Skopos theory: a retrospective assessment


Andrew Chesterman


[2010a      In W. Kallmeyer et al. (eds), Perspektiven auf Kommunikation. Festschrift für Liisa Tittula zum 60. Geburtstag. Berlin: SAXA Verlag, 209-225.]


1. Introduction



It is often said, especially by laymen, that translation does not really have a theory. Not true: it has lots! (Well, it depends what you want to call a theory; but still...)  But at least it does not have a general theory, right? Translation Studies has produced at best only a mixture of fragmentary theories. – This claim is not quite true either: we have several candidates which present themselves as general theories of translation. One them is skopos theory.

It is now about a quarter of a century since the publication of Reiß and Vermeer’s classic work, Grundlegung einer allgemeinen Translationstheorie (1984), and even longer since the earliest publications on a functional approach to translation. Skopos theory, as a particular type of general functional theory, seems fairly well established on the map of translation studies, and is duly mentioned in all the textbooks. But how well has it stood the test of time? My aim here is to offer a general retrospective assessment of the theory, also taking account of some more recent criticism.


2. Axiomatic assumptions


Any theory rests on basic assumptions that are not tested within a given research paradigm, but are taken as given, self-evident, based on common sense and logic. We must start from somewhere, after all. But of course we can always query these assumptions if we wish, standing outside the paradigm. Some of them may be only implicit, hidden. But good theories aim to make all the relevant assumptions as explicit as possible, for instance as axioms from which the rest of the theoretical claims can be deduced. Skopos theory is unusual among other theories of translation, in that it has this form of a deductive, “syntactic” theory based on a small number of explicit axioms. In the 1984 version, these are called “rules” (Regeln). I give them here in summarized form (in the original German, from Reiß and Vermeer 1984: 119), followed by some brief initial explications and comments.


1. Ein Translat ist skoposbedingt.

2. Ein Translat ist ein Informationsangebot in einer Zielkultur und –sprache über ein Informationsangebot in einer Ausgangskultur und –sprache.

3. Ein Translat bildet ein Informationsangebot nicht-umkehrbar eindeutig ab.

4. Ein Translat muß in sich kohärent sein.

5. Ein Translate muß mit dem Ausgangstext kohärent sein.

6. Die angeführten Regeln sind untereinander in der angegebenen Reihenfolge hierarchisch geordnet (“verkettet”).


              Ad 1: Skopos theory thus assumes that a translation always has a skopos (a purpose), even though this may not always be clear (ibid.: 21). This skopos may often differ from that of the source text (surely a useful point). The skopos is the highest determining factor influencing the translator’s decisions. Elsewhere (ibid.: 96), the rule is phrased: “Die Dominante aller Translation is deren Zweck.” The theory assumes that the skopos is oriented towards to the intended target recipients: all translations have such a readership; even if you cannot always specify them, there are always “there” (ibid.: 85). – I will return below to problems of definition.

              Ad 2: The theory assumes that language is embedded in culture. Translation is seen as a subtype of more general cultural transfer (Reiß and Vermeer 1984: 13). The “information offer” concept relates to the underlying theory of communication, whereby a sender “offers” information to a receiver. This information is assumed by the sender to be “interesting” to the receiver (ibid.: 76, 103), and, if the communicative act is successful, it will be interpreted by the receiver in a way that is compatible with the sender’s intention and does not give rise to a “protest” (ibid.: 67, 106).

              Ad 3: Translations are not normally reversible; and a given source text has many possible translations.

              Ad 4: Intratextual coherence is assumed to exist to the extent that the text makes sense to the receiver, that it is compatible with the receiver’s cognitive context, as in any form of communication. Note that rules 4 and 5 have a clear prescriptive form, unlike the others.

              Ad 5: This fidelity rule assumes that the translation represents the source text, in some way which is relevant to the skopos. The theory recognizes a range of equivalence types.

              Ad 6: This rule is of a different status from the others, and, as part of a general theory, problematic. We might at least want to query the order of rules 4 and 5 as being universally valid.

              Immediately after giving this summary, the authors claim that these rules are “probably” the only general rules of translation (ibid.: 120). All further development of the theory would then be filling in more detail, providing rules for the analysis of the target situation, establishing conditions for the selection of different translation strategies, and so on (ibid.: 85).

              A last initial comment: at the very beginning of the book, the authors define “theory”, quite reasonably, thus: “Unter ‘Theorie’ versteht man die Interpretation und Verknüpfung von ‘Beobachtungsdaten’” (ibid.: vii). This definition nevertheless seems to be rather at odds with the way they actually present their theory. The argument of the book does not start with empirical observations or inductive generalizations, but proceeds deductively. Examples are given to support claims, but many of them seem to be invented.

