Portal de Periódicos da CAPES

Generalized Event Knowledge in Logical Metonymy Resolution

2011; Wiley; Volume: 33; Issue: 33 Linguagem: Inglês

1551-6709

Alessandra Zarcone, Sebastian Padó,

linguistics and terminology studies

Generalized Event Knowledge in Logical Metonymy Resolution Alessandra Zarcone (alessandra.zarcone@ims.uni-stuttgart.de) Institut fur Maschinelle Sprachverarbeitung, Universitat Stuttgart, Azenbergstr. 12 70174 Stuttgart, Germany Sebastian Pado (pado@cl.uni-heidelberg.de) Seminar fur Computerlinguistik, Universitat Heidelberg, Im Neuenheimer Feld 325 69120 Heidelberg, Germany Abstract metonymic contexts. Being part of the mental lexicon, qualia are supposed to model linguistic knowledge – in opposition to world knowledge and pragmatic inferences. Psycholinguistic work has identified experimental correlates for the lexical hypothesis (see Pylkkanen and McElree (2006) for a review), however processing studies have focused predominantly on the existence (or non-existence) of the type mismatch and have largely ignored the question of the range of CEs accessible to metonymic interpretation. It is clearly plausible to assume that we associate some sort of event knowledge with lexical items referring to entities, and that this knowledge is part of our mental lexicon. However, the particular account provided by the lexical hypothesis seems to undergenerate the range of potential interpretations for logical metonymy. Consider the following examples: The interpretation of logical metonymies like begin the book has traditionally been explained by assuming the existence of complex lexical entries containing information about event knowledge (qualia roles: reading the book/writing the book). Qualia structure provides concrete constraints on interpretation, which are however too rigid to be cognitively plausible. We suggest generalized event knowledge as an alternative source of interpretation. Results from a first self-paced reading ex- periment, where we capitalize on the verb-final word order in German subordinate phrases to create rich expectations for events, are presented to support this hypothesis. Consequences of this hypothesis for the interpretation logical metonymies are (a), it is primarily driven by pragmatic and world knowledge; (b), it may use the same (rather than distinct) mechanisms and resources as general incremental sentence comprehension does. Keywords: Logical metonymy; generalized event knowledge; qualia structure; self-paced reading. Logical metonymy: lexicon vs. pragmatics Logical metonymy is a phenomenon where an event- subcategorizing verb is combined an apparently incompatible object-denoting nominal object (John began the book, Mary enjoyed the beer). The interpretation of logical metonymy in- volves the integration of an implicit event (covert event or CE: John began the book → John began reading the book; Mary enjoyed the beer → Mary enjoyed drinking the beer). This process has proven to be a topic of enduring interest to differ- ent language research communities: on the theoretical side, as a challenge to the compositionality principle in meaning construction (Partee, ter Meulen, & Wall, 1993); and on the psycholinguistic side, due to its link to increased processing costs (McElree, Traxler, Pickering, Seely, & Jackendoff, 2001; Traxler, Pickering, & McElree, 2002). In theoretical linguistics, where logical metonymy has been investigated notably by Pustejovsky (1995) and Jackendoff (1997), an account has been offered which we call the “lexical hypothesis”: Logical metonymy is a type mismatch between the (semantic) subcategorization of a metonymic verb for an event and the entity denoted by the object on the other side, which requires the integration of an event to be resolved. 1 The event is retrieved from the lexical entry of the object: each noun is associated with a complex structure (qualia structure) in the mental lexicon. It notably contains an “agentive quale” (the event that brings about the object) and a “telic quale” (the main purpose of the object), which can be retrieved as CEs in 1. My goat eats anything. He really enjoys your book. (→ eating) (Lascarides & Copestake, 1998) 2. John is a famous wrestler. He really enjoys a good fight. (→ fighting) 3. John is a wrestling fan. He really enjoys a good fight. (→ watching) Sentence 1 cannot be accounted for, since the qualia structure for book does not include eating. For sentences 2 and 3, we can probably assume that the nominal entry for fight lists both interpretations: fighting as the agentive quale, and watching as the telic quale, but it is difficult to imagine how the lexical knowledge provided by qualia can “select” the role played by the agent (wrestler, fan) in selecting one CE interpretation. An alternative approach, which we call the “pragmatic hy- pothesis” (Fodor & Lepore, 1998; Asher, 2010), argues that metonymy resolution is driven by dynamic inferences based on context and world knowledge rather than static lexicon entries. This take on metonymy grounds the retrieval of a CE in a post-access inferential process which would have a later onset than lexical access. This is an attractive hypothesis that combines well with insights about human language processing in general, and has been verified experimentally (De Almeida & Dwivedi, 2008; Almeida et al., 2009). However, it currently lacks a concrete characterization of the type and organization of knowledge involved in metonymy interpretation. The aim of this paper is contribute to making the pragmatic hypothesis more precise by investigating metonymy in the con- text of the recent framework generalized event knowledge, or 1 The term ’metonymy’ is used since the coercion of the object into an event is seen as an extension of its meaning.