Safety occupies a foundational position across the depth-psychology corpus, yet the term is far from univocal. Two major streams converge here without quite merging. The first, inaugurated by Herman and systematized by Najavits, treats safety as the indispensable first stage of any recovery from trauma or addiction: before mourning, before reconnection, the patient must achieve stabilization, acquire coping skills, and cease the self-destructive behaviors that perpetuate harm. Here safety is essentially a clinical objective — concrete, behavioral, and sequentially prior to all deeper therapeutic work. The second stream, most fully articulated by Porges and translated into clinical practice by Dana and Winhall, grounds safety in neurophysiology. On this account safety is not primarily a cognitive judgment but a felt state produced by the autonomic nervous system through the mechanism of neuroception — a continuous, largely pre-conscious scanning of environment and relationship for cues of threat or connection. The ventral vagal circuit is the anatomical substrate; the Social Engagement System is its behavioral expression. The therapeutic environment, the prosody of the clinician’s voice, even the proximity of food or plants, all function as cues that regulate whether a client’s autonomic system permits the engagement necessary for healing. Between these two streams lies a productive tension: Najavits emphasizes volition and skill-building, Porges emphasizes that safety cannot be willed but must be neurophysiologically signaled. Lanius and Courtois occupy a middle position, insisting that the establishment of safety in the therapeutic relationship is both the most complicated and the most essential therapeutic task, especially for survivors of disorganized attachment.