DefMoN-Syn v1 (formerly DMN-Syn v1) with Supplementary Materials v1.1: A Theory-Grounded Multilingual Synthetic Corpus for Defense Mechanism Inference

This record contains the DefMoN-Syn v1.0 dataset (formerly DMN-Syn v1.0) and the Supplementary Materials v1.1 package, designed for AI and NLP research on psychological defense mechanisms and affective motivations. The dataset operationalizes Vaillant’s hierarchy of ego defenses and Plutchik’s psychoevolutionary theory of emotion into a structured 10 × 8 conceptual grid, generating synthetic utterances across English, Korean, French, and Georgian. 🔑 Key Features Theory-Grounded Framework: Each utterance is generated under explicit psychological theory constraints. AI/ML Ready: Includes manifests for template/scenario cross-validation, cue-ablation sets, and leakage audits. Multilingual Coverage: Quadri-lingual data enables robust cross-lingual transfer experiments. Reproducibility: Seeds, manifests, schema, and code utilities included for byte-level replication. Ethical Safeguards: Explicit guidelines restrict clinical/diagnostic misuse and encourage cultural localization. 📂 Contents Core package (DMN-Syn_v1_package.zip) DMN-Syn_v1.csv (300 synthetic utterances) schema.json, distribution_report.json dataset README and license Supplementary package (DefMoN_SUPPLEMENTARY_v1.1.zip) Template/scenario CV manifests Cue-ablation sets Leakage audit logs Extended reproducibility scripts 🚨 Limitations Dataset is fully synthetic; not based on real individuals. Contains two Korean rows with blank/punctuation-only text (to be patched in v1.0.1). Requires external validation for deployment in real-world NLP/AI tasks. 📖 Citation If you use this dataset, please cite: Kim, R. S. (2025). DefMoN-Syn v1 (formerly DMN-Syn v1) with Supplementary Materials v1.1: A Theory-Grounded Multilingual Synthetic Corpus for Defense Mechanism Inference. Zenodo. https://doi.org/10.5281/zenodo.17101927