|
Psycholinguistics and Computer-aided content analysisResearchers: Yves Bestgen, Sophie Piérard, Robert Hogenraad Our main research domain lies within the framework of the study of discourse production and comprehension. At the root of discourse analysis is the idea that a text is more than a sequence of sentences. It can be divided into segments, clustering sentences that are locally coherent. Global coherence is founded on the relations between these segments. The structure of a discourse is determined by these two kinds of coherence. This structure provides a bridge between the content that the author wants to express and the representation the addressee must build. The cognitive processes underlying these linearization and delinearization steps remain largely unknown. To study them, we focus on the linguistic devices that underlie the discourse structure, such as punctuation, connectives, adverbials and referential expressions. This lead us to investigate the use of these devices in production, their impact on comprehension, and several cognitive mechanisms underlying their action in L1 (native speakers) and L2 (foreign language learners). We also develop techniques for automatic text analysis such as Latent Semantic Analysis, automatic analysis of text coherence and analysis of the emotional content of texts. (For more information and downloadable papers, please see: "www.psor.ucl.ac.be/personal/yb/") Main collaboration :
Previous Lab Members :
It is possible to predict wars from the analysis of the words contained in documents describing the moments before war breaks out (Hogenraad, 2003, 2005, 2007). I analyze these documents using the PROTAN suite of content analysis programs. For McClelland (1975), an increase of the gap between the need for power and the need for affiliation often precedes the outbreak of major conflicts. In plain language, “we tend to deprive of their faces any person whom we believe to be at the mercy of our will” (Auden, 1962, p. 62). Stories are similar in many respects to the real world (Black, 1984). I also content analyze stories that describe conflicts. Whether fiction or history, I need to transform McClelland’s theory into a content analytic filter. This is the “motive dictionary”. Reader, it works! Not that I have much merit. All it takes to predict is a correlation (Scriven, 1959). Some will even say (after the fact) that predicting the March 2003 war in Iraq was a slam dunk. To explain is more difficult. In science, you can often predict without explaining. The reverse is possible too. Collaborations:
References:
Auden, W. H. (1962/1989). The dyer's hand and other essays. New York: Vintage Books.
|
17/12/2007
|
|