Paris 2011

Session 1 - Problems and Solutions on Baltic Shores / Problématiques et exemples sur les rivages de la Baltique

Bookmark and Share
Version imprimable
Léa Huotari

Few data, few corpora, the case of French-Finnish language pair

Voir la video


This paper presents an on-going study on the change of the syntactic subject from inanimate in the source text to animate in the target text with two typologically different languages: Finnish and French. This syntactic change is called humanization of the subject in translation and it is proposed as a candidate to the so-called and criticized translation universal hypothesis. Firstly, the article presents the pilot study designed to empirically test the possible humanization of the subject in translation proposed by Chevalier (1995). It has been conducted on a corpus of translated French and Finnish taken from contemporary fiction and contains 531 clauses and 809 subjects in ST. The corpus study shows a few patterns in the translation of the syntactic subject that appear in four contexts: animation involving a perception that can be auditory or visual, animation in free indirect discourse, animation with an acting inanimate subject, and, lastly, animation involving a psychological or physical human characteristic. These preliminary results will be discussed in terms of cognitive linguistics. In the conclusion, the problems arose by the research findings will be examined and it will be argued that a new bi-directional corpus of translated texts from Finnish-French-Finnish needs to be built. This will enable a short discussion concerning representativeness and the type of data to collect.

Full text/Texte intégral


1This paper presents a short overview of an ongoing doctoral dissertation project concerning the change of the subject from inanimate in the source text to animate in the target text, as in Chevalier [Chevalier, J. C. (1995: 43)]:

(1) ST: Lee’s mouth was full and benevolent.
TT: “Il avait une bouche large et bonne”
[Literally: He had a big and pleasant/good mouth.].

2This syntactic change is called humanization of the subject in translation and it is proposed as a candidate to the so-called and criticized translation universal hypothesis. Firstly, the pilot study, its methodology and findings will be shortly presented. The preliminary results of the pilot study will be discussed in terms of cognitive linguistics. In the conclusion, the problems arose by the research findings will be examined and it will be argued that a new bi-directional corpus of translated texts from Finnish-French-Finnish needs to be built. This will enable a short discussion concerning representativeness and the type of data to collect.

3The hypothesis of our on-going PhD research takes root in corpus-based research in the paradigm of descriptive translation studies and comes within the hypothesis that there would be linguistic features, which typically occur in translated texts [Toury, G. (1980), (1995) and Baker, M. (1993), (1996)]. Baker [Baker, M. (1996)] proposes four features of translation referred to as translation universals for investigation: Simplification, Explicitation, Normalization or Conservatism and Levelling out. As it will be showed further, the humanization of the subject in translation within the language pair French and Finnish is studied in terms of a special case of simplification and normalisation in translation.  

4It has to be said that studies on translation universals haven’t been conclusive and their study has even been considered vain by some scholars [cf. House, J. (2008) and Tymoczko, M. (2005)]. Secondly, it is very likely that, as House [House J. (2008: 6)] notes, “[t]he quest for specific translation universals is futile for several reasons, among them the undeniable fact that since translation is an operation on language, general linguistic universals also apply to translation”. This is very probably the case; explicitation and implicitation, for instance, are not specific to translation but also appear in non-translated communication. However, as Chesterman [Chesterman A. (2004a: 43)] notes,

[...] any claim about a translation universal can really only be an approximation. But this does not matter, as long as scholars are aware of what they are claiming. After all, what these corpus scholars are basically doing is seeking for generalizations. [...] Any level of generalization can increase understanding.

5And further,

[w]hat ultimately matters is perhaps not the universal, which we can never finally confirm anyway, but, new knowledge of the patterns, and patterns of patterns, which helps us to make sense of what we are looking at. [Chesterman, A. (2004b: 11)]

6In other words, generalization increases understanding of translation and its patterns, and therefore it is important to my opinion to continue to carry on tests on translation patterns. Furthermore, even though many researchers now state that the explanation for translation universals is likely to be cognitive, very few corpus studies have actually tried to explain them from a cognitive perspective.

7Increase of prototypical human subject as a so-called translation universal

8I propose that the subject slot in translations will be more frequently filled by prototypical human subjects than in their source texts. I also suggest that prototypical human subjects will be more frequent in translated texts than in non-translated texts in the target language.

