Dependency and realisation of EDUs

In this section, the two main aspects of text structure under investigation shall be defined, namely dependency and realisation. At the same time, a review is provided of the most important literature on the two aspects. Bearing in mind the six dimensions of the cognitive-functional communication model presented in Chapter 2 regarding linguistic structures and cognitive dimensions, this chapter shall demonstrate how the text structure dimension is not only deeply intertwined with that of syntax but also closely related to the lower level of morphology and to the higher level of context (see Figure 2.1). Both dependency and realisation are traditional grammatical parameters which have been described in the literature on syntax.

The first parameter shall be referred to as dependency to account for EDUs that are syntactically coordinated with other EDUs and for EDUs that are syntactically subordinated to other EDUs. Coordination and subordination are well-established grammatical terms, and:

[I]n what is probably their most widespread application, ‘subordination’ and

‘coordination’ – along with their adjectival cognates ‘subordinate’, ‘coordinate’, etc. – are syntactic notions denoting relations between parts of a complex syntactic unit. That is, they concern the structure of sentences or clauses and their parts.

(Fabricius-Hansen & Ramm, 2008, p. 2).

An example of two coordinated EDUs is found in 50), and a subordinate EDU is found in the second EDU in 51), both underlined. In both examples, the coordinator (but) and subordinator (as) have been highlighted in bold-faced type.

50) [Mr President, I am firmly of the belief that people have a right to smoke,] [but my personal taste is that they should do so only in the privacy of their own homes.] <ep-00-12-11.txt:42>

51) [Mr President, this is not reform,] [as the Commission itself points out.] <ep-02-05-29.txt:37>

In the first example, the two EDUs are coordinated by the coordinator but, which usually expresses some kind of contrast in terms of rhetorical relation. The two EDUs are also sentence-like in that they are grammatically independent of each other and can easily stand alone, that is, the comma after smoke can be replaced by a full stop, resulting in two independent sentences. In

89

the second example, the second EDU is subordinated to the first EDU, a matrix clause this is not reform, due to the presence of a subordinator as, and cannot stand on its own. As a rule of thumb, it is the connectives – coordinators or subordinators – that determine the dependency between EDUs placed inside the same sentence. But in cases where there is no explicit connective, it is the morphological realisation of the main verb in the subordinate EDUs that determines the dependency. This is the case in example 52) where a matrix clause is followed by a subordinate construction with two English –ing forms (causing and fuelling). It is also the case in example 53) where two coordinated main clauses are separated by a semicolon expressing a juxtaposition of the two EDUs (cf. Ferrari & Zampese, 2000, p. 236). In EU texts, the recommended usage of colons and semicolons between EDUs advises that colon should be used to separate subordinate EDUs and semicolons to separate coordinate EDUs (see English Style Guide. A handbook for authors and translators in the European Commission, 2012, p. 15).

However, the results of the next chapter suggest that this advice is not always followed.

52) [It would make the world a much more dangerous place,] [causing more resentment]

[and fuelling more terrorism.] <ep-03-01-29.txt:28>

53) [Relative stability is based on track record;] [equal access is based on fleet tonnages.]

<ep-02-05-29.txt:37>

The concept of dependency has mainly been examined in the field of syntax, but in discourse linguistics, the notion of coordination has also attracted much attention. Some scholars believe that syntactic coordination equals rhetorical coordination, that is, dependency across sentences (e.g. Asher & Vieu, 2005; Tomlin, 1985; Txurruka, 2000), whereas others claim the opposite, that is, syntactic coordination does not necessarily correspond to rhetorical coordination (Blühdorn, 2008; Haspelmath, 2004, p. 3f; Ramm & Fabricius-Hansen, 2005; Skytte & Korzen, 2000). This also holds true for syntactic versus rhetorical subordination.

Undoubtedly, the idea to link syntactic subordination to discourse nuclearity [i.e.

rhetorical subordination] has intuitive appeal, and moreover it is not difficult to accumulate evidence in support of such a link: Quite often, a syntactically subordinate clause is in fact ‘less central to the writer’s purposes’ than the matrix clause is. But on the other hand, one can also gather evidence for the opposite position – the writer’s purposes running against syntactic subordination, which

90

altogether suggests that matters are more complicated … (Stede, 2008a, pp. 47–

48)

The disagreement on syntactic versus rhetorical coordination and subordination probably arises as a result of the stereotypical and vague coordinator and together with its counterparts in other languages. In an attempt to separate the two kinds of linguistic dependencies from each other, Lehmann (1988, p. 182) refers to syntactic parataxis and hypotaxis versus rhetorical or discourse coordination and subordination. In this thesis, however, I apply only the coordination versus subordination distinction, always specifying the syntactic or rhetorical nature of this. In general terms, I shall refer to syntactic dependency and rhetorical hierarchy.

