German universal concessive conditionals with wh-clause-initial and wh-clause-medial marking

DOI

These are the data analysed in Chapter 6 of Vander Haegen's dissertation entitled "Konstruktionsgrammatik und Variation. Eine Mikrotypologie universaler Irrelevanzgefüge im Gegenwartsdeutschen". The dataset includes an annotated sample of N = 3000 German written universal concessive conditionals (e.g. "Was immer auch passiert, ich bin für dich da!" 'Whatever happens, I will be there for you!') from the German Reference Corpus DeReKo. Concessive conditionals (CCs) can basically be defined as conditionals with quantification, as they express a (usually exhaustive) set of antecedent values in the protasis, instead of only one antecedent value as in prototypical conditionals. In universal concessive conditionals (UCCs), the antecedent set is evoked through a combination of a free-choice-expression (here: the complex particle "immer auch" '-ever') and a variable in the guise of a wh word (here: "was" 'what'). German UCCs can either be marked wh-clause-medially (using the particles "auch" and/or "immer" as above – both mean '-ever') or wh-clause-initially (with an expression of irrelevance like "egal" 'no matter', "gleichgültig" 'no matter', "wurscht' 'no matter (lit.: sausage) etc.). The wh-clause-initial and -medial marking strategies are in near-complementary distribution. The dataset was compiled in the context of a doctoral dissertation project which sought to, inter alia, determine functional and formal differences (apart from the position and lexical specification of the free-choice marking) between wh-clause-initially and -medially marked UCCs in contemporary German. It comprises n = 1500 tokens of wh-clause-initially and n = 1500 tokens of wh-clause-medially marked UCCs. The data are annotated for over 20 formal and functional variables. The full dataset (including the corpus data) and a dataset including only ID numbers and annotations (to facilitate processing with statistical software) are shared in two separate .csv-files. An R Markdown file with the data analysis and an html file with the R code and output are shared as well.

COSMAS II web, 2.4.5.4

Microsoft Excel for Mac, 16.103

R (aarch64-apple-darwin20), 4.4.2

R Studio, 2025.09.1+401

The data source is publicly available after logging in on https://cosmas2.ids-mannheim.de/cosmas2-web/ or https://korap.ids-mannheim.de/.

Identifier
DOI https://doi.org/10.18710/FVA2YV
Metadata Access https://dataverse.no/oai?verb=GetRecord&metadataPrefix=oai_datacite&identifier=doi:10.18710/FVA2YV
Provenance
Creator Vander Haegen, Flor ORCID logo
Publisher DataverseNO
Contributor Vander Haegen, Flor; TROLLing team; Fonds Wetenschappelijk Onderzoek – Vlaanderen; Ghent University; The Tromsø Repository of Language and Linguistics (TROLLing)
Publication Year 2025
Funding Reference Fonds Wetenschappelijk Onderzoek — Vlaanderen 1197324N
Rights CC0 1.0; info:eu-repo/semantics/openAccess; http://creativecommons.org/publicdomain/zero/1.0
OpenAccess true
Contact Vander Haegen, Flor (Ghent University); TROLLing team
Representation
Resource Type manually annotated corpus data; Dataset
Format text/plain; text/comma-separated-values; text/x-r-notebook; text/html; type/x-r-syntax
Size 27003; 1681007; 425414; 17812; 2506091; 147; 382; 193
Version 1.0
Discipline Humanities