Let us initially assume that every language is a cultural fact. It integrates and articulates a whole range of behaviours of the social groups that use it, and it constitutes one of the most elaborate instruments for thinking, saying and acting on the world within intra- and inter-group social relations. If so, in principle, everything in the language - and every expression of the language in speech -contains within itself one or more markers that reveal this cultural bond, traits that refer to sets of values, behavioural, linguistic and extra-linguistic patterns that, as much as the pertinent phonological, grammatical and semantic traces, individualize and characterize or typify a given language/culture complex in relation to other languages/cultures, which may be close or distant (by any criterion of proximity or distance that one wishes to adopt).

Even aspects apparently restricted to the grammatical dimension still contain these markers as they witness, among many other aspects, the interlingual differences in the marking of gender, number, degree and the definite/indefinite, the expression of tense and aspect (linked more or less narrowly to verb conjugation), forms of treatment, and prepositions that mark spatiality (static or dynamic). On the discursive level, markers of this particularizing nature can be observed in the intertextualities that make sense in a given language/culture complex but make another sense (or no sense at all) in other language/culture complexes. Other markers are more directly related to the referential dimension of languages, concerning different areas, such as ecology (flora, fauna, topography, hydrography, etc.), material culture (objects and spaces created by man), social culture (social relations of any order) and religious culture (in the terms of Nida, 1945), or, perhaps more precisely, ideology (references to belief systems) (see, e.g., AUBERT, 1998, 2003). All these markers will be called cultural markers here, and we admit, as far as translation and translating is concerned, that they represent, alongside the poetical function of language, the main difficulties both in translating and in thinking about translating. They can even imply the admission of the intrinsic unfeasibility or the profound relativization of the translational act.

This finding, even though admissible or unquestionable, clashes, however, with certain empirical evidence and with some presuppositions of related disciplines.

Thus, in the case of terminological studies - and, in particular, of bilingual terminology (understood as an auxiliary discipline of translation) - the very raison d'être of the discipline demands that one accept, as a possible hypothesis, the viability of finding (or establishing, by a normative act) biunivocal relationships of meaning in the terminologies of specialty languages expressed in different idioms (interlingual synonyms), and which find their model expression in the nomenclatures of the basic sciences, in measurement systems, in international or internationalizedtechnical standards, among others. It suggests therefore, that, under certain circumstances or conditions of production and reception of texts in a normative or standardized context (specialty languages), cultural differences will be neutralized or annulled.

Somewhat in the same direction, the admission of a permeating presence of particularizing cultural markers seems to conflict with the perception resulting from numerous studies based on corpora of originals and their respective translations, which point out - in varying degrees according to the pair of languages in question - high rates of word-for-word translation and transposition (literal translation, in the broad sense). In fact, and in general terms, we have enough data to affirm, with a high degree of certainty, that, in the relation between Brazilian Portuguese and the other Neo-Latin languages, the literality indices vary between 60 and 80%, while that between Brazilian Portuguese and the Anglo-German languages, the indices vary between 40 and 55%.² Apparently, it is difficult to question this finding, which places us in front of a dilemma that is theoretical, methodological, and descriptive.

An initial attempt to clarify the issue might suggest that these different strands of thinking, investigation and practices are based on differing views of the study of language, the different languages spoken, and translation. The dilemma would not be, therefore, in the “thing in itself” but rather in the different approaches to the phenomena, which ultimately configure, through the force of the perspectives themselves, objects of investigation that are also diverse.

In fact, the cultural vision is established - with greater or lesser methodological rigour- as an anthropological approach; the terminological perspective derives, in part, from lexicological and lexicographic studies but more fundamentally from an epistemology of basic and applied sciences; the contrastive or comparative point of view embedded in the concept of modalities or technical procedures of translation is rooted in descriptive linguistics and in the structuralist perspective, which adopts, as an axiom, the autonomy of languages and especially language as an object of study.

It is possible, therefore, to simply accept that each of these approaches is defensible on its own terms; that this is not a dilemma, but only a scientific parallel, for which it would be pointless to seek a compromise.

