To what extent can statistical language knowledge account for the effects of world knowledge in language comprehension? We address this question by focusing on a core aspect of language understanding: pronoun resolution. While existing studies suggest that comprehenders use world knowledge to resolve pronouns, the distributional hypothesis and its operationalization in large language models (LLMs) provide an alternative account of how purely linguistic information could drive apparent world knowledge effects. We addressed these confounds in two experiments. In Experiment 1, we found a strong effect of world knowledge plausibility (measured using a norming study) on responses to comprehension questions that probed pronoun interpretation. In experiment 2, participants were slower to read continuations that contradicted world knowledge-consistent interpretations of a pronoun, implying that comprehenders deploy world knowledge spontaneously. Both effects persisted when controlling for the predictions of GPT-3, an LLM, suggesting that pronoun interpretation is at least partly driven by knowledge about the world and not the word. We propose two potential mechanisms by which knowledge-driven pronoun resolution occurs, based on validation- and expectation-driven discourse processes. The results suggest that while distributional information may capture some aspects of world knowledge, human comprehenders likely draw on other sources unavailable to LLMs.

We tested whether the proportion of typical sentences in a series of auditory sentences would lead people to adjust the strength of activation of world knowledge (i.e., retrieval rules adaptation) during comprehension. This issue is important because it could help clarify how people efficiently integrate different memory information in cognitive processes. In two experiments, all task materials were presented to participants as a whole package, in which proportions of typical sentences, with typical final locations, varied under different conditions. In Experiment 1, the proportion of typical sentences was equal to the atypical ones (i.e., 50% typical vs. 50% atypical), whereas in Experiment 2, the proportion of typical sentences was not equal to the atypical ones (i.e., 75% typical vs. 25% atypical, and 25% typical vs. 75% atypical). Visual fixation on the critical area in a visual display before/while hearing the critical words was compared across conditions, and across-condition differences were used as an index of the adaptation of the retrieval rule in the activation of world knowledge. The findings indicated that the adaptation of retrieval rules occurs throughout the whole test package of sentence comprehension, and the strength of activation of world knowledge in sentence comprehension can be adjusted.

In this study we explore whether world knowledge (WK) processing differs between individuals listening to their native (L1) or their non-native (L2) language. We recorded event-related brain potentials in L1 and L2 speakers of Spanish while they listened to sentences uttered by native speakers of Spanish. Sentences were either congruent or incongruent with participants’ WK. In addition, participants also listened to sentences in which upcoming words could not be anticipated on the basis of WK. WK violations elicited a late negativity of greater magnitude and duration in the L2 than the L1 group. However, sentences in which WK was not helpful regarding word anticipation elicited similar N400 modulations in both groups. These results suggest that WK processing requires a deeper lexical search in L2 comprehension than in L1 comprehension.