Gideon Toury. 1995. "The Notion of 'Assumed Translation' - An Invitation to a New Discussion". H. Bloemen, E. Hertog & W. Segers, eds. Letterlijkheid, Woordelijheid / Literality, Verbality). Antwerpen/Harmelen: Fantom, 1995. 135-147.

This text was scanned from the original version. Layout (pages and line breaks) preserved.
Unlawful distribution prohibited.

[Page 135]
 

GIDEON TOURY

The Notion of 'Assumed Translation'

An Invitation to a New Discussion1

For Raymond, on his sixtieth birthday:
and the wing of friendship never moults a feather!
Charles Dickens

Anyone wishing to approach translations descriptively is faced with the
delimitation of the object of study: what would be taken up and what would be
left out? The current state of translation studies makes this question impossible
to answer in any straightforward way. Not only is today's discipline a remarkably
heterogeneous series of paradigms which are, at best, loosely connected, but
there still reigns an overriding tendency to regard the different paradigms as
mere alternative ways of dealing with 'the same thing'. Which they are not, nor
should they be expected to be.

Far from being a neutral procedure, establishing an object of study is
perforce a function of the theory in whose terms it is constituted, which is
always geared to cater for a particular set of needs. Its establishment and
justification are therefore intimately connected with the questions one wishes
to pose, the possible methods of dealing with the objects of study with an eye
to those questions - and, indeed, the kind of answers which would count as
admissible. The question is not really what the object is, then, but rather what
would be taken to constitute a proper object, in pursuit of a certain goal, such
that any change of approach would entail a change of object. This is so even if
all objects superficially fall under the same heading; be it even 'translation' and
'translating' themselves. It is not the label that counts, but the concept it applies
to; and concepts can only be established within conceptual networks.
Unfortunately, the fallacious rejection of somebody else's concepts on the
grounds that they are untenable within one's own frame of reference, which
was designed to serve a completely different purpose, is still very much with us.


The mainspring of my own scholarly endeavour has long been the conviction
that the position and function of translations (as entities), and of translating
(as a kind of text-generating activity), in a prospective target culture, the form
a translated text would have (and hence the relationships which would tie it to
its original), and the strategies resorted to during its generation, should not be
approached as distinct facts. Rather, it is first and foremost the interdependencies
of them all which should be tackled, the intention being to uncover the

[Page 136]

regularities which mark the relationships assumed to obtain between function,
product and process; not only for individual cases, but also towards ever
higher generalizations.

In an attempt to pursue this particular goal, it has been suggested that
translations be regarded as facts of the culture which hosts them, with the concomitant
assumption that whatever their function and identity, these are determined
within that same culture and reflect its own constellation. It was by virtue of
this methodological starting point that this theoretical framework earned the
nickname of 'target-oriented'.

When it was first formulated, back in the seventies, this approach to the
study of translations and translating in their immediate contexts was considered
outrageous, and its initiator a sworn enfant terrible. At that time, translation
studies was indeed marked by extreme source-orientedness. Most of its
paradigms were application-ridden too: Whether concerned with teaching or
quality assessment, their preoccupation was mainly with the source text and
with the proclaimed protection of its 'legitimate rights'. Target constraints,
while never totally ignored, often counted as subsidiary; especially those which
would not fall within any kind of linguistics. Many factors which govern real-
life translational behaviour, and the fact that these factors have indeed resulted
in a variety of very different translation practices, were resented, or, at best,
relegated to the realm of 'mere' history.2

Meanwhile, most translation scholars, while not abandoning the seemingly
safe base of the source text, have at least come to integrate many more target-
bound considerations into their reasoning. In addition, a second paradigm
which was heavily target-oriented, so-called Skopostheorie, gradually emerged
as an alternative. It even managed to gain considerable ground, albeit mainly
in German-speaking circles. Thus, target-orientedness as such no longer
arouses the same antagonism it used to less than twenty years ago.

