Use fader automation in the audio mixer, push the fader up when he turns his head and pull it back when he is looking forward. It would probably need eq as well. You could duplicate the audio track add a bit of gain and HF eq to this second track and then just fade up this track on the head turns.
The super-lazy way to do this is to put a compressor/limiter on the track. Set a high make-up gain value, a fast attack and release, and a threshold around -2. That would slam the quiet portions of your audio, and squash the loud parts. It probably won't sound great, but it will be easy.
Richard posted the "right" way.
There are several ways to adjust Volume, or Gain, in Audio in PrPro. This ARTICLE goes into a little detail on using each.
Personally, I like Richard's suggested method, but one must remember that the Audio Mixer adds Track Keyframes for the automation. If one were to later edit the Clips on that Audio Track, those Track Keyframes remain, where they were, and do NOT follow Clips. Still, as I do the final Volume adjustments AFTER my editing, having Track Keyframes is not an issue.
You didn't mention what version you are using, but if you have CS5.5, you can right click the audio and select edit in Audition. Once there, put the filter "speech volume leveling" which is found in effects/amplitude and compression/speech volume leveler. That will level the low and high volume. You may need to remove noise that gets amplified by this process and other filters are available for that.
Thanks everyone, this all looks good. Started doing it Richard's way, still more time consuming than I had hoped, but I do need it to sound good. I will also be trying the Audition method, as I am working with CS5.5, although I also have CS5 on my system when CS5.5 loops audio tracks. Does anyone know if Adobe has fixed the audio conforming issues with CS5.5?