ARTFEED — Contemporary Art Intelligence

Private Etymology: A Design Model for Shared Symbols in Human-AI Interaction

ai-technology · 2026-08-15

A recent concept-and-prototype paper presents 'Private Etymology,' a machine-representable relational provenance that captures the evolution of dyad-specific symbolic expressions over time in long-term interactions between humans and AI. Available on arXiv (2608.08443), it introduces 'relational reuse' as a technique to reactivate specific expressions in future sessions without needing to re-explain their meanings. Prior research has indicated that individuals create shared symbols, partner-specific phrases, personal idioms, and inside jokes, leading to relational microcultures. Although recent studies have explored how humans and conversational AI negotiate symbolic meanings, existing long-term systems lack a comprehensive design model for documenting the meaning-making process. This paper proposes a model to track the lifecycle of shared symbols, aiming to enhance the depth and continuity of human-AI relationships.

Key facts

  • Paper title: 'Private Etymology: Designing Relational Reuse of Shared Symbols in Long-Term Human-AI Interaction'
  • arXiv ID: 2608.08443
  • Announce type: replace-cross
  • Introduces 'Private Etymology' as a relational provenance model
  • Proposes 'relational reuse' for reactivating dyad-specific expressions
  • Addresses lack of design model for recording meaning and acceptance in long-term human-AI systems
  • Builds on prior work on shared symbols and relational microcultures
  • Paper is a concept-and-prototype paper

Entities

Institutions

  • arXiv

Sources