Interestingly, the first formulations of Skopostheorie (e.g. Vermeer 1978)
almost coincided with the beginnings of my own target-orientedness (Toury
1978 [11976]) - which sheds interesting light on how changes of scholarly
climate occur, especially considering that for quite a while, proponents of the
two approaches remained practically unaware of each other's work.3 To be
sure, even now, there is at least one major difference between the interests of
the two respective paradigms, which also accounts for the different assumptions
each of them proceeds from: whereas mainstream Skopos-theorists see the
ultimate justification of their frame of reference in the more 'realistic' way it
can deal with problems of an applied nature, the main object being to 'improve'
(i.e. change!) the world of our experience, my own endeavours have been
geared primarily towards the description and explanation of whatever has been
regarded as translational within particular target cultures, the ultimate object
being to formulate a series of interconnected laws of a probabilistic nature,
along with their conditioning factors.

Recent attempts to conduct historical studies within Skopostheorie (most
notably Vermeer 1992), on the one hand, and to apply some of the basic
assumptions of the other target-oriented paradigm to translation didactics
(e.g. Toury 1980b, 1984, and especially 1992), on the other, indicate that the
 

[Page 137]
 

gap may be narrowing. This tendency is also manifest in the recent work of
some second-generation Skopos-theorists, most notably Nord (1991), who has
made a bold attempt to integrate a version of the notion of 'translational
norms', so central to the evolution of my own way of reasoning, into an
account which is basically Vermeerian.4

This narrowing of gaps notwithstanding, I will be the first to admit that my
research program has not fared all that well. The main obstacle, as I see it, was
the prevailing tendency to read it through the (often rather dark) glasses of
other approaches to translation rather than as a self-contained paradigm. Even
the terms I used were often interpreted as if they were still part of (an)other
theoretical paradigm(s), in spite of their having been introduced in a completely
different conceptual framework and hence assigned a very different sense. The
inevitable result was gross misconception of many of the claims, including the
very target-orientedness of my approach. In view of the misunderstandings
which have cropped up, it seems advisable to put some of the basic assumptions
of my approach to the test again, thus opening the door to a fresh discussion.


Strange as it may sound to the uninitiated, there is nothing too perverse in
claiming that a text's position (and function), including the position and
function which go with a text being regarded as a translation, are determined
first and foremost by considerations originating in the culture which hosts
them. In fact, this is the most normal practice of the 'persons-in-the-culture'
themselves. Thus, when a text is offered as a translation, it is quite readily
accepted bona fide as one, no further questions asked. Among other things:
this is the reason why it has often been that easy for fictitious translations - i.e.
texts which are presented as translations with no corresponding source texts in
other languages ever having existed - to pass as genuine ones. By contrast,
when a text is offered as having been originally composed in a language,
reasons will often manifest themselves to suspect, correctly or not, that the
given text has in fact been translated into it; among other things, through
certain features of textual make-up and verbal formulation, which persons-in-
the-culture have come to associate with translations and translating. Adopting
culture-internal distinctions as a starting point for the descriptive study of
translation as it is conceived of, and executed, within the conditioning framework
of a culture, has the big advantage of not imposing on its object-level any
distinctions which may prove alien to that culture. It thus allows one to
proceed with as few assumptions as possible which could be difficult to
maintain, in the face of real-world evidence.

There is no way a translation could share the same systemic space with its
original, thus belonging to two cultures at the same time; not even when the
two are physically present side by side. This is not to say that, having been
severed from it, a translation would never be in a position to bear on the source
culture again, on occasion even on the source text itself. Texts, and hence the
cultural systems which host them, have been known to have been affected by
translations of theirs. It is nonetheless significant that any such practice

[Page 138]

involves a reversal of roles, in full accordance with our starting point: while
genentically a translation, the affecting entity ceases to function as one. When
observed from the point of view of the new target culture, that is. To be sure, it
is not just any translation that would exert an influence on the original, for the
simple reason that it is its translation. Rather, what is picked up is picked up
because it has become a fact of a particular (target!) culture, which is, more-
over, regarded as privileged for that precise reason.

