Devin Robert Wright

We compare how humans retell stories to how ChatGPT retells stories in chains of three retellings by different people or different accounts on ChatGPT. ChatGPT provides competent summaries of the original narrative texts in one step of retelling. In subsequent retellings few additional changes occur. Human retellers, by contrast, reduce the original text incrementally and by creating 55–60% of novel words and concepts (synsets) at each iteration. The retellings by both ChatGPT and humans show very stable emotion ratings, which is a puzzle for human retellers given the high degree of novel inventions across retellings. ChatGPT maintains more nouns, adjectives, and prepositions and also uses language later acquired in life, while humans use more verbs, adverbs, and negations and use language acquired at a younger age. The results reveal that spontaneous retelling by humans involves ongoing creativity, anchored by emotions, beyond the default probabilistic wording of large language models such as ChatGPT.

Quantifying *identity fusion*—the psychological merging of self with another entity or abstract target (e.g., a religious group, political party, ideology, value, brand, belief, etc.)—is vital for understanding a wide range of group-based human behaviors. We introduce the Cognitive Linguistic Identity Fusion Score ([CLIFS](https://github.com/DevinW-sudo/CLIFS)), a novel metric that integrates cognitive linguistics with large language models (LLMs), which builds on implicit metaphor detection. Unlike traditional pictorial and verbal scales, which require controlled surveys or direct field contact, CLIFS delivers fully automated, scalable assessments while maintaining strong alignment with the established verbal measure. In benchmarks, CLIFS outperforms both existing automated approaches and human annotation. As a proof of concept, we apply CLIFS to violence risk assessment to demonstrate that it can improve violence risk assessment by more than 240%. Building on our identification of a new NLP task and early success, we underscore the need to develop larger, more diverse datasets that encompass additional fusion-target domains and cultural backgrounds to enhance generalizability and further advance this emerging area. CLIFS models and code are public at [https://github.com/DevinW-sudo/CLIFS](https://github.com/DevinW-sudo/CLIFS).