Utterances As Grammatical Tangles. Using Subject-Centered Sentence Models In Translation

Preliminary Notions

Our syntactic vision of translation includes deverbalisation, which consists in stripping down the text in the source language (SL) for meaning, and reverbalisation, or (re)translating the meaning into words and grammatical / stylistic structures in the target language (TL). The translator uses structural reference points, first, to understand the message and second, to build complete statements. What might these points be? We suggest starting from the subject of the sentence, to which would be added one by one, and this in a precise order, all the other terms or phrases.

We will therefore distinguish between three generic structure types, using Ogden and Richards’ semiotic terminology. To introduce this classification amounts to saying that we consider grammatical structures as signs:

the material structure . In their work, translators are confronted with material structures;

the conceptual structure , an abstract model since it belongs to the language and is studied by linguists;

the referential structure , dematerialised insofar as it refers to a logical relation. This type of structure is comparable to the deep structures put forward by Noam Chomsky (see also State of the Matter hereafter).

State of the Matter

Half a century ago, Lucien Tesnière formulated his flagship theory of the “little drama” represented by the phrase taken in its semantic and syntactic fashion. The verb would then express a process to which beings or objects designated by the nouns (actants, in Tesniere terminology, known as arguments in today's linguistics) would be subjected to define the circumstances of the process conveyed by the sentence (circumstants, or adjuncts). The number and quality of the links that the verb maintains with actants and circumstants is recorded in the linguistic system. Tesnière calls them “valencies” like chemical valency, a notion that we will be using in a broader sense, such as the ability of an element of a given sentence to associate itself with other specific elements.

Tesnière's legacy gave birth to the school of dependency grammar, distinguished by the importance it gives to the hierarchy of words (“stemma” in L. T.'s writings, “dependency tree” in later works). Dependency grammars, already popular at their onset, are now taken up by researchers working at the crossroads of linguistics and information and communication technologies.

However, if these approaches are effective for language learning, and therefore for the understanding, or deverbalisation of the message expressed in a foreign language, they seem less convenient for the reverse operation, reverbalisation, because they leave aside the ontogenesis of the sentence, that interests translators.

Another French linguist, Gustave Guillaume, spoke of ontogenesis in his numerous courses and essays (e. g. Guillaume, 2003). He gave a narrower meaning to this concept of “development of a linguistic object”, applying it to the semantic and grammatical conception of the word. According to him, the ontogenesis of the sentence reproduces the ontogenesis of the word. This leads him to examine the noun as the only part of speech that is “self-incident”, i.e. referring to itself, unlike the verbs and the adjectives that are only used to say something about the object or concept designated by the noun. At the phrasal and textual levels, self-incidence is represented by the subject of the sentence.

For Tesnière, the subject was nothing more than the “prime actant”; for Guillaume, it becomes the main focus. However, neither will be a reference for translation scholars, who have been emancipating themselves from linguistics since the 1960s, because “translation studies interest themselves in texts, while linguistic studies interest themselves in systems of signs” (Pergnier, 2004).

From a linguistic point of view, translation seems to be more easily understood by means of textual analysis tools, which are studying the type of text, its function or purpose, its cohesion and coherence (Ballard, 2013). The taxonomy of texts has been systematically theorized by Katharina Reiss (Reiss, 2014) who comes up with three main types - informative, expressive and operative, whose identification is based on the text's genre - but also hybrid forms and “multi-semiotic variants”. Reiss's Skopos theory is essentially semantic and situational. Other, yet more functionalist works of translational text analysis date back to J. R. Firth. This British linguist promotes the syntactic dimension of translation and adopts the genre framework. But it takes to work in fields other than pure translation to address the question of global syntax equivalences.

Today, a structural approach to the sentence has been adopted either in psycholinguistics or in the IT. While we endorse the transformational approach, which is an essential basis for research in Natural Language Processing (NLP) when it comes to the deep structures and rules of rewriting (Association pour le traitement automatique des langues http://atala.org/), one still has to describe in a concise and practical way the relationship between equivalent surface structures in both languages, explaining the possible choices of the translator.

