait dans la chair…se meurtrissant
davantage contre ses jambes, comme s’il avait voulu entrer en elle. (299)
A: with his face between her knees, which he was pressing against his
breast…he pressed harder against her, as though he wished to become a
part of her. (248)
B: pressing his face hard against her knees…wildly, savagely he pressed
his face against her knees as though he had been anxious to force
through her flesh. (278)
C: with his face between her knees, burying it in her flesh…beating him-
self more and more against her legs, as if he wished to enter into her.
D: pressing his face hard against her knees…savagely pressing his face
against her legs as if he wanted to force his way into her flesh. (297)
E: pressing his face hard into her flesh between her knees…pressing his
face harder and harder against her legs as though wanting to force his
way into her, between her thighs. (260)

3.4. Heterogeneity and Individuality

Translations A and B appear to present a significant case of heterogeneity in

that they belong to the same time period, late 19th century, Translation B being
produced only 11 years after Translation A. In principle they should both be
subject to Victorian mores, but as can be seen in the examples above, Transla-
tion B does not shy away from sexual topics or unflattering religious references,
and seems therefore surprisingly modern. The explanation for this is to be found
in the context of production of Translation B. This translation was produced by
the Lutetian Society, a secret literary society which had a restricted membership
composed of the elite ruling classes. The aim of this society was to produce un-
expurgated translations of Continental literature for the limited number of its
members; 310 copies were published (Merkle 2003). Private societies were able
to subvert the dominant ideologies and norms, since they were not subject to

Across Languages and Cultures 7 (2) (2006)


censorship. Quite different conditions elicited the two translations, even though
they were produced in the same time period, and it is the different contexts
which explain the divergence between the translations.
A further significant source of heterogeneity is when a translation is incon-
sistent with its own regularities. As mentioned in the previous paragraph, Trans-
lation B is a close translation of the source text and does not comply with Victo-
rian niceties. However, on occasion the Victorian way of expression infiltrates
the translation, as in Example 3 above where enceinte [pregnant] is translated
by the Victorian expression, ‘in the family way’. Furthermore, like Translation
A, Translation B is sometimes toned down to be less direct or explicit than the
original text (see Examples 1 and 4).
There are also instances of past translations ‘haunting’ present ones.9 In
Example 2 Translation C seems to be a throw-back to Translation A in its
bowdlerization. The rendering of Translation E in Example 9, which brings a
Victorian expression ‘ladies of easy virtue’ into the most recent translation (to
translate filles), could be interpreted as a haunting from the past, and is certainly
heterogeneous, since elsewhere in the translation the translator uses ‘tart’ to
translate the French word.
‘Ladies of easy virtue’ has an old-worldly quaintness about it, just like
Translation E’s expression ‘eminently bedworthy’ in Example 2. This could il-
lustrate a case of individuality, showing Translator E’s particular style of trans-
lation. Although Translators D and E belong to the same contemporary period,
Translator E’s style contrasts markedly with Translator D’s, thus illustrating an-
other type of heterogeneity within the same time period. The following example
not only displays a different interpretation on the part of Translator D, but also
an idiosyncratic usage: he uses the word ‘fuck’ elsewhere in the translation:

Example 13
Mais baise-moi donc! Oh! Plus fort que ça, mon Mimi! (410)
A: But kiss me! – oh! More than that, my Mimi! (343)
B: But kiss me, kiss me! Oh, harder than that, Mimi dear! (385)
C: But do kiss me! Oh, harder than that, my own Mimi. (332)
D: Come on, fuck me then! Oh, harder than that, Mimi! (407)
E: Oh, give me a kiss…no, harder than that, Mimi darling! (365)

Perhaps the most striking example of heterogeneity is a case where the normally
prudish Translation A contains a rendering (‘which became more intimate else-
where’), which is more explicit than the French, and which resembles the most
recent translation, Translation E. In this case it is as if the present translation is
haunting the past one:

Across Languages and Cultures 7 (2) (2006)


Example 14
ébauchant là des connaissances, qui se dénouaient ailleurs (434)
[beginning acquaintances which ended up elsewhere]
A: forming acquaintances there which became more intimate elsewhere
B: beginning acquaintances which ended elsewhere (408)
C: getting to know people with whom she ended up elsewhere (352)
D: making acquaintances which ended up elsewhere (430)
E: picking up men whom she then got to know far more intimately else-
where later on (387)

The examples in this section display an almost contradictory state: the particu-
larity of specific contexts and styles, and their non-particularity due to the fluid-
ity between time periods, contexts, and retranslations.

3.5. ‘Basic story’, Improvement, Retranslation Hypothesis

From the examples given above readers will be able to conclude that a basic
story for the original text Nana and its five translations may be able to be con-
structed, but it would be fairly general, since at a more detailed level there is a
certain amount of divergence of renderings and of interpretations. In any case,
the possibility of constructing a basic story does not mean that it is an inherent
structure but, rather, that it would be constructed by readers for a particular pur-
pose, such as formulating a plot summary of the novel and its translations.
With regard to ‘improvement’ we could say that Berman would prefer the
later translations, which in general are source-oriented as compared with the
very first translation. However, for this set of translations there does not appear
to be a movement towards the production of a canonical translation, and it is
difficult to say that the translations improve in any way. With regard to the re-
translation hypothesis it is true that the first translation is adaptive to the target
system, and the later translations are more source-oriented. However, the first
and second translations belong to the same time period: together they could be
called the first translations. The stark difference between these two translations
is explained by the specificity of the two very different contexts of production.
This implies that rather than through reference to a general historical progres-
sion, the nature of translations and retranslations is best explained through par-
ticular contextual conditions.

