A new arXiv paper examines language dynamics in AI agent communities. The abstract reports that these communities sustain lexical persistence: popular terms spread, recur, and achieve broad adoption.

The open question, according to the abstract, is whether such communities also develop the temporal semantic structure characteristic of human linguistic communities. The paper's title suggests a finding of lexical persistence without semantic entrenchment, though the abstract itself does not state the conclusion explicitly.

The distinction matters because human language communities not only reuse words but also embed them in evolving semantic structures. Whether AI agents do the same has implications for understanding emergent communication in multi-agent systems.