The fact that translations often serve as a basis for further acts of translation
is no refutation of our target-oriented assumption either: while a translation
does function as a source text in such instances, it is still a fact of a former
target culture now turned into a mediating one; and it is picked up and
assigned the role of a source text not because of anything it may inherently
possess, but because it concurs with the concerns of a new prospective recipient
system.

On the other hand, translation activities and their products not only can,
but do cause changes in the target culture. By definition, that is. After all,
cultures resort to translating as a major way of filling in gaps, whenever and
wherever such gaps manifest themselves; either as such, or (very often) from a
comparative perspective, i.e. in view of a corresponding non-gap in another
culture that the prospective recipient culture has reasons to look up to and try
to exploit. Semiotically, then, translation is as good as initiated by the target
culture. In other words, the starting point is always one of a certain deficiency in
the latter, even if sometimes - e.g. in a 'colonial' situation - an alleged gap may
be factually pointed out for it by a patron of sorts who also purports to 'know
better' how that gap may best be filled. Even here, the more persuasive
rationale is not the existence of something in another culture/language as such,
but rather the observation that something is 'missing' in the target culture
which should have been there and which, luckily, already exists elsewhere and
can be taken advantage of.

In the simplest of cases, both deficiency and fill-in consist in mere textual
entities: a text which hasn't been there before is now introduced into the target
culture. Being an instance of performance, every individual text is of course
unique; it may be more or less in tune with prevailing norms and models, but
in itself it is a novelty. As such, its introduction into the receiving culture
always entails some change, however slight, of the latter. To be sure, the
novelty claim still holds for the nth translation of a text into a language: it is the
resulting entity, the one which would actually be incorporated into the target
culture, which is decisive here; and this entity will always have never existed
before. Even alternative translations of the same text are not likely to occupy
the exact same position in the culture which hosts them, not even if they all
came into being at the same point in time.

In more complex cases, models may be imported into the recipient culture
as well (that is, sets of rules for the generation of texts pertaining to a
recognizable type). Such a migration normally involves groups of texts which
embody a recurring pattern or else are translated in a similar fashion, even
though high-prestige translations of individual texts are known to have had a
similar effect.

[Page 139]

The likelihood of causing changes in the target system beyond the mere
introduction of the new text itself stems from the fact that, while translations
are indeed intended to cater for the needs of a target culture, they also tend to
deviate from its sanctioned patterns, not least because of the postulate of
retaining invariant at least some features of the source text, which seems to be
inherent in any culture-internal notion of translation. This tendency often
renders translations quite distinct from non-translational texts, and not
necessarily as a mere production mishap either; it is not unusual for a certain
amount of deviance to be regarded not only as justifiable, or even acceptable, but
as actually preferable to complete normality, on all levels at once. Moreover,
even if they are not culturally favoured, deviations - even when they manifest
themselves in the very make-up of the texts - do not necessarily disturb the
persons-in-the-culture. Thus, more than one writer has observed that the
(tentative) identification of a text as a translation "'protects' the reader, as it
were, from misinterpreting the writer's intentions... [It] implies that deviations
from cultural norms are not judged as intentional, and therefore are not
assigned any 'hidden' meaning" (Weizman and Blum-Kulka 1987: 72).

In fact, as has been suggested time and again, there are often good reasons
to regard translations as constituting a special system, or 'genre' of their own
within a culture. What is totally unthinkable is that a translation may hover in
between cultures, so to speak: As long as a (hypothetical) interculture has not
crystallized into an autonomous (target!) systemic entity, e.g. in processes
analogous to pidginization and creolization, it is necessarily part of an existing
(target!) system.


Bringing all this to bear on our basic assumption, the latter could be reformulated
to read as follows:

translations should be regarded as facts of target cultures; on occasion facts of a
special status, sometimes even constituting identifiable (sub)systems of their own,
but of the target culture in any event.
This formulation implies that, while certainly indispensable, establishing the
culture-internal status of a text as a translation does not in itself provide a
sufficient basis for studying it as one. Any attempt to offer exhaustive descriptions
and viable explanations would necessitate a proper contextualization, which is
far from given. Rather, its establishment forms part of the study itself which is
applied to a text assumed to be a translation. In an almost tautological way it
could be said that, in the final analysis, a translation is a fact of whatever target
sector it is found to be a fact of, i.e. that (sub)system which is found to be best
equipped to account for it: function, product and underlying process.
Consequently, the initial positioning of an assumed translation, which is a sine
qua non for launching a meaningful analysis (see e.g. Lambert and Van Gorp
1985), may be no more than tentative, it may often have to undergo revision as
the study proceeds, namely, on the basis of its interim findings.

