a b s t r a c tThe novelty of scientific or technological knowledge has a paradoxical dual implication. Highly novel ideas are subject to a higher risk of rejection by their evaluating audiences than incremental, "normal science" contributions. Yet the same audiences may deem a contribution to knowledge valuable because it is highly novel. This study develops and tests an explanation of this dual effect. It is argued that the recognition premium that highly acclaimed authors' work enjoys disproportionately accrues to work that is consistent with the authors' previously developed identity. Because high novelty is a salient identity marker, authors' past recognition for highly novel work helps same authors' new highly novel work earn positive audience valuation. It is further argued that, because recognition for novelty is partly inherited from mentors, disciples of highly acclaimed producers of novel work are more likely to have their work prized for its novelty. In contrast, the authors' or their mentors' recognition earned for relatively less novel work does not trigger similar spillover effects and leaves the authors vulnerable to the novelty discount. Unique data on the productivity, career histories, and mentoring relations of academic electrical engineers support these arguments.