Revisiting the concepts of translation studies: equivalence in linguistic translation from the point of view of Peircean universal categories

1 Introduction

Translation studies as an autonomous discipline has struggled to introduce its own terminological apparatus: in the course of its development, translation scholars have managed to list several concepts which have been considered key ideas of translation. They all seem to pertain to the fundamental mechanisms that operate within the translation process and include the concepts of the source and the target (text/audience/culture) along with the related idea of directionality as well as the question of transferring the invariant and retaining the necessary degree of similarity, which would most probably differentiate translating from adapting or rewriting.^[1]

Presenting a synthetic overview of translation paradigms, Chesterman (1997) uses the word “meme” to address the ideas that seem to be recurrent across various translation theories, arguing that they “affect the way translators think and translate” (1997: 12) and, as a result, they survive or replicate in a fashion similar to genes. Should a meme become a central idea at a given time and become generally accepted, it becomes a translation norm, which constrains and regulates the translation process. Among his “supermemes,” which are limited to the most fundamental ideas within translation studies, he includes the source-target relationship, untranslatability, the literal-faithful opposition, the postulate of “all writing is translating” and, last but not least, the concept of equivalence. It follows, then, that equivalence, though being a concept that has proven difficult to be defined, has always been central to linguistic translation.

2 Struggling to define equivalence across translation studies

The concept of equivalence has ebbed and flowed in the academic discourse on translation,^[2] which can only prove its “problematic” nature. Accordingly, it has been defined, rewritten, replaced or simply rejected by translation scholars, who have been highlighting different aspects of the same process, i.e. translation. That is why in order to understand equivalence in linguistic translation it is necessary to provide a brief overview of paradigms, turns or fashions that have organised the body of knowledge of translation studies and indicate the common threads that not only separate different approaches, but also join them and thus come into constructive dialogue.

The paradigm of equivalence includes mainly the linguistic turn of translation studies, which would embrace the status of the source text and the need for equivalence. The term is believed to be introduced in the 1950s, be it thanks to Vinay and Darbelnet (1995; originally published in 1958) or Jakobson (1959). The former understand equivalence as one of translation techniques or procedures, as they refer to it, aimed at replicating “the same situation as in the original, whilst using completely different wording” (Vinay and Darbelnet 1995: 342). The latter follows the idea of “equivalence in difference” and argues that translation involves two equivalent messages expressed by means of different code units (Jakobson 1959: 233), with the reservation that “there is ordinarily no full equivalence between code-units” (1959: 233). The part about different wording is obvious in both definitions, as this is the sine qua non of interlinguistic translation, which includes two languages with their own lexical resources. What is striking is the idea of the same situation in the first case – Vinay and Darbelnet indicate that it is about the stylistic impact of the text that should be retained. It follows, then, that equivalence should be determined against the context (i.e. the situation). But this sameness of situation seems the most problematic, as we may pose a question whether it is ever possible to reconstruct the same situation. Jakobson’s stance is therefore more reasonable and is developed further on in his paper. Languages are different in what they must convey – he writes (Jakobson 1959: 236), indicating that languages are resources limited by some conventions and rules on the one hand, but on the other hand, open to dialogue with other signs. In other words, Jakobson underlines the semiotic dimension of translation, but since he is a representative of the linguistic turn, his understanding of equivalence has been generally associated with this perspective and may not be very popular currently.

A similar comment may be made in the context of Catford (1965) and his ideas. Again, the main problematic aspect of his theory, apart from the no longer popular linguistics-orientedness, is the so-called “sameness of situation” (Catford 1965: 52). Catford claims that equivalence exists at the level of specific texts, which by means of some replacement of textual material refer to the same extralinguistic reality. Though the wording may be as unfortunate as in the case of Vinay and Darbelnet, Catford underlines the fact that equivalence is about some kind of contextual meaning and, what is more, he points out that the level of correspondence is a matter of different ranks and does not have to be attained at all of them simultaneously.

