Across the Levels of Analysis: Explaining Predictive Processing in Humans Requires More Than Machine-Estimated Probabilities

2026-04-10

AI Summary

This study addresses the prevailing tendency in current research to rely excessively on the output probabilities of large language models (LLMs) as proxies for human predictive language processing, often overlooking the theoretical distinctions among Marr’s three levels of analysis—computational, algorithmic, and implementational. Critiquing and extending two dominant perspectives, this work proposes moving beyond mere probability matching by integrating LLMs with psycholinguistic models across all three Marr levels. By constructing a cross-level explanatory framework, the research not only clarifies the limitations of LLMs in simulating human language processing but also offers a systematic pathway for future studies aiming to unify computational and cognitive models of language.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Cognitive Modeling & Cognitive Systems: Analogy

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Under the lens of Marr's levels of analysis, we critique and extend two claims about language models (LMs) and language processing: first, that predicting upcoming linguistic information based on context is central to language processing, and second, that many advances in psycholinguistics would be impossible without large language models (LLMs). We further outline future directions that combine the strengths of LLMs with psycholinguistic models.

Problem

• predictive processing
• language models
• psycholinguistics
• Marr's levels of analysis
• human language processing

Innovation

• predictive processing
• language models
• Marr's levels of analysis
• psycholinguistics
• cognitive modeling