This study introduces a quantitative framework for analyzing linguistic creativity through the measurement of eight cognitive-linguistic markers. Using natural language processing (NLP), the framework quantifies metaphoricity, synesthesia, associativity, semantic prosody, affective valence, semantic diversity, perceptual detail, and emotional arousal. To ensure the quality of the corpus it’s comprised both human-authored texts (poetry, prose, songs, scientific articles) and Large Language Model (LLM)-generated content. Using a combination of heuristic-based and lexicon-driven methods, the study demonstrates a robust approach to comparative stylistic analysis. We identified specific linguistic patterns for each category, and showed how human and AI-generated texts differ in emotional and lexical metrics. By establishing an empirical foundation for evaluating linguistic creativity, this research advances both computational stylistics and a broader understanding of human–machine creative dynamics.