Further attempts at defining the concept were made by Nida (1964), who differentiated between formal and dynamic equivalence, and Newmark (1981), who seems to have reiterated Nida’s words by introducing a similar division and writing about semantic and communicative translation. In Nida’s theory, the central idea revolves around the closest natural equivalent (Nida 1964: 166), which presupposes the highest possible “degree of approximation” in terms of language and culture. In this way, Nida underlines the role of both culture and the principle of similar effect, but fails to provide objective means to measure the degree of equivalence. However, he clearly signals that equivalence is a relation between the source and the target text that may be slanted towards either of the two, depending on the needs, text types and audiences. Newmark (1981) renames the distinction, following the old-age dichotomy between literal and free translation: accordingly, he provides a list of levels of literalness in the form of a V-diagram, starting with word-for-word translation and finishing with adaptation (Newmark 1988: 45). Apart from the educational and prescriptive values of the books by Newmark, his approach to equivalence does not seem to have moved the debate any further.

Similarly, typologies put forward by other scholars, including the representatives of the Leipzig School (see e.g. Kade 1968 or Jäger 1975) as well as Koller (1979) or Komissarov (1980), though relatively clear and aimed at measuring or indicating types of equivalence, do not seem to have helped to finally define the concept. Instead they were subject to criticism, as was the case with other approaches founded on linguistics. For instance, Snell–Hornby (1988: 20) rejects such mechanistic approaches, which too often restrict the notion of equivalence to the level of single lexical items or the number of potential equivalents. She concludes that such an approach means defining equivalence on the basis of specific texts and cases of translation, which in turn does not allow any generalisations. Despite its fundamental shortcomings pertaining to measuring an abstract concept, the tendency to formalise and operationalise equivalence has continued into the present day, which may be exemplified by the approach put forward by Gerzymisch–Arbogast (2001), who offers a list of equivalence parameters analysed on two levels: a system and a text level, or Izquierdo (2014), who presents a corpus-based analysis of similarities and differences on the level of sentences or sentence constituents and arrives at the list of five parameters used to measure equivalence. Recently, equivalence has been even measured by means of empirical research: a case in point may be the quasi-experiment conducted by Walker (2019), who uses the eye-tracking methodology to measure the “equivalent effect,” with the hopes of establishing “a cognitively-oriented model” of the measured quality (Walker 2019: 137).

Though the linguistic approach is currently considered insufficient, the ideas developed within it seem still valid. The scholars mentioned above understand equivalence slightly differently, but they all underline relevant features of the concept. Equivalence is a relation between two objects: the source and the target text; there is some kind of transfer or replacement that leads to its establishment; this transfer operates on various levels and ranks and may result in different values of equivalence; and finally, equivalence may be measured in terms of some extralinguistic contexts, e.g. the principle of similar effect or culture.

The subsequent turns of functionalist, descriptive and cultural approaches shifted the focus to the target text and, at points, considered the concept to be unnecessary. Reiss and Vermeer (1984: 139–140), who represent the functionalist school, do not reject the concept of equivalence completely, but instead complement it with the concept of adequacy, thus relegating equivalence to cases of the so-called “Funktionskonstanz.” In other words, equivalence is achieved only when both texts have the same function in both contexts, since it guarantees a relatively high degree of semantic fidelity (according to the scholars, an ideal level of absolute sameness is almost impossible to be achieved). It means that equivalence is somehow reflexive and symmetrical: as long as the context is similar, it is enough to copy the content and produce equivalent texts. Otherwise, it is necessary to mediate the meaning, in which case the only acceptable term is the one of adequacy.

Descriptivists do not represent a unified point of view: for instance, Holmes (as quoted by Bassnett 2002: 34–35) dismisses the concept of equivalence as being perverse and not matching the sense of translation. Similarly, Hermans rejects the notion, but not because of elusive semantic fidelity. He criticises the structuralist view, according to which equivalence would mean “equality in value and status” (Hermans 2003: 39). It follows that equivalence does not fit translation, since if the target text is considered equivalent, i.e. equal in status, it is no longer a translation, but simply a rightful copy or version, as is the case with sacred texts accepted as automatically accurate. Translation, on the other hand, is repeatable, and, as such, it is associated with plurality rather than equal status. His argument holds true, as it highlights the fact that a translation is never the same as the original, if only for existing in a target context. But the overall approach seems too strict and based again on the idea of sameness and identity.

