How Researchers Test Animal Minds Without Asking

In the laboratory, animals push levers, pull strings, dig for food, swim in water mazes or respond to images on computer screens, and these tasks supply information for experiments on discrimination, attention, memory and categorization.1 Animals push levers, pull strings, dig for food, swim in mazes or respond to computer images, and the pattern of their choices is the evidence.

Tasks that turn choices into evidence

Reward is a common tool. In a typical categorization experiment, a bird or other animal sees pictures one by one on a monitor and is rewarded for pecking or touching a picture of a category item, but not for non-category items. Memory for places can be probed too: in a radial arm maze an animal starts on a platform with paths leading to goal boxes where food may be found, and a water maze is typically a circular tank of opaque, milky water.

Other tests rely on spontaneous behaviour rather than training. In the mirror mark test, an animal is marked with colour on a body part it can see only in a mirror, and mirror self-recognition is confirmed if it tries to touch or remove the mark only when a mirror is present.2 In a dolphin study of uncertainty, the animal was not trained how or when to use an uncertainty response, yet it used that response selectively on difficult trials near its discrimination threshold.3

Limits of the methods

Failing a test does not settle the question. Some researchers caution that a negative result may reflect the testing context rather than a missing ability; for example, dogs do not appear interested in their mirror image but are reported to pass a supposedly olfactory version of the mirror test.