Source: (2003) University of Pennsylvania Law School. Scholarship at Penn Law. Paper 33. The Berkeley Electronic Press (bepress). Downloaded 14 October 2005.Robinson supports the proposed âpurposesâ? text of the New American Law Institute Report on Sentencing Reform but argues that in practice it will not mean what traditional utilitarians, like those supporting âlimiting retributivism,â? are expecting. First, the proposed text allows reliance upon non-desert distributive principles only to the extent that they serve their stated goals. As the ALI Report concedes, there are limits to the effectiveness one can expect from rehabilitation and, as is now becoming apparent from social science research, our realistic expectations for the effectiveness of deterrence are similarly fading. It is true that incapacitation undoubtedly works to prevent future crime and there is increasing evidence that ârestorativeâ? processes can be effective in achieving their goal, but following these distributive principles can have crime prevention costs (which the textâs proposed distributive principle would not seem to allow to be taken into account). For example, most importantly, reliance upon these principles can undercut the criminal lawâs moral credibility and, thereby, its power to gain compliance by harnessing social norms. Second, the greatest constraint on the influence of the enumerated non-desert purposes is the proposed principleâs demand that no distribution can conflict with the demands of desert. Contrary to the assumption of the original advocates of âlimiting retributivismâ? that desert provides only vague outer limits on punishment, desert has quite specific demands, driven in large part by the demand of ordinal ranking: that a case of greater blameworthiness receive greater punishment than a case of comparatively less blameworthiness. Given the limited range of punishments a liberal democracy ought to be willing to inflict, distinguishing cases of distinguishable blameworthiness means that the deserved punishment will fall within a narrow range on the punishment continuum (a result consistent with what social scientists are lately learning about the rather sophisticated intuitions of justice shared by laypersons). In closing, Robinson offers brief remarks addressed to those who oppose a desert distribution, of which he suggests there are two sorts: those who have an erroneous of what âdesertâ? means, and those who think avoiding crime is more important than doing justice. To the first group, he simply sketches the current consensus view of constitutes the modern conception of desert. To the second group, he argues that a desert distribution may well be the most effective in avoiding future crime, once one takes into account the crimogenic effect of non-desert distributions, when the community comes to understand that the criminal justice system is not in fact in the business of doing justice. In any case, he argues that in the absence of sufficient data to reliably answer the empirical question as to which distribution would best reduce crime, we ought to do at least what we know we can do: do justice. Author's abstract.