Tom, actually that's exactly what the better surround processors(e.g. DPLII, Logic 7)can do with 2-channel material, i.e. they "know" what should stay in the front speakers and what should be extracted and steered to the surrounds(not with 100% perfection, of course). This is because when a sound reflected from a surface reaches the recording mike(or a listener)a split second after the direct sound it's out of phase with the direct sound. The surround processor can measure these phase differences. Slightly oversimplified, direct in-phase material on the left channel stays in the left front speaker, out-of-phase material in the left channel is extracted and sent to the left surround speaker; the right channel is processed similarly. Another process is that direct sound equally strong in both channels is extracted and sent to the center speaker, forming a real center(adjustable with the DPLII center width control)rather than the "phantom" center that we experience with those sounds when listening with two front speakers.
The amount of ambience present in 2-channel material varies depending on how closeup the recording mikes were and whether "ambience mikes" farther back were also used. A very closeup recording with less ambience present will benefit less from surround processing, but to some extent using the dimension control on DPLII to increase the sound directed to the surround speakers can help.
