Multipath parsing in the brain

TL;DR

This study combines neuroimaging data with computational models to investigate how the brain processes syntactic ambiguities during real-time language comprehension, providing evidence for multipath parsing in the brain.

Contribution

It introduces a novel approach linking incremental dependency parsing predictions with neuroimaging data to support the multipath parsing hypothesis.

Findings

Evidence for multipath parsing in the brain

Bilateral superior temporal gyrus involved in syntactic ambiguity resolution

Dependency parser surprisal correlates with neural activity during comprehension

Abstract

Humans understand sentences word-by-word, in the order that they hear them. This incrementality entails resolving temporary ambiguities about syntactic relationships. We investigate how humans process these syntactic ambiguities by correlating predictions from incremental generative dependency parsers with timecourse data from people undergoing functional neuroimaging while listening to an audiobook. In particular, we compare competing hypotheses regarding the number of developing syntactic analyses in play during word-by-word comprehension: one vs more than one. This comparison involves evaluating syntactic surprisal from a state-of-the-art dependency parser with LLM-adapted encodings against an existing fMRI dataset. In both English and Chinese data, we find evidence for multipath parsing. Brain regions associated with this multipath effect include bilateral superior temporal gyrus.