This brings us to linguists of Guillaumian legacy, such as Annick Engelbert (Engelbert, 1998) or Pierre Larrivée (Larrivée, 2004) who endeavoured to “explain ... what are the explicit relationships between, on the one hand, the patterns of linguistic forms such as sentences and discourses and, on the other hand, the mental representations to which the speaking subjects refer when they express themselves through these patterns, in order to transmit them to their counterpart”, leading to a “computational architecture” (Desclés, 2008).

Existing Comparative Data

Let us consider some well-known asymmetries between French and Russian pertaining to the subject, or metamorphoses as we call them (Gavrilova & Frison, 2017). Listed below are normal Russian equivalents for syntactic particularities of French speech (cf. Grevisse, 2013). To simplify, all examples are taken from French newspapers and translated as if these were to be published as such in Russia.

• Personal structure → impersonal structure

• Active voice → passive voice

• Inanimate noun as an agent (personified, often together with the verb) → adverbial (usually an adjunct, temporal or locative)

• Personal pronoun in several successive sentences → repetition of noun, or, more rarely, replacing it with a synonym or paraphrasing it

• Noun preceded by an indefinite article as the subject at the beginning of the sentence → Noun as the subject at the end of the sentence

Being familiar with them, however, does not provide a comprehensive strategy for reverbalising whole sentences, let alone texts, as there are no such thing as absolute equivalency without a context.

All of the above relate to the dialectics between the logical and the grammatical subject. The final translation must be organised around the grammatical subject bearing in mind possible grammar, usage and stylistics issues. The statement’s semantic centre is then the logical subject, be it implicit or explicit.

Procedures

We will try to develop a generic dependency tree based on the lexical valency of noun types that can be used as sentence subjects, in the TL and SL. It will then be applied to specific contexts and translation situations in order to devise a reverbalisation strategy that would enable the transition from one structure to another, all this in an epistemological, communicative and probabilistic framework.

In didactic terms, we adhere to the positions and methods of the interpretive theory of translation for decoding the message (see the four-stage decomposition in Seleskovitch & Lederer, 2014), drawing inspiration from general linguistics (see Guillaume, 2003 for the incidence theory) and the comparative grammar of French and Russian (Roudet, 2016; Gak, 2014) for restitution of the ideas thus deverbalised.

Since translation studies are not an exact science (cf. Reiss, 2014), we will not be looking for absolute matches, but for probable equivalents. In this way, our research will meet the concerns of stylistics and discourse analysis: it is the choices of the message's author that translators interpret and reverbalise not only according to the meaning, but also taking into account the sometimes haphazard preferences of native speakers, commonly called “the genius of the language”.

Indeed, it is in and through the communication of native speakers that translation is verified. It is always meant to be used by them, hence the need to inscribe the translated text in an interdiscourse (Venuti 2013). The structural dimension of the message is determining for the arrangement of lexical elements. With this in mind, we will analyse the thematic progression of the texts studied, both authentic and transitional (see Material of the study below), and highlight regularities that could guide the translator.

To properly theorise our data, we must first validate it. Émile Benveniste, although a comparatist himself, warned against the presumption of “universal” language laws. Our reflection on the importance of syntactic equivalents will not be based on interlinguistic parallels, but rather on the respective structural potentialities of the two languages. We hope that this will help to bridge the epistemological gap between linguistics and translation.

Material of the Study

In the present paper, we will use a pair of two Indo-European languages, French, an analytic language and Russian, a synthetic language. This basic difference relates in particular to:

• the thematising function of the articles in French and the word order in Russian;

• the specificities of the verbal system, elaborate in French and reduced to logical times in Russian;

• the case system, absent from French but essential for the arrangement of the Russian sentence.

Specialised (technical, legal, economic, etc.) documents will be studied in both languages.

We will also look at a number of translations which, being transitional texts written by native speakers and bearing the imprint of syntactic oddities, are an enlightening source of information about the limitations experienced by translators.

