Morpho-syntactico-semantic parafoveal processing: Eye-tracking evidence from word n+ 1 and word n in Russian.

Anastasia Stoops, Jack Dempsey, Kiel Christianson

Abstract: Two experiments compared morpho-syntactico-semantic parafoveal processing of five-letter words n + 1 (Experiment 1) with five-letter regions at the end of longer words n (Experiment 2), understudied cross-linguistically. Earlier boundary-change studies showed that subject/object case assignment in Russian can be extracted from a parafoveally presented but never directly fixated letter when the related preview is the most expected continuation (Stoops & Christianson, 2017, 2019). This study reversed the syntactic expectations for the identical and related previews (Cloze ratings: 94% grammatical identical object vs. 0% ungrammatical related subject). The related preview was read more slowly than the no-change preview in the later measures: go-past for the words n + 1 and n, according to both frequentist and Bayesian analyses. Additionally, the study clarifies the augmented allocation of attention hypothesis—skilled readers process parafoveally visible parts of a longer word faster than length-controlled upcoming word n + 1, yet the message-level contextual linguistic information affected the target words n and n + 1 similarly. The most intriguing finding is the delayed morpho-syntactico-semantic effect: even though the morphologically ungrammatical marking was parafoveally available, the syntactic fit only affected delayed processing, manifested as increased reading of previous text. More cross-linguistic work is needed to understand the role of higher level linguistic information beyond the predictability of individual lexical items on parafoveal processing during reading.