Montague semantics and modifier consistency measurement in neural language models

TL;DR

This paper introduces three novel tests inspired by Montague semantics to evaluate compositional behavior in neural language models, revealing their limitations in capturing certain semantic properties.

Contribution

It proposes a new methodology for measuring compositionality in language models using Montague-inspired tests, highlighting gaps in current models' semantic understanding.

Findings

Neural language models often fail to exhibit expected compositional behavior.

Current models may lack the capacity to capture certain semantic properties.

Linguistic theories from Montague semantics may not align with distributional model capabilities.

Abstract

This work proposes a novel methodology for measuring compositional behavior in contemporary language embedding models. Specifically, we focus on adjectival modifier phenomena in adjective-noun phrases. In recent years, distributional language representation models have demonstrated great practical success. At the same time, the need for interpretability has elicited questions on their intrinsic properties and capabilities. Crucially, distributional models are often inconsistent when dealing with compositional phenomena in natural language, which has significant implications for their safety and fairness. Despite this, most current research on compositionality is directed towards improving their performance on similarity tasks only. This work takes a different approach, introducing three novel tests of compositional behavior inspired by Montague semantics. Our experimental results indicate that current neural language models do not behave according to the expected linguistic theories. This indicates that current language models may lack the capability to capture the semantic properties we evaluated on limited context, or that linguistic theories from Montagovian tradition may not match the expected capabilities of distributional models.