Thus assumed, one of the possible consequences would be the understanding that the operation that seeks the sayability of a text in a language/culture of reception which is different from that which gave rise to the text is subdivided into two operations: a terminological-linguistic (strict sense translation) and another which is cultural (anthropological). This is not a purely speculative hypothesis. Certain contemporary trends in the translation market - notably in the domain of so-called “application” software - suggest a division of labour precisely between “translation itself” (understood as a strictly linguistic operation, in the formal sense of the term) and “localization” (understood as a rewriting taking into account the cultural factors of the community of arrival - including making distinctions, for example, between Portuguese from Portugal and Brazil, between British and American English, and between French from France, Belgium, Switzerland, Quebec, Africa, etc).

Another, more ambitious hypothesis would seek, however, to bring these two strands closer together. Ultimately, it could be argued that, in doing so, anthropology becomes an auxiliary discipline of linguistics or, conversely and more likely (since cultural phenomena are not limited to languages, although always finding expression in them), it would be necessary to include linguistics as one of the subdomains of anthropology.

We do not intend here to force the epistemological limits of the respective disciplines. In fact, we consider it perfectly possible to include in the treatment of the problem initially put forward, the notion of “culture”, always, however, from the perspective of linguistics, as a possible approach, but not with exclusive rights, and, therefore, without harming the scientific autonomy of the background disciplines.

Given, therefore, that separate languages ​​and language itself ​​are cultural phenomena, for the purposes of this article a restricted conceptualization of culture will be adopted, limiting our focus to the linguistically expressed cultural marker. What interests us here is to inquire about the difficulties and approaches to be adopted in identifying, in textual sequences analyzed in the original compared with their translations, the elements (lexical, grammatical, discursive or others) that, in their specific update in a given textual segment submitted to the translation operation, incorporate linguistic, intertextual or referential cultural markers into the text that are relevant to the referred operation and which are potentially subject to losses - or, more likely, to substitutions (changes in referentiality or in the “way of speaking”) - which, inexorably, will result in effects of refraction.

The object of study proposed here is not composed of languages as abstracts systems or structures but rather of the acts of enunciation, speech, and verbal production which take place, by definition, in specific contexts and co-texts. In this perspective, the cultural marker will be seen less as a dictionary fact and more as a discourse fact. Thus, when we observe a text or segment of verbal language, we do so on terms and phrases which are updated in context. And one of the defining characteristics of the objects thus observed resides precisely in their being found in a given updated co(n)text.

Updating in co(n)text (as opposed to certain forms of metalinguistic updating) implies that the meaning potential of a word or expression is only partially realized. For example, the fact that the Norwegian term klokke can mean clock, bell/bell tower and time (in the expression hvaerklokken = what time is it?) is, in theory, a linguistic-cultural peculiarity of Norwegian, but it is highly unlikely that an inquiry asking for information about the time will evoke, in the minds of the interlocutors, the tolling of a bell tower simply because the context excludes the updating of this meaning.

Likewise, the term “chair”, used in a catalogue of a furniture store, will not evoke the academic sense (“chair”, or also “university subject” in Portuguese) although this sense is reflected in the potential meaning of the word “chair” in Portuguese. The translation of the term found in the catalogue can, therefore, be made without the need to include this potential for evocation in the translation solution. Conversely, in the aforementioned academic context, that is, in a textual segment in which this is the consigned meaning (“Chair of Brazilian Literature”, for example), it will not be necessary to include in the translation solution the reference to the meaning of furniture. Only in cases where the text itself evokes the double meaning such confluence should be rescued, in one way or another. Thus, in English and French translations, a textual segment such as

allows for this recovery through the term “chair”, which in English can have, and, in French effectively has, the meaning of “cathedra”, university chair. In other languages, however, the recovery of the double meaning will have to be carried out in a more elaborate way, as in Norwegian:

in which the term stol in the professorstolen compound refers to another, more generic sense, which is shared between the two languages, that of “chair” as a “post”, “position” or “function”, which overlaps with, but does not exclude, the concrete meaning of “piece of furniture”. Taken alone, it does not update the concept of “university chair”, which, however, is implicit in the other component of the compound, professor, a designation reserved exclusively for full professors of higher education institutions (corresponding, in turn, in Brazilian Portuguese, to the old concept of “catedrático”, today “professor titular”). This solution, however, excludes the concept of “university subject”, which the term “cadeira” has in Portuguese.

In other cases, however, it must be admitted, the differentiation seems to occur independently of the specific update.

