CLUNL

Syntactic and lexical factors in processing complexity

Identification

• Project identification: Factores Sintácticos e Lexicais na Complexidade do Processamento [ PTDC/CLE-LIN/114212/2009 ]
• Group: LiFE – Formal and Experimental Linguistics
  
• Principal investigator: Antonino Grillo
• Duration: Feb. 2011 – July 2014
• Funding entity: FCT/MCTES

Description

The project investigates the contribution of syntactic and lexical factors in modulating the processing complexity of different varieties of dependencies (Agreement over embedded material, A and A movement) in unimpaired adult speakers, children acquiring language and agrammatic Brocas aphasics. The project extends the scope and depth of the Generalized Minimality approach to agrammatic aphasia (Grillo 2005, 2008) and the event structure based approach to the representation and acquisition of passivization developed in Gehrke and Grillo (2007, 2008).

A pervasive line of research in psycholinguistics has shown that object relatives are harder to process than subject relatives, as measured by reading time, comprehension accuracy, and eye tracking (King and Just 1991; Gibson 1998, 2000; Gordon et al. 2004; Van Dyke 2007). This asymmetry is often characterized as intervention or locality effects since the subject NP of the relative intervenes between the moved NP and its trace, in the case of object but not subject relatives. The distinction is not limited to healthy processing, as it is similarly found in the language acquisition (Friedmann et al. 2008) and language breakdown (Grodzinsky 2000; Garraffa and Grillo, 2008) literature. This robust distinction that affects multiple domains would benefit from an overarching theoretical understanding. In attempting such a goal, Grillo 2009 has claimed that the asymmetry can be explained as a special case of Relativized Minimality (Rizzi 2004). Several factors contribute in modulating minimality effects in processing, providing preliminary support for a similarity based account: the more similar the intervening NP to the moved one, the higher the complexity of a sentence. Attraction errors “the boat of the mariners leave” (Bock & Miller 1991) display similar properties and call for a unified explanation. A primary goal of this project is to investigate the role that similarity, in terms of grammatical and non-grammatical (e.g. lexical) features between the moved NP and intervening NP(s), plays in modulating minimality effects in both unimpaired adult speakers and agrammatic aphasics and critically whether a distinction is observed between them. Another syntactic complexity contrast observed in healthy subjects, aphasics, and children is Passives compared to Actives. However, unlike in relative clauses where the Cross Modal Priming paradigm (Swinney et al. 1979) has successfully shown that antecedent (i.e., moved NP) reactivation occurs at the trace position, Osterhout and Swinney 1993 found no reactivation of the moved NP at the postverbal position, where traditional analyses of passives place the trace (Baker et al. 1989). Rather, reactivation was found only after the by-phrase. These results question the existence of NP traces in passives and/or the validity of traditional analyses. For independent theoretical reasons, Collins 2005 and Gehrke and Grillo 2007, 2009 have proposed that passives are derived through movement of part of the predicate VP and not only the NP. Among other consequences, this account correctly predicts NP reactivation to occur only after the by-phrase. It is important to re-run the experiments to test for the exact position of the trace with consideration of these different assumptions. This account makes yet another prediction that needs to be tested.
Gehrke and Grillo claim that passives are derived on complex predicates only, and that to passivize states (noncomplex), a computationally costly coercion operation is required, coercing the simple state into an inchoative state, a state having come into existence. They therefore predict that the stative-eventive complexity asymmetry found in actives (Gennari and Poeppel 2003) is reversed or attenuated in passives. The well known asymmetry between comprehension and production of “actional” (~eventive) and “non-actional” (~stative) predicates, in the language acquisition (Borer and Wexler 1987; Fox and Grodzinsky 1998) and aphasia (Grodzinsky 1995) literature, offer preliminary support to this idea. Further evidence especially in the context of passives is required to assess Gehrke and Grillo’s theory of passives. From this state of affairs, the following issues emerge:
1. It must be established if grammatical and non-grammatical factors modulate minimality effects in different ways within unimpaired speakers and across populations;
2. It must be established if, contrary to what was found in the active form, passive of states are harder to process than passives of eventive predicates and whether VP movement approaches are more adequate to explain these results;
3. It must be established if VP movement approaches can predict antecedent reactivation patterns in passives.
This project aims at exploring these three issues through testing unimpaired adults and aphasic speakers. The project benefits from the results achieved by the team in the offline investigation of similar constructions in young children in the project PTDC/CLELIN/099802/2008, the project COST-A33 as well as an international team of researchers specialized in the theoretical-experimental study of these structures.

Understanding the moment-by-moment processing of movement, and how the interface conditions behind it are applied in real time, is crucial to both theoretical and psycho-linguistics and an essential step towards a proper characterization, and an improved treatment, of language impairment such as SLI and aphasia.

Team

Researchers:
Antonino Grillo (principal investigator)
Arnout Koornneef (Utrecht Institute of Linguistics OTS)
Ana Castro
Andrea Nicole Santi (University College of London)
Antónia Estrela
Berit Gehrke (Universitat Autònoma de Barcelona)
Fernanda Pratas
João Costa
Maria Garraffa (Fundação Marica de Vincenzi)
Michael Wagner (McGill University)
Bruno Fernandes (research fellow)
Margarida Tomaz (Linguistics Centre, Universidade de Lisboa, research fellow)

Consultants:
Yosef Grodzinsky
Lewis P. Shapiro