In The Economics of Rights, Co-operation and Welfare (1986), Robert Sugden follows the tradition from Thomas Schelling and David Lewis in ascribing a central role to the notion of salience within his theory of convention. However, against this, Bruno Verbeek (2002) has argued that an empirically adequate notion of salience may not be incorporated into a generic account of convention without circularity. This paper examines Verbeek's argument against a background of experimental as well as theoretical work on coordination problems. It finds that the argument fails to consider at least two candidate theories of salience that may be incorporated into the theory of convention without circularity: cognitive hierarchy theory that iterates Sugden’s notion of psychological salience and Schelling salience. Thus, in the end Verbeek's criticism of the role ascribed to salience in theories of convention may be dismissed, though its discussion draws a fruitful perspective.