[Page 140]

Above all, it shouldn't be assumed that the identity of the (sub-)culture
which hosts an assumed translation is known just because one sees what
language it is formulated in. Seemingly an easy way out, the assumption of a
one-to-one relationship between culture and language often proves misleading,
the more so as the exact identity of the target language itself may have to be
reconsidered in the course of the study. To take an extreme example:
 

The following text, one of the versions of the note which warns passengers on
German trains against improper use of the emergency brake, seems to have
been intended as an English utterance:

Emergency brake
Full brake only in
case of emergency
Any misuse will
be punished

In spite of the fact that it is only presented as parallel to three other versions of
the same note, in German, French and Italian, there are sufficient indications
for tentatively regarding it as a (rather literal) translation, and of one particular
version, at that: the German one. At the same time, this note does not pertain
to any of the institutionalized cultures which have English as a 'national'
language; and not just because each one of these cultures already has a codified
version of the warning in its repertoire, all different from the present one.
Ignoring the ridiculous possibility that it is not attributable to any culture,
there is no escape from regarding the English version of the note as situated in
the German culture, albeit in a very specific section of it which does take the
intended language into account.

Thus, the system which may be said to host this translation - and to have
governed its generation - is the artificial sub-culture shared by the speakers of
various languages who also have English, for as long as they are in Germany
(or at least on board a German train). This is the only contextualization which
would ensure a satisfactory explanation of the linguistic make-up of the text
and the (reconstructed) practice resorted to by its translator (which are dealt
with in detail in Toury in press: Chapter 4).
 

Proper contextualization also involves a heightened differentiation be-
tween translational items pertaining to one and the same culture; namely, in
terms of their respective positions within it. As already indicated, not even two
translations of a single text are likely to occupy exactly the same position, least
of all as a mere reflection of the position of the original in the source culture. If
differences of textual-linguistic make-up, or of relationship to the shared
source text, are to receive a viable explanation, the position appropriate to each
translation will have to be established - and taken into account in all seriousness.5

The position most relevant to the kind of questions we wish to pursue is of
course the one a translation was designed to occupy when it first came into being.
After all, this is the only position which may be claimed to have actually

[Page 141]

governed its generation and the decisions made in its course. However, it is
only towards the end of a study that an intended position can be established
with reasonable certainty. This would be achieved by weighting the position
the text initially had against the findings concerning its make-up and formulation,
and the way it represents its original, while taking into account what is already
known about the translational tradition in which it came into being.
Consequently, the intended position - which cannot be pointed to in any real
sense - would always have the status of an explanatory hypothesis rather than a
'fact'; not even a reconstructed one.

Also significant is the possibility that translations which have retained their
identity as cultural facts, even as translations, may nevertheless have changed
their exact position within the target culture over time. Of course, changes of
this kind can have no bearing on either the intended position of a translation,
or even the one it initially had. On the other hand, they may often shed
considerable light on preferences of later periods of time, at least some of them
pertinent to translation as performed in those periods.



In most paradigms, some kind of a definition of translation would have been
expected by now - a list of (more or less) fixed features which, if accepted as a
starting point and framework for research, would entail a purely deductive
mode of reasoning. However, the obsession with restrictive definitions proves
counter-productive precisely when the aspiration is to account for real-life
phenomena in the circumstances which gave rise to them; they tend to hinder
rather than advance descriptive-explanatory work.

