Seminars and Panels

Atypical Combinations and Scientific Impact

Professor Brian Uzzi

Dec 8, 2014
11:00 A.M. - 12:30 P.M.
232 Philip G. Hoffman Hall

Often purported but rarely tested is the claim that science is spurred on when atypical ideas are united, inspiring fresh thinking to problems. Yet, many scientific ideas and innovations intentionally build in convention, rather than remove it. Here, we analyzed all 17.9 million research papers in the web of science, circa 1945–2005 using a methodology that characterizes each paper’s conventional and novel combinations of prior work. We find that the premium often expressed for papers with novelty is at odds with the reality that most scientific work typically draws on highly conventional, familiar mixtures of knowledge. Especially virtuous combinations are not characterized by novelty or conventionality alone. Rather, the highest impact papers interject novelty into otherwise unusually conventional combinations of prior work, and remarkably, are twice as likely to top the citation distribution. Finally, teams are more likely than solo scientists to interject novel combinations into their papers, suggesting that the exceptionalism of teams is an ability to incorporate novelty. Finally, these empirical regularities are largely universal, appearing across fields and decades, suggesting fundamental rules about creativity in science. At root, our work suggests that creativity in science appears to be a phenomenon of two extremes. At one extreme is conventionality and at the other is novelty. Curiously, advancing to the frontier of science appears best served not by efforts along one boundary or the other but with efforts that reach toward both frontiers. Read more...