Experiments test history-dependent semantic change

This proposed study asks whether stories retold between people and language models change because of their interaction history, even when source and framing match. Its proposed mechanism treats an event boundary, a marked narrative break, as a timing cue. It could reset each partner's phase, their measured position in an interpretation cycle, and shift the chance that a meaning becomes fixed. Chains would receive regular or phase-jittered cues while timing, reading exposure and source content stay matched. Phase would be measured independently. Support would require phase-specific meaning changes beyond measured component responses, plus two-way resets: human cues altering model segmentation, how it divides a narrative, and model cues altering human interpretation.