«Discourse and Text Structure»

Ted J.M. Sanders and Wilbert Spooren
«Discourse and Text Structure»
Chapter 35 from: The Oxford Handbook of Cognitive Linguistics, edited by Dirk Geeraerts and Hubert Cuyckens. Oxford University Press. Print Publication Date: Jun 2010. Online Publication Date: Sep 2012.


Extracto del apartado en páginas 27 a 32 de la publicación en PDF. Véanse las referencias en la publicación original del texto.

«Looking into the Future: Integration of Different Approaches

»At the end of this chapter, we have reached the point where we can stop and ask about the avenues that lie ahead of us. We see several interesting developments that may set the research agenda for the coming years. We focus on issues that follow from our analysis of the state of the art in the preceding sections. A first and very basic issue is the question of discourse segmentation: What are the building blocks of discourse? To what extent do they correspond to traditional units of analysis such as the clause, sentence and—in the spoken mode—the turn? Are discourse units in spoken and written language comparable? To what extent are grammatical and discourse structure isomorphic (see Verhagen 2001 for a discussion of similar topics)? A second important issue is the linguistics–text linguistics interface. As noted in section 1, we see a growing exchange or sharing of ideas between grammarians, (formal) semanticists, and pragmaticists on the one hand, and text linguists on the other.

»Questions that can be asked are: What is the relationship between information structuring at the sentence level and at the discourse level? And, how do factors such as tense and aspect influence discourse connections (Lascarides and Asher 1993; Oversteegen 1997)? For instance, discourse segments denoting events that have taken place in the past (The duke fell of his horse. He died) will be typically connected by coherence relations linking their content, whereas segments whose events take place in the present or future typically contain many evaluations or other subjective elements (I am sure I saw the duke fall of his horse just now. He may die), and are prototypically connected by epistemic relations.

»Another promising topic related to the sentence–discourse interface is that of intraclausal and interclausal relationships: are the types of causality found at the intraclausal level (John made him pay the bill vs. John let him pay the bill; Verhagen and Kemmer 1997) similar to the types of causality found at the discourse level (Stukker 2005)? For instance, can The headache caused the soprano to cancel the concert be (insightfully) compared to Because she had a headache the soprano cancelled the concert? A final topic related to the linguistics–text linguistics interface is the relationship between discourse and grammar.

»In the more functionally oriented literature, there is a rich tradition of corpus studies of linguistic structures in a discourse context. A good example is the work on the discourse function of subordinated clauses (Tomlin 1985), more specifically if/when–clauses (Haiman 1978; Ramsay 1987) and purpose clauses (Thompson 1985; Matthiessen and Thompson 1987). Thus, the discourse function of purpose clauses appears to depend on their placement in relation to the main clause. In medial or final position, their role is one of local elaboration, but in initial position, their role becomes one of foregrounding information. They signal how to interpret the following clause, and how to relate it to the preceding text. Hopefully, such studies will inspire more (cognitive) linguists to look at linguistic structures as vehicles built by language producers to enable interpreters to understand what they have in mind. Recently, Langacker (2001) has presented a framework for the further integration of discourse and Cognitive Grammar. A third, obvious issue is the relationship between the principles of relational and referential coherence. Clearly, both types of principles provide language users with signals during discourse interpretation. Readers and listeners interpret these signals as instructions for how to construct coherence.

»Therefore, the principles will operate in parallel, and they will influence each other. The question is: how do they interact? This issue can be illustrated with the simple example in (25).

»(25) John congratulated Pete on his excellent play.

»a. He had scored a goal.

»b. He scored a goal.

»At least two factors are relevant in resolving the anaphoric expression he in (25a) and (25b): the aspectual value of the verb in the sentence and the coherence relations that can be inferred between the sentences. At sentence level, the verb in (25a) is in the perfect tense; at the discourse level, there is one straightforward interpretation of coherence relation is available, namely, the backward causal relation CONSEQUENCE-CAUSE. In (25b), the verb is in the imperfect tense, and, at the discourse level several coherence relations can exist, e.g. TEMPORAL SEQUENCE (of events), or ENUMERATION/LIST (of events in the game).The resolution of the anaphor-antecedent relation is related to these two factors.

