Volume 69, Issue 1 - Special Issue: Non-central usages of datives, Issue Editor: Gabriele Diewald

Non-central usages of datives are defined in terms of increasing deviation from the prototypical, central usage of the dative, i. e. its usage in ditransitive constructions with ‘give’ verbs, on several parameters. The introduction offers an exemplification of the clines of (non-)centrality of dative usages, and specifies the aim of the volume. It also provides an outlook on the individual perspective each contribution takes on this issue.

The paper discusses various constructions of ancient Indo-European languages that have been described as featuring a dative of agent. The occurrence of the dative can be explained either through its beneficiary meaning, or as indicating an experiencer. In fact, a number of passages that have been taken as evidence for the reconstruction of a dative of agent do not contain agent phrases at all. Thus, while different constructions have parallels in two or more ancient languages, it is impossible to reconstruct a dative of agent as a unified category for PIE, except in the case of nominal verb forms denoting obligation.

One of the functions of the dative is to mark non-prototypical subjects, i. e. subjects that somehow deviate from the agentive prototype. The Germanic languages, as all subbranches of Indo-European (cf. Barðdal et al. 2012. Reconstructing constructional semantics: The dative subject construction in Old Norse‐Icelandic, Latin, Ancient Greek, Old Russian and Old Lithuanian. Studies in Language 36(3). 511–547), exhibit structures where the subject or the subject-like argument is not in the nominative case, but in the accusative, dative or genitive, for instance. The focus of this article is on the dative, leaving accusative and genitive subjects aside, in particular homing in on lexical semantic similarities and differences between the individual Germanic languages. We compare Modern Icelandic, Modern Faroese, and Modern German, on the one hand, and the historical Germanic languages, i. e. Gothic, Old English, Old Saxon, Old High German, Middle English, Middle Dutch, Middle German, Old Norse-Icelandic and Old Swedish, on the other. The goal is to document the semantic development of the construction across time. This, in turn, is a part of a more general research program aiming at reconstructing the origin and the development of the Dative Subject Construction in Germanic and Indo-European. As the Germanic languages are both genealogically and areally related, we suggest a computational model aiming at disentangling genealogical and geographical factors, in order to estimate to which degree the two interact with each other across languages and across historical eras.

By integrating dialect data into typological theses on dative-marked nominal possessive constructions of the type dem Vater sein Hut , this article will continue recent approaches of both dialectological and typological research. By means of an empirical study, this article intends to not only describe the surface phenomena, but to uncover the underlying structure of this construction in East Franconian dialects. Language maps illustrate that instances of microvariation concerning the case-marking of the possessor ( Wem vs. Wen vs. Wer seine Weiden sind das ?) and the form of the possessive pronoun ( meiner Tante ihr vs. sein Wirsing ) are diatopically conditioned.

German inalienable possession constructions with a PP-embedded possessum come in three variants: (i) with external possessor (EP), (ii) with internal possessor (IP), and (iii) with doubly-marked possession (DMP). Choice of dat (ive) versus acc (usative) case adds two more options. To capture this variation with a formal-syntactic account, this contribution posits (a) possessor raising from Spec DP of the possessum to affectee v P, triggered by lack of case in Spec DP, for dat EPs, (b) a base-generation possessor-as-direct-object analysis of acc EPs, (c) gen (itive) as last resort, triggered by the lack of a case licensor in the verbal domain, for IPs, and (d) a combined base-generation and gen -as-last-resort analysis of DMPs.

There seems to be no construction that would code specifically the possession relation of external possessors. Instead, various host constructions such as the free-affectee construction (a subtype of which is the free-dative construction), monotransitive or applicative construction can – to a different degree – accommodate participants that are bound by a possession relation. The binding procedure identifying the possessor operates at the pragmatic-semantic interface and takes into account the semantic roles of the event participants, their discourse saliency and lexical properties (such as animacy), world knowledge, properties of the possessum (such as the degree of (in)alienability), etc. The external-possession relation and the meaning of the hosting construction are orthogonal to each other, but there is a strong interplay between them. Positing a dedicated external-possessor construction faces the following problems: the same binding procedure is found in other constructions as well (such as the monotransitive or ditransitive constructions); at the same time, the very possession relation is only inconsistently found in constructions referred to as external-possessor constructions, and, what is more, the possession relation may sometimes be canceled (even if it is inalienable). To account for this terminologically, I introduce the term non-thematic affectee construction , a subtype of which is the free (non-thematic) affectee construction particularly spread in European languages. The latter is found with different types of coding (accusative, dative, different prepositional phrases) which have different diachronic sources.