What Language Does and What the Evidence Supports: A Functional Role Taxonomy and Evidence Audit of Language Grounding in Embodied Agents

2026-08-04

AI Summary

This work addresses a critical gap in embodied intelligence research, where the purported role of language often lacks empirical grounding, making it difficult to assess its actual contribution to agent behavior. To bridge this disconnect, the paper introduces a functional-role-based analytical framework that categorizes language’s roles in embodied systems into five distinct types: Specification, Embodied Representation, Action Orchestration, Grounding Regulation, and Execution Coupling. This framework enables a systematic evidence audit of existing approaches and is the first to uniformly apply across both modular and end-to-end architectures, facilitating fine-grained evaluation. The analysis reveals that most current studies fail to effectively leverage language as a mediating mechanism or provide sufficient evidence linking linguistic components to behavioral improvements, with performance gains frequently lacking rigorous attribution to language use.

Abstract

Foundation models place language throughout embodied agents, but its presence does not show what it contributes or how well that contribution is grounded. This survey separates these two questions. We define five non-exclusive functional roles for language: Specification, Embodied Representation, Action Orchestration, Grounding Regulation, and Execution Coupling. For each role, we trace the path from linguistic content to its embodied consumer and identify the observations or interventions that can test the claimed responsibility. Applying this framework to the reviewed literature reveals a recurring gap between functional use and evidential support. Interpretable or revised linguistic intermediates may be incorrect, go unused, or fail to affect later behavior. Even when actions are directly conditioned on language, system-level success does not by itself isolate language's contribution. We therefore evaluate grounding claim by claim, asking whether the reported evidence supports the specific responsibility assigned to language. Using role claims rather than architectures as the unit of comparison allows us to compare modular and end-to-end embodied agents without extending conclusions beyond the reported evidence.

Problem

• language grounding
• embodied agents
• functional roles
• evidence audit
• foundation models

Innovation

• language grounding
• functional role taxonomy
• embodied agents
• evidence audit
• grounding evaluation