»In (25a); he must refer to Pete, while in (25b), several antecedents are possible: John, Pete, or even an actor mentioned earlier. Interestingly, the interrelationship of sentence and discourse levels turns up again: How does the sentence-internal property of aspect interact with the discourse property of coherence relations in the process of anaphor resolution? Is the anaphor resolved as a consequence of the interpretation of the coherence relation? Questions of this kind have already been addressed in the seminal work of Hobbs (1979) and have recently been taken up again in a challenging way by Kehler (2002). A fourth specific issue is the refinement of the relationship between the central concepts of subjectivity, perspectivization, and the typology of coherence relations, which needs to be explored in much greater detail (Sanders and Spooren 2001b).

»The starting point for these studies consists of corpus-based accounts of connectives in terms of subjectivity and speaker involvement (Pander Maat and Degand 2001; Pander Maat and Sanders 2001), discussions of perspective and subjectivity (Sanders and Spooren 1997; Pit 2003), Mental Space analyses of perspective (Sanders and Redeker 1996) and connectives (Dancygier and Sweetser 2000; Verhagen 2005). A fifth issue and area for further research is the interrelationship between spoken and written discourse.

»Results from text-linguistic and psycholinguistic studies presented here are largely based on the study of written discourse. To what extent can they be generalized to spoken discourse? And what will the specific insights from the linguistic analysis of spoken discourse add to the picture we have so far? These questions become especially important when claims concerning cognitive reality are at stake. After all, our most natural and spontaneous way to communicate is not simply in discourse, but in spoken discourse. Finally, there is an important methodological issue on the road ahead.

»A traditional forte of Cognitive Linguistics is its determination to provide cognitively plausible analyses of linguistic phenomena. A less well developed aspect of Cognitive Linguistics is the empirical study of language in use, aiming either to find regular patterns that feed the theories, or to actually test theories against language use. Plausible theoretical ideas regularly have to be revised after serious empirical testing. And even though there are more and more examples of studies combining linguistic theorizing with some kind of testing either in corpus examinations or in language processing experiments, these studies do not dominate the field. Still, to balance the picture of the actual situation, we are happy to find that there is indeed a growing tendency towards quantitative, usage-based studies in Cognitive Linguistics in general.

»We will mention three fields where we see this tendency.

»First, there is the field of lexical studies where Geeraerts, Grondelaers, and Bakema (1994) have shown how lexical salience can be operationalized on a corpus of actual language use, and can then be employed to explain the actual choices of lexical construal that language users make.

»More recently, the same quantitative approach has been extended to more grammatical fields of research. Bybee (2001) epitomizes the use of the quantitative analysis of salience in the phonological (and to some extent morphological) domain; specifically, she uses type and token frequencies to explain diachronic phonological changes (see also chapter 36 of the present Handbook).

»Second, in the field of syntax, Grondelaers’ work on Dutch er is an excellent illustration of how the work by Ariel can be extended and incorporated into quantitative studies of syntactic variation. Building on corpus data and experimental findings, Grondelaers (2000) extends Ariel’s Accessibility Theory of definite reference to indefinite reference, to explain and predict the distribution of er ‘there’ in sentences like Op de hoek van de straat is (er) een bakker ‘At the corner of the street (there) is a bakery’. Grondelaers’ work is especially interesting in that it uses offline corpus data to generate hypotheses that are subsequently tested in a psycho-experimental design.

»Third, in the area of language acquisition, the work of Tomasello (2000) and his co-workers generates many new insights and further questions: Do we want to explain the acquisition order of connectives only in terms of the input provided by the parents? How would such a usage-based account relate to theories of increasing cognitive complexity (see section 3, and Evers-Vermeul and Sanders 2001)?

»In conclusion, it seems that, especially on the level of discourse, the integration of cognitively plausible theories with empirical testing is the ultimate aim, rather than a situation that has already been realized. Therefore, we consider the level of discourse a ‘new frontier’ for Cognitive Linguistics.»