An external or internal “predictive adaptive response” (PAR) can be defined as an adaptive change in long-term behavior or development due to an environmental exposure that triggers it. A PAR can lead to differential development among initially similar individuals, and increase evolutionary fitness. Despite many theories and empirical observations of PAR-like changes in depressive tendencies, clear empirical findings on human personality changes following depressive symptoms are lacking, possibly because these changes take a long time to develop and most follow up studies have been short. Here we show that in sufficiently long (5- and 15-year) clinical and general-population follow ups, increases can be observed in the Temperament and Character Inventory's personality trait harm avoidance as a function of temporally accumulating major depressive episodes (132 depression patients from Vantaa Depression Study) and depressive symptoms (3105 participants from Young Finns general-population sample). Personality changes did not occur in the other six personality traits of the inventory, but did in a highly similar neuroticism trait from another inventory. Even when controlling for concurrent changes in depressive symptoms from the baseline to the endpoint, depressive symptoms that occurred during the follow-up period associated with harm-avoidance changes, rendering individuals more fearful and anticipating harm. This study provides consistent, specific, and plausible dose–response and temporal gradients between accumulated depressive episodes and personality change. Effect sizes were between small to moderate, though. Altogether, the findings support the feasibility of using existing systems of personality assessment (i.e., self-report questionnaires) to study PARs, despite the multiplicity of the systems.