Re: Object Constancy

From: Harnad, Stevan (
Date: Thu May 01 1997 - 19:16:06 BST

> From: Taylor, Karl <>
> Isn't the size of an object one "cue" that we use when we see to help
> determine the distance of an object? (ie. as Della says, aren't the two
> things related?)

Correct. And distance cues, together with retinal size, determine the
apparent distance of an object.

Distance cues are things like the horizon: The moon "illusion" is
actually a case of a partial FAILURE of size constancy: The moon does
NOT look the same size (or distance) when it is mid-sky versus on the
horizon, even though the retinal shadow (and the distance) is the same:

When the moon is on the horizon, distance cues from further and further
objects make it look big, and also near (because the distance cues end
at the horizon, so the moon looks as if it's just "past" the horizon).

When it is in mid-sky, with no distance cues, it looks smaller and
further: smaller because the distance cues are not there to "scale" its
apparent size, and further for the same reason.

Have a look at:
and look at the rest of Tim Klitz's pages on perception at University of
for some other excellent perceptual explanations and demo's.

Here's some valuable advice: You can use Altavista as an Encyclopedia.
Just go to
and select "Advanced Search"

Then you can do worldwide searches on any keywords you like.
I found Tim's page by just searching for
"moon with illusion"
and in the second box, I put "perception"

Try it! There are some brilliant people putting terrific resources on
the Web and all you need to do is surf to them!

> For example: your growing/ moving sphere, if seen in a pitch black room
> (I guess it would have to be a glowing growing/ moving sphere) would be
> ambiguous without any further "cues" such as shadowing etc., wouldn't
> it?

That's right; so you need the depth cues for size constancy to work.

> So, although some distinction is fine generally, aren't there examples
> where there is no clear distinction?

Size constancy definitely fails when there are no depth cues, because
then the only event to go by is a growing retinal image. What is amazing
is how depth cues overpower this growing retinal shadow and lock onto an
object that stays the same size and moves closer.

This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:52 GMT