              In a later publication, Vermeer (1996: 12f) contextualizes skopos theory explicitly as a form of action theory. Here too he sets out a number of axioms (now called, in English, “theses”), as follows, ending at about the point where the previous list (above) began:


1. All acting presupposes a “point of departure”, i.e. an actor’s position in space and time, convictions, theories, etc., including their respective history.

2. All acting is goal-oriented.

3. From a variety of possibilities [...]  that action will be chosen which one believes one has the best reasons for choosing under the prevailing circumstances. The reason(s) may not be conscious for the actor.

4. Given the prevailing circumstances, an actor tries to reach the intended goal by what seem to him the/an optimal way, i.e., for which he believes he has the best overall reason(s).

5. Translating is acting, i.e. a goal-oriented procedure carried out in such a way as the translator deems optimal under the prevailing circumstances.

6. Thesis 5 is a general thesis valid for all types of translating [including interpreting].

7. In translating, all potentially pertinent factors (including the source text on all its levels) are taken into consideration as far as the skopos of translating allows and/or demands. [Emphasis original]

8. The skopos of (translational) acting determines the strategy for reaching the intended goal.


One might wonder about the apparent underlying assumption here that human behaviour is necessarily always rational – if these axioms are supposed to be descriptive (on which more below). Another underlying assumption, to which we shall return, is the assumption of optimality: that the translator (always) acts in an optimal way.


3. Conceptual contribution


Quite apart from any other merits, a theory may contribute new concepts to a field. These may aid theoretical thinking in general, as well as description and explanation, and may be taken up and adapted by other theories. New theoretical concepts are interpretive hypotheses, to be tested pragmatically in use (see further e.g. Chesterman 2008). Two aspects of this potential conceptual contribution will be mentioned here, beginning with the central concepts themselves.


3.1. Key terms and conceptual distinctions


Some of the earliest criticism of skopos theory had to do with some of its definitions, or the lack of them (see Koller 1990, on functional theories more generally; Kelletat 1986; Hebenstreit 2007). We can also ask whether the relation between the set of terms and the set of necessary concepts is an appropriate one. Are there too many terms, or too few?

             Skopos is said to be a synonym of Zweck (purpose) or Funktion  (Reiss and Vermeer 1984: 96), but “function” itself is not explicitly defined in the same context. Perhaps it could be glossed as “intended effect”. But: effect on whom? Intended receivers, or any and all receivers? And intended by whom? Is it only the client’s intention that counts? What about the source author’s? The publisher’s? When does an effect begin, and end? What about heterogeneous effects? How do we actually measure effects? Furthermore, if skopos equals function, we may wonder why a new term is needed. Confusingly, the German term Funktion is used in several senses, including the mathematical one. Two of these senses do indicate an interesting distinction: “external function” is said to denote the translator’s general objective of making a living, whereas “internal function” refers to the skopos of a given translation (or translation process) (ibid.: 4). This external function seems very close to the term telos proposed in Chesterman and Baker (2008), to describe a translator’s ideological motivation for working as a translator, either generally as a career or on some specific, perhaps chosen, assignment.

              Later, Vermeer (1996: 7-8) seeks to distinguish three related concepts as follows: the intention is what the client wants to do; the skopos is what the translation is for; and the function is the “text purpose as inferred, ascribed by recipient”. But there remain problems here. Are these distinctions necessary? When might an intention clash with a skopos? Function, in particular, remains an unclear concept. Recipients are not a homogeneous set, and may well ascribe very different functions. Even a model reader may react differently on different occasions. And besides, actual reception should surely be distinguished from intended function. Both intentions and functions may be virtually impossible to access, particularly if the translations studied are distant in time or space. – The conceptual and terminological confusion here has not been resolved (see e.g. Nord 1997: 27f; Sunwoo 2007).

              Another problematic term is that of coherence, used both to refer to the similarity relation of equivalence between source and target, and to the intratextual interpretability of the translation itself. These seem very different concepts, and one wonders why the theory uses the same term. Since we already have “equivalence”, and this term is used in skopos theory too, why do we need a new term? We also already have “similarity”, if something looser than “equivalence” is wanted.

              A translation is defined in the second axiom as an offer of information about a source text (which is itself another such offer, about something else). This interpretation of the relation between source and target is much weaker than any notion of equivalence, weaker even than relevant similarity (although Reiss and Vermeer do refer occasionally to the offer as being a “simulating” one, e.g. p. 80, 105). It does not appear to constrain the “offer” in any way, except insofar as the offer is assumed to be “interesting” to the receivers and is “coherent” with the source text. Here again we can ask: does this term really earn its place?

