Figure 10-2 outlines how I think the process of speech works for most people. We know what to say slightly before we actually say it. The process of generating words runs before the process that controls words for appropriateness. We are able to censor anything we deem inappropriate before it is said. This is when we need to either generate something else very quickly or abstain from saying anything. This is when the awkward pauses happen.
It's best to relax your quality control and instead rely on how the words sound when you actually say them and on the feedback from the audience. Here are my top three reasons for avoiding the approach outlined in Figure 10-2 :