A Network Theory of Patentability: Towards an Empirical Measure of Patent Nonobviousness

Abstract

The nonobviousness inquiry is one of the most notoriously complex and contingent of all legal questions. The challenges faced by those tasked with determining an invention’s nonobviousness include understanding unfamiliar technologies and seeing the world from the perspective of an objective inventor, all while avoiding issues of hindsight bias. This research project attempts to simplify the obviousness inquiry by proposing an empirically-grounded network theory of patentability that integrates the dominant competing theoretical approaches to obviousness. We use this network theory of patentability to motivate the definition of a “network measure of nonobviousness” and subsequently demonstrate our measure by using the prior art record to approximate “technological space.” We are then able to situate claimed inventions within technological space and measure the preceding inventive activity in the area, providing insight into how likely it is that the inventions are obvious. Our network measure of nonobviousness offers valuable objective insight into the obviousness analysis that is based on the empirical patenting record and robust to hindsight bias. It shows promise as an aide for patent application assessment, patent portfolio management, and patent dispute resolution. From a methodological perspective, this project is the first to use network analysis to answer doctrinal questions.