Regarding the German term Translation itself, we can appreciate the way in which skopos theory (following a German tradition in Translation Studies) uses this to cover both written and oral translation: this is a neat solution we have not managed to imitate in English, and which has subsequently been widely accepted. There will, however, always be argument about the appropriate extension of the term. Kelletat (1986) and Koller (1990) think the skopos notion of translation is too broad because of the way it downgrades the importance of the source text and thus allows very free translations, adaptations etc., within the concept. Kelletat (1986: 15) even suggests the Reiss/Vermeer definition would include the whole of Latin literature! In my view, on the contrary, it is too narrow, if it is taken to exclude non-optimal translations.

              The theory’s use of the term “adequacy” (Adäquatheit) also merits a comment. The term was already familiar from other approaches, particularly Toury’s (e.g. 1980). But skopos theory defines it differently, not as a retrospective relation of closeness between target and source but as a prospective one between the translation, the source text and the skopos (Reiss and Vermeer 1984: 139). This skopos-sense of adequacy is so easily confused with the Toury-sense that scholars now either have to specify which sense is intended or give up using the term altogether. It is risky to give a new sense to an already established term.

              Skopos theory, like other functional approaches, has also contributed to a more differentiated conceptualization of the agents involved in the translation process. Instead of simply having a sender and a receiver, we have learned to distinguish between writer, client, translator, publisher, recipient, addressee and so on. In this sense, skopos theory has helped to shift the discipline towards a more sociological approach.


3.2.  Underlying metaphorical structure


A good theory’s concepts do not exist in isolation, but in a network of relations. This network may be more or less consistent in terms of its metaphorical structure. Martín de León (2008) has recently drawn attention to some interesting problems in the underlying metaphorical conceptualization of skopos theory. She argues that the theory combines two different metaphors: TRANSFER and TARGET. This suggests a lack of conceptual consistency, insofar as the metaphors are incompatible.

The TRANSFER metaphor describes the movement of an object from A to B, and assumes that the object (or some essence of it) does not change en route. This means assuming some kind of equivalence, of course. As an underlying metaphor for translation (visible in the etymology of this word), it normally needs to reify some notion of meaning (referred to as the message in Holz-Mänttäri 1984). The client’s intention might also be regarded as an “object” that is to be preserved. However, the view of a translation as merely an “offer of information” about the source text appears to go against the TRANSFER metaphor.

The theory’s notion of intertextual coherence also relates to this metaphor, albeit loosely. But how valid is this assumption that meaning is “there” in the text? Several contemporary models of cognition would argue that meaning always emerges via a process of interpretation, a process which depends on multiple variables and is not completely predictable (see e.g. Risku 2002). – In my view, both these positions are overstated. Surely some meanings are more obviously, objectively “there” in a text, while others are much less so and are open to interpretive variation. If no meanings were objectivizable at all, there would be no work for terminologists and no-one would dare to step into a plane.

The TARGET metaphor on the other hand describes a process from a source along a path to a goal. It does not assume an unchanged, reified message. It implies that the translator can participate in constructing the meaning of the message and thus highlights notions of intentionality and rationality. Skopos theory stresses the expertise and responsibility of the translator to select what needs to be translated and to translate it in the most appropriate manner. But this metaphor also prompts questions. Suppose a given process or action does not have a single goal but multiple ones, perhaps regarding heterogeneous receivers? And where actually is a goal located? Strictly speaking, the goal is not in the text but in the mind of the initiating agent, for whom the translation is merely a means to achieve a goal or goals. Further: where in the theory is there any space for an assessment of the goals themselves? Is it really enough to say that any end justifies the means? – We will take up the ethical dimension of this argument below.



4. Ontological  status of the theory


Perhaps the most debated problem of skopos theory has been its unclear ontological status. Does it aim to be a descriptive theory (of what is) or a prescriptive one (of what should be)? Does it describe a real world or an ideal, optimal one?

              This ambiguous status is already apparent in its axioms: axioms four and five are openly prescriptive, but the others are not. Reiss and Vermeeer say that there is no such thing as “the best” translation for a given source text. “Es gibt nur das Streben nach Optimierung unter den jeweils gegebenen aktuellen Bedingungen” (1984: 113). – This is an interesting formulation. The “es gibt” looks like an existential, descriptive claim: it is a fact that translators strive, that they do their best. Well, how valid is this fact? We could reply that good translators do indeed do their best, most of the time, but surely there must also be many translators who merely do the minimum, at least sometimes. Professionals must often satisfice, after all. And there are many bad translators, of course (if a translator is anyone who does a translation, as a general theory should surely assume).