9My research hypothesis is that if translation universals at all exists, translated texts would contain more prototypical animate or human subject than non translated text as for instance from my corpus in (2) where the inanimate “les doigts de Luc” of the source text (ST) becomes animate in its Finnish target text (TT) “Hän” (he):

(2) ST: En lissant ses cheveux, dans un geste d’adieu désolé, les doigts de Luc ont rencontré quelque chose de bizarre. (E. Carrère, L’Adversaire, P.O.L éditeur, coll. Folio, Paris.) [Literally: While smoothing down her hair in a distressed farewell gesture Luc’s fingers came across something weird.]

TT: Kun Luc lohduttomana silitti Florencen hiuksia hyvästiksi, hän tunsi äkkiä sormissaan jotain outoa. (Transl. M. Haapio, Valhe, Like Kustannus, Keuruu) [Literally: While distressed Luc smoothed down Florence’s hair as a sign of farewell, he felt suddenly in his fingers something weird.]

10And in (3), where “un juron” becomes “joku kiroaa” (someone curses):

(3) ST: “De l’autre bout de la gare parvient un esclaffement sourd, puis le crissement d’un éclat de verre sous un pied, un juron.” (Makine)[Literally: From the other side of the railway station, a guffaw is coming, then a screech of a fragment of glass under a foot, a curse].

TT: “Hallin toisesta päästä kuuluu hohotusta, lasinsiru rasahtaa saappaan alla, joku kiroaa” [Literally: From the other side of the hall, a guffaw is being heard, a fragment of glass rustles under a boot, someone curses].

11An increase in the frequency of the animate subject makes a good candidate for the so-called translation universal hypothesis and more specifically for the simplification and normalization in translation hypothesis because the animate human referent is shorter in length than the inanimate one of the source text and it is also more prototypical than the inanimate one. Thus, it seems cognitively more salient than the inanimate one. As Johansson [Johansson, S. (2004: 32)] notes it:

12There are good reasons for focusing on the subject in translation. […] According to Halliday (1994), the subject is not just a grammatical function, but has a semantic basis: it functions as a “resting point” of the argument; together with the finite verb, it forms “the nub of the proposition” (p.77). The subject typically expresses central participant roles, such as agent and experiencer, although there are also dummy subjects and subjects expressing other participant roles. It is commonly realised by a noun phrase which is also the theme of sentences, i.e. to speak with Halliday (p. 38), “the starting-point of the message” or “the ground from which the sentence is taking off”. The theme, i.e. frequently the subject, is central in the development of texts.

13Because the subject is essential in the building of both sentences and texts, we might expect translators to preserve the subject of the source text, unless there are good reasons to do otherwise.

14Furthermore, as Laury [Laury, R. (2006: 153)] notes, “a wide range of scholars have made the observation, robustly supported by empirical evidence, that human referents tend to manifest features prominence on the level of both discourse and grammar. Namely, human referents [i.e. animate subjects], are likely to be topical and agentive (Kuno 1976, Dixon 1979, Comrie 1978, Silverstein 1976, 1981) and they are consequently likely to appear in core grammatical roles, especially as subjects (Du Bois 1987, Asby and Bentivoglio 1993, Thompson 1997, Nakayama and Ichihashi-Nakayama 1994, Kärkkäinen 1996, Helasvuo 1997, 2001)”. See also Bock (1986); Chafe (1976, 1994), Tomlin, (1995, 1997) cited by Parrill [Parrill, F. (2008)].

15Unlike most of the studies on so-called translation universals, this study is not conducted on a comparable corpus, which consists of separate collections of texts in the same language but on a bi-directional parallel corpus, which consists of translated text in French and Finnish aligned to their source texts in Finnish and French. I am doing so because to my opinion it is essential to look at the role of source texts forms and structures in the occurrence of so-called translation universals. As Bernardini [Bernardini S. (2011: 12)] has noted it:

MCC [Monolingual Comparable Corpora]  analyses highlight differences, and as such are ideal tools for making strong hypothesis about norms operationg in a given target culture. But they have little explanatory power and rely too much on the problematic notion of textual comparability. The parallel corpus approach is still indispensable to confirm that observed differences are indeed due to the translation process […].

