The paradox of innovation: New concepts and disruptiveness in science

Abstract

This study examines the relationship between the introduction of new concepts and scientific disruptiveness using more than 42 million publications published between 1950 and 2024. We find that papers introducing new concepts are, on average, less disruptive than papers without new concepts. Historical analyses show that the relationship shifted from positive in earlier decades to consistently negative after the 1980s. We further show that team diversity strengthens the negative association between introducing new concepts and disruptiveness. Decomposition analyses indicate that papers introducing new concepts receive more citations overall, but disproportionately more consolidating citations than disruptive citations. Restricted cubic spline analyses reveal a statistically significant non-monotonic association. Low positive concentrations of new concepts are associated with lower disruptiveness, whereas the association reverses only in the sparsely populated upper tail of the distribution. Taken together, the findings reveal a complex relationship between conceptual change and scientific disruptiveness.