Researchers have proposed a new model for how humans maintain information during sentence comprehension, suggesting that syntactic structure guides selective information retention. This model posits that the cost of maintaining information is influenced by the number of predicted heads and incomplete dependencies, which are not reducible to each other. Experiments using a Japanese reading time dataset support this hypothesis, revealing a trade-off where readers who invest more in maintenance benefit more from predictability. AI
Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →
IMPACT Provides a cognitive model that could inform the design of more efficient natural language processing systems.
RANK_REASON This is a research paper published on arXiv detailing a new hypothesis and experimental findings on sentence comprehension.