A frame-based approach to implicit arguments
General information
Research project supported by the Deutsche Forschungsgemeinschaft (DFG)
Project team:
- Prof. Dr. Gert Webelhuth (PI)
- Dr. Heike Baeskow
Project duration:
May 2024 – Apr 2026
Keywords
Implicit arguments, Frame Semantics, Construction Grammar, verbs, polysemy, nominalizations, context-dependent information
Summary
The argument structures of lexical units can be modified in various ways. Arguments can be added (Jane sneezed the napkin off the table), and in the reverse case – which will be the focus of the project – an argument that is supposed to be part of a lexical unit’s argument structure can be omitted (Jane has been baking all afternoon). Although no such entity is explicitly expressed, it is intuitively obvious that some pastry was involved in the baking event. But where does this information come from, and how is it licensed or constrained? This question is complicated by the fact that implicit arguments are neither restricted to verbs nor to the complement position of lexical units. The aim of this project is to perform large-scale hypothesis-guided empirical corpus studies in order to deduce whether null instantiation is generalizable at least to some degree.
We adopt a Frame Semantics approach which takes into account the interaction of lexical information, cognitively grounded knowledge, syntactic constructions and pragmatic aspects. The frame-based approach has certain advantages for this research. In particular, it allows us to go beyond individual examples of argument omission and to analyse the behaviour of semantically related lexical units which evoke the same frame (e.g. eat, devour) with respect to the different types of null instantiation they allow (i.e. definite, indefinite, generic or constructional). These analyses are expected to reveal not only idiosyncrasies but also generalizations as to the omissibility of arguments. Moreover, polysemous lexical units will be examined with respect to the omissibility of arguments in their different readings – including metaphorical meaning extensions, which have been largely neglected so far.
Polysemy will be examined contrastively for English and German. Analyses of null instantiation in the context of nominalizations are expected to reveal the degree to which these lexical units exploit (or deviate from) the argument structures of their verbal bases. While the focus is on verbs and nominalizations, analyses of adjectives, prepositions, and adverbs will be initiated. Since the interpretation of implicit information displayed by lexical units of these categories is highly context-dependent, the restrictions they impose on the types of context in which potential antecedents occur and cues for the identification of antecedents will be analysed. The analyses proceed from the FrameNet database, which organizes semantically related lexical units in frames and provides syntactic patterns. Additional queries performed for lexical units
of selected frames in the Concretely Annotated English Gigaword Corpus will enrich our database considerably. The use of this resource is an innovation in research on implicit arguments because it allows us to state search patterns over syntactic trees and thus to find a large number of examples in context in which lexical units occur with or without overt complements