Camn, you got the right idea except for the actual values - Since sound travels (in air at sea level at 70 degrees) at about 1135 feet per second, that means that it travels just over one foot per millisecond, so two milliseconds would be about 2.26 feet. Your example of 1/500 of a foot, would cause a delay of about 1/500 of a millisecond, or roughly 2 microseconds.
At 15 kHz, one cycle takes 66.6 microseconds to complete, so a 2 microsecond delay (in only one channel of a stereo source) would cause a phase shift (only at 15kHz, other freq's would be different) of just under 11 degrees. I doubt too many people would notice this small a phase change, at least in the context of an actual mix; perhaps while listening to test tones...
Still, the point is very valid that very small changes in mic location, whether micing in stereo or using two mics and blending the tracks, can cause phasing problems. These would usually be in the form of comb filtering, where different frequencies are affected differently. Some would be attenuated nearly to zero, while others would be either doubled or not affected, or anywhere in between. Small changes in the location of one mic can make large changes in the effect.
One important thing to remember is that phase relationships which are caused by timing delays, are DIFFERENT at EVERY frequency. This means that there is no constant condition that can compensate for phasing problems, such as the (mistakenly labeled) PHASE switch on a console strip.
For an example of what phasing sounds like, borrow an old guitar stompbox like
the MXR Phase 90, and run your entire mix thru it. Cool for some effects, but not something you'd want to happen un-intentionally... Steve