The 'delayed-reaction method'* has been used in animal psychology to test for higher thought processes in animals. The basic idea was first put forth by H. A. Carr, but its first use in a published study was by W. S. Hunter in 1913.
The basic setup is thus: You have a number of boxes (traditionally three, in Hunter's studies), an animal, some food, and 'delay chamber' from which the animal can see the boxes but cannot get at them. You spend some time training the animal that a when a light is shown in a box, this means that there is food in that box (making sure that the food appears in a randomly assigned box). Then you place the animal in the delay chamber, turn on the light, turn off the light, wait a certain amount of time, and then open the gate to the delay chamber.
What you're looking at is the amount of time you can leave the gate closed after the light goes out, and still have the animal go to the correct box. A rat will loose interest in the correct box after being made to wait for 10-40** seconds. A dog will remember which box the food is in for up to five minutes. A raccoon will remember for 25 seconds.
The theory is that once the light is gone, the animal must have some sort of symbolic abstraction, presumably an 'idea' or 'thought' of some kind, to mark which box the food is in. This is a good theory, but it turns out the real world is slightly more complicated. In the case of rats, dogs, and cats, the symbolic abstraction isn't wholly in the mind; they often use something called 'overt orienting attitude' (these days it's usually just overt orientation ). This is when the animal holds a certain posture, specifically one in which it remains staring at the box that had the light in it. This posture seems to correlate with the animal's attention span, and may very well take the place of abstract thought (that is, they aren't looking at it because they're thinking about it; instead they go to it because they've been looking at it).
Some animals (raccoons, young humans (and older ones too), and chimpanzees, among others) don't use overt orientation; others (cats and rats) use it only part of the time.
There are two basic forms of this experiment; the indirect, given above, and the direct, first used by W. Kohler in 1926. In the direct method, the animal is allowed to see the food being placed in the box (or buried in the ground, etc.) This allows you to cut out the conditioning necessary to pair the stimulus of the light with the placement of the food. Usually this direct method will cause the animal to remember the position of the food for a longer period of time than the indirect method will.
There is a problem with this type of experiment, in that the length of time any given species will remember where the food is changes, sometimes drastically, from one experiment to another. Part of this may be the amount of reinforcement an animal receives for connecting the light (or what have you) with the food, but even with the direct method there is immense variation which has not been accounted for (last I heard, anyway. If you've come across any theories, node them!). Even so, it does seem to indicate that something's going on in those fuzzy little heads.
* There may be some other term for this type of experiment. if you know of any other terms, please let me know.
** Different experiments have gotten different times depending on various factors.