Elaboration in descriptive texts

As Table 7.1 shows, elaboration accounts for 24% of the total relations in the corpus (n = 447). elaboration is also the most common relation in 17 out of 30 brochures. Consequently, we may ask why the relation of elaboration is so frequent in the tourist brochures? At this point, it is necessary to recall that Stede (2008, p. 318) has criticised elaboration as a particularly vague and imprecise relation. He writes:

[M]any annotators seem to resort to elaboration as a “default”, so that the presence of this relation in some RST tree can result ei-ther from the spans being in a genuine, “good” elaborationrelation (and the annotator confidently assigned it) or from a perceived un-clear relationship between the spans, which is somehow also covered by elaboration.²

I would argue that any judgement regarding elaboration — or any other RST relation — needs to be sufficiently informed by the context where the relation is considered to hold. After all, this is a key principle of RST as a theory of text organisation. If there is a risk of using elaboration more often than necessary, a case-by-case study of the relation should be able to detect the behaviour of

“defaulting” to elaborationin problematic analyses.

As I will show below, the relation of elaboration appears to have several functions in the data. elaborationdoes not only add information to a previously introduced topic by establishing a relation between two RST segments: the relation also occurs between a segment and a span. By incorporating another span in the RST structure, elaboration facilitates the introduction of rhetorical relations to the structure. Let us now look at an example below.

Figure 7.2 shows the rhetorical structure of an instance of text-flow in WEH 1998, which is made up of several paragraphs. This is reflected in the layout structure, as the text consists of a layout chunk with multiple child nodes: the header (s-3.24) and three paragraphs (s-3.25,s-3.26-9 and s-3.30-2). In terms of the rhetorical structure, I would suggest that this text is a coherent entity with a specific function that describes a common topic in the tourist brochures: the climate (Valde´on 2009, p. 26). Yet to establish how such coherence is achieved, the rhetorical structure of the text needs to be subjected to a closer examination.

To begin with, the paragraph in the segments-3.25is central to the text. All other paragraphs elaborate this paragraph, as indicated by the topmost elab-oration span in Figure 7.2. This topmost span embeds three other spans: an antithesis span and two elaboration spans. Together, the antithesis and

2Note that I have applied the typographic conventions used in this dissertation to this quote.

page-4-column-1 ELABORATION ANTITHESISELABORATIONELABORATION EVALUATION

s-3.24: Meet the four seasonss-3.25: Helsinki is definitely a city of four different seasons, which means warm sunshine in summer, cool weather and beautifully coloured leaves in autumn, varying amounts of snow and a frozen sea in winter, and bright and clear evenings and mornings in spring. s-3.26: Normally, the hottest month is July with an average temperature of +17.0 C, the recorded maximum in thirty years being +30.8 C.

s-3.27: The coldest month is usually January, the average temperature being -5.7 C.

s-3.28: Helsinki is located on the southern coast of Finland, very close to 60 North Latitude, which explains the abundance of light in summer months.

s-3.29: After the longest day, around the 22th of June, the darkness lasts only 1 hour 20 minutes.

s-3.30: The varying weather conditions have turned Finns into real experts at building and heating houses, as well as in winter traffic. s-3.31: Many special solutions, like snow removal machinery and central heating, have been developed to guarantee uninterrupted traffic and comfortable living.

s-3.32: Helsinki keeps going all year round.

Figure 7.2: Examples of elaboration inWEH 1998

the firstelaboration span constitute a paragraph. What these embedded spans do is they endow the description with a certain ‘depth’, which allows the intro-duction of additional topics related to climate. This is further illustrated by the third elaboration span, which embeds anevaluation span that explains how the citizens of Helsinki have learned to live with the local climate.

One possible explanation for the rhetorical structure in Figure 7.2 arises from the evaluative function of the tourist brochures. Considering that the brochures usually aim to provide a positive appraisal of the destination (Francesconi 2011, p.

344), they require a mechanism for introducing and highlighting selected aspects of the destination. From the perspective of discourse structure, the paragraphs may be considered macrostructures with specific functions that provide the required mechanism.

In this case, the paragraphs introduce (1) the general topic and describe (2) the climate, (3) the location (4) and their combined effect on the city and its inhabitants. These paragraphs may be further split into microstructures that consist of RST segments and their interrelations (Longacre 1992, pp. 112-114).

From this position, the investigation could obviously be extended all the way down to the stratum of lexicogrammar (cf. Hiippala 2007). Together, the paragraphs and the segments that constitute them illustrate the meaning potential and flexibility of text-flow in this aspect: the same structural pattern could be used to introduce a completely different topic.

Consequently, I set out to investigate how often another span acts as a satellite in an elaboration relation by querying the annotated corpus for the instances that match the aforementioned criteria. The result showed that in 81 out of the 447 elaboration spans (18%), the satellite consisted solely of another span. In 197 elaboration spans (44%), the satellite consisted of a number of RST segments and another span. I would thus suggest that the purpose of elaboration is to do more than simply add detail to the description. While also adding detail to the topics introduced in the text, the relation simultaneously enables the introduction of other rhetorical relations by embedding additional RST spans.

It may be suggested that this kind of recursive deep elaboration is a particu-lar rhetorical configuration to enrich and liven up the descriptions in the tourist brochures. Finally, if the recursive use of elaboration were to be studied fur-ther, the application of classical RST could be beneficial (see Section 3.2.3.1).

Determining the relations between clauses — not sentences — would provide the appropriate detail for focusing exclusively on the structure of linguistic discourse.

Keeping in mind the multimodal goals of this dissertation, I will now move on to discuss the next relation: enablement.