When the polar bear encounters many polar bears: event-related potential context effects evoked by uniqueness failure

Abstract

When essential context information is not given in a discourse, understanding mostly fails. By means of event-related brain potentials we investigated what happens when the definite determiner does not find a unique referent in the context or when the appropriate use of the indefinite determiner is violated by the presence of an already introduced referent. The misuse of the determiners evoked an N400/P600 complex immediately upon reading the determiners. Since this result was the same for both determiners, we suggest that initial cognitive processes are similar and include the detection of a mismatch, triggering an N400, followed by a top-down monitoring process, indexed by a P600. This latter process presumably involves the evaluation of possible alternative sentence interpretations which might preserve the sentence's meaning. We hypothesise that theoretically assumed processing differences between the two determiners might occur at a later time point in processing.

Keywords:

• discourse processing
• presupposition
• uniqueness
• definiteness
• ERP

Acknowledgements

This study was supported by the German Research Foundation (SFB 833, Project B2). We thank Anna-Elisabeth Burkard for her help in data acquisition.

Notes

1. We are aware of the disagreement in the linguistic literature as to whether the definite determiner functions as a PSP trigger (Heim, 1991; Krahmer, 1998) or whether its assumptions are merely part of the assertion (Russell, 1905; see also Abbott for the Strawson-Russell debate, 2010). Similar views also pertain to the indefinite determiner (Alonso-Ovalle et al., 2009; Heim, 1982, 1991; Krahmer, 1998). However, since the pre-suppositional view on definites is in the majority in the literature, for the sake of simplicity, we have categorised the definite determiner and the indefinite determiner as PSP triggers and focused on the processing differences between these two determiners.

2. Since several authors used slightly different and longer time windows to calculate the amplitudes of the N400, the Nref, or the LAN, we performed an additional analysis on the time interval capturing mean amplitudes of the time interval 300–500 ms following the trigger. This analysis revealed the same pattern of results as reported for the time interval of 350–450 ms.

3. To make sure that the component which we found really corresponds to the N400 component associated with semantic difficulties, we compared the synonyms and the repetitions of the NPs in the test sentences and contexts to evoke the well-known semantic priming effect (Anderson & Holcomb, 2005; Bentin, 1989; Rolke et al., 2001). Since both N400 components resemble each other strongly in localisation and temporal onset, we are certain that the component found for the mismatching condition also reflects the N400 associated with semantic processes.

4. We analysed ERPs evoked by all words in the test sentences to look for potential later effects which might come along with different PSP processing of the definite and the indefinite determiner, respectively. There were, however, no meaningful effects as all interactions of Matching × Definiteness or interactions with these two factors were above the significance level (N400: p > .09; P600: p > .18).