Edited By Ana Diaz-Negrillo and Francisco Javier Diaz-Pérez
Conjunctive relations across languages, registers and modes: semi-automatic extraction and annotation: Ekaterina Lapshinova-Koltunski, Kerstin Kunz
EKATERINA LAPSHINOVA-KOLTUNSKI / KERSTIN KUNZ
Conjunctive relations across languages, registers and modes: semi-automatic extraction and annotation
In the present paper, we focus on the analysis of conjunctions as intra- and intersentential links in texts which play an important role in text organisation. Our research goal is to explore a broad range of cohesive conjunctive relations across languages, registers and with varying mode of discourse (spoken vs. written). We aim to analyse how the resources for establishing cohesive conjunctive relations provided by English and German language systems are instantiated in naturally occurring texts of English and German. More specifically, we intend to explore contrasts between the two languages in form, frequency, function and relation. For this purpose, we develop semi-automatic extraction and annotation procedures. In a first step, we extract instances of selected conjunction types on the basis of available lexico-grammatical knowledge (lexical lists, context restrictions, etc.). In a second step, we annotate the extracted results with the types that they instantiate (e.g. semantic: additive, adversative, causal, etc., or syntactic: connects, subjuncts and adverbs). This methodology facilitates our corpus-based contrastive analyses as it provides easy access to information on comparable types of cohesive conjunctions for both languages under investigation. Now, we can query our corpus for different aspects or properties of conjunctive relations encoded as abstract categories – on both a general and a more fine-grained level, or in a specific context. The frequency lists extracted are used to analyse the distribution of conjunctive relations...