These voices may be contrasted with the approach represented by Toury (1980, 1995, who retains the concept and argues that it is both abstract and concrete. The abstract understanding of equivalence assumes that it is “a theoretical term, denoting an abstract, ideal relationship, or category of relationships between TTs and STs” (Toury 1980: 39), which actually epitomises his postulate of relation (Toury 1995: 35). The concrete or empirical understanding assumes that equivalence is a descriptive term given in every translation, i.e. the actual relation between both texts: “[r]ather than being a single relationship, denoting a recurring type of invariant, it [equivalence] comes to refer to any relation which is found to have characterized translation under specified set of circumstances” (Toury 1995: 61). Equivalence becomes thus a defining and inherent feature of translation.

Pym (1992) echoes this idea, as he states that equivalence is indeed a category that is used to define translation. Translation, in turn, should exemplify equivalence that would be describable and definable – a vicious circle with no way out. Therefore, Pym (2007) suggests another perspective: he understands equivalence as a negotiable concept determined by the translator. Pym differentiates between natural equivalence, which resembles the concept of formal correspondence, and directional equivalence, created by the translator. The latter may be thus used “to justify quite significant textual expansion or reduction” (Pym 2007: 286).

The question of significant departures from the source text is one of the central topics within the cultural turn. Bassnett (2002: 34) recognises that “once the translator moves away from close linguistic equivalence, the problems of determining the exact nature of the level of equivalence aimed for begin to emerge.” She argues that equivalence is not a quality that should be associated with sameness, as the latter is rarely achieved even between texts in one language. However, she does not reject the term: instead she acknowledges more semiotics-oriented perspectives by referring to works of Lotman, Popovič or Neubert. Lefevere (1992: 10) is stricter in this respect: he rightly points out that both theoreticians and practitioners are unable to agree on one definition of equivalence, which means that the concept should be abandoned. He justifies this relatively strong claim by indicating the unclear scope of the concept: it is either too broad, and therefore meaningless, or too narrow and specific, which makes it difficult to be applied in research or practice.

Similar tendencies are observed also in the field of multimedia translation. Because of the role of the non-linguistic elements (e.g. images or auditory signs) and omnipresent constraints (e.g. spatial or temporal ones) the concept of equivalence is threatened. That is why it has been questioned and replaced with other concepts, such as accessibility, relevance or usability (Gambier and Gottlieb 2001). Again, the main problem lies in the assumption that equivalence exists at the level of linguistic fidelity and, ideally, should mean close similarity rather than any relative concept of relation between the source and target text. If the assumption cannot be met, then the concept becomes problematic and may prove unnecessary.

The paradigm of the non-linguistic turn underlines the problems with the understanding of equivalence. It makes clear the fact that a mathematical understanding of equivalence as a kind of symmetrical and reflexive sameness does not match the process of linguistic translation. For these reasons, scholars representing this paradigm claim that equivalence should be rejected, questioned or replaced by other terms, depending on how critical they are, or it should be viewed from a different angle: as a global, relative and negotiable concept.

3 Nature of equivalence: some assumptions

The previous section proves that equivalence is indeed a problematic concept, defined differently by scholars depending on their background and research goals. Halverson (1997) rightly observes that equivalence is what divides scholars slanted towards linguistics and those representing descriptive and cultural turns. Along the same lines, Gambier (2016) claims that the paradigm of equivalence may be juxtaposed with the subsequent paradigm of the cultural turn, currently challenged by the digital one. Though the differences are fairly clear and easily observable, there are common threads woven somewhere between contradictory ideas about equivalence.