16In the case of the subject, even if like just mentioned above many studies have shown that animacy and discourse relevance play a central role in determining which subject will be selected, Tomlin [Tomlin, R. (1997)] and Parill [Parill, F. (2008)] among others have also shown that during utterance formulation the entity on which speaker’s attention is focused while she is planning her utterance shapes what she says. Thus, I would expect the source text to have the same effect on the choice of the subject in the target language and therefore it is important to work on a bi-directional parallel corpus of French-Finnish-French texts and their translation.

Testing the humanization of the subject in French-Finnish-French translations

17My purpose is to find out whether prototypical animate subjects are more frequent in Finnish and French translations than in their French-Finnish source texts. Looking at translations between a language pair from different typological groups, like French and Finnish, presents the advantages of revealing more clearly interference in a field largely dominated by indo-European languages as pointed out by Mauranen [Mauranen, A. (2000)]. For instance, in our case, Finnish is a pro-drop language and the pronoun especially in first and second person is very commonly left out especially in written standard language [Vilkuna, M. (2000: 131)]. In French on the contrary, the pronoun is obligatory (unless it is an imperative). Thus, a higher usage of pronouns of the first and second person in Finnish translations would constitute a feature of the law of interference proposed by Toury [Toury, G. (1995)] or of the unique item hypothesis put forward by Tirkkonen-Condit [Tirkkonen-Condit, S. (2004)].  

Presentation of the corpus and preliminary results

18The pilot corpus consists of three original French novels (works by Carrère, Houellebecq and Makine) and their translations into Finnish and three Finnish novels (works by Lander, Paasilinna and Tervo) and their translations into French. They have all been published between 1991 and 2002. The syntactic subjects of the corpus were collected from three equal sections in each novel: themore or less first 30 clauses from the very beginning of each novel, 30 clauses from the middle, and 30 from the beginning of the last chapter. At the moment, the corpus contains 531 sentences in STs with a total of 809 subjects in STs (Finnish and French) and 835 subjects in TTs. Unlike, Johansson’s study on subject change in translation from English into Norwegian (2004), which included only the subject of the first main clause of each sentence, I have included also the subjects of subordinate clauses because subject changes appear also there.

Agrandir Image1

Figure 1. Percentage of animate “subjects” in French (represented in blue) and in Finnish (represented in red) in both STs and in TTs

19As Figure 1 shows, the total percentage of animate “subjects” is increases for both languages in TTs (lighter colours). I use quotation marks for the term subjects because, as I will explain below, the figure does not take into account the mere syntactic subjects for Finnish, but also the implicit subjects in the sentences lacking the syntactic subjects. The two columns on the left represent the percentage of animate “subjects” in Finnish STs (dark red) compared to French TTs (light blue), whereas the two columns on the right represent the percentage of animate subjects in French STs (dark blue) compared to the translated texts in Finnish (light red). From the 65.5% of animate “subjects” in Finnish STs we get 69.5% of animate subjects in French TTs, and for 62.1% of animate subjects in French STs there are 68.6% of animate “subjects” in Finnish TTs.

20Figure 1 shows all animated “subjects” appearing in the corpus regardless of what might be the motivation for their change. In other words, I have included also those animations that have been mandatory in the TT due to morphosyntactic differences between the two languages. For the Finnish “subjects”, as I mentioned above, I have included both explicit and implicit ones, because I wanted to take into account also those sentences that lack the syntactic subject. That is to say, I regard the implicit subject of the sentence as the actual “subject” of the sentence, which is comparable to the equivalent French sentence. I shall mention here that in the definitive version of the thesis, the focus will be first on the syntactic subjects and thus won’t include the zero person of Finnish and the lack of personal pronoun. As mentioned above, when languages like Finnish and French that are so different from a morphosyntactic point of view are compared to each other, we have to make a very detailed contrastive analysis to determine the nature of the change in translation and to include some typological adjustments in order to rule out the mandatory humanizations and keep only optional ones. In my corpus, I have found 33 “subjects” – out of the 835 translated “subjects” that could have been translated by using an inanimate “subject” in TT without any deviation from the “subject” of ST – which were changed to animate ones in their translation. The figure corresponds to 3.9% of the cases where the “subject” changed in TT from inanimate to animate. These animations of the “subjects” were higher in Finnish with 20 changes from an inanimate subject in ST to an animate one in TT, whereas there were 13 animations of the subject in French TT.

