Abstract
The Socratic method has demonstrated, in studies across law, medicine, and nursing, that dialectic structured questioning improves expert reasoning. Cybersecurity currently teaches tools and techniques but lacks Socratic dialogic leaving graduates underprepared for novel scenarios requiring cross-domain reasoning. Because law, medicine and nursing cultivate similar reasoning as security there is principled reason to expect reasoning benefits could transfer to security students. This paper argues that the correspondence warrants investigation. Investigation requires a named, teachable reasoning target analogous to clinical reasoning and we specify that target as adversarial cognition, a framework integrating adversary features, system architecture, implementation logic, governance structure, human factors, and systemic entropy into attack or defense pathways. These five echo the causal analysis of the Socratic and Aristotelian tradition and modern failure analysis in engineering, information security, and quality management. Our method adapts the Socratic elenchus as nursing and legal education have, pairing questioning with narration and productive misdirection. We evaluate the method to cognitive apprenticeship, situated learning, and narrative pedagogy, demonstrate it through a documented Active Directory attack case, and specify an open, replicable protocol testing whether Socratic-trained practitioners differ from procedurally trained students. The protocol is designed to reveal whether Socratic training improves domain-specific reasoning, general reasoning, or both. Finally, because the same measurement applies to LLM reasoning, we propose comparing Socratic student reasoning and LLM reasoning now that AI-assisted offensive tools make understanding differences in adversarial reasoning increasingly consequential.
Included in
Education Commons, Information Security Commons, Management Information Systems Commons, Philosophy Commons, Technology and Innovation Commons