Recently I got the following poster and found it very succinct and insightful. And it said something about corpus linguistics. I'm putting it here in case you may also find it interesting--
The rhetorical power of causal connectives, or: why cognitive semantics needs discourse analysis and vice versa.
Professor Dr. Paul van den Hoven,
Utrecht Institute of Linguistics OTS
Utrecht University, The Netherlands
As far as we know every language has a surprisingly rich set of elements to express intra clause causality. In English we have for example since, therefore, thus, because, that is why, so and several others. Cognitive semantics tries to explain the distribution of these linguistic elements in discourse: which factors explain the choices language users make? The basic hypothesis in this approach is that these choices coincide with variables in the way human beings conceptualize causality.
So far promising relations have been found (Pit 2003, Stukker 2005 and many others). I will mention two conceptual variables.
1. The conceptual domain of the causal relation. Is the relation a content relation (rain causes wed streets), a epistemic relation (wed streets cause that we know it has been raining) or a pragmatic relation (rain causes that we propose to stay at home). Different kinds of relations seem to go with a preference for different connectives. The conceptual distinctions come from the work of Sweetser.
2. The kind of actors involvement. Is the main causal participant human, animated or non-human? If animated or human, is the ‘effect’ a result of a voluntary decision (the birthday of my child causes that I stay home) or non-voluntary (my child being hit by a car causes that I was alarmed). Different kinds of actors involvement seem to go with a preference for different connectives. The conceptual distinctions come from Tarski’s force dynamics.
The dominant methodology in this research is corpus analysis. In a corpus of discourse fragments the text surrounding the connective is analyzed to determine the conceptual characteristics of the causal relation. Correlations between these surrounding text elements and the connectives are investigated (for example, in Dutch the connective doordat correlates highly with non animated, content relations; the connective want correlates highly with human, voluntary epistemic relations). The C often implicit C basic assumption of this kind of research is that all information about the conceptual characteristics of the causality is already available from the surrounding text, apart from the connective, and that the connective just represents, indicates, confirms this information; the function of the connective is then purely a help for the reader to interpret.
Discourse analysts and rhetoricians should question this basic assumption. Is it not more plausible that these connectives constitute meaning? Connectives do not just indicate the kind of causality that is at hand, but they are a means in the way the writer intends to construct this causality. Using a different connective builds a different rhetorical structure. Causal connectives C small and insignificant as they might be C turn out to be powerful rhetorical means.
I will argue that this seems to be the case. I will present a number of challenging examples and show that often a deep, very broad analysis is needed to determine the causal structures that the writer constructs by means of the choice of the connective.
My conclusions are:
1. These tiny little text elements are (as many other particles) extremely interesting from the perspective of critical discourse studies. They play an important role in the way responsibilities and accountabilities are rhetorically modeled. They are for example means to suggest that not actor A is accountable for a certain effect but another force outside of his will, or to suggest the entire opposite, just by the choice of the connective. Or they can bring the reader in a very close relation to actor A and very distant from actor B, or the entire opposite, just by the choice of a connective.
2. Corpus based cognitive semantic research needs insights, analytical categories, theories from argument theory and other discourse and communication theories.
3. Causality is said to be one of the three most fundamental structures in the human conception of the world (the other being time and space). Cognitive semantics often focuses on comparative linguistic research (for example Pit 2003 compares backward causal connectives in Dutch, French and German). Our observation will show that such comparative linguistic research is a very form of cultural discourse analysis.