Probabilistic, Reformative Justice
3 hours ago
- #Probabilistic Sentencing
- #Justice Reform
- #Legal Philosophy
- The article critiques the binary nature of criminal justice outcomes through a hypothetical scenario involving two indistinguishable individuals (one guilty, one innocent), arguing that compromise is intuitively better than all-or-nothing verdicts.
- It introduces a taxonomy of justice systems based on two axes: punitive vs. reformative (why justice is enforced) and binary vs. probabilistic (how, defined by loss functions). This creates four quadrants: punitive-binary (ranked 3rd), reformative-binary (2nd), punitive-probabilistic (4th/worst), and reformative-probabilistic (1st/best).
- The author argues that in reformative systems focused on minimizing crime and rehabilitating offenders, guilt/innocence distinctions are less crucial, so sentences should be based on the probability of guilt rather than binary verdicts, making verdict and sentence conditionally independent.
- The probabilistic approach is not widely implemented due to cultural resistance (public preference for clear-cut judgments), existing de facto probabilistic elements (like mental health aid), and concerns it could erode standards like 'innocent until proven guilty'.
- Concludes that the loss function of justice is a design choice, not a direct result of its purpose, and advocates for probabilistic sentencing within reformative systems as optimal.