Cognitive Processes 403nest-building activity as totally irrelevant to matters of ani-
mal cognition?
As indicated, current animal-learning textbooks often treat
Pavlovian and instrumental conditioning separately from ani-
mal cognition. However, there are interpretations of Pavlovian
conditioning in terms of attention (e.g., Mackintosh, 1975)
and information processing (e.g., Pearce & Hall, 1980). To
mention a final example, many orthodox instrumental learning
phenomena, ranging from reward schedule effects to bright-
ness discrimination learning, have been said to involve com-
plex memorial processes (e.g., Capaldi, 1994). Thus, just as
the distinction between hardwired behavior and cognition
may be too sharply drawn, so too might the distinction be-
tween learning and cognition.
COGNITIVE PROCESSES
Perception
Interestingly, built into the perceptual systems of animals are
decision processes of the sort that could otherwise be medi-
ated by learning or cognition. For example, the eye of a ver-
tebrate is a complicated mechanism shaped by evolution to
solve problems of importance to a given species in its partic-
ular environment. The senses, therefore, may be regarded as
information-processing devices. Consider some examples.
Frogs have retinal “bug detectors.” The retina of the rabbit
contains several specialized mechanisms, including a “hawk
detector.” Different species of birds have different retinal dis-
tributions of photoreceptors shaped by their particular envi-
ronments. As one example, birds of prey, which tend to hunt
from above, have the densest array of photoreceptors in the
section of the retina that views the ground. Moreover, the
placement of eyes in the head varies according to an animal’s
lifestyle. In some animals, our species included, the eyes face
toward the front. In other species, the eyes are placed more to
the side of the head so as to better view stimuli from the sides
and behind. To consider still another example, bees and some
species of birds are able to detect ultraviolet light.
Some ant species send out foragers who follow more or
less random paths in their explorations. On the way out the
scouts lay down a train of scent molecules, or pheromones.
When a scout finds food it returns to the colony. A scout find-
ing more nearby food returns to the nest sooner and thus lays
down a stronger scent path. Other ants follow the stronger
path. A longer path leading to food, discovered by any other
scout, gets less traffic and its scent fades as the pheromones
evaporate. This apparently simple sensory solution to a prob-
lem of importance to the survival of ants has, according to
Peterson (2000), suggested to engineers and computer scien-
tists “powerful computational methods for finding alternative
traffic routes over congested telephone lines and novel algo-
rithms for governing how robots operating independently
would work together” (p. 314). Moreover, some computer
scientists have devised software to solve complex problems
by mimicking the pheromone-following behavior of ants.
All of the previously cited examples, from ants to bees to
frogs to birds to rabbits (not to mention echolocating in bats),
indicate that sensory systems of animals have evolved to
solve significant problems. Thus these systems, if not cogni-
tive themselves, are at least in some instances the gateways to
cognition, and they solve problems that would otherwise in-
volve cognition. Moreover, a better understanding of these
sensory systems, whether it be of pheromone-sensing ants, or
of echolocation-using bats, may provide important clues to
the operation of higher level cognitive processes.Discrimination Learning and CategorizationIn a discrimination learningstudy a hungry rat might be re-
warded with food for responding to one stimulus, say, black
(B), and nonrewarded for responding to another stimulus,
say, white (W). The two stimuli may be presented separately
on different trials (successive training) or together in the
same trial (simultaneous training). In successive training,
discrimination learning might be indexed by more vigorous
responding to B (called the positive cue,in this case, B+)
than to W(called the negative cue,W–). In simultaneous
training, B might appear irregularly on the left (B+W–) on
half the trials and on the right (W–B+) on the remaining
half. Discrimination learning might be indexed by the ani-
mal’s selection of B+when it is either to the left or to the
right of W–.
Discrimination learning has been and continues to be a
major battleground between theories that stress associations
and theories that stress other processes such as cognition or
perception. Spence’s (1936, 1937) theory of discrimination
learning is a good example of a more or less orthodox asso-
ciative theory that has battled successfully with various
nonassociative views. Spence’s theory suggests that all stim-
uli falling on the receptors when a response is made become
excitatory when rewarded (i.e., such stimuli elicit responding)
and that all stimuli falling on the receptors when a response is
made become inhibitory when nonrewarded (e.g., such stim-
uli oppose responding). Both excitation and inhibition gen-
eralize to similar stimuli (stimulus generalization), andnet
excitation(excitation minus inhibition) regulates responding.
This deceptively straightforward and simple theory is quite
powerful. First, it can explain many discrimination learning