Here is Professor Ernst's abstract:
Proponents of legislative patent reform argue that the current patent system perversely impedes true innovation in the name of protecting a vast web of patented inventions, the majority of which are never even commercialized for the benefit of the public. Opponents of such legislation argue that comprehensive, prospective patent reform legislation would harm the incentive to innovate more than it would curb the vexatious practices of non-practicing entities. But while the “Innovation Act” wallows in Congress, there is a common law tool to protect innovation from the patent thicket lying right under our noses: the reverse doctrine of equivalents. Properly applied, this judge-made doctrine can be used to excuse infringement on a case-by-case basis if the court determines that the accused product is substantially superior to the patented invention, despite proof of literal infringement. Unfortunately, the reverse doctrine is disfavored by the Court of Appeals for the Federal Circuit and therefore rarely applied. It was not always so. This article is the first comprehensive study of published opinions applying the reverse doctrine of equivalents to excuse infringement between 1898, when the Supreme Court established the doctrine, and the 1982 creation of the Federal Circuit. This “lost precedent” reveals a flexible doctrine that takes into account the technological and commercial superiority of the accused product to any embodiment of the patented invention made by the patent-holder. An invigorated reverse doctrine of equivalents could therefore serve to protect true innovations from uncommercialized patents on a case-by-case basis, without the potential harm to the innovation incentive that prospective patent legislation might cause.Interestingly, according to Ernst, "the Second, Sixth, and Ninth Circuits had precedent requiring that the district court must always consider reverse equivalents prior to determining infringement," and the standard was only whether the accused product was "substantially changed," not whether it was a "radical improvement" (a standard that emerged from scholarly articles, not case law).
I don't have high hopes for the revival of this doctrine, but the Federal Circuit has made clear that it is not dead yet; for example, Plant Genetic Systems v. DeKalb (2003) quoted an earlier case as saying that "the judicially-developed 'reverse doctrine of equivalents' . . . may be safely relied upon to preclude improper enforcement against later developers." So litigators should keep this in their toolkits, just in case.