Thus, any a priori definition, especially if couched in essenrialistic terms,
allegedly specifying what is 'inherently' translarional, would involve an untenable
pretense of fixing once and for all the boundaries of an object which -
culturally speaking - is characterized by its very variability: difference across
cultures, variation within a culture and change over time. Not only would the
field of study be considerably shrunk that way, in relation to what cultures
have been, and are willing to accept as translarional, but research limited to
these boundaries may also breed circular reasoning: to the extent that the
definition is indeed adhered to, whatever is studied - selected for study
because it is known to fall within it, in the first place - is bound to reaffirm the
definition. Unless one is willing to transcend the arbitrarily set boundaries,
that is; which is what any research practice applied to existing translations
seems to have involved, even if performed within an essentialistic, and hence
classificatory frame of reference.

Thus, the unpleasant truth of the matter is that even those who have sworn
by the need to proceed deductively, on a well-formulated initial definition,
never hesitated to pick texts (or other phenomena) for study on the grounds
that they had been presented, or otherwise regarded as translational. Within a
particular culture, that is. It appears that they too were adopting the pre-
systematic attitude of the persons-in-the-culture, but as no more than a
necessary evil: they were never willing to follow it, or the procedures it may

[Page 142]

have suggested, in any consequential way. What they did instead was to
tamper with their data; e.g. by imposing on it distinctions between 'fuller' and
less full' realizations of the framing definition, which was thus elevated to a
kind of maximum, or else by introducing additional (a priori, and hence non-
cultural and ahistorical) distinctions, e.g. between 'translation' and 'adaptation'.
Of course, the number of these distinctions could be multiplied almost
indefinitely, and, at any rate, they do not offer very much by way of explanatory
power, when it comes to culturally contextualized phenomena. Not for a
moment did the possibility cross their minds of giving this forced target-
orientedness of theirs a systematic status, which would have entailed an
inductive attempt to derive general principles from the facts themselves rather
than speculation within a more or less rigid frame of reference.6

To be sure, the principles we formulated ourselves have never been put
forward as an alternative definition of the 'Gegenstand der Übersetzungs-
wissenschaft' as such, as wrongly posited by some critics, most notably Koller
(e.g. 1990). Rather, what they have always constituted was a working hypothesis,
designed to provide guidelines for the establishment of corpuses for studies of
one particular kind, sharing one particular set of goals. Within our frame of
reference, the assumption is applied to all utterances which are presented or
regarded as translations within the target culture, on no matter what grounds,
which I have now referred to as assumed translations. Under such observation,
there is no pretense that the nature of translation is given, or fixed in any way.
What is addressed, even in the longest run, is not even what translation can be,
in principle, but what it proves to be in reality, and hence what it may be
expected to be under various specifiable conditions.

There may of course be any number of reasons for regarding a target-
language utterance as a translation. On the other hand, there is also the
possibility of encountering phenomena, which could have plausibly been
regarded as translations but which were not - whether they were regarded as
something else or whether the distinction between translations and non-
translations was simply non-functional, and hence a non-fact, in the culture in
question. Items of this kind can of course be studied too, but an account will
have to be given precisely of the fact that they were not presented/regarded as
translational within the culture which hosts them; and as part of the study
itself, not just as some kind of 'background information'.

The whole point here is precisely to tackle questions such as why something
was, or was not, presented/regarded as translational, and not why it should
have been (much less why it should not have been) presented in that way.
Whatever the justification for this hypothesis, and for the heuristics deriving
from it, it is thus intimately connected with particular interests and no others.
Adopting our assumption as a working hypothesis thus involves two important
benefits: a considerable extension of the range of objects of study, in full
agreement with those real-life situations that we set out to account for, and
functional operarivity even in cases where the basic principle might have seemed
factually inapplicable.

[Page 143]


Proceeding from culture-internal notions often involves local, pre-systematic
nomenclature as well: It stands to reason that many of the distinctions recognized
as functional within a culture would also find expression in language, labelling
being one important indication of cultural institutionalizarion. This possibility
notwithstanding, our principles have not been put forward with respect to the
English word 'translation', as strangely posited by Gutt (1991: 7), so that there
is hardly room for his doubt an to their applicability to German Übersetzung,
Amharic tïrgum, or any other 'ethnic' label. It is the notion of (assumed)
translation that is at stake here; and no matter what name it goes under, this
notion can be accounted for as a cluster of (at least three interconnected)
postulates:

(1) A Source-Text Postulate;
(2) A Transfer Postulate;
(3) A Relationship Postulate.

