Pre-registered study of neural markers of unexpected speech: Re-evaluating the PMN, N400 and P600 in Malay.

At this stage, it is unclear whether the neurological correlates of a change in register would be qualitatively different from a change in language, or whether both are effectively indexing a generalized outcome of deviation from expectations. While older psycholinguistic interpretations of the PMN, N400 and P600 effects tend to characterise these effects as mismatches in phonology, semantics and syntax, respectively, it is possible that all three effects index a general property of unexpectedness during real-time language processing, with the timeline reflecting how readily the language processing system is able to integrate current information with expectations arising from recent context. Hence, an unexpected speech sound may be evident shortly after the sound is heard, while a semantic mismatch necessarily takes longer to process, as more of the word must be processed before its expectedness/ease of integration with prior context becomes evident. For example, some papers show the probability of a word in a given context predicts the N400, whether measured by cloze probability (e.g. Federmeier and Kutas, 1999; Thornhill & Van Petten, 2012; Kwon, Sturt & Liu, 2017) or distributional information (e.g. Michaelov, Coulson & Bergen, 2022); and other papers show that different kinds of semantic violation are detected on different timescales (e.g., N400 for thematic violations vs P600 for taxonomic violations: Savic, Savic & Kovic, 2019). Within this framework, the ERP recorded when switching from one language to another - or from one register to another - likely indexes the probability of encountering particular linguistic content at a particular time. Languages with discrete registers allow us to investigate the characteristics of expectation-dependent ERP effects (the N400 and P600), by varying the likelihood of particular linguistic content without altering the lexical semantics or the syntax of an utterance. Furthermore, as a switch from the formal register to the informal register is more likely than the obverse in natural linguistic contexts (i.e., ‘slipping’ into the informal), the study allows us to investigate both the type of violation, and the degree.