In Search of Missing Arguments: A Linguistic Approach

Proceedings of Recent Advances in Natural Language Processing, pages 331–338, Hissar, Bulgaria, 12-14 September 2011. In Search of Missing Arguments: A Linguistic Approach Josef Ruppenhofer Dept. of Information Science and Language Technology Hildesheim University ruppenho@uni-hildesheim.de Philip Gorinski and Caroline Sporleder Computational Linguistics Saarland University {philipg,csporled}@coli.uni-saarland.de Abstract Semantic argument structures are often in- complete in that core arguments are not lo- cally instantiated. However, many of these implicit arguments can be linked to ref- erents in the wider context. In this pa- per we explore a number of linguistically motivated strategies for identifying and re- solving such null instantiations (NIs). We show that a more sophisticated model for identifying definite NIs can lead to notice- able performance gains over the state-of- the-art for NI resolution. 1 Introduction Semantic Role Labeling (SRL) is traditionally concerned with identifying the overtly realized ar- guments of a predicate. However, in a natural dis- course only a relatively small proportion of the theoretically possible semantic arguments tend to be locally instantiated in the same clause or sen- tence that contains the target predicate. The other arguments are so-called null instantiations (NIs). Even core arguments of a predicate, i.e., those that express participants which are necessarily present in the situation which the predicate evokes (see Section 2 for a more detailed explanation of core vs. peripheral arguments), are frequently not in- stantiated in the local context. While null instan- tiated arguments are not locally realized, they can often be inferred from the context. Consider examples (1) and (2) below (taken from Arthur Conan Doyle’s “The Adventure of Wisteria Lodge” and part of the SemEval-10 Task- 10 corpus (Ruppenhofer et al., 2010)). We use A and B in the examples to indicate speakers.1 In a frame-semantic analysis of (1) interesting evokes the Mental stimulus stimulus focus 1We provide this information for clarity, it is not explicitly marked in the corpus. (Mssf) frame. This frame has two core semantic arguments, EXPERIENCER and STIMULUS, as well as eight peripheral arguments, such as TIME, MANNER, DEGREE. Of the two core arguments, neither is actually realized in the same sentence. Only the peripheral argument DEGREE (DEG) is instantiated and realized by most. To fully un- derstand the sentence, it is necessary to infer the fillers of the EXPERIENCER and STIMULUS roles, i.e., the reader needs to make an assumption about what is interesting and to whom. For humans this inference is easy to make as the EXPERIENCER (EXP) and STIMULUS (STIM) roles are actually filled by he and a white cock in the previous sentence. (Note that the two utterances in (1) are spoken by the same person.) Similarly, in (2) right evokes the Correctness (Corr) frame, which has four core arguments, only one of which is filled locally, namely SOURCE (SRC), which is realized by You (and co-referent with Mr. Holmes). However, another argument, INFORMATION (INF), is filled by the preceding sentence (spoken by a different speaker, namely Holmes), which provides details of the fact about which Holmes was right. (1) …

Content truncated to 500 words