While all three may look familiar, their status within a target-oriented frame of
reference such as ours is very different from the one they may have had in any
other paradigm of translation studies: regarded as postulates, they are all
posited rather than factual; at least not of necessity. Therefore, rather than
constituting ready-made answers, they give rise to a series of questions to be
addressed by anyone wishing to study translation in context.
Let us look briefly into these lower-level assumptions in an attempt to
clarify their posited status and the way they combine to form the overall
culture-internal notion of assumed translation.
 

(1) The Source-Text Postulate
Regarding a target-language text as a translation entails the obvious assumption
that there is another text, in another culture/language, which has both
chronological and logical priority over it: not only has such an assumed text
assumedly preceded the one taken to be its translation in time, but it is also
presumed to have served as a departure point and basis for the latter.

The crucial thing is that it is not the source text as such, nor even the
possibility of actually pointing to it, that is at stake here, but the assumption
that one must have existed. Therefore concrete texts in languages other than
the target's are not part of the necessary equipment for launching research
either: even if none is used, the study can still pertain to translation studies,
namely, as long as the assumptions of their temporal preexistence and logical
priority are taken to bear on the issues.

To be sure, a target fact which was tentatively marked as a translation, with
the Source-Text Postulate implied, may then turn out to lack a corresponding
text in any other language/culture, and not only when one has simply failed to
locate it. A concrete source text may never have existed, to begin with. This is
why fictitious translations form legitimate objects of study within our paradigm:
until the mystification has been dispelled, the way they function within a

[Page 144]

culture is no different from the way 'genuine' translations do.

On the other hand, an assumed translation may later on be found to have
had more than one source text, being a case of compilative translation, or a single
source text which, however, differs from the one it was initially assumed to have (as
in the case of indirect translation).
 

(2) The Transfer Postulate
The Source-Text Postulate also entails the assumption that the process whereby
the assumed translation came into being involved the transference from the
assumed source text of certain features that the two now share. This assumption
is a clear result of bringing two different kinds of knowledge to bear on one
another: knowledge about products, on the one hand, and about (cross-
linguistic and cross-cultural) processes, on the other.

When regarded from a target-oriented point of view, transfer operations -
their very existence as well as their exact nature (and that of the transferred
features) - first manifest themselves as posited too. Both aspects can be
submitted to examination, but they would remain distinct in their very es-
sence: even if recourse to transfer operations would have been confirmed,
these will not necessarily be found to match the posited one. Needless to say,
either one can be submitted to real examination only after the appropriate
source text has been secured, which may well involve an element of confrontation
with the assumed translation in question. (See Toury in press: Chapter 3.)
 

(3) The Relationship Postulate
Finally, adopting the assumption that a text is a translation also implies that
there are accountable relationships which tie it to its assumed original, an
obvious function of that which the two texts allegedly share and which is taken
to have been transferred across the cultural-semiotic (and linguistic) border.

A target-internal Relationship Postulate can be highly intricate, down to
very specific hypotheses as to the level(s) where intertextual relationships may
be expected to occur. Some of these relationships may even be postulated as
necessary and/or sufficient, within the normative structure of the culture in
question, as are many of the notions of 'equivalence' in source-oriented
approaches to translation. However, all this needn't be reflected in reality
either: upon examination, relationships actually tying together pairs of texts, or
parts thereof, may well be found to differ from the postulated ones. Since there
is no inherent need for intertextual relationships to always be of the same kind
or intensity, the nature and extent of these relationships, as well as their
correspondence to the culture's attitudes, constitute just another set of questions,
to be settled through concrete research work.