In fact, a transitional text tends to be written in haste and without being given importance to the subtleties of grammar and expression. Even experienced translators may not think directly in the TL. Hence the search, conscious or unconscious, for an impossible parallelism, while translators should think of other possible equivalences instead.

Translating from French into Russian: the extent to which verbs depend on nouns in a collocation perspective

Let us examine the following example, a quote from a speech by President Macron on 16 January 2018:

L’alternative est claire et ouverte : elle passe par l’accueil dans des centres d’hébergement où la situation de chacun sera examinée .

1. Extract from “Transcription du discours du Président de la République auprès des forces mobilisées”, http://www.elysee.fr/declarations/article/transcription-du-discours-du-president-de-la-republique-aupres-des-forces-mobilisees/, 16.01.18, last access: 11.03.2018.

In order to deverbalise this statement, it is important to figure out whether we really understand the verb “passe”. It has lost its spatial meaning and generally its meaning of action verbs. The function of the verb consists in the logical relation of the subject “l’alternative” (A) to the object phrase “l’accueil…” (B) as A=B. Thus translators first separate the subject and its object and only then look for a verb which could express the relation between them. In this case, it is better to start with the subject’s referent as it is the topic of the statement; thus it is connected to the preceding text.

In his speech, President E. Macron suggests an alternative to the refugee camp in Calais. The implicit logical subject is not the State, but the refugees themselves. It is be impossible to say so judging only by the general meaning of the word “accueil” (“obrashchenie” - reporting - rather than “priëm” - receiving (them)).

But the logical subject may not coincide with the grammatical one. In the Russian text, the subject may remain unchanged, as its position at the beginning indicates the topic: “the alternative is evident / clear / easy to understand / well-known and accessible / simple…” The syntagmatic relationship with the preceding sentence could be made explicit: “The alternative which they could choose [a campsite]...” Note that if translators preferred to avoid borrowings, they hardly could succeed: it would be difficult to find a synonym for “alternative”. But trying to render this meaning with another part of the statement (e.g. “Vmesto togo” (= except…) would change the sentence’s syntax, which in its turn would prevent us from putting the characteristics of suggested actions (“claire et ouverte” = clear and open) in the comment of the first part.

The second part, the one that contains the troublesome predicate, is based on the same subject as the first one, although it is expressed by a pronoun. As this two-part structure is undoubtedly a figure of speech, it should be preserved in translation. But a more common structure in Russian consists in using a second pronoun (“eto” = this). This way, if the object is still a noun, translators can omit the verb: “*... eto obrashchenie v punkty / tsentry vremennogo soderzhania migrantov….” = This is reporting in points / centres for migrant provisional support)... The translator can choose a simple solution, where the object absorbs the verb, taking over its function: “*... obratit’sya v punkty / tsentry” (= report in points / centres).

Both solutions are acceptable, but they leave out a connotation. In fact, “passer par” not only infers something similar, but has also an instrumental meaning. In other words, work in temporary detention centres for people asking to be recognized as refugees is only a precondition for solving the problem of migrants, but not the solution itself. This semantic nuance can be explicited via:

• Lexical compensation, as in “...eto, prezhde vsego, obrashchenie….” (it is first reporting...). It is the simplest tactic, but it requireы adaptation within the limits of a rigid structure, resulting in something cumbersome.

• Ad hoc verb selection: “podrazumevat’ / predpolagat’…” (imply / suppose). In theory, it is the better solution as it allows structural symmetry in translation. In reality, it often takes too much time and, worse, one cannot be certain that an appropriate verb exists.

• Both lexical and grammatical means: replacing the structure and adding words expected by the valency of the main structure: “chtoby uzakonit’ svoë polozhenie, im sleduyet, prezhde vsego...“ (in order to legitimise their situation, they should first...). Here the additional meaning of “passer par” is expressed by the “chtoby” conjunction. All remaining parts are reconstituted on this basis. The disadvantage is that the translator might have to rewrite the whole sentence.