Across Languages and Cultures 7 (2) (2006)



With regard to the case study data, certain theories regarding narrative versions
and retranslation do not have explanatory force: these are the idea of the transfer
of a basic structure underpinning a set of narrative versions; the notion of im-
provement over time towards the eventual production of a great (canonical)
translation; and the retranslation hypothesis according to which there is a natu-
ral progression from target-oriented to source-oriented translations.
The theories which best explain the data are the following. The retransla-
tions are narrative versions which are elicited and constrained by specific condi-
tions. It is those conditions which can explain the similarities and differences
between the different translations. The conditions comprise broad social forces:
changing ideologies and changing linguistic, literary, and translational norms;
as well as more specific situational conditions: the particular context of produc-
tion and the translator’s preferences, idiosyncrasies, and choices. Specific con-
texts of production may result in very different translations being produced dur-
ing one time period (as with Translations A and B). This is a source of hetero-
geneity within the same time period, as are individual translators’ styles and in-
consistencies. Individual translators (possibly influenced by social mores) may
interpret sections of the source text differently, and these different interpreta-
tions affect the translation. The issue of interpretation brings together the role of
the literary text as an open-ended text, and the role of the translator as inter-
preter. More generally a post-structuralist approach provides the possibility of
combining both textual and contextual forces. Retranslations can be seen not
only as the product of contextual forces discussed above, but also as the subset
of a textual system, the narrative matrix with its own manner of functioning. In
particular, narrative marks are limitless: they may be reiterated (a source text is
reiterated in a different context), and they call forth further narratives (including
(re)translations). Influenced by the new context, retranslations retain or cancel
aspects of the narrative mark(s) they derive from (Translation A, notably, can-
celled various sensual and sacrilegious references of the original text). Narrative
marks are related to all other narrative marks, such that they haunt each other.
Thus a retranslation may be permeated by aspects of other (re)translations, in-
cluding norms and expressions prevailing at another time period or in another
context, and this is another factor that creates heterogeneity. As well as express-
ing power relations in society (e.g. the influence of the powerful British middle
class), the narrative matrix and subset of retranslations have their own logic and
power of operation (e.g. the continuing production of retranslations).
Studies of different sets of retranslations and the use of different theoretical
perspectives may provide some different conclusions as to the forces influenc-
ing retranslations. However, other studies of data such as Paloposki and Koski-
nen (2004) concur with my finding that there are multiple sources of explana-

Across Languages and Cultures 7 (2) (2006)


tion for what is going on in retranslations. Paloposki and Koskinen propose the
following sources of explanation for variations between retranslations in their
corpus: stage of development of a literature; the relationship between target and
source culture; historical and ideological contexts; publishers’ requirements; the
nature and expectations of intended readers; material aspects such as illustra-
tions; and the translator’s profile, preferences and interpretations.
As well as producing reflections on retranslation, I hope to have made a
methodological point in this paper by showing how bringing together two sets
of theories from different but related disciplines can be productive in conceptu-
alizing and shedding light on a particular translational phenomenon.

I have not seen the expression ‘Retranslation Theory’ used in the literature. What I am referring
to is theoretical discussions and observations concerning the phenomenon of retranslation.
Of course the target language is never really the ‘same’ because a language evolves over time,
and the changing language is precisely one of the reasons a retranslation is undertaken.
The very first word-for-word rendering (Goethe and Berman) is not included in the discussion as
its status as a translation is debatable: Berman calls it an ‘introduction’.
For his part, Chesterman does not undertake an empirical study, but says with regard to the re-
translation hypothesis that “the jury is still out on this one” (Chesterman 2000:22).
Abstracts of articles from the database Linguistics and Language Behavior Abstracts were sur-
veyed. These abstracts had been brought up by the keywords ‘translation and reinterpreta-
Recall how above the Romantic Idealist notion of retranslation stages was taken into a Natural
Sciences-based paradigm to become a hypothesis: the theory was both the same and very
different when framed in a new context.
The translations are labelled A to E, as explained earlier. Page numbers are given in round
brackets. My gloss of the source text phrase or sentence is given in square brackets.
Note that Zola himself was an atheist.
This raises the issue of a translator referring to previous translations. There is evidence that the
translator of Translation D used Translation B: a number of particular renderings in Transla-
tion B are used verbatim in the later translation. This is not necessarily a source of hetero-
geneity, since the translator of D only adopts the B options if they suit his contemporary

Bensimon, P. 1990. Présentation. Palimpsestes. Retraduire. Vol. 4. IX–XIII.
Berman, A. 1990. La retraduction comme espace de la traduction. Palimpsestes. Retraduire. Vol.
4. 1–7.
Berman, A. 1995. Pour une critique des traductions: John Donne. Paris: Gallimard.
Brownlie, S. forthcoming. Literary Theory and the Translator: Gathering Together the Transla-
tor’s Multiple Roles. Translation Studies in the New Millennium.
Chatman, S. 1978. Story and Discourse: Narrative Structure in Fiction and Film. London: Cor-
nell University Press.