A general perspective is provided by Halverson (1997: 209–210), who indicates prototypical features of the term equivalence in general. It pertains to at least two objects, which means that it is founded on comparing and contrasting; it presupposes the quality of sameness, likeness or similarity; it should be used along with a set of qualities that are to be compared. Halverson then indicates problems pertaining to these components: the question of the appropriate unit of translation, the nature and degree of the required sameness and the quality or criteria according to which relevant units are compared, respectively. As evidenced in the previous section, translation scholars have addressed the problems from various perspectives. It is vital therefore to indicate the most important features of the concept of equivalence and assumptions that may help to defend its status. The list may read as follows:

1. First of all, it needs to be admitted that equivalence is central to linguistic translation, as it somehow defines this activity (which in turn defines equivalence as noticed by Pym 1992). The term has been used across numerous definitions of translation and has been recurrent in names of techniques or strategies. It may not be the preferred concept, yet it is what differentiates translation from other operations and what is encountered across encyclopaedia entries. Accordingly, the concept should be retained and probably redefined rather than rejected or replaced with other terms.

2. It seems apparent that the greatest difficulty connected with equivalence lies in the precision of its definition. It needs to be highlighted that even though equivalence may stand for the quality of equal value, in the case of translation this equality does not mean perfect, bilateral symmetry, absolute sameness or identity. Rather, it represents “an appropriate relation of relevant similarity … established and maintained between the source text and the target text” (Chesterman 1997: 69).

3. It follows, then, that equivalence is a relative and functional term, depending not so much on language similarities or differences, but contexts that surround translation and the composition of texts. In other words, it may be used in Toury’s terms as a concept denoting a genuine relation of relevant similarity between the source text and the target text. This relation is described individually at the level of concrete texts, thus making equivalence a concrete quality.

4. Texts become then one of the potential levels of reference: as indicated in the previous section, equivalence may be attained (or not) at different levels, including, the level of words, texts, content, situation, imagery, etc. What is more, specific parts of one text may have different “degrees” of equivalence, which means that it is a descriptive, gradable and relative category, as it is conditional on various factors (Baker 1992). This assumption is indicative of another burning problem of translations studies, viz. the meaning of the concept of a text, which should be analysed in a relevant context.

5. It becomes crucial, then, to define this relevant context of translation with all potential constraints, as they determine the translation process. Equivalence, as an inherent feature of translation, should be consequently defined in most generic terms if it is to be applied to different cases of linguistic translation as one of its core concepts. In addition, it seems unfeasible (and probably unnecessary) to determine the range or the desired degree of equivalence, since it is relative by nature and the actual degree depends on numerous factors, identified individually for each translation task. Equivalence is not a golden standard of ideal sameness providing prescriptive instructions. It is a quality of translation that is negotiable, changeable and case-dependent. It seems, therefore, reasonable to reduce the concept to its essential components, which should be reflected in an abstract definition.

6. Equivalence may be perceived as a category of assessment used mainly by translation scholars, critics and students as well as proofreaders, commissioners, clients and other agents who may have the wish to carry out such an analysis profession-wise. While in the first group the analysis will most probably be connected with the problem of the ontic status of equivalence and its definition, in the second group of stakeholders it will most frequently pertain to the pragmatic and business dimension of translation and will be aimed at indicating translation errors, highlighting inappropriate equivalents or determining the strategy.

                Although it is a quality negotiated by various social actors, equivalence lies mainly within the competence of the translator. As pointed out by Chesterman (2017 [2012]: 74), “equivalence is produced by translators, and then implicitly claimed by them, a claim that is (or is not) then accepted by the client and later readers.” It is therefore conditional on the subjective opinion of the competent translator (Catford 1965; Pym 2004; Toury 1995). It means, then, that equivalence should be assessed in a wide context that involves not only texts, but also a relevant portion of the extralinguistic reality, including the immediate context, media and modes or social actors.

7. Equivalence, as a general feature of linguistic translation, may cover a number of criteria, ranging from pure linguistic correspondence, content invariance or the pragmatic profile to functional or system meaning. It is a relative concept not only because it is sensitive to other factors, but also because it may be assessed from a number of vantage points, depending on the main focus of the researcher. In other words, it is not to be associated with any absolute, full or ideal values: even though the concept may be considered gradable, there are no precise rating scales. We may only attempt to say that a particular target text seems more equivalent (being relevantly more similar to the source text in terms of the syntagmatic meaning), than another target text, for it is a more literal rendering of the latter; it addresses the expectations of target social actors (i.e. is more relevant, adequate or accessible, as it highlights the functional meaning); it is more acceptable in terms of some rules or conventions (i.e. is more acceptable in terms of its system meaning); or it fits the context of translation better (i.e. reflects contextual features on the level of its textual meaning). However, only a context-based analysis can produce a complete image of equivalence in linguistic translation.