Take, for example, the closest equivalent in Norwegian to the sense of saudade (lengsel). In Norwegian, the adverb pair hjemme/hjem refers to “house” in the sense of “home” (including “paternal home”). The first - hjem-me - is used in a static sense (“at home”, as in hun er hjemme nå = she is at home now), while the second - hjem - implies movement (“from/to” home”, as in huninviterte meg hjemtilseg = she invited me to her home). It turns out that, in order to express the notion of <missing home> “Portuguese <tersaudadedecasa>”, in Norwegian the form hje³² is used, as in hun lengter hjem til foreldrene sine = she misses her parents’ house/home, which indicates that lengte (to miss/ter saudade in Portuguese) as well as the corresponding noun lengsel (the state of missing/saudade) should be interpreted as a movement verb/noun (something like <her emotion/her desire is projected towards her parents' house>), and not a state verb/noun, as in Portuguese. Thus, in any update, lengsel for saudade and lengte for saudade (the state of missing) will be, in a situation of differentiation/contrast (typically, in interlingual translation), perceived as bearing a specific cultural marker.

In any case, it is clear that the existence of the cultural marker is only revealed in the confrontation through differentiation; or, in other words, the notion of cultural marker refers to a distinctive element, that is, to something that differentiates a specific linguistically formulated expressive solution from another solution considered to be partially or totally equivalent.

Another issue concerns referentiality. In the approach adopted here -that of the cultural marker resulting from differentiation - referentiality appears in at least three different (and possibly complementary) aspects: (i) intralinguistic referentiality (in the sense of “linguistic culture”, apud Nida, 1945);(ii) intertextual referentiality; and (iii) extralinguistic referentiality.

An illustration of the intralinguistic referentiality - which well demonstrates the complexity of the matter - can be found in the typical texts of English legal language: herein, hereinafter, etc. In the Anglo-speaking community itself, these are marked terms as they are characteristic of legal discourse or similar to legal discourse. In other discursive typologies, their equivalents are in this ..., as from this point ..., etc.

Translated into Portuguese, the options neste (contrato/instrumento, etc.) and daqui para a frente (from now on) are unmarked constructions, that is, they are not exclusive to legal language. If used for the translation of the deitics herein and here in below, they establish a difference between the original and the translation, and, although the translated text is discursively unmarked within the reception community⁴, the translation as such is marked (thus configuring a modulation, see Aubert, 1998). The options no presente (contrato, instrumento, etc.) and doravante (henceforth) constitute marked constructions, more typical of legal discourse (and, possibly, bureaucratic discourse in general), and are much less frequent in other discursive situations. In this sense, they are intralingually marked as they help to distinguish the specific text from other textual typologies current in the Portuguese-speaking community, but they are translationally unmarked, as they rescue precisely the [+ legal] marker already present in the original.⁵

The intertextual dimension also represents a set of specific cultural markers but which, unlike the previous one, are not anchored in the lexico-grammatical structure; that is, it is not a phenomenon crystallized in the linguistic code but rather in the collection of sayings, modes of speaking, speeches that, for whatever reason, are included in the repertoire of the relevant socio-linguistic group. Widely known literary works, certain religious texts, titles, advertisements, popular songs, films, soap operas and television series, phrases attributed (although not always correctly) to recent or more remote historical characters, television personalities’ catchphrases, etc. are normally part of this repertoire, to which are added more localized repertoires, specific to certain subgroups (regional, institutional, familial). As a whole, just as idioms (ways of saying) signal to interlocutors that they belong to a specific socio-linguistic and cultural group, they constitute signs of mutual recognition, and, due to their particularizing effect, they can generate specific difficulties in the translation process. If, on the one hand, to be or not to be, or If they have no bread, let them eat cake are sufficiently internationalized (universal or universalizing repertoires), on the other hand, any references to Dona Flor e seus Dois Maridos (Dona Flor and her Two Husbands), X não é nenhuma Brastemp (X is no Brastemp), a saudosa maloca (the nostalgic poor shed), vamos logo aos finalmentes, (let’s quickly get to the end), cachorro também é gente (dogs are also people), pimba na gorduchinha (hit the target), etc., are much more specific (in this case, to the Brazilian language-culture complex).⁶ In any case, they only acquire their own fullness of meaning if perception of the origin of the intertext is guaranteed.

