1. Introduction Equivalence and equivalent effect

2. Roman Jakobson: the nature of linguistic meaning and equivalence

In Chapter 1 we saw how, in his paper ‘On linguistic aspects of translation’ (1959/2012), structuralist Roman Jakobson describes three kinds of translation: intralingual, interlingual and intersemiotic, with interlingual referring to translation between two different written sign systems. Jakobson goes on to examine key issues of this type of translation, notably linguistic meaning and equivalence.

Jakobson follows the theory of language proposed by the famous Swiss linguist Saussure (1857–1913). Saussure distinguished between the linguistic system (langue) and specific individual utterances (parole). Central to his theory of langue, he differentiated between the ‘signifier’ (the spoken and written signal) and the ‘signified’ (the concept), which together create the linguistic ‘sign’. Thus, in English the word cheese is the acoustic signifier which ‘denotes’ the concept ‘food made of pressed curds’ (the signified). Crucially, the sign is arbitrary or unmotivated (Saussure 1916/1983: 67–9). Instead of cheese, the signifier could easily have been bread, soup, thingummyjigor any other word. Jakobson also stresses that it is possible to understand what is signified by a word even if we have never seen or experienced the concept or thing in real life. Examples he gives are ambrosia and nectar, words which modern readers will have read in Greek myths even if they have never come across the substances in real life; this contrasts with cheese, which they almost certainly have encountered first-hand in some form.

Jakobson then moves on to consider the thorny problem of equivalence in meaning between words in different languages, part of Saussure’s parole. He points out (1959/2012: 127) that ‘there is ordinarily no full equivalence between code-units’. Thus, the Russian syr is not identical to the English cheese (or, for that matter, the Spanish queso, the German Käse, the Korean chijeu, etc.) since the Russian ‘code-unit’ does not include the concept of soft white curd cheese known in English as cottage cheese. In Russian, that would be tvarog and not syr. This general principle of interlinguistic difference between terms and semantic fields importantly also has to do with a basic issue of language and translation. On the one hand, linguistic universalism considers that, although languages may differ in the way they convey meaning and in the surface realizations of that meaning, there is a (more or less) shared way of thinking and experiencing the world. On the one hand, linguistic relativity or determinism in its strongest form claims that differences in languages shape different conceptualizations of the world. This is the famous Sapir-Whorf hypothesis that had its roots in the behaviourism of the 1920s and in the anthropological study of the native American Hopi language, which, according to Whorf (1956), had no words or grammatical categories to indicate time. Another claim that is often made is that Eskimos have more words for snow because they perceive or conceive of it differently. This claim, and indeed linguistic determinism itself, is firmly rejected, amongst others, by Pinker (1994: 57–65; 2007: 124–51), who points out that the vocabulary of a language simply reflects what speakers need for everyday life. The absence of a word in a language does not mean that a concept cannot be perceived – someone from a hot climate can be shown slush and snow and can notice the difference.

Full linguistic relativity would mean that translation was impossible, but of course translation does occur in all sorts of different contexts and language pairs. In Jakobson’s description (ibid.), interlingual translation involves ‘substitut[ing] messages in one language not for separate code-units but for entire messages in some other language’. Thus, a translation of cottage cheese would not be the TT unit for cottage plus the unit for cheese; the message cottage cheese would be considered and translated as a whole. For the message to be ‘equivalent’ in ST and TT, the code-units will necessarily be different since they belong to two different sign systems (languages) which partition reality differently (the cheese/syr example above). In Jakobson’s discussion, the problem of meaning and equivalence focuses on differences in the structure and terminology of languages rather than on any inability of one language to render a message that has been written or uttered in another verbal language. Thus, Russian can still express the full semantic meaning of cheese even if it breaks it down into two separate concepts.¹ The question of translatability then becomes one of degree and adequacy (see Hermans 1999: 301).

For Jakobson (ibid.: 129), cross-linguistic differences, which underlie the concept of equivalence, centre around obligatory grammatical and lexical forms: ‘Languages differ essentially in what they must convey and not in what they may convey’. Examples of differences are easy to find. They occur at:

  the level of gender: e.g. house is feminine in Romance languages, neuter in German and English; honey is masculine in French, German and Italian, feminine in Spanish, neuter in English, etc.;

  the level of aspect: in Russian, the verb morphology varies according to whether the action has been completed or not;

  the level of semantic fields, such as kinship terms: e.g. the German Geschwister is normally explicated in English as brothers and sisters, since siblings is rather formal. Similarly, in Chinese it would be 兄弟姐妹 (‘xio¯ng dì jieˇ mèi’, literally meaning ‘elder brother, younger brother, elder sister, younger sister’).

Even what for many languages is a basic relational concept such as to be (English), être (French) and sein (German) is broken down in Spanish to ser and estar, while Arabic, Russian and many others do not use such a verb explicitly in the present tense. These examples illustrate differences between languages, but they are still concepts that can be rendered interlingually. As Jakobson (ibid.) puts it, ‘[a]ll is conveyable in any existing language’. For him, only poetry, with its unity of form and sense and where ‘phonemic similarity is sensed as semantic relationship’, is considered ‘untranslatable’ and requires ‘creative transposition’ (ibid.: 131).