Another indication of the posited nature of translation relationships, under
the present observation, is the fact that they can often be (tentatively) accounted
for even in the absence of a source text, namely, on the basis of certain features
of the assumed translation itself, and on the concomitant assumption that it
was indeed translated. Again, fictitious translations often manage to pass for
genuine ones without arousing too much suspicion by making manipulative
use of this fact: not only do their producers present them as translations, but

[Page 145]

they actually disguise them as ones, e.g. by embedding in the texts features
which have come to be associated with genuine translations in the culture in
question. On the other hand, this possibility often serves as a sound basis for
establishing the immediate source(s) of an assumed translation, those which
would then be taken as a basis for comparative study, e.g. when it is suspected
of being compilative, or indirect.


If we now proceed to take the three postulates together, an assumed translation
would be regarded as

any target-culture text for which there are reasons to tentatively posit the existence
of another text, in another culture and language, from which it was presumedly
derived by transfer operations and to which it is now tied by certain relationships,
some of which may he regarded - within that culture - as neceisary and/or
sufficient.
As should have become clear by now, neither source text nor transfer operations
and transferred features, nor even translation relationships, would have been
excluded within such an approach. They are just given a different status. This
is also to say that 'orientedness' is far from tantamount to 'exclusiveness', as
wrongly interpreted by many:7 the present paradigm can be characterized as
target-oriented because this is where its observations start. By no means should it
be taken to mean that this is where these observations would also be exhausted.8

Looking at it from another angle, it is only reasonable to posit that a study
in translation activities which have already yielded their products would start
with the observables, first and foremost, the translated utterances themselves,
along with their constituents, as situated within their immediate contexts.
From there on, the study could proceed to facts which are observational 'in the
second order' (i.e. facts which need (re)construction before they can be
submitted to scrutiny), most notably the relationships which tie together the
output and input of individual acts, the ultimate intention being to end up
reconstructing the non-observables at their root, particularly the exact processes
whereby they came into being.
 

Notes

1.   A slightly different version of this article forms part of my forthcoming book, Descriptive
Translation Studies - and Beyond (Toury in press).

2.   And see Delabastita's recent argument (1991) that the opposition between theoretical
and historical approaches to translation is a false one through and through.

3.   An interesting attempt to associate target-oriented thinking on translation, especially of
my brand, and some of the basic ideas of Deconstruction, was made a few yean ago by
Van den Broeck (1988). While I wouldn't endorse all his claims, it is certainly an
intriguing article for anybody interested in the way scholarly paradigms change.

[Page 146]

4.   Unfortunately, while doing so, Nord (re)introduced the concept of 'loyalty', and as an
a priori moral principle at that, which may well be opening a new gap between the two
approaches as the old one seems to have been closing.

5.  To my mind, failure to do precisely this is a major flaw in much of the otherwise
impressive work emanating from the Göttingen group researching literary translation,
especially as they purport to arrive at a cultural history of translation into German, and
hence claim to carry out their research within a strict cultural-historical framework. All
too often, all translations are treated on a par, and there is no way of knowing whether
a particular case is representative, or significant in any other way. In fact, in spite of
their proclaimed intention, the textually interesting has often been given precedence over
the historically significant.

6.    Menachem Dagut was probably the first to have put his finger on the basic differences
between various scholarly treatments of translation in terms of whether their orientation
was deductive or inductive. In fact, he made this distinction a main line of argumentation
in a lengthy review of my 1980 book (Dagut 1981). He himself was all in favour of
deductive work, but even he could not but involve bits and pieces of inductive
reasoning in the descriptive part of his work (Dagut 1978).

7.   Thus, the line of reasoning which I have been following is a far cry from a mere
variation of good old 'literary reception', the only difference being that we wish to "deal
with translations rather than original works", as claimed, e.g. by Mary Snell-Hornby
(1988: 24, 25). Even when it is literary texts which are involved (which is not
necessarily the case anyway), the locus of study is not the texts themselves, much less so
their actual reception in the target culture, but rather what the texts reinstated in the
positions they had initially occupied can reveal as concerns the constraints under which
they came into being,

8.   In fact, I cannot see why such an approach would concern itself with transfer any less
than, say, the position the Göttingen group claims to have adopted (e.g. Frank 1990:
Section II), to a great extent as an (over)reaction to my own program. It is my firm
conviction that, even though transfer can be addressed in other frameworks too, it
remains at best only partly explainable unless all target constraints are taken into
consideration, which can only be done within a target-oriented frame of reference.
 