Interpretation of the results

21So far, I have been able to extract four contexts that seem to favour the humanization of the “subject”. These are: clauses involving a perception that can be auditory or visual as in (3); free indirect discourse; clauses with an acting inanimate subject as in (2); and, lastly, clauses involving a psychological or physical human characteristic. For time restrictions, I provide here examples for only three type of humanization of the syntactic subject. The third example involves a psychological or physical human characteristic. This type is by far the most systematic in my corpus, with 11 occurrences out of the 33 humanizations that were found. This corresponds to 33.3% of the humanization cases:

(4) ST: “Ranskattaret lihoivat hyvää vauhtia ja heidän hipiänsä alkoi loistaa.” (Paasilinna 1992).

[Literally: French women had put on weight quite fast and their skin began to shine.]

TT : “Les Françaises engraissaient à belle allure et commençaient à avoir bonne mine. ” (Paasilinna 1998)

[Literally: French women had put on weight quite fast and began to look well.]

22The disappearing subject of the source text is again an inanimate part “heidän hipiänsä”, which means skin that is especially nice and smooth. The French translation prefers the animate whole, “French women”.

23From a cognitive perspective, it is interesting to note that Langacker [Langacker R. (1998: 193)] states that:

(i) a whole is generally more salient than the individual parts; (ii) discrete physical objects are generally more salient than abstract entities; and (iii) humans and (to a lesser extent) animals are generally more salient than inanimate objects (other things being equal)

24The cognitively most salient entity of this hierarchy seems to play a role in the translation of the subject in French and Finnish and vice versa. This is not so surprising if we think of metonymy as a major source of prototype effect [Rosh, E (1978); Lakoff, G (1987)]. What is interesting is that in the cases found in my pilot corpus, translators seem to reconstitute the whole and thus operate a reverse metonymy.

25To be able to test for the impact of this particular prototype effect in translation we need to rule out the fact that it would be bound by the genre of literary texts. Metonymy is in fact a rhetorical or stylistic device that is largely used in literary texts.

[…] metonymy is best known for its place in literature such as in Balzac’ novels. David Lodge (1981) has argued that it is possible to map the literary history of the 20th century based on the movement back and forth between literature that is essentially metaphoric (e.g. James Joyce and T.S. Eliot) and literature that is basically metonymic (e.g. W.H. Auden and George Orwell). [Gibbs Jr R. (1999)]

26We could then speculate that the deletion of metonymy by the shift in translation to the more salient part (the whole) could be due to normalization that occurs in translation and has been explained by invisibility [Venuti, L (1995)] and risk avoidance [Pym, A. (2008)] that characterises translations.

27In order to make sure that metonymy is due to prototype effect and not to normalization in the particular genre of literary texts, it is crucial to enlarge the corpus to other genres.


28This article illustrates the importance of theoretical research and empirical investigations prior to final, larger scale corpus design [Biber, D. (1993)]. This is especially true when there are no pre-existing corpora to work on because building a large-scale corpus requires a large amount of work.

29The next corpus will have to contain a bi-directional parallel corpus of French and Finnish source texts from various genres from popular science to art and press articles.

30Additionally, in order to have a deeper image of the humanization of the subject in French-Finnish-French translation the analysis of the bidirectional corpus should be combined by an analysis on a comparable corpus. “While in the first instance the analysis focuses on the features specific to translated texts as a result of the translation process, in the second instance the analysis focuses on the deviation of translated texts from the reference norm” [Zanettin, F. (2011: 22)]. For practical reasons, the comparable corpus will consist of the original texts in French and in Finnish used in the bi-directionnal corpus. This means that when compiling the corpus, the different texts in Finnish will have to be as close as possible to the French ones in terms of genre, publication (journal, book, webpage…), process (editing or not), language of the translator (native language or not). This is a challenge between two languages that don’t share the same status.

31Even if working with a poorly resourced language pair like Finnish and French can be challenging as far as comparability of the texts included is concerned, there are not only disadvantages. The advantage is that the number of translators is rather small and thus its community can be more easily reached. The other advantage is that the corpus can be quite representative in terms of number, because it requires proportionally a much smaller dataset than building a corpus for two languages that are largely resourced like French and English where the total number of published translations is often even difficult to trace. When building up a translational corpus, one needs also to reflect on the proper definition of translation. Do we only take professional translations? Published one or also non professional available on the Internet for instance? Do we only incorporate translations made by natives when in practice very often times translations from Finnish into French are often done by Finns?

