Heavy NP shift is the parser’s last resort: Evidence from eye movements☆

Abstract

Two eye movement experiments explored the roles of verbal subcategorization possibilities and transitivity biases in the processing of heavy NP shift sentences in which the verb's direct object appears to the right of a post-verbal phrase. In Experiment 1, participants read sentences in which a prepositional phrase immediately followed the verb, which was either obligatorily transitive or had a high transitivity bias (e.g., Jack praised/watched from the stands his daughter's attempt to shoot a basket). Experiment 2 compared unshifted sentences to sentences in which an adverb intervened between the verb and its object, and obligatorily transitive verbs to optionally transitive verbs with widely varying transitivity biases. In both experiments, evidence of processing difficulty appeared on the material that intervened between the verb and its object when the verb was obligatorily transitive, and on the shifted direct object when the verb was optionally transitive, regardless of transitivity bias. We conclude that the parser adopts the heavy NP shift analysis only when it is forced to by the grammar, which we interpret in terms of a preference for immediate incremental interpretation. KeywordsSentence comprehension; Parsing; Heavy NP shift; Argument structure; Eye movements Consider the comprehension of sentences in which a direct object is separated from the verb by intervening material, as in (1) below: (1) Lucy ate t i with a fork [the extremely delicious, bright green broccoli] i . This construction is known as heavy NP shift. According to the classic syntactic account (Ross, 1967; but cf. Kayne, 1998;Rochemont & Culicover, 1997), heavy NP shift involves a movement operation that displaces the verb's direct object from its underlying position adjacent to the verb, where it receives both accusative case and a thematic role, to an adjoined position to the right of other verbal arguments and/or adjuncts. As in other movement operations, a trace of the moved constituent remains in the pre-movement position. Notably, heavy NP shift is typically able to apply only when the direct object noun phrase is relatively long or "heavy" relative to the intervening constituent, as evidenced by the comparative awkwardness of (2a) and the unacceptability of (2b):

Minimality as vacuous distinctness: Evidence from cross-linguistic sentence comprehension

“…These results are in line with a range of observations from several languages suggesting that verb transitivity information crucially determines the processing of arguments encountered in post-verbal positions (e.g. Ferreira and McClure, 1997;Pablos, 2006;Staub et al, 2006;Staub, 2007). However, Pickering et al (2000) provided compelling evidence for a general direct object preference in the processing of NP-S ambiguities, by demonstrating effects of direct object plausibility in the ambiguous region even for verbs that were biased towards an intransitive reading.…”

Learning to order words: A connectionist model of heavy NP shift and accessibility effects in Japanese and English

“…In production, Stallings et al (1998) found that heavy NP shift was lower with verbs which had high expectations for a noun phrase complement compared to more promiscuous verbs. In comprehension, Staub, Clifton, and Frazier (2006) found that pairings of verbs that were biased towards noun phrase complements (transitive-biased condition) with post-verbal adjuncts or prepositional phrases that signaled heavy NP shift did not lead to immediate slowing due to reanalysis. Therefore, comprehenders were not biased to expect heavy NP shift in these ambiguous situations and verbs which might provide clear signals for heavy NP shift in language learning did not appear more in shifted structures.…”

Expectation and Locality Effects in German Verb-final Structures

“…Although earlier work documenting these effects focused primarily on expectations at the lexical level, more recent work has provided evidence for expectations at the level of syntactic constituency on the basis of grammatical analysis of prior linguistic content (Lau, Stroud, Plesch & Phillips, 2006;Staub, Clifton & Frazier, 2006;Jaeger, Fedorenko, Hofmeister & Gibson, 2008;Levy, Fedorenko, Breen & Gibson, 2012).…”