A Model of Online Temporal-Spatial Integration for Immediacy and Overrule in Discourse Comprehension.

Takahisa Uchida,Hiroshi Ishiguro,Peter Ford Dominey,Nicolas Lair

doi:10.1162/nol_a_00026

Abstract

During discourse comprehension, information from prior processing is integrated and appears to be immediately accessible. This was remarkably demonstrated by an N400 for "salted" and not "in love" in response to "The peanut was salted/in love." Discourse overrule was induced by prior discourse featuring the peanut as an animate agent. Immediate discourse overrule requires a model that integrates information at two timescales. One is over the lifetime and includes event knowledge and word semantics. The second is over the discourse in an event context. We propose a model where both are accounted for by temporal-to-spatial integration of experience into distributed spatial representations, providing immediate access to experience accumulated over different timescales. For lexical semantics, this is modeled by a word embedding system trained by sequential exposure to the entire Wikipedia corpus. For discourse, this is modeled by a recurrent reservoir network trained to generate a discourse vector for input sequences of words. The N400 is modeled as the difference between the instantaneous discourse vector and the target word. We predict this model can account for semantic immediacy and discourse overrule. The model simulates lexical priming and discourse overrule in the "Peanut in love" discourse, and it demonstrates that an unexpected word elicits reduced N400 if it is generally related to the event described in prior discourse, and that this effect disappears when the discourse context is removed. This neurocomputational model is the first to simulate immediacy and overrule in discourse-modulated N400, and contributes to characterization of online integration processes in discourse.

Highlights

An astonishing aspect of human language comprehension is that it brings to bear a vast variety of information sources to the ongoing interpretation of words in language, and it does so without temporal penalties for accessing this information
The model simulates lexical priming and discourse overrule in the “Peanut in love” discourse, and it demonstrates that an unexpected word elicits reduced N400 if it is generally related to the event described in prior discourse, and that this effect disappears when the discourse context is removed
Word embeddings for English in all three experiments were generated using Wikipedia2Vec (Yamada et al, 2020), which learns word embeddings based on the Wikipedia corpus, which includes over 3 billion words in approximately 5.5 million articles

Summary

Introduction

An astonishing aspect of human language comprehension is that it brings to bear a vast variety of information sources to the ongoing interpretation of words in language, and it does so without temporal penalties for accessing this information. Search takes time, when using artificial arrays of stimuli (Treisman, 1982). Under naturalistic conditions, attentional mechanisms are recruited and search time is significantly reduced (Peelen & Kastner, 2014). The comprehension system appears to have immediate access to diverse sources of stored information. This “immediacy assumption” posits that during comprehension, an attempt is made to relate each content word to its referent as soon as possible (Just & Carpenter, 1980). The immediacy assumption was initially developed in the context of reading

Objectives

Results

Conclusion