Bibliography

BROECK, Raymond van den. 1988. Translation Theory after Deconstruction'. Linguistica
Antverpiensia XXII: 266-288.

DAGUT, Menachem. 1978. Hebrew-English Translation: A Linguistic Analysis of Some
Semantic Problems. Haifa: University of Haifa.

DAGUT, Menachem. 1981. 'Review Article of Toury 1980a'. Hebrew Computational
Linguistics 18:69-75. [Hebrew]

DELABASTITA, Dirk. 1991. 'A False Opposition in Translation Studies: Theoretical
versus/and Historical Approaches'. Target 3/2: 137-152.

FRANK, Armin Paul. 1990. 'Forty Years of Studying the American-German Translational
Transfer: A Retrospect and Some Perspectives'. Amerikastudien/American Studies 35/1:
7-20.

[Page 147]

GUTT, Ernst-August. 1991. Translation and Relevance: Cognition and Context. Oxford:
Blackwell.

KOLLER, Werner. 1990. 'Zum Gegenstand der Übersetzungswissenschaft', R. Arntz &
G. Thome, eds. Übersetzungswissenschaft: Ergebnisse und Perspektiven. Festschrift für
Wolfram Wilss zum 65. Geburtstag. Tübingen: Narr. 19-30.

LAMBERT, Jose & Hendrik van Gorp. 1985. 'On Describing Translations'. T. Hermans,
ed. The Manipulation of Literature. Studies in Literary Translation. London/Sydney:
Croom Helm. 42-53.

NORD, Christiane. 1991. 'Scopos, Loyalty, and Translarional Conventions'. Target 3/1:
91-109.

SNELL-HORNBY, Mary. 1988. Translation Studies: An Integrated Approach. Amsterdam-
Philadelphia: Benjamins.

TOURY, Gideon. 1978 [11976]. 'The Nature and Role of Norms in Literary Translation'.
J.S. Holmes, J. Lambert & R. van den Broeck, eds. Literature and Translation. New
Perspectives in Literary Studies. Leuven: Acco. 83-100.

TOURY, Gideon, 1980a. In Search of a Theory of Translation. Tel Aviv: Porter Institute.

TOURY, Gideon, 1980b. "The Translator as a Nonconformist-to-be, or: How to Train
Translators so as to Violate Translational Norms'. S.-O. Poulsen & W. Wilss, eds.
Angewandte Übersetzungswissenschaft: Internationales Übersetzungswissenschaftliches
Kolloquium an der Wirtschaftsuniversität Århus/Dänemark, 19-21 Juni 1980. Århus. 180-
194.

TOURY, Gideon. 1984. 'The Notion of 'Native Translator' and Translation Teaching'.
W. Wilss & G. Thome, eds. Die Theorie des Übersetzens und ihr Aufschlußwert für die
Übersetzungs- und Dolmetschdidaktik. Tübingen: Narr. 186-195.

TOURY, Gideon. 1992. ''Everything Has Its Price': An Alternative to Normative
Conditioning in Translator Training'. Interface 6/2: 60-72.

TOURY, Gideon, in press. Descriptive Translation Studies - and Beyond. Amsterdam/
Philadelphia: John Benjamins.

VERMEER, Hans J. 1978. 'Ein Rahmen für eine allgemeine Translationstheorie'. Lebende
Sprachen 23:99-102.

VERMEER, Hans J. 1992. Skizzen zu einer Geschichte der Translation 1-2. Frankfurt: IKO -
Verlag für Interkulturelle Kommunikation.

WEIZMAN, Elda & Shoshana Blum-Kulka. 1987. 'Identifying and Interpreting Translated
Texts: On the Role of Pragmatic Adjustment'. G. Toury, ed. Translation Across
Cultures. New Delhi: Bahri Publications. 61-73.