In his comment, A. Roelofs claimed that a feedforward-only theory of spoken word production (WEAVER++) can account for certain basic facts of spoken word production that B. Rapp and M. Goldrick (2000) argued could not be accounted for by feedforward-only theories. Rapp and Goldrick argued that to account for these facts, mechanisms such as feedback and cascading activation are required. In this reply, the authors discuss difficulties with Roelofs's argumentation and point to recent relevant empirical evidence. The authors conclude that Roelofs failed to disprove Rapp and Goldrick's claims and, furthermore, failed to provide strong evidence of the specific feedback characteristics in WEAVER++.