Establishing reference in language comprehension: An electrophysiological perspective

Abstract

The electrophysiology of language comprehension has long been dominated by research on syntactic and semantic integration. However, to understand expressions like "he did it" or "the little girl", combining word meanings in accordance with semantic and syntactic constraints is not enough-readers and listeners also need to work out what or who is being referred to. We review our event-related brain potential research on the processes involved in establishing reference, and present a new experiment in which we examine when and how the implicit causality associated with specific interpersonal verbs affects the interpretation of a referentially ambiguous pronoun. The evidence suggests that upon encountering a singular noun or pronoun, readers and listeners immediately inspect their situation model for a suitable discourse entity, such that they can discriminate between having too many, too few, or exactly the right number of referents within at most half a second. Furthermore, our implicit causality findings indicate that a fragment like "David praised Linda because…" can immediately foreground a particular referent, to the extent that a subsequent "he" is at least initially construed as a syntactic error. In all, our brain potential findings suggest that referential processing is highly incremental, and not necessarily contingent upon the syntax. In addition, they demonstrate that we can use ERPs to relatively selectively keep track of how readers and listeners establish reference. IntroductionWords reliably mean things. In fact, something about word meaning is invariant enough to be listed in a dictionary. Here, for example, is what the Cambridge Advanced Learners Dictionary lists for the word "girl":girl (noun): 1. a female child or young woman, especially one still at school: "Two girls showed us round the classrooms."Of course, we all know that when a word is placed in the context of other words, shades of meaning can emerge. For "girl", some of the most familiar ones are listed in the dictionary too: girl (noun): ……… 2. a daughter: "We have two girls.", "My little girl is five." 3. [usually plural] a woman worker, especially when seen as one of a group: "shop/office girls" 4. [always plural] one of a group of female friends: "I'm going out with the girls tonight.", "The girls at work gave it to me"

On sense and reference: Examining the functional neuroanatomy of referential processing

“…In line with earlier ERP research (e.g., Nieuwland and Van Berkum, 2006a;Van Berkum et al, 2007), the present findings thus show that the language comprehension system relies on different neural mechanisms to deal with referential ambiguity, referential failure and semantic anomaly. Moreover, a closer examination of relative activation increases and decreases across different brain regions for all conditions revealed an intriguing pattern of results.…”

“…Referential failure and sentence-external anaphoric inferences Also consistent with our predictions, referentially failing pronouns elicited increased BOLD responses in medial and bilateral parietal, and left middle frontal brain regions (e.g., Hammer et al, 2007;Kuperberg et al, 2003;Newman et al, 2001;Ni et al, 2000), supporting the notion that readers initially ascribed referential failure to a problem with the morpho-syntactic gender of the pronoun (e.g., Osterhout and Mobley, 1995;Van Berkum et al, 2007).…”

“…These regions include medial frontal (BA 8/9/10) and parietal (BA 7/23/31) cortices (e.g., Ferstl andvon Cramon, 2001, 2002;Ferstl et al, 2005;Kuperberg et al, 2006;Maguire et al, 1999;Volz et al, 2006a,b;Zysset et al, 2002Zysset et al, , 2003, but also right lateral prefrontal (BA 8/9/ 10/46) and bilateral parietal regions (BA 39/40) involved in monitoring and evaluating of memory retrieval outcomes (e.g., Buckner and Wheeler, 2001;Wagner et al, 2005). 1 For referential failure (compared to referential coherence), we predicted that the brain treats referentially failing pronouns as morpho-syntactic violations in the first instance (e.g., Osterhout and Mobley, 1995;Van Berkum et al, 2007), eliciting activity increases in left middle/superior and inferior frontal (BA 6/8/44, Hammer et al, 2007;Newman et al, 2001;Ni et al, 2000), and medial and bilateral parietal regions (BA 7/23/31/40; Kuperberg et al, 2003;Ni et al, 2000). However, because the pronouns as used here are 'free' in a Chomskian sense (Chomsky, 1981), there are actually alternative resolutions of the referential failure at hand.…”

Event-related brain potentials index cue-based retrieval interference during sentence comprehension

“…Given that the gender of the determiner (-o/-a) was grammatically correct or incorrect given that of the antecedent (camiseta), grammatically incorrect determiners in our study may thus initially be perceived as a morphosyntactic "dead end", triggering re-analysis (Gouvea et al, 2010;Kaan and Swaab, 2003;Osterhout and Holcomb, 1992). Perhaps for similar reasons, pronouns with unexpected or incorrect gender-marking will also evoke P600 effects (e.g., Nieuwland and Van Berkum, 2006;Osterhout and Mobley, 1995;Van Berkum et al, 2007;Xiang et al, 2009).…”

“…Critically, under any neurocognitive account of sentence comprehension (Bornkessel-Schlesewsky and Schlesewsky, 2009a;Friederici, 2002;Hagoort, 2005), this outright syntactic violation should have evoked a P600 effect, which is the ERP effect most commonly associated with syntactically problematic or dispreferred linguistic expressions Van Berkum et al, 2007;Kutas et al, 2006). Moreover, a multitude of electrophysiological studies have reported P600 effects for related gender mismatch violations across different languages (Barber and Carreiras, 2005;Molinaro et al, 2011;Wicha et al, 2004;Van Berkum et al, 2007;Xiang et al, 2009). The fact that the main manipulation in the current study elicited a sustained negativity rather than a P600 effect suggests that the processing consequences of incorrectly gender-marked ellipsis can be qualitatively different from the repair or reanalysis processes as assumed to be indexed by P600 modulations, and perhaps that participants did not treat failure to retrieve an antecedent as a consequence of the gender agreement violation.…”