32Lastly, when dealing with cognitive translation studies, one needs to be aware of the necessity to combine different research methods. Halverson [Halverson, S. (2010)] suggests that cognitive linguistics and bilingualism studies should be combined. She also stresses that [op. cit.: 363]:

”Given the offline nature of corpus data, they are not suited to verify hypotheses concerning the psychological let alone the neurological processes underlying language use” (Tummers et al. (2005: 233))

33In other words, if cognitive linguistic theory is to be adequately tested empirically on translation, then corpus data alone is not sufficient. Halverson continues [op. cit.: 363]:

“The motivation for combining corpus and experimental data in this investigation is the recognition that corpus data does not speak directly to issues of cognition. Corpus data must be validated through experimental data”

34I would argue that corpus data can also be combined with more process oriented metadata. For instance, if the corpus concerns recent translational data set then the translators of the texts selected in the corpus could be contacted and asked to fill up a questionnaire concerning the translation process. They could be asked questions like: is the language you translated this particular text your mother tongue? Has the translation been edited by a native speaker? How much time did you have to translate? etc. This additional information would make possible to generate new hypothesis that would get us closer to the hard to find causal explanation to translation.

35Finally and by way of conclusion, even if we manage to find patterns in the humanization of the subject, further studies in other languages will be needed in order to be able to really determine whether translations have a tendency to animate the subject and is affected by prototype effect. Indeed, Finnish would often prefer an impersonal or intransitive expression rather than an animate subject and a transitive verb. The humanization of the subject in Finnish translation from French could therefore be due to the influence of French. More collaboration in corpus design and in testing different language pairs is needed when testing claims about the so-called translation universals.


Research material

Carrère, Emmanuel (2002) [2000], L’Adversaire, P.O.L éditeur, coll. Folio, Paris.

Carrère, Emmanuel (2000). Valhe, transl. by Marja Haapio, Like Kustannus, Keuruu.

Houellebecq, Michel (1998). Les particules élémentaires, Flammarion, Paris

Houellebecq, Michel (2000). Alkeishiukkaset, transl. by Ville Keynäs, WSOY, Juva.

Lander, Leena (1997). Iloisen kotiinpaluun asuinsijat, WSOY, Juva.

Lander, Leena (2000). Les rives du retour, transl. by Anne Colin du Terrail. Arles: Actes Sud.

Makine, Andrei (2002) [2001]. La Musique d’une vie. Paris: Points Seuil.

Makine, Andrei (2002). Elämän musiikki, transl. by Annikki Suni. WSOY, Juva.

Paasilinna, Arto (1992) [1991]. Elämä lyhyt, Rytkönen pitkä, WSOY, Juva.

Paasilinna, Arto (1998). La Cavale du Géomètre, transl. by Antoine Chalvin, Denoël, Paris.

Tervo, Jari (1999). Pyhiesi yhteyteen, WSOY, Juva.

Tervo, Jari (2002). Bienvenue à Rovaniemi, transl. by Paula and Christian Nabais, Denoël, Paris.

Works cited

Baker, Mona (1993), “Corpus Linguistics and Translation Studies: Implications and applications” in Text and technology: In honor of John Sinclair, John Benjamins, Amsterdam, pp. 233–250.

Baker, Mona. (1996), “Corpus-based translation studies: The challenges that lie ahead” in Terminology, LSP and translation. Studies in language engineering in honour of Juan C. Sager, John Benjamins, Amsterdam, pp. 175–186.

Bernardini, Silvia (2011), “Monolingual comparable corpora and parallel corpora in the search for features of translated language” in SYNAPS – A Journal of Professional Communication 26/2011, pp 2-13

Biber, Douglas (1993), “Representativeness in Corpus Design” in Literary and Linguistic Computing, Vol. 8, No. 4, pp. 243-257

Bock, J. Kathryn (1986), “Syntactic persistence in language production” in Cognitive Psychology 18, pp. 355–387.

Chafe, Wallace (1976), “Givenness, contrastiveness, definiteness, subjects, topics, and point of view” in Subject and Topic, Academic Press, New York, pp. 25–55.

