Speakers
Description
Speakers may need to consciously assess which expressions are appropriate for a communicative situation. One such dimension is indexicality: linguistic expressions may evoke the situational contexts in which they are typically employed. In a usage-based understanding, these indexical potentials emerge from recurrent co-occurrence between linguistic forms and situational features and become part of speaker knowledge through entrenchment and schematization (Schmid, 2020). Corpus analysis is methodologically important because natural language frequency patterns can be treated as a model of the input from which speakers may abstract linguistic knowledge (Stefanowitsch & Flach, 2017; Ellis, 2017). However, existing lexicographical resources do not systematically represent usage-based indexicality, and studies that triangulate corpus-derived patterns with speaker knowledge at the vocabulary level are missing. The project introduced in this paper aims to describe systematically and model usage-based indexicality in German communication verbs (CVs) (Harras et al., 2004), drawing on data from both language use and speaker knowledge. Its research approach can provide a basis for conceptualizing the representation of indexicality in lexicographical resources. At the current stage, the paper focuses on the methodological design of the project (first results will be available when the conference takes place).
The project builds on a preliminary corpus study of spoken German (Meißner, 2025), in which CVs were investigated in the FOLK corpus across private, institutional, and public interaction domains (Kaiser, 2018). Analyses of variance and distinctive collexeme analyses (Gries & Stefanowitsch, 2004) were used to identify patterns of over- and under-representation of CV lexemes across domains (indexical potential types (IPTs)). The project asks to what extent such IPTs correspond to speakers’ associations, both with respect to fields of CVs with similar indexical profiles and in terms of granularity, i.e., whether indexical meaning attaches not only to lexemes but also to specific grammatical forms. To this end, three forms are selected: first-person singular present, third-person singular present, and impersonal passive, representing speaker-centered, reporting, and agent-detached perspectives.
To answer the question, corpus-based evidence is triangulated with experimental data on speaker knowledge. The corpus component is based on twelve subcorpora, each containing 100,000 full-verb occurrences, designed to represent contemporary German across private, institutional, and public communication and across spoken, written, computer-mediated, and selected special types of communication. Existing corpus resources are used where possible and are supplemented with new web data where necessary. After preprocessing, analysis of variance and distinctive collexeme analysis are used to derive IPTs for lexemes and grammatical forms.
The speaker-knowledge component mirrors this corpus-based perspective through online questionnaires. For lexemes, 605 CVs from Harras et al. (2004) are distributed over ten item lists and tested in five questionnaire designs: one general association-strength rating, three domain-specific ratings for private, institutional, and public communication, and one free-association task. The rating tasks use labelled seven-point scales; the free-association task asks which communicative situation first comes to mind. This combines scaled judgment data with less pre-structured elicitation of prototypical contexts. A form-level study mirrors these designs for a 20% subset of 121 CVs for the grammatical forms. The rating data are analyzed with ordinal mixed models, with rating type, form, and their interaction as fixed effects and random intercepts for participants and items.
By comparing corpus-derived IPTs with rating and association profiles, the study identifies convergences, divergences, and complementary patterns between language use and speaker knowledge. It refines a usage-based account of indexicality by examining how frequency-based distributional patterns relate to metalinguistic associations. Thereby, the study provides an empirical basis for conceptualizing a representation of situational-indexical information in lexicographical resources for German CVs. More generally, the project outlines a transferable method for modelling how expressions are associated with communicative situations. A lexicographical representation of such associations can support users in choosing forms appropriate to specific communicative settings.