摘要：Text and precedent are an old pair. So old it should frighten us that we do not have a theory of their interaction. Precedent is the device by which a sequence of cases dealing with the same problem may be called law rather than will, rules rather than results. (In a system of civil law there may be no "precedent" at all, on the conceit that the Code contains all rules and the gloss none.) To have a theory of precedent is to have a theory of the extent to which judges' acts are law. Yet we do not have such a theory. Veteran judges such as Cardozo can proclaim that no theory is possible, that adherence to precedent is simply a matter of trial and error, and that when adherence is too dissonant with other rules or too harmful because of effects of the rules, we shall stop adhering. 1 There we have it -- a grand balancing test, with neither a maximand nor weights to produce a decision when the criteria conflict, as they always do. Few Justices hint at a theory of precedent; no Justice has produced a consistent theory; although the academy is awash with competing theories of substantive law, there is no contest in the theory of stare decisis. Not because one candidate has swept the boards, but because no one has a principled theory to offer. This backwater of the law is nonetheless incalculably important for the theory of adjudication as well ...