Extralinguistic referentiality,in turn, which concerns only the terms, words and expressions whose meaning designates a non-linguistic referent, can be broken down, as we have seen and following the model proposed by Nida (op. cit.), into the domains of ecology, material culture, social culture and ideological culture.

This breakdown, however, is not always very clear. Certain fruit (açaí, jabuticaba, etc.) do not only belong to the ecological domain but also to the eating habits of certain segments of the Brazilian population. Spatial organization (for example, the garden/backyard distinction) concerns both material culture (architectural/landscape aspect) and social culture (open public area - facing the street - private open area - at the back of the house). A vela de 7 dias (7-day candle) is both part of material culture and ideological culture. A pai-de-santo (a candomblé priest) and a Christian priest are part of both social and ideological culture. Only the updating co(n)text of the term will be able to determine which domain, at a given moment of a given discourse, can be regarded as dominant; however, the co(n)text itself may prove to be, deliberately or not, ambiguous.

From the perspective of translation, let us consider the following examples, linguistic expressions of specific institutional realities of Brazilian Portuguese:

In the first example, the maintenance of the acronym CNPJ, without other explanations (without explanatory affixing, translator's note or similar resource), typifies, in the descriptive-contrastive analysis, a case of loan. Seen from the point of view of the production space of the source text, therefore, there is no differentiation, which, in the hypothesis analyzed here, would exclude the identification of a cultural marker. However, seen from the perspective of the reception space of the meta text, the acronym in question (probably preceded by the name/corporate name of the entity and, most certainly, accompanied by a sequence of 14 digits, with a diagonal bar after the first 8 digits and a dash preceding the last two digits), will be perceived as “an official registration identifier of the legal entity in question”, typical of the extralinguistic reality of the source text, and which, in some reception spaces will very possibly be re-interpreted as corresponding to the entry in the Registry of Commerce of the legal entity in question (and not in the Federal Revenue, as is the case).

In the second example, the initial comparison suggests that the fundamental difference resides in the replacement of “N” (nacional) by the adjective Brazilian, and that this is an adequate way to ensure the perception that it is an institution of the cultural space of the source text, avoiding the risk of confusion with similar institutions of the cultural space of reception of the meta text. It thus constitutes an intentional difference - the “national” option would not be inconceivable - which amounts to a request for the reader of the translation to perceive the specificity of the extralinguistic reality referred to by the acronym INSS.

Note, however, that if the acronym were translated as National Institute of Social Security or Institut Nationale de Securité Sociale, and therefore with a more literalizing translational intention (transposition, in the English version, word-by-word, in the French version), the English-speaking or French-speaking reader would still notice the cultural difference. In the US, Social Security is a System, not an Institute; in France, Securite Socialé is institutionally categorized as a Direction (directorate) and it also comprises a large number of bodies, coordinated by a central agency, ACOSS - Agence Centrale des Organismes de Securité Sociale.⁷

The specific issue of identifying cultural markers shares the nature of the translational phenomenon in general: it is not an operation limited to the interlingual situation although the contrast that makes them noticeable is indeed typical of a translational situation. It also covers the perceptible variations on the intralingual plane. Thus, for example, in the domain of the Portuguese language, the option for the familiar pronoun tu instead of você can refer to various cultural markers, of a situational, stylistic or dialectal nature, but which are characterized as such only if factors such as the interlocutors themselves, their place in time and space, and the like are taken into account (or if they are relevant to the discursive and/or analysis situation in question).

The above consideration accepted, the cultural marker is not perceptible in the linguistic expression taken in isolation, nor is it confined within its original discursive universe. The cultural marker only becomes visible (and therefore updated) if this original discourse (a) incorporates within itself a differentiation, or (b) if it is placed in a situation that makes differentiation stand out. Returning to the previous illustration, the distinction tu/você produced in a South Brazilian Gaucho or Portuguese from Portugal discursive universe takes on a certain function within the linguistic subsystem under consideration (marking the degree of intimacy or hierarchical relationship between the interlocutors), but it will not be perceived as containing cultural markers; the distinction only becomes a cultural marker (in this case, dialectal) if one or more of the interlocutors/receivers identify themselves sociolinguistically as belonging to another linguistic subsystem (from São Paulo or Rio de Janeiro, for example), or if the text itself incorporates such different interlocutors (in which case the differentiation arises as the result of a metalinguistic effect).

