I found these just by using the searchbox on my browser.
http://en.wikipedia.org/wiki/Sound_stage
http://en.wikipedia.org/wiki/Stereo_imaging
Basically, the mic set-up during a recording mimics how ears would hear the instrument, so they are recorded as such. A properly recorded track should have a relatively easy to pick out image of the soundstage, where the vocals are in front, the drums and bass guitar behind, and guitars are to the left and right of the vocals. If there's a symphony or a quartet backing up a modern band, they should be somewhere behind, along with the choir. You should be able to hear a relative position along the X (width) and Z (depth) planes, and each instrument should be coming from a clear spot.
Headphone listening however differs in two ways. First, the relative distance of the drivers to the ears means that it won't project an image much larger than your head. Even the best headphones still have some part of the soundstage (like the vocals and guitars) inside the listener's head. That said, in terms of size alone, a speaker set-up is also "to scale," unless that speaker set-up is the same size as the intended stage the studio recording mimics. The other problem is that you don't get true stereophonic listening with headphones. With speakers the left ear can hear the right speaker and vice versa, making for a deeper soundstage and more precise imaging.
Further reading here:
http://en.wikipedia.org/wiki/Crossfeed
http://www.meier-audio.homepage.t-online.de/crossfeed.htm
http://www.head-fi.org/t/220770/describing-sound-a-glossary
