Eye-Tracking Study Shows Nominalizations Make Children Reread to Answer Correctly

Event nominalizations—structures that compress complex events into nominal form, increasing informational density while reducing the transparency of thematic-role assignment—are a central feature of academic language. They pack more information into fewer words, but they also hide who did what to whom. A new eye-tracking study from the University of Verona examines how children learning to read handle this trade-off, and the results show that even when young readers answer comprehension questions correctly, the way their eyes move through the text reveals extra processing effort beneath the surface.

The study, published in Frontiers in Language Sciences, recruited 79 fifth- and sixth-grade children from three groups: 28 typically developing monolingual Italian children, 21 monolingual Italian children with developmental dyslexia, and 30 typically developing bilingual children with a migrant background for whom Italian was the societal language. Each child silently read short passages in which the same target event was expressed three ways: as a nominalization, as a passive construction, or as an active construction. After each passage, the children answered a comprehension question while their eye movements were continuously recorded.

The researchers analyzed three kinds of data: comprehension accuracy, standard eye-movement measures such as fixation times, and scanpaths—the ordered sequences of fixations that trace how a reader’s gaze travels across the text.

The headline finding concerns accuracy versus process. Nominalizations did not significantly reduce how often children answered the comprehension questions correctly. Yet they clearly changed how the children arrived at those answers. During the initial reading of the passages, scanpath dissimilarity was higher for nominalizations than for either verbal construction, meaning children’s gaze patterns became more varied and less uniform when the event was packed into a noun.

The effect persisted into the question-answering phase. In trials where children answered correctly, scanpath dissimilarity was again higher for nominalizations than for passives. More strikingly, nominalizations increased the probability of partial rescanning and decreased the probability of minimal rescanning relative to active texts. In plain terms, children answering questions about nominalized events were more likely to go back and reread parts of the passage before committing to a correct answer, rather than relying on a single, efficient pass through the text.

Correct answers, in other words, do not mean effortless reading.

The three reader groups did not behave identically. Children with developmental dyslexia showed a generalized increase in processing effort, visible in both the online reading of the passages and the question-answering phase. Bilingual children displayed an intermediate profile, with increased reliance on rereading during question-answering.

The design compared three expressions of the same events within the same children, which strengthens the claim that the construction type itself—not the underlying content—drove the processing differences. Still, the measures capture online behavior and immediate comprehension questions; the authors’ conclusions concern processing patterns during reading and answering, not long-term learning outcomes.

The broader significance lies in what the results imply about academic language. The study suggests that comprehension assessments based on accuracy alone may underestimate the effort that complex morphosyntax demands of developing readers, and that eye-movement measures can expose the hidden work behind seemingly correct answers.

Subject of Research: How developing readers, including bilingual children and children with dyslexia, process event nominalizations during reading comprehension.

Article Title: Processing event nominalizations in developing readers: eye-tracking evidence from bilingual and dyslexic readers

Article References: Vender, M., Venagli, I., Nardon, A., Piccinin, S., Dal Maso, S., & Melloni, C. (2026). Processing event nominalizations in developing readers: eye-tracking evidence from bilingual and dyslexic readers. Frontiers in Language Sciences, 5, Article 1912839. https://doi.org/10.3389/flang.2026.1912839

Image Credits: AI Generated

DOI: 10.3389/flang.2026.1912839

Keywords: eye-tracking, reading comprehension, nominalizations, developmental dyslexia, bilingual children, developing readers, scanpaths, morphosyntactic complexity, question answering, Italian children, academic language, Processing