[Martin Taylor 2003.12.19.2210 EST]
[From Bruce Gregory (2003.12.19.1919)]
I read somewhere how to reverse to Necker cube by concentrating on one
vertex or another. It seems that this illusion is based on where your
attention happens to be located.
About 40 years ago Bruce Henning and I did a few experiments on
ambiguous figures. It turns out to be relevant to this thread!
The "fact" that the Necker Cube can be seen in two configurations
turns out not to be a fact at all. It is what happens when someone is
_told_ that they will see one of two configurations. If a naive
subject is just asked to report what they see, they see a myriad of
different configurations. Not only that, but the rate at which new
configurations appear is very closely related to the number of times
a configuration change is seen (A-B-A-B is three changes).
This holds for all kinds of such illusions, in motion, in repeated
spoken phrases, in almost anything that we tried (which didn't
include the "witch-beauty" figure or the Dali jewel vase-face (which
we did try, but not in that context). The only figure we could invent
that seemed truly to show only two configurations was a flat sheet of
grey plasticene dented with a ping-pong ball, edge-lit and seen
through a peep-hole.
In the case of speech repetitions, we were able to influence directly
what people perceived by telling them (a) that all the changes they
would hear would be in English, or (b) that they might hear English
or nonsense. What they heard conformed to what they were told. We
were able to prove (or at least strongly support) that it was not
just that they reported what they were told they would hear, because
we got exactly the same functional relation between number of
configuration changes and number of different configurations under
both conditions. It was what they actually perceived that changed
with the different instructions. (Verbal transformations and an
effect of instructional bias on perception.Canad. J. Psychol., 1963,
17, 210-223 (with G.B.Henning), and Transformations of perception
with prolonged observation. Canad. J. Psychol., 1963, 17, 349-360
The point relevant to the present thread is that what people perceive
in a given pattern or movement, even at the apparently very near
physical level of configuration, depends strongly on their
expectation (call it imagination, if you will).
I suspect that is true of the lady/hag
illusion. When you discover, or are told, what to pay attention to,
control is not particularly difficult.
I think this may be an example of the same thing. What I know,
however, is that when the ambiguous figure really has only two
plausible configurations it may be easy to cause a switch voluntarily
once or twice, but it's very hard to keep being able to do it. I
wonder if this is true of the Necker Cube when you have been trained
to see only two configurations in it. I no longer can.