In a multi-element speaker (2-way, 3-way) it is entirely possible that your brain will interpret a lack of time-alignment between the drivers as height. In other words, if you hear HF elements of the sound slightly before MF frequency elements, the brain should decode that as a source higher than you because the tweeter is located above the midrange. If your speaker is a single-element or perfectly time-aligned, how would the brain possibly get a signal that provides height information? I know that I have many good sources of music that are 2-track direct mic'ed recordings (my favorite type) and have excellent imaging, sources well distributed in their original locations, ... but unless I close my eyes and picture those sources higher than each other I don't actually hear it. When I close my eyes and visualize the performance, "cognitive bias" will convince me that I hear height.
The enitre imaging perception is created by phase-shift in the music, or the difference in timing between the left and right ear and volume/spl.
I really haven't thoroughly researched it but I am not aware of any way that two sources of sound can create any more than 2-D imaging, can be nearer/further and left/right well beyond the sources (speakers), but I remain skepical on 3-D imaging from 2-channel.
A turntable OTOH, now there's a way to create more false soundstage as the cartridge moves in sympathetic motion to the stylus' movement and creates out-of-phase crosstalk, ... another subject entirely.