8. Even though equivalence is a feature of linguistic translation, or its descriptive category, it does not guarantee that a target text, which because of some criteria seems more equivalent than another target text, will automatically be the preferred or the accepted text. Equivalence does not inform us whether a particular target text is better or worse, since this is a category of description rather than assessment. Although it is necessary in the case of linguistic translation, high equivalence is not the prime or the only aim of translation, as it is simply an intrinsic quality of the former having different values. A specific context of translation may produce a target text with limited equivalence. Otherwise, a high degree of equivalence observed on the level of the syntagmatic meaning may be necessary.

This list of the most important observations concerning equivalence demonstrates that there are features that may be considered recurrent and essential. They may be summarised as follows: equivalence carries the idea of relevant similarity or resemblance between two texts joined by some kind of transfer; it is a describable concept with several levels of reference existing in a real and specific context, but numerical scales are rarely applicable; it is established by professionals, who usually conform to and act within a specified framework of contextual features; it is a concept inherent in translation; it is aimed at describing rather than assessing translation; and finally, it is a relative concept that depends on the adopted perspective.

Equivalence seems the true essence of every task of linguistic translation, which in turn determines its profile and degree. This claim is visible in, for instance, Toury’s postulate of relation (1995: 35) or Chesterman’s relation norm (1997: 69), both of whom argue that translation presupposes a relationship of similarity between the source and the target text, which secures the translation-status. This circularity may be ambiguous and questionable, but it proves that equivalence is one of central concepts of translation (in the sense of translation studies) that most probably needs redefining. A general definition of this concept should reflect its fundamental, relative and dynamic nature, which may be done with the help of Peircean universal categories.

4 Equivalence and universal categories: attempt at revision

Peirce dedicated a considerable amount of attention to creating a system of triads, which would be reflected in his later writings and ideas. Accordingly, Peirce listed Firstness, Secondness and Thirdness as three levels with monads, dyads and triads (CP 1.293); he would consider all sciences to be divided into Science of Discovery, Science of Review and Practical Science (CP 1.181); his general theory of signs, i.e. semeiotic, would have three divisions, including speculative or pure grammar, critical logic and speculative or pure rhetoric (CP 2.229); he distinguished three types of inferences, including logical reasoning (further divided into deduction, induction and abduction; e.g. CP 1.65–1.74), perceptual judgments of sensation and emotions as well as habits. The assumingly tantalising number of three was reflected also in the way Peirce perceived semiosis as a process founded on three elements; he argued that a sign is built on the basis of a triadic sign relation with a representamen, object and interpretant (CP 2.228) and he divided signs into icons, indices and symbols (CP 2.227, 2.247–2.249). The list illustrating Peirce’s penchant for number three seems impressive and almost endless. This specific way of perceiving and explaining the reality may actually confirm the power of a triad-based perspective: Peircean universal categories may prove to be genuinely comprehensive and all-embracing irrespective of the area of research, as they provide a missing link to structuralist binaries.

The universal categories have been used by translation scholars, attempting to explain linguistic translation or other related concepts, as indicated in Section 2. Further examples may be illustrated by, e.g. papers by Stecconi (2004), who develops his model of t-semiosis and indicates fundamental conditions of translation, Gorlée (2008), who examines the problem of translation in opera by establishing links between the Jakobsonian model of communication and Peircean universal categories, Gorlée (2015), in which she presents the theory of transduction, or Robinson’s (2016) revisionist writing on semiotranslation.