Across Languages and Cultures 7 (2) (2006)


Chatman, S. 1981. Critical Response: Reply to Barbara Herrnstein Smith. In: Mitchell. W. J. T.
(ed.) On Narrative. Chicago/London: University of Chicago Press.
Chesterman, A. 2000. A Causal Model for Translation Studies. In: Olohan M. (ed.) Intercultural
Faultlines. Manchester: St Jerome. 15–28.
Collombat, I. 2004. Le XXIe siècle: l’âge de la retraduction. Translation Studies in the New Mil-
lennium Vol. 2. 1–15.
Du-Nour, M. 1995. Retranslation of Children’s Books as Evidence of Changes of Norms. Target
Vol. 7. No. 2. 327–346.
Frank, A. P. 1990. Systems and Histories in the Study of Literary Translations: A Few Distinc-
tions. In: Bauer R. & Fokkema. D. (ed.) Proceedings of the XIIth Congress of the Interna-
tional Comparative Literature Association. Vol I. Munich: Iudicium.
Gibson, A. 1996. Towards a Postmodern Theory of Narrative. Edinburgh: Edinburgh University
Hermans, T. 1996. Norms and the Determination of Translation: A Theoretical Framework. In:
Alvarez R. & Vidal M. C.–A. (eds.) Translation, Power, Subversion. Cleve-
don/Philadelphia/ Adelaide: Multilingual Matters.
Johnson, B. 1980. The Critical Difference: BartheS/BalZac. In: The Critical Difference: Essays in
the Contemporary Rhetoric of Reading. London: Johns Hopkins University Press.
Kujamäki, P. 2001. Finnish Comet in German Skies: Translation, Retranslation, and Norms. Tar-
get Vol. 13. No. 1. 43–71.
Mcquillan, M. 2000. Introduction: Aporias of Writing: Narrative and Subjectivity. In: Mcquillan,
M. (ed.) The Narrative Reader. London/New York: Routledge.
Merkle, D. 2000. L’entrée de Nana ‘en expurgation vizetéllienne’ dans le système littéraire
britannique victorien. Excavatio 13. 25–33.
Merkle, D. 2003. The Lutetian Society. TTR Vol. 16. No. 2. 73–101.
Paloposki, O. & Koskinen, K. 2004. A Thousand and One Translations: Revisiting Retranslation.
In: Hansen, G., Malmkjaer, K. & Gile, D. (eds.) Claims, Changes and Challenges in Trans-
lation Studies. Amsterdam: Benjamins.
Perrin, N. 1969. Dr Bowdler’s Legacy: A History of Expurgated Books in England and America.
USA/London: Macmillan.
Rimmon-Kenan, S. 1983. Narrative Fiction: Contemporary Poetics. London/New York:
Smith, B. H. 1980. Afterthoughts on Narrative: Narrative Versions, Narrative Theories. Critical
Inquiry Vol 7. No. 1. 213–236.
Stam, R. 2000. Beyond Fidelity: The Dialogics of Adaptation. In: Naremore, J. (ed.) Film Adapta-
tion. London: The Athlone Press.
Susam-Sarajeva, S. 2003. Multiple-entry Visa to Travelling Theory: Retranslations of Literary
and Cultural Theories. Target Vol. 5. No. 1. 1–36.
Toury, G. 1995. Descriptive Translation Studies and Beyond. Amsterdam/Philadelphia: John
Vanderschelden, I. 2000. Why Retranslate the French Classics? The Impact of Retranslation on
Quality. In: Salama-Carr, M. (ed.) On Translating French Literature and Film II. Amster-
dam/Atlanta: Rodopi.
Venuti, L. 2004. Retranslation: The Creation of Value. In: M. Faull, K. (ed.) Translation and Cul-
ture. Lewisburg: Bucknell University Press.
Weeks, J. 1981. Sex, Politics and Society. London/New York: Longman.

Across Languages and Cultures 7 (2) (2006)


Zola, E. 1884. Nana: A Realistic Novel. Translated without Abridgment.
London: Vizetelly & Co. (A)
Zola, E. 1971. (first published 1895). Nana. Trans. Victor Plarr.
London: Book Club Associates. (B)
Zola, E. 1972. Nana. Trans. George Holden. London: Penguin Books. (D)
Zola, E. 1973. (first published 1956). Nana. Trans. Charles Duff.
London: Folio Press. (C)
Zola, E. 1998. (first published 1992). Nana. Trans. Douglas Parmée. Oxford: Oxford University
Press. (E)
Zola, E. 2002. (first published 1880). Nana. ed. Henri Mitterand. Paris: Gallimard.

Across Languages and Cultures 7 (2) (2006)