Let us suppose that the translator has quickly found an appropriate verb. Now this verb should be included in the sentence to replace “alternative”. To solve this problem, the translator should go back through the sequence subject → object → verb to the object “obrashchenie” or, if converted into a verb, “obratit’sya”. Our utterance can be thus represented as follows:

Al’ternativa = A + B:

Х podrazumevat’ | predpolagat’ ... obrashchenie |obratit’sya v C , gde … D

where Х is the element we are looking for, the concurring solutions, among which potential synonyms of our verb, are separated by a vertical bar, and the words in italics are to be declined or conjugated. The part of the sentence that we deduce from the others is in bigger characters.

It is easy to see that in Russian the nominative object expects a verb, third person singular, while a verbal object would be governed a verb too, but would probably shift to a subordinate clause (such as in *...podrazumevayet, chto…). But we also want to avoid Х = “ona”|“eta al’ternatva”|“eto”. The latter would cause inaccuracy, because the “eto” pronoun would not refer to the alternative, as shown above for the copula, but to the certainty of it. Therefore, no verb in third person singular can be used. At the same time, we have to leave behind accueil = “obrashchenie”.

Which form should be chosen for the verb? A solution can be the indeterminate personal form “podrazumevayetsya”, “predpolagayetsya”. Note that, thanks to the transition to such a form, the logical link, expressed by “passer par” disappears. On the one hand, such a phenomenon could be considered as being a side effect of our transformations; on the other, it corresponds to the general logics of the French and Russian syntaxes. In fact, whereas the French language makes explicit any logical conditionality with the help of conjunctions, parenthetical phrases, articles, and verbs of motion, personifications, or paraphrasing, the Russian language prefers to try and keep these meanings implicit.

Returning to the example (1), the inanimate verb would expect a relative clause. Whenever adding a subordinate clause, the translator might want to make sure that the original structure does not have to be simplified, by replacing another subordinate within the same complex sentence with an object or an adverbial.

In consequence, our indicative translation may look something like this:

*Альтернатива очевидна и проста: предполагается, что мигранты обратятся в специальные центры, где ситуация каждого из них будет тщательно рассмотрена.

Or, after some stylistic editing:

(1) (fr) Существует очевидная и простая альтернатива: предполагается, что мигранты обратятся в специальные центры, где ситуация каждого из них будет тщательно рассмотрена.

Looking for a subject : a method for translating from French into Russian

But is perhaps the subject-centered approach only relevant for translating from a language with determinants and a rich verbal system such as French? Let us consider example (2) in Russian, which had to be translated into French. It is an extract from President V.V. Putin’s speech. Let us try and see how our model applies to it.

(2) Из них [учёных мирового уровня] 11 иностранцев, это учёные с мировым именем, лидеры в своих отраслях, два российских, мы выбирали их по конкурсам, и пять наших соотечественников, которые в своё время уехали работать за границу, вернулись и работали здесь .

Extract from V.V. Putin, “Telemost s veduscimi rossiyskimi vuzami”, http://kremlin.ru/events/president/news/56712, 25.01.18, last access: 11.03.2018.

On the English parallel page of the Kremlin site, this extract was officially translated in the following way:

(en) Of them, 11 were foreign scientists of international standing, leaders in their industries, two Russian researchers, whom we selected on a competitive basis, and five our compatriots who once left to work abroad but returned and worked here.

http://en.kremlin.ru/events/president/news/56712, last access: 11.03.2018.

For our part, we would suggest the following translation into French:

(fr) Onze d’entre eux étaient des étrangers de renommée internationale, qui occupaient une place de premier plan dans leur domaine, deux étaient des chercheurs russes, sélectionnés sur concours, et les cinq derniers, des universitaires russes partis autrefois à l’étranger, qui étaient revenus pour travailler dans notre pays.

The very first issue that arises at deverbalisation time is: what is the actual structure of this statement, what is the main idea and the new author’s idea (comment) and what is already known and merely adorning this idea (topic)? The message is that there is a relatively high number of Russians among them.