The following discussion presents yet another attempt at integrating Peircean thoughts with translation theory, which will hopefully help arrive at a general definition of equivalence. Though it refers mainly to linguistic translation, it is argued here that interlingual translation is only one operation carried out by translators, who in the course of the process paraphrase and translate linguistic signs into non-linguistic ones in order to produce the target text. Hence, Jakobsonian division does not stand for three separate types of translation, but rather three types of operations on which linguistic translation is based. Accordingly, it is assumed that a text stands for an orderly composition expressed in a specific code (which in the case of linguistic translation has to include linguistic signs) and by specific means aimed at communicating integral meaning (see e.g. Lotman and Piatigorsky 1978, Lotman 1990), and the process of translation takes place in a semioticised portion of reality referred to as the context of translation (for details see Rędzioch-Korkuz 2021). The discussion is centred on indicating specific parallels mainly between the universal categories as introduced by Peirce and the linguistics-and translation-oriented features of equivalence. It highlights the potential of a semiotic perspective in terms of rewriting and revisiting fundamental concepts of translation studies. However, this attempt is rather tentative in character and is not intended as a refutation of earlier or contemporary understandings of equivalence.

The idea of Firstness is associated with difference (as in two different sign spaces and texts), a pure quality and something original. The category of Firstness is:

predominant in the ideas of freshness, life, freedom. The free is that which has not another behind it, determining its actions; but so far as the idea of the negation of another enters, the idea of another enters; and such a negative idea must be in the background, or else we cannot say that the Firstness is predominant (CP 1.302).

Further on, Firstness is predominant because it is peculiar and idiosyncratic and finds its reflection in feeling, thought and free will rather than objective perception. This may resemble the notion of equivalence as being an individual value of a given pair of texts and context, open to a relative approach and interpretation adopted by the translator (the entering another). Perceiving equivalence as a category of Firstness underlines two important features of the concept: firstly, it is relative, which means that it is given but it is not an automatic pattern that may be repeated according to step-by-step instructions. This suggests that equivalence, theoretically, defies any fixed or quantitative classifications. Secondly, this perspective underlines the autonomy of the translator and his or her competence. The translator, as an expert, is the one who exercises creative freedom and utilises a specific vision of the target text. Equivalence is given as a feature identifying a translation event (Firstness as being or potentiality; equivalence existing a priori as argued by Toury 1995), having its own quality pre-determined to a certain extent by the translator.

This first dimension of equivalence means also a necessary degree of difference or originality, which leads us to the rejection of the postulate of “the same quality,” and moves towards the postulate of, e.g. Hönig and Kussmaul (2003: 63), who argue that translation is characterised by a necessary degree of difference. In other words, it helps to substantiate the claim that there is only a relevant degree of resemblance between the source and the target texts and that “equivalent” does not mean perfect symmetry or equality in status and value.

Firstness does not exist in isolation and so does equivalence, which needs to be contextualised or embodied. What is more, there must be a struggle in Peircean terms, which presupposes a reciprocal action between two entities outside any laws or conventions. Secondness (i.e. existing) means two various ideas, i.e. “action, where our modification of other things is more prominent than their reaction on us, and perception, where their effect on us is overwhelmingly greater than our effect on them” (CP 1.324). There is an obvious conflict: on the one hand, there is the necessity to take action aimed at modifying the source text (moving towards domesticating strategies), but on the other hand, there is the source text with its powerful effect, which makes us retain this effect and subsume to this power (moving towards foreignising strategies; this analogy seems in line with Pym’s concept of directional equivalence determined by the competent translator). This struggle and conflict both underline the fact that equivalence is not a static value that is to be ascribed to a translation event, but rather its potential is established in the context of Secondness, which means conflict and constraints.

The following analogy may be suggested: the translator is responsible for modifying the source text, in that he or she renders it in a target text, being under the impression it makes on him or her. There is the ongoing conflict between two ends of translation: the source text evokes powerful effects, which are subsumed under the urge to, e.g. make it suitable for target audiences. According to Peirce, Secondness is predominant “in the ideas of causation and of statical force. … Constraint is a Secondness” (CP 1.325). The idea of “statical force” does not refer to equivalence, which, as already argued, is a dynamic and negotiated quality, but to the context of translation and creative freedom of the translator, since pre-defined constraints are given within this context. The fact that there is the link between the source and the target text creates the quality of Secondness and the context of translation, which brings along objective and hence static constraints. Eco (2001: 16–17) claims that:

[s]imilarity in meaning can only be established by interpretation, and translation is a special case of interpretation, in Peirce’s sense. To substitute a given expression with a series of interpretants means that the substituting expressions are never equivalent to the one substituted, since they can say more under a certain profile and less under another. ‘Under a certain profile’ means according to a given context.

