THEORIES ABOUT SENTENCING HAVE LONG BEEN DOMINATED BY TRADITIONAL ASSUMPTIONS - THAT PRISONS REHABILITATE THE CRIMINAL OR RESTRAIN HIM IF HE IS DANGEROUS; AND THAT TO ACCOMPLISH THESE AIMS, JUDGES AND OTHER OFFICIALS SHOULD BE GIVEN THE WIDEST DISCRETION IN THEIR DECISIONS. THIS TEXT POINTS OUT THE FLAWS IN THIS REASONING BY DOCUMENTING THE FAILURES OF REHABILITATION AND THE FUTILITY OF PREDICTING RECIDIVISM, AND THEN PRESENTS AN ALTERNATIVE. IT ARGUES THAT UNDER A JUST SYSTEM THE LENGTH OF A SENTENCE WOULD BE BASED PRIMARILY ON THE OFFENDER'S DESERTS: THAT THE SERIOUSNESS OF THE CRIME SHOULD DICTATE THE PUNISHMENT, AND WITHIN DEFINED LIMITS. OTHER REFORMS OF THIS SYSTEM WOULD INCLUDE STRINGENT LIMITATIONS ON INCARCERATION AS PUNISHMENT; ALTERNATIVES TO INCARCERATION FOR THE BULK OF CRIMINAL OFFENSES; SHARPLY SCALED-DOWN PENALTIES FOR FIRST OFFENDERS; REDUCTION IN SENTENCING DISPARITY; NARROWING OF SENTENCING DISCRETION; AND ELIMINATION OF THE INDETERMINATE SENTENCE. THIS VOLUME IS THE RESULT OF A FOUR YEAR INVESTIGATION BY THE COMMITTEE FOR THE STUDY OF INCARCERATION. (AUTHOR ABSTRACT MODIFIED)