Position Paper: Logical Soundness Is Not a Reliable Criterion for Neurosymbolic Fact-Checking with LLMs

Forum topic · 小凯 · 2026-04-07

Summary

This position paper by Jason Chan, Robert Gaizauskas, and Zhixue Zhao argues that formal logic is structurally inadequate as the sole criterion for neurosymbolic fact-checking systems that use large language models (LLMs). Many such systems translate natural language claims into logical formulae and verify whether conclusions can be validly derived from verified premises. The authors show that this approach fails to detect misleading claims because logical soundness systematically diverges from the inferences humans typically make and accept. Drawing on cognitive science and pragmatics research, the paper presents a typology of cases where logically sound conclusions elicit human inferences unsupported by the underlying premises. As a remedy, the authors advocate a complementary approach: treating LLMs' human-like reasoning tendencies as a feature rather than a bug, and using LLMs to validate the outputs of formal components in neurosymbolic pipelines against potentially misleading conclusions.

Position: Logical Soundness Is Not a Reliable Criterion for Neurosymbolic Fact-Checking with LLMs

Field: NLP Authors: Jason Chan, Robert Gaizauskas, Zhixue Zhao

Abstract

As large language models (LLMs) are increasingly integrated into fact-checking pipelines, formal logic is often proposed as a rigorous means by which to mitigate bias, errors and hallucinations in these models' outputs. For example, some neurosymbolic systems verify claims by using LLMs to translate natural language into logical formulae and then checking whether the proposed claims are logically sound, i.e. whether they can be validly derived from premises that are verified to be true.

We argue that such approaches structurally fail to detect misleading claims due to systematic divergences between conclusions that are logically sound and inferences that humans typically make and accept. Drawing on studies in cognitive science and pragmatics, we present a typology of cases in which logically sound conclusions systematically elicit human inferences that are unsupported by the underlying premises.

Consequently, we advocate for a complementary approach: leveraging the human-like reasoning tendencies of LLMs as a feature rather than a bug, and using these models to validate the outputs of formal components in neurosymbolic systems against potentially misleading conclusions.

Key Points

Formal logic is commonly proposed as a safeguard against bias, errors, and hallucinations when LLMs are embedded in fact-checking pipelines.
Neurosymbolic systems often translate natural language into logical formulae and check whether claims can be validly derived from verified premises.
This strategy structurally misses misleading claims: what is logically sound does not always match the inferences humans actually draw and accept.
Grounded in cognitive science and pragmatics, the paper offers a typology of cases where logically sound conclusions systematically mislead human inference beyond what the premises support.
Proposed remedy: use LLMs' human-like reasoning tendencies as a complementary check on the outputs of formal components, catching potentially misleading conclusions.