              It seems to me to be clear that skopos theory is essentially prescriptive, although it has some descriptive assumptions. It aims to describe how good translators, expert professionals, work; what good translations are like. It describes an ideal world (see also Chesterman 1998). Vermeer has acknowledged this (Chesterman 2001), saying that the theory seeks to describe optimal cases. Elsewhere, however, he also seems to suggest that functional theories in general are both descriptive and prescriptive:


Skopos theory is meant to be a functional theoretical general theory covering process, product and, as the name says, function both of production and reception. As a functional theory it does not strictly distinguish between descriptive and (didactic) prescription. (Vermeer 1996: 26n)


Although the term “functional” remains problematic, I find this claim curious. Consider for instance the analysis of the reception of translations in a given culture in a given period. This would be an analysis of how the translations “functioned” in the target culture (data might include all kinds of responses, critical reviews, library loans, size and number of editions published, allusions to the translations in other writings, use of the translations as source texts for further translations, or as literary influences; sales of commercial products advertised by the translations; changes in the social, political, religious or ideological conditions; and so on). The analysis would not need to be prescriptive in any way. Even if the analysis compared the reception with the inferred intentions of different clients, this would not imply a prescriptive approach.

              On the other hand, there is one obvious way in which prescriptive claims can be viewed descriptively, and that is by formulating them as predictive hypotheses, as argued in Chesterman (1999). Vermeer actually does precisely this at one point (1996: 31): “if you translate in such and such a way then y will happen”. Such predictions can then be tested in the normal way, and the results can be generalized in the form of guidelines which, if followed, are reliably assumed to lead to translations which do not give rise to negative feedback (“Protest” in skopos-theoretical terms). This, of course, is precisely what translator training courses teach. It is also what skopos theory aims to do. If you keep the skopos in mind, and translate accordingly, the result will be better than if you neglect the skopos.



5. Empirical status of the theory


As presented, skopos theory is not founded on a search for empirical regularities. This point has been made by many critics (e.g. Koller 1995: 215). We can nevertheless consider how its various assumptions and claims might be tested empirically. It is striking that very little such testing has actually been done. What kind of evidence would falsify or weaken its claims? I will first consider the theory’s descriptive adequacy, then comment on its explanatory adequacy and possible testable consequences.


5.1. Descriptive adequacy


Axioms two and three in the original list above are descriptive. Axiom two, on translation as an offer of information, is definitional. It is an interpretive hypothesis, which can be glossed something like this: ‘in this theory, we claim that a translation is usefully interpreted as ...’. As such, the claim is not falsifiable, but is testable pragmatically, i.e. in use (see further Chesterman 2008). Has this interpretation been widely adopted and led to further hypotheses? Not notably, it seems. On the contrary, it has aroused some criticism, as it seems to allow the concept of translation to expand too far (e.g. Kelletat 1986).

              Axiom three states that translations are not reversible. This claim can certainly be tested empirically, via back-translation. In my view, the claim is too extreme. It would surely be more accurate to say that the smaller the unit of translation, the more reversible it is; that in cases of standardized translations – e.g. in multilingual  glossaries of special fields or in the names of institutions, or in many idioms and proverbs, in numbers, etc. – reversibility may well be the norm. In other words, the claim needs to be restricted, made subject to other conditional factors such as size of translation unit, text type, skopos, and so on.

              There have been a few empirical studies recently which question some of the other basic assumptions of skopos theory. Koskinen (2008) examines the working conditions of EU translators. One of her findings is that in many cases, EU translations that are not intended for the general public are not directed at a target culture at all, but are oriented by the needs of the source institution (99-100). This goes against the skopos theory assumption that a translation should have optimal functionality for target culture addressees. However, this type of EU case is not evidence against the idea that a translation is primarily determined by its skopos.  Here, the skopos is simply not a target-oriented one. Interestingly, Koskinen points out that the special requirements of this kind of translation are experienced as particularly problematic by translators who have been trained in a functional approach: their translation brief seems to conflict with the target-oriented way in which they have been trained to think.

Furthermore, many professional translators do not work as autonomous individuals but as members of a team of experts, including terminologists, subject specialists, revisers, copyeditors and so on. Such conditions do not always support the skopos theory assumption that it is the translator who ultimately decides how to translate, as the expert. (“Er entscheidet letzten Endes, ob, was, wie übersetzt/gedolmetscht wird.” Reiß and Vermeer 1984: 87.) One recent study illustrates this well: Nordman (2009) examines the complex process of Finnish-Swedish translation in the bilingual Finnish Parliament, and highlights interesting disagreements between the translator’s preferences and those of revisers or legal experts, and how these are resolved. The translators and revisers she studied seem to have different norm priorities. It is not always the translator’s views that prevail.

Even in some literary translation the priority of the translator’s expertise has been questioned. In a questionnaire study dealing with poetry translation, Flynn (2004) queries the status of some of the factors which skopos theory assumes, including that of the dominance of the translator’s own expertise. Flynn found that the situational factors affecting the final form of the translation are more like sites of confrontation between the various agents involved, including publishers and proof-readers as well as translators. The translator does not necessarily always have the final say. Flynn’s results admittedly concern a particular type of translation only, in a particular (Irish) context; but again, we can point out that a general theory should be able to cope with all types.

