Emergence of Non-Canonical Case Marking in Indo-European

Several languages of the world exhibit verbal predicates where the subject is not in the nominative case, but is instead marked in an oblique case, like accusative, dative or genitive. Remnants of such structures exist in English and Dutch, like e. methinks, meseems and d. me dunkt. In several other Indo-European languages, constructions like that not only occur as remnants that are confined to the lexicon, but instead they exist as a meaningful part of the grammar, even belonging to the productive component of language use. Languages like that are German, with compositional predicates like mir ist gleich, and corresponding Icelandic examples, mér er sama, both meaning 'I don't care'. These constructions clearly deviate from the typical nominative case marking of subjects, since there may not be any argument present in the nominative case, while at the same time the subject-like oblique shows several behavioral properties of subjects. Looking back to earlier language stages, constructions with subject-like obliques are found in all the early and ancient Indo-European languages, including Latin, Ancient Greek, Sanskrit, Hittite, Tocharian, Gothic, Old Irish, Old Russian, Old Lithuanian, Old Albanian, and Classical Armenian. Within the NonCanCase project, data from Ancient Greek, on the one hand, and from Gothic, Old English, Old Saxon, Old High German, Middle English, Middle High German, Middle Dutch, Old Norse-Icelandic, Old Swedish, Modern Icelandic, Modern Faroese, and Modern German, on the other, were collected and inserted into an electronic Mysql database. On the basis of these data several specific research questions were addressed, namely whether these subject-like obliques behave syntactically as subjects or objects, to which degree the semantics of the oblique subject construction overlaps across the Indo-European subbranches, whether the origin and emergence of these construction can be identified, whether the historical development of these constructions may be revealing with regard to the origin of the construction, and which possible grammaticalization paths may be identified. Our findings, published in leading international journals in the field, show that subject-like oblique arguments behave syntactically as subjects in both the earliest Germanic languages and Ancient Greek. Thus, the hypothesis that oblique subjects have developed from objects cannot be corroborated. A comparison in terms of lexical-semantic verb classes has also documented a major overlap between the semantic fields across the languages under investigation, which clearly speaks for inheritance. Moreover, Latin and Ancient Greek turn out to be most similar, while Old Norse-Icelandic, Old Russian and Old Lithuanian all deviate from the Latin/Greek prototype and also from each other, suggesting that the construction is old in the Indo-European languages. On the basis of cognate stems, cognate case morphology, cognate predicate structure and the same semantic fields, we have reconstructed argument structure constructions with oblique subjects for both Proto-Germanic and Proto-Indo-European. These reconstructions are modeled with the formalism of Sign-Based Construction Grammar and involve predicate-specific constructions, partial predicate-specific constructions, as well as abstract schematic constructions, all with oblique subjects.

Alignment and argument structure lies at the heart of all current theoretical models in linguistics, both syntactic and typological models. Case marking, alignment and argument structure has been the topic of several recent research projects funded by nat ional research councils in both Europe and the United States. In spite of the overwhelming weight of argument structure and alignment in current linguistic research, no large-scale, comprehensive study of the historical development of case marking and ali gnment systems has been carried out in modern times, using modern linguistic approaches and frameworks, and covering a whole genetic language family from its first documentation until modern times. NonCanCase aims to investigate argument structure and a lignment from a historical perspective, or more precisely non-canonical subject and object marking, focusing on its development through the history of the Indo-European languages. Several modern IE languages contain structures where either the subject(-li ke argument) or the object is not canonically marked, like Icelandic, German, Russian, Lithuanian, Armenian, Hindi and Urdu, and corresponding structures are readily found in Old Norse, Gothic, Old Church Slavic, Latin, Ancient Greek, Hittite, Tocharian a nd Old Persian. The alignment system of the Indo-European proto-language has also been a matter of substantial debate, although this debate is based on the ancillary properties found with certain types of alignment systems and does not take case marking a s its point of departure. The current alignment typology, moreover, is only based on canonical intransitive and transitive patterns, and excludes several sentence patterns with non-canonical case marking, like dative subject constructions and others, from the classification. The fact that the classical alignment typology is only based on a subset of the relevant structures severely undermines its value and calls for a new approach to case marking and alignment.