Further, what is the logical subject of the original statement? Obviously, the scientists working in the so called “metagrant” system (see above in the speech text). The fact is President Putin distinguishes within this group Russian scientists, on whom he places special emphasis. It makes sense to organise the translation in such a way as to use the logical subject as a grammatical one. But there is also the problem of transforming the asyndeton in Russian, which prevails in the original utterance and forms a sequence of nouns: “inostrantsy - ychënye - lidery - rossiyskie [uchënye] - our sootchestvenniki” . The fact that the Russian language has cases does not help at all. What is more, deverbalisation may become delicate because of categorisation of the subjects: in fact, scientists (“uchënye”), who worked at one point a time abroad are also Russian citizens. However President Putin wanted to underline that they were faithful to their homeland. He therefore used the rather pleonastic term “sootechestevenniki” (= compatriots). This corresponds to the casual Russian usage of the possessive determinants ”nash, nashi” (=our), much wider then in French. We chose to avoid the pleonastic construction, contrary to the official translation into English which kept it in order to influence the reader, creating in them a feeling of unity.

At the reverbalisation stage, translators therefore have to replace the asyndeton, preferring subordinate clauses, and probably including more verbs in the sentence. Terms such as “rossiyskie uchënye” (= Russian scientists) and “nashi sootechestvenniki” (= our compatriots) are completed, first, with a phrase linked by asyndeton, and second, with a subordinate clause, and within the subordination with a coordinate clause. Numerous animate nouns, which cannot be changed into verbs, are likely to become appositives. As nouns and substantivals prevail in the original, they would turn into participial structures: “sélectionnés sur concours, partis autrefois à l’étranger”. Subordinate clauses were kept, but moved according to the syntactic usage of the French language in order to avoid monotony. Lastly, we added subordination at the beginning (“qui occupaient une place de premier plan dans leur domaine”), as soon as it became clear that the main part of the message would be organised, syntactically, around appositives.

The subject-centered model of translation seems even more useful for official texts. Here is a quote from a note by Russian Vice-Minister Arkady V. Dvorkovich published in the “Kommersant”:

Вопрос ослабления эмбарго в отношении стран, которые поддерживают антироссийские санкции, не рассматривается .

Extract from “Pravitel’stvo ne rassmatrivayet vozmozhnost’ snyatya prodembargo s Turtsii i stran ES”, Kommersant, 25.04.2017, https://www.kommersant.ru/doc/3282081, last access: 11.03.2018.

The grammatical subject “vopros” is typically something that is difficult to translate into French. Therefore, a translator should start from that, gradually, adding the remaining parts of the sentence.

A word-to-word translation (*La question de l’assouplissement de l’embargo frappant les pays qui soutiennent les sanctions antirusses n’est pas examinée) is doomed to failure. In French, a sentence starting with “question” nearly loses all meaning. It seems awkward and unnatural.

Meanwhile, the verb “ne rassmatrivayetsya” (=is not examined) depends on the grammatical subject. It plays a double auxiliary function: it bears negation as the comment and helps, along with the subject, to organise the whole sentence. While deverbalising, it is important to see that the verb forms a circular structure. Therefore the translator must understand with which noun the verb is associated. Grammatically speaking, the subject-verb part of the sentence is “vopros ne rassmatrivayetsya” (=the question is not addressed), but the local subject is “oslablenie embargo” (=reducing the embargo). Using a collocation approach the translator chooses an equivalent for “ne pas être envisagé” (=not planned).

As for reverbalisation, translators must once again add verbs or verbals. Namely, the logical subject will most probably be changed into a verb. And we added yet another verbal “frappant” (=striking).

But who is the acting subject? He/she is not named explicitly. Depending on whether the translator prefers to imply or refer directly to they, two translations are possible, an implicit one:

(fr-1) Il n’a pas été envisagé d’assouplir l’embargo frappant les pays qui soutiennent les sanctions antirusses.

or a personal one:

(fr-2) Les autorités n’envisagent pas d’assouplir l’embargo frappant les pays qui soutiennent les sanctions antirusses.