Synthetic Quadri-lingual DMN Corpus (v1)

Synthetic Quadri-lingual DMN Corpus (Version 1.0) This dataset contains 300 fully synthetic utterances annotated with the Defensive Motivational Nodes (DMN) framework. DMN assigns each utterance one defense mechanism (10-class taxonomy based on Vaillant, 1976) and one affective motivation (8-class taxonomy derived from Plutchik, 1980). Language distribution • English 100 utterances (EN) • Korean 100 utterances (KO) • French 50 utterances (FR) • Georgian 50 utterances (KA) All examples were generated by a rule-based template script using ChatGPT (July 2025 release). No real user data are included; therefore, the corpus is free of personal or sensitive information and falls outside human-subjects research regulations.