Neural Network Analysis of Fiction Characters: Assessment of Adaptive Intelligence

The article discusses the results of neural network qualitative analysis of artistic narratives as models of adaptive intelligence activation. R. Sternberg, who develops the concept of adaptive intelligence, suggests looking at it as a construct implemented in cultural contexts. This approach formed the basis for the study of narratives, where the hero finds himself in a dangerous situation and uses ingenuity and resourcefulness. The analysis tool was the Chat GPT language generative model, which served as an expert in the field of intelligence psychology and narrative research. The data obtained in the form of a text analysis of 100 characters and assigned values on the scales of adaptive intelligence, narrative and cultural context were subjected to statistical and mathematical processing and presented as a cluster tree. The critical values of the r-Pearson correlation coefficient combined the characters into two superordinate clusters and set the polarization of groups of leaders with high adaptive intelligence and a group of outsiders, whose values on all scales turned out to be low. This allowed us to draw a conclusion about the factors of the organization of hero groups: the pressure factor of an acute situation, the factor of social recognition of productivity, the factor of subjectivity, the factor of integration into culture. Thus, a universal mechanism for activating adaptive intelligence and unique survival narratives are revealed, uniting the characters in groups of contexts. The conducted research opens up prospects for studying adaptive intelligence as a social construct, identifies the factors of its activation and expands the understanding of the impact of cultural narratives on the intellectual forces of humanity, shifting the focus from the problems of self-identification to the plot of adaptation and survival.