This article contends that it is time to take a critical look at a series of binary categories which have dominated the scholarly and reform epistemologies of the sentencing decision process. These binaries are: rules versus discretion; reason versus emotion; Offence versus offender; normative principles versus incoherence; aggravating versus mitigating factors; and aggregate/tariff consistency versus individualized sentencing. These binaries underpin both the 'legal-rational' tradition (by which I mean a view of discretion as inherently suspect, a preference for the use of philosophy of punishment justifications and an explanation of the decision process through factors or variables), and also the more recent rise of the 'new penology'. Both approaches tend to rely on 'top-down' assumptions of change, which pay limited attention to the agency of penal workers. The article seeks to develop a conception of sentencing craftwork as a social and interpretive process.' In so doing, it applies and develops a number of Kritzer's observations (in this issue) about craftwork to sentencing. These craftwork observations are: problem solving (applied to the rules-discretion and reason-emotion dichotomies); skills and techniques (normative penal principles and the use of cognitive analytical assumptions); consistency (tariff versus individualized sentencing); clientele (applied to account giving and the reality of decision making versus expression). By conceiving of sentencing as craftwork, the binary epistemologies of the sentencing decision process, which have dominated (and limited) the scholarly and policy sentencing imaginations, are revealed as dynamic, contingent, and synergistic. However, this is not to say that such binaries are no more than empty rhetoric concealing the reality of the decision process. Rather, these binaries serve as crucial legitimating reference points in the vocabulary of sentencing account giving.