Accordingly, syntactic coordination is considered a syntactic relation between two EDUs of the same syntactic construction, which may consist of all the EDU categories described in Chapter 3. Syntactic subordination, on the other hand, is used to refer to EDUs that are syntactically dependent on another EDU but not vice versa (Ferrari & Zampese, 2000, p. 141).

Again, I refer to the types of subordinate EDUs presented in Chapter 3. However, such a coarse distinction is by no means unproblematic. In fact, one of the other categorisation issues in this area of grammar is that of treating coordination and subordination as two independent groups or as two poles on the same scale. Cristofaro (2003, p. 15ff) is one of the more recent scholars who suggests that no sharp binary distinction between coordination and subordination should be drawn. In fact, coordination and subordination are the two extremes of a gradient where many coordinated constructions show various degrees of subordination (cf. also Cosme, 2008, p. 109f;

Johannessen, 1998, p. 237ff; Kortmann, 1996, p. 56ff; Quirk, Greenbaum, Leech, & Svartvik, 1985, p. 927f).

Lehmann (1988, p. 182) describes coordination as involving the linking of EDUs of the same syntactic construction. The same viewpoint is taken by Quirk et al. (1985, p. 918), who specify that in coordination ‘the units are constituents at the same level of constituent structure, whereas in subordination they form a hierarchy’, with the subordinate EDU being a constituent of the superordinate EDU. There are various ways of distinguishing between and identifying coordination and subordination, a common one being by means of the explicit indicators of both types of linking constructions, i.e. coordinators and subordinators respectively. Another indicator, which holds at least for Danish, is word order: in most subordinate clauses, a negation must be placed between the subject and the verb and not between the verb and the object in sentences and other types of clauses (Heltoft & Hansen, 2011, p. 314).

91

The main difference between coordination and subordination, then, is a matter of the degree of syntactic dependency (see Givón, 2001, p. 327). Below, in 54), we see an example of how two coordinated EDUs are syntactically independent of each other, each having their own subject and main verb. The second EDU in 55), by comparison, exhibits a higher degree of dependency, because it depends on the first EDU for the expression of one argument, namely the subject, and lacks illocutionary force of its own. The same phenomenon is apparent in non-finite subordinate constructions like the one found in 57), which exhibits a higher degree of dependency than the subordinate finite construction in 56), since it lacks tense and an illocutionary force of its own. Lehmann (1988, p. 184ff) refers to this phenomenon as the

‘hierarchical downgrading’ in which subordinate EDUs integrated into the matrix EDU are found at one end of the downgrading continuum.

54) [The plane arrived late in London,] [and John went straight home.]

55) [John arrived late in London] [and went straight home.]

56) [John went straight home,] [because he arrived late in London.]

57) [Having arrived late in London,] [John went straight home.]

In most cases of coordination, it is also possible to add another connective to the coordinator. In example 58), the content from 55) has been added a connective (therefore) after the coordinator and. In this way, there is a more explicit highlighting of the consequential rhetorical relation between the two EDUs than could be seen 55).

58) [John arrived late in London] [and therefore went straight home.]

The notion of realisation has already been introduced in the previous chapters, but it has not been properly defined yet. Realisation, in the context of this thesis, is mainly related to subordinate EDUs and to whether these contain a finite verb form, a non-finite verb form or no explicit verb form at all. The non-finite verb forms, including nominalisations, do not mark tense, mood or aspect at all (Korzen, 1998, pp. 68–69). This is shown in the following examples; in 59) by means of an English –ing form (having arrived) and in 60) by a nominalisation (arrival).

92

59) [Having arrived late in London,] [John went straight home.]

60) [After his late arrival in London,] [John went straight home.]

This lack of subject marking in non-finite constructions generally entails an inherent subject and topic continuity, which means that the situation or event in question is processed and interpreted as related to the on-going topic but as less important than the situation or event of the matrix EDU, realised with a finite verb. Therefore, the rhetorical relation between an EDU in which the main verb is non-finite is entirely dependent on the matrix EDU, and such structures express a particularly strong integration of the EDU in question into the matrix EDU, as also stated by Lehmann (1988, p. 214):

[A]dvanced hierarchical downgrading of the subordinate clause implies a low syntactic level for it. We will thus be justified if in the following we take advanced downgrading as a sufficient condition for high integration. High integration of the subordinate into the main clause correlates positively with its desententialisation.

Lastly, it is also possible to find an EDU expressed as a subordinate clause or phrase in which the verb has been omitted. Here, the verb is completely absent, but can, nevertheless, be inferred from the co-text of the adjacent EDU: [Once {John arrived} in London], [he went straight home.].

61) [Once in London,] [John went straight home.]

The example above also shows how subordinate EDUs can be placed before the matrix EDU.

This is not the case in coordinated EDUs, except when these are of the same syntactic class and not linked by any discourse cue; in this case, the EDUs may shift position without altering the meaning (Ferrari & Zampese, 2000, p. 237).