Likewise, to return to the example of chair in the sense of cátedra, (university chair), in the translation situation involving Brazilian Portuguese as the source language and French or English as the target language, there will be no perceptible cultural marker, except when the way in which the individual obtained the chair is also relevant in the text/discourse in question (public competition of academic degrees and examinations vs. hiring, foreseeing the so-called tenure track in the US). In the translational relationship which has Norwegian as its target language, it is presented as bearing a cultural marker, thus requiring a translational procedure that cannot be classified as literal, formal or semantic.

The situation of differentiation (of contrast) constitutes, therefore, a sine qua non condition for the perception of the existence of the cultural marker. It could be argued that the very existence of a cultural marker depends fundamentally on such differentiation or contrast; that it would not be something pre-existing - inherent or immanent - but conditioned, and dependent, in order to exist, in each specific situation of differentiation and contrast. Even if we do not want to be so emphatic, the fact remains that the identification of the cultural marker is linked (a) to external linguistics, and (b) to the discursive situation. In this sense, it is not a phenomenon of language (langue), but of speech (parole), and, even on this level, not of any speech situation, but only of those that comprise an element of differentiation/contrast.⁸

These considerations point to a question which is important for research in the area of intercultural studies in general, and translation studies in particular. The perception of the cultural mark (marke of alterity) is produced and takes shape from the viewpoint of the observer: again, not only does it not pre-exist before this viewpoint (or, at least, it does not have its existence confirmable without this viewpoint); but also the (linguistic, intertextual or extralinguistic) reference is different in each different linguistic-cultural reception space, the perception of the difference, when it occurs, will also be different in each case.

Here, there is a supplementary problem and a possible limitation for cultural studies linked to translation. Indeed, it is necessary to ask whether the researcher can systematically identify all cultural markers. However aware the researcher is of cultural issues, and even if he or she is endowed with a reasonably advanced bilingualism and biculturalism, it is almost inevitable that, in the original/translational comparison, he or she will identify with greater clarity those cultural markers that stand out in the perspective of his or her own dominant culture.

Thus, the perception of these cultural markers will undoubtedly have been made from the perspective of differentiation, but, quite possibly, by marking the terms that appear to the researcher as differentiated because this researcher belongs to another linguistic sub-community (e.g., a researcher from São Paulo analyzing texts of Brazilian literature from a regionalist perspective, anchored in the realities of Minas Gerais (Guimarães Rosa) or Bahia (Jorge Amado). The procedure that can, at least, reduce the effects of loss in the observation will be, therefore, to carry out the comparison, and only then, in light of the results of the comparison, determine which terms in this translational situation appear as bearers of cultural marks. This recommendation is - but only in part - put into perspective by the fact that the researcher is quite possibly already aware of the potential cultural contrasts and may, even with a certain margin of risk, “anticipate” which terms, in the translation in question, may be presented as contrasting with the linguistic, cultural and discursive universe of the target text.

By way less of conclusion than a provisional synthesis, the thoughts contained in this essay suggest that a) the very identification of the presence of cultural markers requires, as a prerequisite, the comparison - in this case, the comparison between the original and the translation -, without which there will be no safe parameter for such identification;

b) in each presumed occurrence of a cultural mark, its confirmation and the delimitation of its extension will require identifying which components of the virtual sense (the semes) are actually employed in the co(n)text under consideration. Only those in which it is possible to verify and delimit the effective presence (updating) of a specific cultural mark will it be relevant to assess the level of its sayability in the target culture;

c) the extralinguistic dimensions, as presented in the classification proposed by Nida (op. cit.), contain an excess of overlaps and ambiguities, and it is worth asking whether their four categories: ecology, material culture, social culture and ideological culture could not, with some advantage, be reduced to two: the material dimension (ecology and material culture), and the socio-ideological dimension;

d) the dimension of linguistic culture is more complex than might be supposed at first sight, and involves, at least, two distinct facets: the structural (detectable cultural markers in the lexical, morphosyntactic and semantic structure), and the discursive (including the markers that are manifested in intertextuality and, more generally, in the collection of linguistic “uses and customs” of the linguistic-cultural community in question).