International Journal of Corpus Linguistics - Volume 14, Issue 4, 2009

This paper presents an argument for agent-oriented modeling (AOM) as a research methodology and a metalanguage for corpus linguistics. It is triggered by three closely related issues arising from compiling multimodal corpora such as the Spoken Chinese Corpora of Situated Discourse (SCCSD). Given a real-life situation, there are three types of representation: (i) the Written Word representation, (ii) audio recording, and (iii) video recording. It is shown that the three types are all data-transformative and involve data loss, and that they are intrinsically flawed. The current multiple-layered approach to data integration is also shown to be inadequate. AOM is proposed to be a potential solution to the problems. Modeling decision tree, levels of modeling, and modeling schema written in XML are demonstrated. The philosophical basis of AOM, and its theoretical implications are also discussed.

The analysis in the present study examines the variable use of ne in 2,459 tokens of verbal negation from a corpus of electronic French discourse including moderated chat, non-moderated chat, and discussion fora. These data from computer-mediated communication (CMC) environments were analyzed in order to demonstrate that although both non-moderated and moderated chat are essentially synchronous (i.e. involving real-time interaction), they are on opposite ends of a continuum of French sociolinguistic variation. In non-moderated chat, which is clearly less formal, ne is used productively as a discursive-pragmatic marker. However, in moderated chat, which resembles more formal written French discourse, ne deletion rarely occurs. Although discussion fora are somewhere between the two types of chat on the continuum of sociolinguistic variation, there is a clear tendency for them to be slightly closer to the formal/written side, yet this mode of CMC displays the greatest amount of mode-internal variation.

Corpus linguistic analyses reveal meanings and structural features of data, that cannot be detected intuitively. This has been amply demonstrated with regard to non-fiction data, but fiction texts have only rarely been analysed by corpus linguistic techniques. This is the case even though it has been shown by previous analyses that corpus stylistic analyses reveal literary meanings of the data that are left undetected by the intuitive analyses of literary criticism. The analysis of the keywords and most frequent phrases of Jane Austen’s novel Pride and Prejudice presented in this article confirms this claim by uncovering meanings that are not discussed in literary critical secondary sources. This constitutes evidence for the large potential of corpus stylistics for the analysis of literature and its meanings.

We report on a project investigating the lexico-grammatical properties of English scientific texts. The goal of this project is to gain insight into the linguistic effects of two scientific disciplines coming into contact with one another (e.g. computer science and linguistics) and possibly forming a merged, new discipline (i.e. computational linguistics). The crucial question to be addressed is how such merged disciplines construe their own, distinctive identity and which kinds of linguistic means they employ to this end. To approach this question, we apply the notion of register, i.e. functional variation or variation according to context of use. On the basis of a corpus of scientific research articles from nine scientific domains, we explore selected lexico-grammatical patterns and assess their contribution to register formation.