Abstract

This Article takes a critical approach to the assumptions underlying current practices of violence risk assessment. The Article first explicates a fundamental difference between how mental health researchers now interpret the phrase violence prediction and how they understood that phrase in the 1970s. Applying a critical approach, the Article then shows that courts and mental health professionals may need to abandon the hope that more accurate methods of predicting violence will help clinicians make better decisions about potential violence. Although this result may seem disappointing, it can liberate mental health professionals from regarding patients as statistical sources of risk and can encourage them to treat patients instead as sources of initiative and moral worth. The Article then provides a moral and legal framework for reformulating the Tarasoff rule. Taking the view that consequentialist, predictive approaches to formulating the duty to protect are ethically questionable (as well as scientifically impractical), the Article suggests that assigning therapists a duty to intervene in response to a patient's explicit, credible threat is consistent with the view that patients (along with other humans) ought to be treated not merely as means, but always as ends in themselves.