This article responds to the reply by Cooper and Shallice to Botvinick and Plaut's previous comment . In the authors' view, the reply by Cooper and Shallice left their original responses more or less intact. The present exchange has strongly highlighted the degree to which any account of the habit system will interact with how the goal-directed system is understood. Many of the points the authors of this article have debated have turned out to depend on what assumptions are made about the goal-directed system and about how it interfaces with the habit system. In view of this, the authors believe that what seems necessary is to place their present accounts of routine sequential action within a larger computational framework that explicitly addresses both systems and their interrelations.