Ad hoc conventions generalize to new referents

2025-09-05

AI Summary

How do humans establish generalizable, shared naming systems for entirely novel objects? Method: We conducted dyadic repeated communication experiments in which participants negotiated ad hoc names for >1,000 abstract tangram images from the KiloGram dataset and subsequently tested name generalization to unseen stimuli. Contribution/Results: Naming conventions were not arbitrary labels but reflected deep conceptual alignment—supporting the “conceptual space remodeling” hypothesis. Semantic consistency significantly increased for novel objects and decayed nonlinearly with visual similarity, aligning with Shepard’s universal law of generalization. This generalization effect remained robust across varying levels of object “namingness” (i.e., intrinsic nameability). Critically, this study provides the first controlled experimental evidence that ad hoc linguistic conventions can generalize systematically beyond training instances. These findings offer key empirical support for theories of language emergence, conceptual representation, and symbol grounding—particularly in human–AI collaboration contexts where shared referential meaning must be rapidly co-constructed.

Technology Category

• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Humans and AI: Learning Human Values and Preferences

Application Category

• Economics, Online Markets and Human Computation: The sharing economy
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

How do people talk about things they've never talked about before? One view suggests that a new shared naming system establishes an arbitrary link to a specific target, like proper names that cannot extend beyond their bearers. An alternative view proposes that forming a shared way of describing objects involves broader conceptual alignment, reshaping each individual's semantic space in ways that should generalize to new referents. We test these competing accounts in a dyadic communication study (N=302) leveraging the recently-released KiloGram dataset containing over 1,000 abstract tangram images. After pairs of participants coordinated on referential conventions for one set of images through repeated communication, we measured the extent to which their descriptions aligned for undiscussed images. We found strong evidence for generalization: partners showed increased alignment relative to their pre-test labels. Generalization also decayed nonlinearly with visual similarity (consistent with Shepard's law) and was robust across levels of the images' nameability. These findings suggest that ad hoc conventions are not arbitrary labels but reflect genuine conceptual coordination, with implications for theories of reference and the design of more adaptive language agents.

Problem

• Testing generalization of ad hoc naming conventions to new referents
• Comparing arbitrary labeling versus conceptual alignment accounts
• Investigating shared referential systems through dyadic communication experiments

Innovation

• Shared naming system with conceptual alignment
• Generalization to new referents via semantic reshaping
• Nonlinear decay with visual similarity per Shepard's law