              As another example of evidence against the assumption that it is the expert translator who makes the final decisions I cite an ongoing PhD project by Julia Lambertini Andreotti at Tarragona. She is a qualified court interpreter working with Spanish and English in California. The ethical code there requires that interpreters make no alteration to the register of the legal jargon as they translate. But since many of the clients are not well educated, they simply do not understand the legal terminology, and so do not understand what they are asked. As communication experts, the interpreters naturally wish to adapt the register so that the clients can understand, but this is not allowed. The interpreters are simply not permitted to act as skopos theory assumes.

One might argue that all such examples are cases where a translator is forced to act under duress, against the council of his own expertise, and thus in non-optimal conditions. They would thus fall outside the scope of skopos theory. Reiß and Vermeer explicitly exclude instances of “Translation unter Zwang” (1984: 101). – But there are multiple kinds and grades of duress, including  unrealistic deadlines, legal constraints etc., which characterize much real-world translation and interpreting. Indeed, if there are in fact more non-optimal cases than optimal ones, skopos theory itself would deal only with special cases – surely not the intention of the skopos theorists. A general theory should be general enough to encompass all cases.

              From another point of view, note should be taken of studies on how translators perform under time pressure (e.g. Hansen 2002). These studies suggest that when professional experts work under unusual time pressure, they tend not to waste time pondering about the skopos or the target audience but simply stay on the surface of the text, translating fairly literally, without reformulations or other major shifts which might actually be appropriate for the readership. Here again we have professionals working in a non-optimal situation, without sufficient time for normal working procedures. Under these conditions, the skopos assumptions seem not to represent what actually happens.

Research such as these studies underlines the way in which skopos theory relates more to an ideal, optimal world than to the real and often suboptimal world of everyday translation. In this sense, some of the general descriptive claims and assumptions of the theory can easily be falsified, or forced into more conditioned formulations – if they are supposed indeed to apply to all translation, not just optimal translation done in optimal working conditions. And what about the undeniable existence of a great many really bad translations? These are nonetheless also translations, of a kind; but they are completely excluded from skopos theory. From the point of view of descriptive adequacy, then, the theory is inadequate. But if it is taken as a prescriptive theory, of course, this is not a valid criticism.



5.2. Explanatory adequacy


The first axiom (in the German list above) is a causal one. From the point of view of the production of a translation, it states that the skopos is the most important conditioning factor. This has obvious prescriptive relevance. But retrospectively, as an answer to the question “why is this translation like this?”, the axiom formulates a causal explanation. The most important reason why a translation (or, modified: an optimal translation) is as it is, is its skopos. This claim gives rise to several points.

              Implied in the axiom is the assumption that every translation has a skopos in the first place. Is this assumption testable? Could we in principle refute the claim by finding a translation that does not have a skopos? Perhaps not. Skopos theorists would say that the skopos is always already there, even if not well defined or even definable. Some scholars (e.g. Kohlmayer 1988) have argued that literary translations do not have a skopos; it is certainly not easy to be precise about the skopoi of literary texts.

              The theory does not assume that the skopos is the only cause, merely that it is the dominant one. How might this dominance be measured? What other causes might be relevant? One might counter-argue that under some conditions the skopos dominates, under other conditions some other cause (perhaps these other conditions would be non-optimal ones?). Kelletat (1986) points out that Reiß and Vermeer fail to show how different translation traditions, as well as different skopoi, are themselves affected by historical conditions.

Recall Aristotle’s four causes: the skopos cause corresponds to his “final”, teleological one. It has been argued (e.g. by Pym 1998: 148f) that skopos theory overvalues this teleological cause at the expense of the others. The skopos does take account also of Aristotle’s formal cause (the target-culture norms), but what about the efficient cause (the mind and body of the translator plus computer etc.) or the material cause (the source text, perhaps also the constraints of the target language itself)? Pym argues that by imputing purpose to a text, skopos theory neglects the socially determined individuals who are actually doing the translation. Skopos theory also avoids the problem situation in which client and translator differ about the skopos, or where there might be a conflict of loyalty. True, Nord’s version of skopos theory (Nord [1988] 1991 and later versions) gives more weight to the material cause by focusing strongly on the analysis of the source text. But Vermeer does not agree with Nord’s introduction of the notion of loyalty (see further below).

              It is often pointed out in textbooks that skopos theory helped to dethrone the source text as the dominating causal factor determining the form of a translation. It has now placed the skopos on the vacant throne. But perhaps we would achieve greater explanatory adequacy, more nuanced explanations, if we recognized a range of explanatory variables including norms, personal preferences, special situational constraints and so on, all of which impinge on the translators decisions (see e.g. Brownlie 2003). The relative strength of the different variables in a given case would then depend on the circumstances. This would also allow for non-optimal working conditions, and indeed non-optimal translators. Any generalization about the overall dominance of one independent variable or another could only be based on a wide range of comparative empirical studies.


