The benefits of multilayer corpus coding for cross-linguistic discourse analysis: Connectives in English and French
(2017) Corpus Linguistics in the South 14 — Location: Birkbeck, University of London (4.March.2017)
Files
CLS2017.pptx
Closed Access - Microsoft Powerpoint XML
- 4.15 MB
Details
- Authors
- Abstract
- When carrying out a cross-linguistic study, the first step is “to make sure that you are comparing like with like” (James 1980), i.e. to establish a common, language-neutral basis for comparison, or tertium comparationis. In other words, as underlined by a number of researchers in contrastive linguistics (e.g. Connor & Moreno 2005; Lefer & Cartoni 2011), any contrastive study must start with the rigorous definition of the object of study, circumscribed by a series of precise criteria of definition and categories which hold across the languages compared. The aim of the presentation is to demonstrate the benefits of a system of multilayer coding of the corpus data based on these rigorous criteria. The advantages of the system are twofold: (i) it allows for maximal operationalization and systematization of the tertium comparationis; (ii) for the cross-linguistic analysis proper, it makes it possible to draw a fine-grained picture of the differences between languages, and paves the way for the application of multifactorial statistics. The focus of the study is English and French connectives, analysed in a comparable corpus of English and French quality newspaper editorials (c. 2 million words per language). To do justice to the specificities of each language, connectives are defined in a fairly broad sense as a category of lexical items which have in common that they signal a logical relationship (e.g. contrast, causality) between two discourse segments at clause level or above. Those lexical items may come from various grammatical categories (i.e. adverbs, subordinators, coordinators), and the clausal segments linked can be of various types (finite, non-finite, verbless clauses) and ranks (main, hypotactic, embedded clauses). Based on this common platform of comparison, I designed a multilayer, hierarchical coding scheme made up of three tiers, viz. (i) part-of-speech of the connective; (ii) degree of finiteness and (iii) degree of embedding of the host clause, themselves subdivided into a number of subtiers. The results show that English and French display striking differences in connective use. For instance, while English tends to use conjunctions of subordination in verbless dependent clauses (e.g. Though not an uncommon attitude, it is arrogant), in French this pattern is quite infrequent. Instead, the French data displays a significantly more frequent use of adverbs in verbless clauses (e.g. Une certitude cependant: les autorités souhaitent redorer l’image de leur police) than English. References Connor, Ulla and Ana Moreno. 2005. Tertium comparationis: A vital component in contrastive research methodology. In Paul Bruthiaux, Dwight Atkinson, William Eggington, William Grabe and Vaidehi Ramanathan (eds.) Directions in Applied Linguistics. Essays in Honour of Robert B. Kaplan. Clevedon: Multilingual Matters, 153-164. James, Carl. 1980. Contrastive Analysis. London: Longman. Lefer, Marie-Aude and Bruno Cartoni. 2011. Prefixes in contrast: Towards a meaning-based contrastive methodology for lexical morphology. Languages in Contrast 11(1), 86-104.
- Affiliations
Citations
Dupont, M. (2017). The benefits of multilayer corpus coding for cross-linguistic discourse analysis: Connectives in English and French. Corpus Linguistics in the South 14, Birkbeck, University of London. https://hdl.handle.net/2078.5/50525