I agree entirely. Here are some examples:
Example 1: Maggie Thatcher image
There are 2 important features in the example:
The image is of a famous person (or was). That provides part of the “previous data”.
The real Maggie is on the LEFT. So readers in countries whose language is written left to right will see the real Maggie first. Again, that becomes “previous data”
Example 2: How many cubes?
Variations of this example are given in kid’s books of puzzles and brain teasers.
The principle is much more important than in the previous two examples. If the brain is unable to make a definite decision on one of two possibilities, the possibilities will be presented alternately. In other words, the cubes will “flip” so that sometimes you are looking at the TOP surface of the cubes, and sometimes you are looking at the underside. (BTW: I have a sneaking suspicion that there might be some people whose mental image does not flip. But I couldn’t find any link.)
For young children, there is no “previous data”. So the cubes will flip with each attitude being given the same amount time.
Adults are more likely to “see” the pile of cubes from the TOP. That’s based on experience. Pyramids are vaguely similar, but in that case only two surfaces of a block are visible - the front and top.
On the other hand, we hardly ever see cubes piled “upside down”. They would need to be fastened or glued together, because the structure is not self-supporting.
I have clear memories, as a kid, of the cubes “flipping” periodically, with an equal amount of time given to each attitude. Now, I need to force myself to see the cubes upside down. In fact, if I don’t blink, I can hold the normal attitude almost indefinitely. So my “previous knowledge” now is different from what it was when I was a kid.
Example 4. Tigers or just shadows at dusk?
At school we learnt about countries where there are dangerous, stripy animals such as tigers. With bushes and undergrowth at dusk, certain combinations of light and shadow can create an image of a dangerous animal.
The important feature (according to these reports) is that the brain will ALWAYS present the dangerous image first, and then either flip to a non-dangerous image (bushes and undergrowth) or persist.
If the image flips to non-dangerous, then in almost 100% of cases, the human can rotate her head, or move it slightly from side to side, and never succeed in re-creating the dangerous image. In the cubes example, the brain passes the buck by giving you first one image and then the other. The same will happen with a dangerous image - if the brain is unable to decide.
On the other hand, if a dangerous image is obviously incorrect, it presents that image once, then never again - however much you move your head. Duty done. I think I would prefer that. I would not bash my head against a rock just to punish my brain for sloppy misinterpretation.