Should subjective probabilities be sharp?

Publication
Episteme