What is "rivaling" during binocular rivalry, competing stimuli or competing eyes? Phenomenologically speaking, the answer seems obvious: you "see" a given pattern without any knowledge of which eye is dominant at any given moment during rivalry - in other words, you "see" a stimulus without regard to its eye of origin. But while viewing rivalry suppose you click a button when a given rival pattern is dominant exclusively, such that the other one is completely suppressed from vision. And suppose this button click causes the two rival patterns to be swapped between your two eyes. In other words, the pattern dominant at the moment of button press will now be seen by the eye that was viewing the suppressed pattern, and vice versa. Do you continue to see the dominant pattern, which is what you'd expect if it was the stimulus itself that was dominant? Or do you suddenly see the previously suppressed pattern, because it's now being viewed by the currently dominant eye? Try the demonstration below to learn the answer.
Instructions: This demonstration consists of a QuickTime movie that swaps two orthogonally oriented rival patterns between the eyes. First, make sure the animation is "rewound" meaning that the pointer is at the left-hand side of the controller. Next, position the cursor over the "play" button (the right-pointing triangle) but don't click it yet. Now free fuse the two rival patterns and let them alternate through a few cycles of rivalry. Then, when one pattern achieves complete dominance (such that you see no hint of the other one) click the "play" button; this will exchange the patterns between the eyes. Do you continue seeing the currently dominant pattern or do you now see the previously suppressed one? You'll want to repeate this excercise several times, and to do this you'll have to "rewind" the animation after each exchange. The swap is made "gradually" to eliminate abrupt transients which themselves can disturb rivalry (Walke & Powell, 1979).
When this experiment was along with control conditions to rule out response bias and expectation, the outcome was clear (Blake et al 1979): the dominant pattern abruptly becomes invisible and the previously suppressed pattern becomes dominant, implying that it was the region of an eye that was dominant, not a particular stimulus. Control trials confirm that this outcome is not due simply to turning the patterns "off" and then back "on."
The same outcome is observed when the rival patterns consist of meaningful objects, as you can confirm using the house/face patterns in the animation below.
Blake, R., Westendorf, D, and Overton, R., 1979: What is suppressed during binocular rivalry? Perception 9, 223 231.
Walker, P. and Powell, D.J. 1979: The sensitivity of binocular rivalry to changes in the nondominant stimulus, Vision Res., 19, 247-249.