A rhetorical approach

Owing to their dual focus on production and processing of discourse, the three linguistic structures studied in this thesis – text structure, discourse structure and information structure – are closely related to three of the five canons in classical rhetoric, namely inventio, elocutio and dispositio (the other two being memoria e actio) (Lo Cascio, 1991, p. 236). I do not claim that there is a one-to-one relationship between the three linguistic structures and the three canons; on the contrary, I assume that the three canons are integrated parts of all three structures simultaneously. As a result, linguistic structures can also be approached from a rhetorical point of view, rhetoric being defined as the study of persuasive and convincing communication.

In a different but comparable view to the cognitive-functional one, Renkema (2004, p. 145ff) relates the concept of register to that of style. Thus, the notion of register as understood from a stylistic viewpoint is that variation in style can be explained by taking into account the situational factors or constraints in the SPEAKING model (Hymes, 1972) of ‘setting’,

‘participants’ and ‘norms’. In this way, Korzen and colleagues (Korzen & Gylling, 2012b;

Korzen & Lundquist, 2003; Korzen, 2003) talk about a correlation between formality and linguistic structures: in a formal or formalised setting such as the one of the texts employed in this study, the participants are placed in a hierarchical role structure, and a mental distance between them is created. This implies a more formal register for the purpose of the depersonalising communication. This formal register expresses an intellectualisation or

‘logification’ of the content; in other words, an interpretation of the discourse units and of the rhetorical relations between these is required. The depersonalising ‘filter’ between the participants reduces the spontaneity and the personal dimension and sensitivity. In this way, we may talk about an opposition between the social and the personal dimensions, between social status and personal involvement. A high degree of socio-cultural hierarchy will – all other things being equal – cause a high degree of logification, interpretation and intellectualisation; on the other hand, a low degree of formality is characterised by a higher degree of personal spontaneity and personal involvement. Typically, these phenomena have been investigated under the

29

sociolinguistic subfield of diasystems (diachronic, diatopic, diastratic, diaphasic and diamesic variations) (Bazzanella, 2008; Coseriu, 1956; Koch & Oesterreicher, 2001; Völker, 2009;

Weinreich, 1954). Logification and interpretation are reflected in various ways in language, but a feature that is particularly sensitive to the different levels of formality is the morphological codification of the verbal content in discourse units, or the explication – for instance, due to the verbal conjugations – of distinctions between various pragma-narrative levels, that is, between the foreground and the background (Tomlin, 1985). The foreground-background distinction is more or less equivalent to the nucleus-satellite distinction in Rhetorical Structure Theory (Mann

& Thompson, 1988), which I shall return to in Chapter 5. The morphological explication of verb forms in discourse units reflects, or grammaticalises, a cognitive hierarchy of the content in question, and may be seen as a particular effect of the intellectualisation and of the logification of a text. In this way, seeing linguistic structures from a rhetorical perspective is closely related to the viewpoints taken by the cognitive-functional approaches.

In other words, it is a matter of adapting, consciously or unconsciously, one’s linguistic style to the given situation by varying factors such as ornamentation, sentence structure and word choice (Gabrielsen & Juul Christiansen, 2010, p. 43). A number of scholars (e.g. Biber, 1991;

Johansson, 2007; Stubbs, 1996) have demonstrated how texts of different genres, types and languages vary across stylistic choices in terms of lexis and grammar, stressing the so-called

‘Humboldt principle’ (Abraham, Givón, & Thompson, 1995) of a one-to-one relationship between form and meaning. Applied to stylistics, this principle prescribes that each formulation has its own stylistic meaning, and that there is no such thing as free variation, even though some formulations may appear equivalent. Renkema (2004, p. 148) recognises three possible views on style, which, in this context, corresponds to the notion of register:

 Style as a possible form for a specific content

 Style as a choice of specific patterns

 Style as a deviation from expectations

In this thesis, I will adopt a combination of the two latter views, seeing, firstly, style and register as choice patterns available to the writer in phrasing what she would like to say. The choices may not be the same in all languages, and perhaps more important, the preferences for certain patterns may not be the same across languages. Secondly, style and register are seen as being influenced by contextual factors and by language-specific routine patterns developed to meet

30

certain expectations from the readers. This means that particular contexts or situations entail a specific style in order to be conceived as appropriate or fitting. Consider the following example, which in many contexts would probably be perceived as stylistically awkward (cf. Neff, Dafouz, Díez, Prieto, & Chaudron, 2004; Nir & Berman, 2010; Slobin, 1996a), even though it is grammatically correct and perfectly coherent.

1) Larry does not eat meat, and he is a vegetarian, and he lives together with Annie, and she eats a lot of meat.

Instead of the many coordinated clauses employed in example 1), readers would expect subordinate clauses expressing causal and elaborative relationships between the clauses. As argued in Chapter 3, the genre of the texts studied in this thesis, namely parliamentary discourse, is expected to be characterised by a somewhat formal and jargon-like register quite different from that of example 1). Thus, both writers and readers have expectations concerning the lexical complexity and informational density.

Even though Renkema’s two views seem to differ in terms of approaching style from an objective (style as specific patterns) and a normative perspective (deviation from expectations), I consider the study of this thesis purely objective in describing and explaining differences in formulation patterns. In contrast, a normative approach would entail a study of whether the speeches by the members of the European Parliament had been able to meet the expectations of the audience. This is beyond the scope of this thesis, but it is interesting to note how much (normative) effort the European Union as an institution puts into converging the language of its politicians, employees and institutions. In a style guide from the European Commission (How to write clearly/Scrivere chiaro/Skriv klart, 2011), ten recommendations on how to write clearly are listed almost identically across the 23 official languages of the EU. The hints are largely similar to those provided for Danish by Pontoppidan (2013), and for English by Thornbury (2005; see below), and closely related to the Gricean Maxims (Grice, 1975), but are, of course, specifically related to writing text in the EU institutions. An interesting difference between the different languages is that the Danish and English versions recommend a mean sentence length of 20 words, while the equivalent Italian guide (p. 6) does not recommend specific sentence length.

The recommendation, which appears in some style guides, not to use sentences containing more than 20-25 words on average is likely to be counterproductive to

31

our documents. The documents risk having too fragmented sentences, with the subsequent possibility of ambiguous sentences or sentences without any sense. In fact, the Italian language allows you to coordinate and/or subordinate clauses in a flexible and at the same time clear manner.¹

In another non-language specific style guide from the European Union (Writing for translation, 2010, p. 7), the recommendation for translators reads as follows:

Long and unwieldy sentences create many problems for the reader, so avoid squeezing too much information into each sentence. This does not mean that you should write very short sentences throughout the text. It is, in fact, not the length itself that creates reading and translation problems, but rather a surplus of contracted sentences, subordinate clauses or other intrusive phrases which hamper the readability of the text.

It would appear that approaching style from a normative perspective is much more a matter of personal taste, ideology or policy than approaching style from an objective perspective. Neither the formalised descriptive apparatus of Cognitive Linguistics, nor the systemic analytical framework of Functional Linguistics is employed in any strict sense in this thesis. Yet, the conception of language as a non-independent faculty plays a central role in the understanding of linguistic structures presented in the chapters to come.