The current culture of measuring citation of entire articles alone tends to favor the articles describing techniques and resources or long-standing hypotheses with lasting predictive value. It may even contribute to the fixation of initially tentative hypotheses by repetition rather than by replication. The prospect of having citation metrics for semantic assertions is exciting because it might now be possible to recognize and reward efficient hypothesis construction, testing and rapid falsification. These complementary skills are particularly valuable to researchers when they navigate new conceptual territory.