Pragmatic corpus analysis, exemplified by Turkish-German bilingual and monolingual data
The paper presents a methodology for empirical multilingual data analysis that combines quantitative and qualitative research. The data is a bilingual Turkish-German and a monolingual Turkish corpus of spoken child language. The methodology proceeds in several steps: (1) description of transcribed data (PartiturEditor) and of the concepts of ‘constellation’ and ‘Evocative Field Experiment’ (EFE), (2) the methodological role of the linguistic unit ‘utterance’, its marking as ‘segment’ in transcriptions and its importance for corpus formation (CoMa), (3) search procedures and frequency assignment of the findings (EXAKT), (4) classification according to constellative features of the data, (5) contextual interpretation of the items, (6) consultation of the transcript where needed, (7) contextually based categorisation of the items resulting in an empirical determination of their varieties. The objective of the methodological stages is an empirical foundation of discourse-based linguistic analysis of multilingual corpora, which we call ‘Pragmatic Corpus Analysis’ (PCA).