Isolating the Effects of Word’s Emotional Valence on Subsequent Morphosyntactic Processing: An Event-Related Brain Potentials Study

Abstract

Emotional information significantly affects cognitive processes, as proved by research in the past decades. Recently, emotional effects on language comprehension and, particularly, syntactic processing, have been reported. However, more research is needed, as this is yet very scarce. The present paper focuses on the effects of emotion-laden linguistic material (words) on subsequent morphosyntactic processing, by using Event-Related brain Potentials (ERP). The main aim of this paper is to clarify whether the effects previously reported remain when positive, negative and neutral stimuli are equated in arousal levels and whether they remain long-lasting. In addition, we aimed at testing whether these effects vary as a function of the task performed with the emotion-laden words, to assess their robustness across variations in attention and cognitive load during the processing of the emotional words. In this regard, two different tasks were performed: a reading aloud (RA) task, where participants simply read aloud the words, written in black on white background, and an Emotional Stroop (ES) task, where participants named the colors in which the emotional words were shown. After these words, neutral sentences followed, that had to be evaluated for grammaticality while recording ERPs (50% containing a morphosyntactic anomaly). ERP analyses showed main effects of valence across tasks on the two components reflecting morphosyntactic processing: The Left Anterior Negativity (LAN) is increased by previous emotional words (more by negative than positive) relative to neutral ones, while the P600 is similarly decreased. No interactions between task and valence were found. As a result, an emotion-laden word preceding a sentence can modulate the syntactic processing of the latter, independently of the arousal and processing conditions of the emotional word.

The Automatic but Flexible and Content-Dependent Nature of Syntax

“…In line with the present results, previous studies also found that positive and negative emotional stimuli behaved similarly at early stages among them, and often found differences when compared to neutral information (Jiménez-Ortega et al, 2012;Espuny et al, 2018a;Padron et al, 2020). In contrast, other studies found differences in the early components to syntactic anomalies between positive and negative emotional information (Martin-Loeches et al, 2012;Jiménez-Ortega et al, 2017;Espuny et al, 2018b). Particularly, in Jiménez-Ortega et al (2017), emotional masked adjectives, also embedded within a sentence, preceded an unmasked neutral adjective containing a gender or a number anomaly.…”

“…Yet, a larger early syntactic component and a reduced P600 have been previously observed for good vs. poor comprehenders (Coulson and Kutas, 2001). Furthermore, similarly to the present study, the same pattern appears when emotional words precede sentences containing a morphosyntatic violation, compared to neutral information (Espuny et al, 2018b). The increased early component/reduced P600 pattern is therefore generally interpreted as a result of a more efficient syntactic information processing (Tanner and Van Hell, 2014).…”

“…However, how emotional information affects syntax is unclear. While some studies observed a similar LAN component between positive/pleasant and negative/unpleasant information (Jiménez-Ortega et al, 2012;Espuny et al, 2018a;Padron et al, 2020), others observed differences at early components to syntactic anomalies between positive and negative information (Martin-Loeches et al, 2012;Jiménez-Ortega et al, 2017;Espuny et al, 2018b). This variability might depend on both the arousal and valence of the emotional material (Citron et al, 2013;Hinojosa et al, 2020).…”

“…For example, moderately arousing pleasant and unpleasant adjectives containing syntactic anomalies with respect to an ongoing sentence have been found to trigger similar LAN components (Padron et al, 2020). In contrast, Espuny et al (2018b), using also moderately arousing words, observed that negative valanced words triggered a larger LAN than positive and neutral valanced words. Nonetheless, in this case, the emotional words were presented preceding sentences containing gender and number anomalies instead within the sentence.…”

Different Influences of Negative and Neutral Emotional Interference on Working Memory in Trait Anxiety

“…Previous similar study also observed no impaired performance in the HA group (Zhang et al, 2019 ). We suggest that this lack of impairment in the group may be due to the following reasons: first, university students are likely high achieving and less representative of general population; second, HA participants may take strategies to complete the task (Espuny et al, 2018 ; Zhang et al, 2019 ) and avoid loss when the task performance may be affected by the interference.…”