Testing Animal Intelligence: Methods and Challenges

[scott1]

How do they test anmials intellgence?I herd of the resarch that dogs could do math,some langue and liture(I think).How do they test that.

 

[Moonbear]

Perhaps you could provide references to those "studies" so we could assess it? There are behavioral tests of things like learning and memory, but they wouldn't be equated with intelligence or math/language skills. I'd really need to see the studies you're talking about to know whether it's credible and explainable.

 

[SciFiGuy]

There's no one test.
For instance, to test how high rats can count, researchers set up a Skinner Box (press the lever, get food) with a certain number of randomly timed electrical shocks delivered to the floor. A nervous rat hits the bar for food much less. Give the animal one shock, and it will quickly learn that only one shock is coming, and start hitting the bar more and eating more after the shock. It can also learn this for two or three shocks, indicating that it can count to three. At four shocks, the rat loses track, and remains nervous (eats less) even after the fourth shock.
For language tests, things get even crazier. The dolphin syntax experiments with Phoenix and Ake are worth Googling. In brief, the dolphins were trained a series of "word" commands - "get" "hoop" "right". Then they were tested to see if they could correctly respond when faced with pairs like "hoop bring basket" (bring the hoop to the basket) and "basket bring hoop" (bring the basket to the hoop". Other tests phrases might include "left bottle bring right water jet" (bring the bottle on the left to the water jet on the right) when the pool contained water jets and bottles on both sides.
There are one or two chimps participating in math experiments, as well. These are all trained to use symbols to communicate numbers.
I believe there have also been a few math tests using startle responses. An animal seeing something wrong will look longer at the thing than if it is right. Imagine that you put one ball behind a barrier, then put two balls behind the same barrier, and raise it to reveal what is behind. If the animal viewing this can do math, there should be a different amount of time that it spends looking at the situation if the barrier raises to reveal three balls than if it reveals four, or two, or one.
Incidentally, that test comes from one invented to test pre-verbal math skills in human infants.
I took an entire course on this in my undergrad work - it's not a small topic.