Surprisal Derives the Recent Filler Heuristic in Mildly Context Sensitive Grammars
نویسنده
چکیده
This paper provides a new account for why online processing of filler-gap relative clause dependencies is more difficult in cases where filler-gap interacts with object control than in cases involving subject control, as reported by Frazier et al. (1983). Frazier et al. (1983) argued for a Recent Filler heuristic in which the parser expects to discharge the most recent filler at every gap site. We observe that statistical subcategorization preferences on the control verb and the embedded verb 'sing' interact , favoring subject control disambiguation. We employ surprisal (Hale, 2001) as a complexity metric on filler-gap structures by construing control as a Movement operation in Minimalist Grammars (Stabler, 1997). We obtain greater surprisals for the Distant Filler condition, deriving the prediction that the Recent Filler heuristic falls out from statistical subcategorization preferences.
منابع مشابه
A Polynomial Time Algorithm for Inferring Grammars for Mildly Context Sensitive Languages
Natural languages are largely context-sensitive, yet context-sensitive grammars cannot be identified in the limit from positive examples [Gold, 1967]. Mildly context-sensitive languages are able to express the most common context-sensitive structures found in natural language. We restrict our view to a class of mildly context-sensitive languages which can be described by simple external context...
متن کاملLambek Grammars, Tree Adjoining Grammars and Hyperedge Replacement Grammars
Two recent extension of the nonassociative Lambek calculus, the LambekGrishin calculus and the multimodal Lambek calculus, are shown to generate class of languages as tree adjoining grammars, using (tree generating) hyperedge replacement grammars as an intermediate step. As a consequence both extensions are mildly context-sensitive formalisms and benefit from polynomial parsing algorithms.
متن کاملInferring Grammars for Mildly Context Sensitive Languages in Polynomial-Time
Natural languages contain regular, context-free, and contextsensitive syntactic constructions, yet none of these classes of formal languages can be identified in the limit from positive examples. Mildly context-sensitive languages are able to represent some context-sensitive constructions, those most common in natural languages, such as multiple agreement, crossed agreement, and duplication. Th...
متن کاملSCTAG: A Mildly Context-Sensitive Formalism for Modeling Complex Intentions in Spatially Structured Environments
The way we represent intentions, behaviors, and the spatial context, is crucial for any approach to mobile intention recognition. Formal grammars are cognitively comprehensible and make expressiveness properties explicit. By adding spatial domain knowledge to a grammar we can reduce parsing ambiguities. We argue that there are a number of mobile intention recognition problems which require the ...
متن کاملMildly Context-Sensitive Languages via Buffer Augmented Pregroup Grammars
A family of languages is called mildly context-sensitive if – it includes the family of all -free context-free languages; – it contains the languages • {anbncn : n ≥ 1} – multiple agreement, • {ambncmdn : m,n ≥ 1} – crossed dependencies, and • {ww : w ∈ Σ+} – reduplication; – all its languages are semi-linear; and – their membership problem is decidable in polynomial time. In our paper we intro...
متن کامل