The suggestion that more severe sanctions be applied to frequent felons is a key issue in criminology. While the essential assumption of the proposal is simple (i.e., a few criminals are disproportionately responsible for the occurrence of crime), the remedy recommended by this realization is not. The identification and long-term confinement of highly recidivistic individuals is fraught with problems in both principle and practice. The notion of selective incapacitation presumes the ability to predict human behavior and also presumes the existence of a reliable data base, two presumptions not borne out by fact. The concept also suffers with respect to ethical issues. The proposal calls for different punishment for the same offense punishes people for anticipated behavior, and may use predictive variables that reflect discriminatory influences of race and class. Further, attempts to punish future dangerousness are at odds with traditional standards of proportionality and justice and the presumption of an individual's innocence until proven guilty. Empirically, prediction studies have given inaccurate results and show high false positive rates. Finally, selective incapacitation may be further criminalizing, limits judicial discretion in sentencing, and is ex post facto in nature. 40 footnotes.