Abstract
A long time ago, in a place that must seem far away, the issue of retroactive application of a new legal rule was easily resolved. At common law, courts generally applied new rules retroactively; to do otherwise was to deny that the judiciary "rather than being the creator of the law was but its discoverer." Retroactivity thus lies near the heart of the debate over the proper judicial role: Do judges make or find law? According to the common law, "making law" entails discretionary policy decisions most akin to the legislative role.' The American federal constitutional system lodges such discretion in Congress. The courts, on the other hand, discover pre-existing principles of law and merely apply them to cases. In this view, when courts overrule a prior decision, they state not only what the law is, but what the law has always been. The overruled decision is not old law, but rather a failed attempt at discovering the true law.
Under common-law jurisprudence, retroactivity is quite simple: always and absolute. Courts, however, weighed down by practical considerations such as a party's substantial reliance on the "old" but now nonexistent rule, have found this lofty ideal hard to reach. Although full retroactivity may stimulate a comforting belief that the judiciary fits nicely within its traditional common-law role, it also defeats settled expectations and reasonable reliance on the stability of the law. A tension between theory and practice drives the retroactivity debate.