5.3. Testable consequences


A deductive theory such as skopos theory, if it claims to be empirical, should also generate hypotheses that can be tested and potentially falsified. The theoretical claims should have testable consequences. Curiously, the theory does not seem to have been very productive in this respect. One reason may be the difficulty of operationalizing its central concepts and claims.

              One predictive hypothesis that does arise very naturally from the theory’s premises is that if the skopos is different, the translation will be different. This can easily be tested: a given source text is translated twice, each time with a different skopos. Nordberg (2003) studied precisely this situation, but focused on differences between professionals and non-professionals rather than skopos-determined differences.

              Other hypotheses can be proposed, which (as far as I am aware) have not been explicitly tested.

              – If a translation is deemed bad, the main reason (a reason?) will be that it does not meet the skopos. (Also obvious, surely.)

              – The more the translator pays attention to the skopos, the better the translation will be. (Another obvious one.)

              – Skopos-type correlates more closely than source-text type with equivalence-type. This one would be more interesting. It would test the relative causal force of skopos vs source text-type on the source-target relation. But there is a problem here. We have typologies of text types, and typologies of equivalence, but, surprisingly, no generally accepted typology of skopoi. Reiß and Vermeer give random examples, but no systematic typology. Would  such a typology necessarily be culture-bound and thus not part of a general theory? Surely not, at least no more than other typologies used in translation research. Prunč (1997a) proposes seven prototype classes, but these are in fact classes of translations, not skopos types. They range from non-translation via pseudotranslation to homologous, analogous, dialogic (involving more interpretation) and triadic (intervening, e.g. feminist) to diascopic translation (gist translation, commentary). Nord (1997) proposes two basic “translation functions”, documentary and instrumental; but here too we are dealing with types of translation rather than skopoi.

In a different article, Prunč (1997b) makes a useful distinction between implicit and explicit skopos. The implicit skopos is defined as the lowest common denominator, the default skopos compatible with the current translation norms and the prevailing translation tradition. The explicit skopos then states, where necessary, any divergence from the implicit skopos. But no typology of explicit skopoi is offered here.

Wagner (in Chesterman and Wagner 2002: 45) does offer one specific skopos typology, for use in the EU. These are: 


              For information, not for publication

              For publication [a very broad  category]

              For advertising and marketing

              For use as a legal document

              For text scanning and abstracting


              The big gap is the lack of research on how a given skopos or skopos type might actually correlate with given translation strategies or techniques. Even within given language pairs, this would be interesting and useful. Reiss and Vermeer (1984: 85) say that a “complete” translation theory would give “rules” for the analysis of the target recipient situation and how this then conditions the translation strategy. But skopos-based research has not developed much in this direction. One wonders why. (For a recent attempt to operationalize the skopos concept, see Sunwoo 2007.) We look in vain for testable hypotheses of the form: if the skopos is (type) X, professional translators tend to choose general strategies ABC. Or: in the case of translations from language S to language T (under conditions PQR), translators tend to realize skopos-determined strategy A as techniques / translation solutions DEF.



6. The ethical dimension


Partly in order to counter criticisms that skopos theory is too target-oriented and underestimates the value of the source text, Nord (e.g. 1997: 123f) introduced the concept of loyalty, which is defined as meaning loyalty to all the people concerned on both side of the communication exchange, including the source author and sender. Vermeer does not like this development, as he thinks it brings in an unwanted ethical dimension. A scientific theory is value-free, he argues.


It has been argued that translation involves an ethical aspect. However, I amof the opinion that ethics must not be mixed up with general theoretical considerations about other subjects. Science should be value-free (wertfrei). (Vermeer 1996: 107)


Toury apparently agrees (1995: 25).

However, I think this position is untenable. In a classic article, Rudner (1953) showed that science can never be value-free. Value judgements are made all the time: in selecting “interesting” research problems (i.e. those that have value...), evaluating the evidence and counter-evidence for hypotheses, prioritizing research goals, and so on. Skopos theory highlights the expertise of the translating agent, who makes decisions. It seems to me that any such theory must include an ethical dimension, at least insofar as ethical issues influence the agent’s decisions. (See also Martín de León 2008.) Furthermore, one could argue that a functional theory of translation should also cover the possible general ends which translations serve, their ultimate functions, as it were: some translations are surely more worth doing than others, for instance, not in terms of pay but in terms of the ethical value of the end product (cf. voluntary, activist translation, for example). It is all very well to say that if we want X we should do A (rather than B, for instance); and we can of course test this hypothesis, checking to see whether A or B is a better means to attain X. The only criterion here seems to be that of effectiveness. But suppose we ask who wants X? And for what reasons? For what further end is X itself a means? And who does not want X?