Translation can rarely (if ever) be fully and perfectly equivalent (meaning the same) if only for a different context of reception. In this quotation, Eco underlines the relative character of equivalence and the significance of the context of translation. He also highlights the fact that equivalence can be analysed within a specific set of variables and becomes therefore a relative and dynamic feature.

The concept of Secondness is connected with the idea of a dyad, which is believed to consist of two subjects united within the dyadic relation. Equivalence becomes a real quality through the process of creating specific relations between the source and the target text, which in turn are determined by this quality. Peirce wrote that a dyad has two sides and a specific mode of union between the subjects, both of which have their special character. Also, the sides determine which subject is considered first and form the second side, i.e. the second subject(s). There is a very visible and clear relation of cause and effect or reciprocity.

This bears resemblance to the general nature of translation, in which there are two sides with the source and the target text, both of which have their own profiles and particular features, involved in the process of translation (the dyadic relation or Secondness), which takes place in accordance with specific chronology and directionality. The source text shapes to a certain extent the target text, being its cause or condition. Equivalence is not accidental and follows from this relation. Even though, as Firstness, it is subject to the translator’s competence and seems a fairly independent quality, it is based on a logical order. Thus, the postulated relevant similarity is highlighted. Equivalence as Secondness does not exist independently from its logical source and target text. It does not exist outside the relevant context of translation, since it may be discussed only when real cases of translation are taken into consideration.

There is also another significant stage or mechanism, which determines translation and, in consequence, shapes equivalence. This is the mechanism of mediation, through which the conflict inherent in Secondness becomes resolved. Thirdness, then, is the means by which Secondness and Firstness are realised, since it “mediates between the two subjects and brings about their connection” (CP 1.328). It is the mechanism in which the sign relation takes place and so translation. Peirce highlights the fact that it is the most complex category and requires careful analysis. Thirdness includes the idea of habit or generality and thus means that the desired effect should comply with general, i.e. the most desired and necessary, rules (the comparison to the process of cooking an apple pie), which clearly excludes all instances of unjustified idiosyncrasies.

Thirdness is then the mediating mechanism between two texts and their relevant sign spaces, allowing dialogue between them. Secondly, this is the process that takes place in the mind of the translator and results in the establishment of final interpretants within the context of translation. The translator may be then associated with the Peircean quasi-mind, i.e. the interpreter of a specific mental image (CP 4.536 and 4.550). However, in the light of current advances in translation studies (often referred to as the sociological turn or the translator studies), the translator is not considered here “merely instrumental,” with signs translating themselves into other signs per se (Gorlée 1994: 191). Rather, the translator is an interpreter working towards identifying and determining the final interpretant, i.e. the destined or the intended meaning (CP 4.356; see also Short 1996: 494). By resorting to the common universe of discourse, the translator is to act as a common mind of the source and target minds, creating other signs and their combinations in the form of a target text. The translator’s role is then active and consists in interpreting and utilising the framework created by Firstness and Secondness, i.e. potential freedom controlled by constraints, mediated by habitual norms inherent in Thirdness.

Thirdness is also the mechanism which brings all of the constraints into dialogue and negotiates the pure and original quality of the Firstness, i.e. difference of the ST, through the conflict embedded in Secondness, i.e. constraints. Equivalence is then a result of mediation and dialogue in the context of translation. Therefore, any general prescriptivism based on the claim that, e.g. “a target text should be ideally the same as the source text,” is a common fallacy. Thirdness is about specific habits or norms, rule-governed patterns (which exclude any potential idiosyncrasies of the translator), but the only habitual thing about equivalence is the fact that it is a matter of mediation within the context of translation as its inherent feature.

