Replication Data for: A corpus-based analysis of the Dat-Nom/Nom-Dat alternation in German

Use email button above to contact.

Dataset abstract

The dataset includes an annotated sample of N = 13292 German written sentences with a Nominative and a Dative argument. The sentences comprise 76 different verbs taking two alternating object orders: 5591 sentences occur with the Dat-Nom order, 8701 sentences occur with the Nom-Dat order. Each sentence is annotated for Object order, the sentence Verb and several features related to both objects, including: (pro)nominality, pronoun type, referentiality, person, number, definiteness, animacy, and length. The sentences and the two objects are shared in a separate .csv-file. An R Notebook with the data analysis is provided as well as an html file with both the R code and output for the analysis.

Article abstract

A subgroup of German Nom-Dat verbs have received considerable attention in the literature due to the propensity of the dative to occur preverbally, which is unexpected on an object analysis of the dative (see references below). Here we argue for an alternative analysis, namely that the relevant verbs alternate between two different argument structures, Dat-Nom and Nom-Dat, and hence that either argument, the dative or the nominative, may be the syntactic subject. Earlier studies have shown that topicalisation of direct arguments is found in ca. 4–12% of the cases in German texts. For comparison, we have extracted 13,000 tokens of 76 verbs from the deTenTen13 corpus and coded them for ten different variables. Our findings support an alternating Dat-Nom/Nom-Dat analysis for these verbs, as 42% of the tokens instantiate the Dat-Nom order and the remnant 58% instantiate the Nom-Dat order. In contexts with full NPs only, the share of Dat-Nom tokens is even higher, 46% compared to 54% Nom-Dat, which altogether excludes a topicalisation analysis of the Dat-Nom word order. In order to throw further light on the alternation, we carry out a multivariate analysis which confirms the effect of topicality, definiteness, length, the animacy of the dative and the inanimacy of the nominative.

The data was retrieved from the deTenTen corpus, accessed through SketchEngine. For more information about the deTenTen corpus, see: sketchengine.eu/detenten-german-corpus. For more information about SketchEngine, see: sketchengine.eu.

References for SketchEngine:

• Adam Kilgarriff, Vít Baisa, Jan Bušta, Miloš Jakubíček, Vojtěch Kovář, Jan Michelfeit, Pavel Rychlý, Vít Suchomel. The Sketch Engine: ten years on. Lexicography, 1: 7-36, 2014.
• Adam Kilgarriff, Pavel Rychlý, Pavel Smrž, David Tugwell. The Sketch Engine. Proceedings of the 11th EURALEX International Congress: 105-116, 2004.

The extracted text fragments included in this dataset only represent insubstantial portions of the source listed above, and they do not represent coherent larger texts. Reuse of such excerpts is permitted under exceptions in IPR and database protection regulations, such as Fair use (cf. US Copyright Act), the EU Database Directive (cf. art 8 Rights and obligations of lawful users), and the Norwegian Copyright Act (cf. § 24 Eneretten til databaser).