The lack of this dimension in skopos theory, where any end is apparently as good as any other, has led to the theory being associated with a kind of robotic, unthinking translation process, which dehumanizes the translator. Baker (in Chesterman and Baker 2008: 21-22) puts this view as follows:


We soon configure something like skopos theory as a narrative in our minds: the theory evokes (for me at any rate) an industrialized, affluent society populated by clients and highly professional translators who belong to the same ‘world’ as their clients, who are focused on professionalism and making a good living, and who are highly trained, confident young men and women. These professional translators and interpreters go about their work in a conflict-free environment and live happily ever after. They do not get thrown into Guantánamo or shot at in Iraq, and they do not end up on the border of Kosovo and Albania in the middle of a nasty war, where they would have to decide whether or not to fulfil their commission at the expense of treating potential victims with compassion and respect.


Prunč (1997b) seeks to counter this criticism that the skopos view of translation risks being interpreted as totalitarian. He stresses how skopos theory can account for cooperation between agents, and expands the loyalty concept to include the translator’s loyalty to him/herself.

              Perhaps Vermeer is equating value-freedom with rationality? Rudner concluded that value-freedom is often confused with objectivity. He argued:


What is being proposed here is that objectivity for science lies at least in becoming precise about what value judgments are being and have been made in a given inquiry – and even, to put it in its most challenging form, what value decisions ought to be made; in short that a science of ethics is a necessary requirement if science’s progress towards objectivity is to be continuous. ([1953] 1998: 497)



7. The competition?


In assessing a theory (or indeed a hypothesis) one also compares it to competing theories. The closest competitor to skopos theory is Justa Holz-Mänttäri’s theory of translational action (1984). Indeed, this is so close, albeit with a different terminology, that they are often considered to be variants of the same framework (Vermeer 1996: 61f). Some of the criticism of skopos theory has also been directed at Holz-Mänttäri’s work (e.g. Koller 1990).

              A more interesting comparison might be made with relevance theory, in its application to translation (starting with Gutt 1991). A detailed analysis would take up more space than is available here, so a few comments must suffice.

              In their discussion of the notion of the offer of information,  Reiß and Vermeer (1984: 76, 103) write that this “offer” is assumed by the sender to be “interesting” to the receiver. I.e. the sender thinks that the information contains something that is “new” to the receiver, something that will have some effect, which will somehow change the existing state of affairs in the receiver’s world. – This formulation suggests the central concept of relevance, as understood in relevance theory. It would seem that skopos theory sees itself as being much broader than relevance theory, including relevance-theoretical insights and much more (Vermeer 1996: 65-68).

              Relevance theory clearly aims to be descriptive and indeed explanatory, not prescriptive. But there does seem to be a conceptual overlap, perhaps more than Vermeer seems to accept (see Vermeer 1996: 47f, and his comments in Chesterman 2001). Both theories assume that human action is rational, and both give importance to the receiver’s side. The two theories differ in the extension of their object of study: relevance theory excludes very free translations, and so-called multilingual descriptions, which skopos theory would include (Vermeer 1996: 63). Skopos theory also allows a deliberate change of skopos from that of the source text, which would seem to be excluded by Gutt’s application of relevance theory.



8. Concluding remarks


Skopos theory has proved itself to be pedagogically invaluable, as a prescriptive theory; I cannot imagine training translators without using the idea of the skopos. But as an empirical theory seeking to describe and explain translation phenomena in the real world, it is weak, precisely because it relies on an optimal set of working conditions with optimally competent translators. One of the most serious problems in the real translation world, however, is the prevalence of poor translations, coupled with poor working conditions and low pay. These issues are not addressed by skopos theory.           

              In terms of its productivity, the theory has been somewhat disappointing. It has not generated a substantial body of research proposing and testing new hypotheses derived from the theory, nor significally developed its conceptual apparatus. True, it has been expounded in numerous textbooks, and applied in numerous training programmes and notably in translation criticism (e.g. Ammann 1990). It has succeeded in giving some theoretical weight to an intuitively sensible idea ­– that translators should consider the purpose of their translations – and certainly helped to shift the focus of theoretical thinking away from equivalence towards other relevant factors and agents which affect the translation process. It has thus helped, perhaps paradoxically, to make translation theory more realistic. And it has prompted some of us to ponder more deeply what we mean by a theory.





Ammann, M. 1990. Anmerkungen zu einer Theorie der Übersetzungskritik und ihrer praktischen Anwendung. TextConText 5, 209-250.

Andreotti, J. L. (in progress). Comprehension of legal discourse in interpreted proceedings. Ongoing PhD thesis, Universitat Rovira I Virgili, Tarragona.