Equivalence becomes then the quality that differentiates translation from other cases of non-translation. However, it does not mean that only fully equivalent texts or the ones which are characterised by a considerable degree of equivalence are instances of translation. Firstly, it is impossible to indicate the levels or degrees of equivalence, so its assessment or analysis is based on approximations rather than any formalised measures. If we argue that equivalence is an inherent feature of linguistic translation, then also texts with a relatively low level of equivalence are translations. In other words, the fact that someone deems a target text an inadequate rendering, since the level of equivalence is very limited according to some assessment criteria, does not mean that it is not a translation (but with a low level of equivalence or of a low quality if the latter is associated with a high degree of the former). It follows that only evident cases of lack of equivalence are instances of non-translation, which is in line with the way Toury (1995) defines translation by means of his postulates.

In order to arrive at a general definition of equivalence it is then necessary to include the features highlighted by drawing the analogy presented above. It may be said that equivalence is a feature which is defined on three different levels, i.e. the abstract level of being a necessary condition of every act of translation (Firstness); the actual level of the relation between the source text and the target text as well as other elements that determine the conditions in which this relation is to be embodied (Secondness); and the level of the dynamic mediation within the context of translation (Thirdness). This framework helps to reiterate the most important features that are listed in Section 2 and highlights the relative and at the same time translation-bound character of equivalence, meaning that it exists in every case of translation but depends on the temporal, spatial, human-related, material, cultural and normative conditions and is determined (to a certain degree) by the competent translator. And since Secondness is an essential part of Thirdness and Firstness is a vital element of both Secondness and Thirdness, it follows that equivalence is not to be defined and described outside these three dimensions, because such a definition would be incomplete and rather ambiguous, not showing the real (and complex) nature of this concept.

What seems significant is the fact that equivalence is a negotiable feature of linguistic translation observed at the level of texts and their surrounding sign spaces: it is a matter of a particular context of translation and should be analysed from the point of view of constraints, which exist and are relevant in a specific case. As it is founded on the universal categories, it is associated with difference, similarity and mediation, which are also relative rather than absolute concepts and depend on the context. Referring these qualities to translation and texts, we may draw a fairly simple and even trivial analogy: in translation some elements are different, some are similar in terms of, e.g. referential meaning, and some are a matter of mediation. Even if we are willing to keep the traditional point of view and argue that equivalence is associated with similarity, we need to acknowledge the fact that similarity presupposes difference and both lead to, or rather, are results of mediation.

Equivalence is, then, a relation between the source text and the target text founded on the three qualities of similarity, difference and mediation, established in the context of translation. It is a category of analysis and description of translation, but not necessarily a category that serves as an assessment criterion. It is a relative and dynamic quality rather than an absolute category of identity and symmetry – it results from the semiotic dialogue between elements building the context of a translation event.

5 Final remarks

The aim of this paper was to attempt to revisit and redefine one of the key concepts of translation studies by presenting it from a wider semiotic perspective. It may be said that the aim has been accomplished in that the concept of equivalence has been discussed by means of drawing parallels between Peircean universal categories and fundamentals of translation. The question as to whether the presented argumentation will defend the status of equivalence and contribute to its revival in translation studies may remain open though.

The argumentation proves, however, that it is feasible to describe the most fundamental and general features of equivalence without providing exact boundaries or strict rules that would mirror the simplistic approach common to linguistics-based theories and ideas of commutation or decoding. Explaining equivalence is possible with the help of universal categories and semiotics, which helps to move beyond semantic fidelity and present linguistic translation as a complex phenomenon taking place within negotiated and dynamic context warranted by relevant similarity, difference and mediation. Equivalence becomes then a defining feature of translation, which means that the concept should be retained rather than rejected.

In general, it seems vital to embrace the traditional concepts of translation studies, especially the ones that define this particular activity. It is true that translation has been influenced by noticeable and profound changes in the area of communication, science and technology, and therefore the field has evolved into inter-or multidiscipline with a strong tendency of being fragmented. However, it is this interdisciplinarity that calls for rewriting. Each science needs general and universal models in order to be able to address its object of study and identify specific regularities that would be applied to each case. Therefore, it is vital to struggle to rewrite the concepts which may seem unpopular or outdated because of a lapse of time rather than replace them with new and frequently ambiguous terms, especially when the concept defines the activity in question. As pointed out by Chesterman (2004: 33), “An interdiscipline like Translation Studies will be doomed to stagnation if this striving towards the general is neglected.”