Chesterman, Andrew (2004a), “Hypotheses about translation universals” in Claims, Changes and Challenges in Translation Studies . John Benjamins, Amsterdam, pp. 1-13

Chesterman, Andrew (2004b), “Beyond the particular” in Translation Universal Do they exist?”, John Benjamins, Amsterdam, pp. 33-49.

Chevalier, Jean-Claude (1995), “D’une figure de traduction : le changement de “sujet” in L’Horlogerie de Saint Jérome, L’Harmattan, Paris, pp. 27–44.

Cruse, D. Alan (1977), “The pragmatics of lexical specificity”in Journal of. Linguistics 13, pp. 153-64.

Gibbs, Jr Raymond W (1999), “Speaking and thinking with metonymy” in Metonymy in Language and Thought, John Benjamins, Amsterdam, pp. 61-76.

Halverson, Sandra L. (2010), “Cognitive translation studies: developments in theory and method” in Translation and Cognition, John Benjamins, Amsterdam, pp. 349-369.

House, Juliane (2008), “Beyond intervention: Universals in translation?” Trans-kom 1 (1), pp. 6–19.

Johansson, Stig (1998), “On the role of corpora in cross-linguistic research” in Corpora and Cross-linguistic Research, Rodopi, Amsterdam, Atlanta, pp. 3-24

Lakoff, George (1987), Women, Fire, and Dangerous Things: What Categories Reveal About the Mind, The University of Chicago Press, Chicago, London.

Langacker, Ronald W. (2002), Concept, Image, and Symbol: The Cognitive Basis of Grammar. Mouton De Gruyter, Berlin.

Laury, Ritva (2006), “Oblique mentions of human referents in Finnish conversation” in Grammar from the Human Perspective. Case, space and person in Finnish, John Benjamins, Amsterdam, pp. 153-171.

Mauranen, Anna (2000), “Strange strings in translated language: A study on corpora” in Intercultural Faultlines. Research Models in Translation Studies 1: Textual and Cognitive Aspects, St. Jerome, Manchester, pp. 119–141.

Parrill, Fey (2008), “Subjects in the hands of speakers: An experimental study of syntactic subject and speech-gesture integration” in Cognitive Linguistics 19 (2), pp. 283-299.

Pym, Anthony (2008), “On Toury’s laws of how translators translate” in Beyond Descriptive Translation Studies, John Benjamins, Amsterdam, pp. 311-328.

Rosch, Eleanor (1975), “Cognitive Reference Points” in Cognitive Psychology 7, pp. 532–47.

Tomlin, Russel S. (1997), “Mapping conceptual representations into linguistic representations: The role of attention in grammar” in Language and conceptualization, Cambridge University Press, Cambridge, pp.162–189.

Tirkkonen-Condit, Sonja (2004), “Unique items over- or under-represented in translated language?” in Translation Universals: Do they exist?, John Benjamins, Amsterdam, pp. 177–184.

Toury, Gideon (1980), In search of a theory of Translation, Porter Institute for Poetics and Semiotics, Tel Aviv.

Toury, Gideon (1995), Descriptive translation studies and beyond, John Benjamins, Amsterdam.

Tymoczko, Maria (1998), “Computerized corpora and the future of translation studies” in Meta 43 (4), pp. 652–660.

Tymoczko, Maria (2005), “Trajectories of research in translation studies” in Meta 50 (4), pp. 1082–1097.

Venuti, Lawrence (1995), The Translator’s Invisibility. A history of translation, Routledge, London & New-York.

Vilkuna, Maria (2000), Suomen Lauseopin Perusteet, Kotimaisten kielten tutkimuskeskus, Edita, Helsinki.

Zanettin, Federico (2011), “Monolingual comparable corpora and parallel corpora in the search for features of translated language” in SYNAPS – A Journal of Professional Communication 26/2011, pp 14-23.

To cite this document/Pour citer ce document

Léa Huotari , «Few data, few corpora, the case of French-Finnish language pair», Tralogy [En ligne], Tralogy II, Session 1 - Problems and Solutions on Baltic Shores / Problématiques et exemples sur les rivages de la Baltique, mis à jour le : 21/05/2014,URL :

Quelques mots à propos de :  Léa Huotari

University of Helsinki