Brownlie, S.  2003. Investigating explanations of translational phenomena. Target 15, 1, 111-152.

Chesterman, A. 1998. Review of Hans. J. Vermeer, A Skopos Theory of Translation. Target 10, 1, 155-188.

Chesterman, A. 1999. The empirical status of prescriptivism. Folia Translatologica 6, 9-19.

Chesterman, A. 2001. Skopos and after. An interview with Hans J. Vermeer. Across 2, 1, 133-138.

Chesterman, A. 2008. The status of interpretive hypotheses. In G. Hansen et al. (eds), Efforts and Models in Interpreting and Translation Research. Amsterdam: Benjamins, 49-61.

Chesterman, A. and M. Baker 2008. Ethics of renarration. CULTUS 1, 10-33.

Chesterman, A. and E. Wagner 2002. Can Theory Help Translators? Manchester: St. Jerome Publishing.

Flynn, Peter 2004. Skopos Theory: An ethnographic enquiry. Perspectives 12, 4, 270-285.

Gutt, E.-A. 1991. Translation and Relevance. Cognition and context. Oxford: Blackwell.

Hansen, Gyde 2002. Zeit und Qualität im Übersetzungsprozess. In G. Hansen (ed.), Empirical Translation Studies: process and product. Copenhagen: Samfundslitteratur, 29-54.

Hebenstreit, G. 2007. Defining patterns in Translation Studies. Target 19, 2, 197-215.

Holz-Mänttäri, J. 1984. Translatorisches Handeln. Helsinki: Suomalainen Tiedeakatemia.

Kelletat, A. F. 1986. Die Rückschritte der Übersetzungstheorie. Vaasa: Vaasan korkeakoulu.

Kohlmayer, R. 1988. Der Literaturübersetzer zwischen Original und Markt. Eine Kritik funktionalistischer Übersetzungstheorien. Lebende Sprachen 33, 4, 145-156.

Koller, W. 1990. Zum Gegenstand ger Übersetzungswissenschaft. In R. Arntz and G. Thome (eds), Übersetzungswissenschaft. Ergebnisse und Perspektiven, Festshcrift für Wolfram Wilss zum 65. Geburtstag. Tübingen: Narr, 19-30.

Koller, W. 1995.  The concept of equivalence and the object of Translation Studies. Target 7, 2, 191-222.

Koskinen, Kaisa 2008. Translating Institutions. An Ethnographic Study of EU Translation. Manchester: St. Jerome Publishing.

Martín de León, Celia 2008. Skopos and beyond. A critical study of functionalism. Target 20, 1, 1-28.

Norberg, U. 2003. Übersetzen mit doppeltem Skopos. Uppsala: Uppsala Universitet.

Nord, C. [1988] 1991. Textanalyse und Übersetzen. Theorie, Methode und didaktische Anwendung einer übersetzungsrelevanten Textanalyse. Heidelberg: Groos.

Nord, C. 1997. Translating as a Purposeful Activity. Manchester: St. Jerome Publishing.

Nordman, L. 2009. Lagöversättning som process och produkt. Helsinki: Institutionen för nordiska sprĆk och nordisk litteratur, Helsingfors universitet

Prunč, Erich 1997a. Versuch einer Skopostypologie. In N. Grbic and M. Wolf (eds), Text – Kultur – Kommunikation. Tübingen: Stauffenburg, 33-52.

Prunč, E. 1997b. Translationskultur . Versuch einer konstruktiven Kritik des translatorischen Handels). TEXTconTEXT 11, 2, 99-127.

Pym, A. 1998. Method in Translation History. Manchester: St. Jerome Publishing.

Reiss, K. and H.J. Vermeer 1984. Grundlegung einer allgemeinen Translationstheorie. Tübingen: Niemeyer.

Risku, H. 2002. Situatedness in Translation Studies. Cognitive Systems Research 3, 3, 523-533.

Rudner, R. 1953. The scientist qua scientist does make value judgements. Philosophy of Science 20, 1-6. Reprinted in E.D. Klemke et al (eds) 1998, Introductory Readings in the Philosophy of Science. 3rd edition. Amherst, NY.: Prometheus Books, 492-498.

Sunwoo, M. 2007. Operationalizing the translation purpose (Skopos). MuTra conference proceedings: LSP Translation Scenarios. Available at (Accessed 16.4.2009).

Toury, G. 1980. In Search of a Theory of Translation. Tel Aviv: Porter Institute.

Toury, G, 1995. Descriptive Translation Studies and Beyond. Amsterdam: Benjamins.

Vermeer, H.J. 1996. A skopos theory of translation. (Some arguments for and against) Heidelberg: TEXTconTEXT.



[Back to Select Publications]