Project Details
Projekt
Speaking - writing - visualizing. Forms, functions and disturbances of modality interdependencies and concept genesis in professional interactions
Subject Area
General and Comparative Linguistics, Experimental Linguistics, Typology, Non-European Languages
Software Engineering and Programming Languages
Software Engineering and Programming Languages
Term
from 2014 to 2023
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 247314954
Professional interaction settings (e.g. team meetings) are often characterized by modality interdependencies (MID), i.e. the combination of modalities such as speaking, writing, visu-alizing. To date, MID were only investigated in parts. There is a lack of approaches, methods and tools for the systematic description and analysis of their forms, functions and disturbances. The project examines MID in professional cross-situational interactions, which are charac-terized by cooperative concept genesis (interactive process of successive mental concept development to real world parts and their representation). The aim of the project is to develop a research approach for MID. It contains three sub-aims:1. (Pattern-related) description of forms, functions and disturbances of MID (generic/as part of a concept genesis)2. Development of a suitable data collection and analysis methodology for (1),3. Development of a supportive annotation tool for heterogeneous data sources (multi-level annotation of text, image, video).The project uses preliminary work (dataset) of the BMBF project IMIP (2008-2011), in which communicative methods of industrial process modelling were studied based on an industrial case study. The case study includes temporally and spatially displaced interaction situations, where the participants use a variety of communicative resources. The complete data set includes primary data (537 minutes video recording, 87 document scans) and secondary data (270 transcript pages). The data indicates a high incidence of MID, which have not been analyzed in IMIP. The research program uses discourse analytic approaches that are phenomenon-related enhanced (e.g. text-linguistically). The development of the annotation tool combines corpus and computational linguistic approaches.Initially, the approaches and methods are developed manually data-driven, then increasing-ly tool-supported. For manual data analysis, a training corpus is elicited of the overall IMIP data set. From the analysis results, tool requirements are deduced. The tool development is based on EXMARaLDA and extends it (including image annotation function). The tool supports step-by-step the semi-automatic data analysis. Functionalities are implemented that allow to set annotated parts (from different data formats) into relation and to visualize them. With a validation corpus, the results (approaches and methods) are tool-supported proved. In workshops central findings are discussed with external experts.The main outcomes and benefits of the project are: an innovative approach for the system-atic description and analysis of MID, a gold standard for heterogeneous data sets, access to rare corpora (industry data), functional extension of EXMARaLDA and open access to the resource.
DFG Programme
Research Grants
