This paper explores phase-internal derivations and considers their consequences for Minimalist theorizing. With the discussion of extraction from subjects, we claim that phase-internal derivations are step-by-step processes, and argue against simultaneous, parallel applications of operations at the phase level. We argue that our claim is theoretically favorable, and show that it can naturally explain extraction phenomena as well as grammatical behaviors of multiple wh-questions. We also take a fresh look at "wellformed" extraction from subjects and subject wh-movement. We propose new analyses and provide empirical evidence in their favor. The discussions in this paper strengthen the argument that step-by-step procedures explain language.