Abstract

This paper proposes a theory of technological invention that is separate but complimentary to our more extensive understanding of commercial innovation. It proposes an explanation for the ultimate source of technological novelty, and why some technologies mature into obsolescence and yet others undergo repeated renewals. The work also suggests a technological resolution to the controversy of discontinuous versus continuous change. ;I propose that breakthrough and incremental invention both result from a cognitively limited process of recombination. While the elements of recombination may be more or less familiar to inventors, the same underlying process generates both phenomena. Inventors drive the tempo of invention by the set of technological components and architectures from which they recombine. When inventors explore unfamiliar components and architectures, their efforts are less useful on average, but the likelihood of a breakthrough increases. Such explorative efforts correspond to periods of technological fervor and radical change. When inventors exploit familiar components and refine previously used architectures, their efforts are more useful on average, but the likelihood of a breakthrough decreases. Such local search and incremental efforts can stagnate, however, into periods of technological stupor. The tension between exploration and exploitation eventually becomes moot if inventors exhaust their recombinant search space. Count models and data from all U.S. patents before 1997 demonstrate support for the theory. ;This dissertation begins by motivating the need for a theory of invention. I continue with a literature review of previous work on the sources of technological change. I then develop my theory of recombinant learning and exhaustion. I move into the empirics with a description of variables, algorithms for their calculation, explanation of statistical models, and analysis of results. In the concluding discussion I offer a recombinant and technological resolution to the controversy of continuous versus discontinuous change. I conclude with organizational and strategy implications