Feeding 39procedures, internal state (e.g., deprivation) seems to act
directly to increase the animal’s tendency to engage in food-
reinforced behavior (Balleine, 1992). In contrast, in operant
conditioning procedures, the effect of internal state on behav-
ior depends on whether the animal has prior experience with
the outcome of its behavior, the reinforcer, in that deprivation
state (Dickinson & Balleine, 1994). In contrast to these ef-
fects, Davidson (1998) has shown in a Pavlovian condition-
ing procedure that the state of food deprivation on test had no
effect on approach behavior unless the animals had had prior
experience with the pellets in the undeprived state. Only rats
that had previously eaten the pellets when undeprived and
then tested undeprived showed a reduction in approach be-
havior. Just as Dickinson and Balleine (1994) interpreted
their results, Davidson (1998) interpreted this as evidence
that motivational control of Pavlovian food seeking by
hunger has to be learned through experience of the reinforcer
in both the deprived and undeprived states.
This analysis is further complicated by two additional find-
ings. The first is that as experience with the instrumental ac-
tion-outcome contingency increases, the motivational factors
underlying performance also appear to shift. Increased train-
ing seems to result in a growing importance of Pavlovian in-
centive factors (i.e., deprivation state) and a decreasing
importance of instrumental incentive learning (i.e., the incen-
tive valuation of the outcome in the animal’s current depriva-
tion state; Dickinson, Balleine, Watt, Gonzalez, & Boakes,
1995). The second is that different instrumental actions in a
chain of responding required for reinforcement appear to
be governed by different motivational factors. Instrumental
actions that occur earlier in a chain of responses seem to
be governed by the animal’s current evaluation of the rein-
forcer. In contrast, instrumental actions that occur immedi-
ately prior to reinforcer delivery appear to be directly regulated
by the animal’s current deprivation state (Balleine, Garner,
Gonzalez, & Dickinson, 1995). This latter finding—of a dis-
tinction in motivational control between proximal and distal
responses—mirrors the common distinction between appeti-
tive and consummatory responding (Craig, 1918; Konorski,
1967) that is also a component of ethological (Leyhausen,
1979; Tinbergen, 1951) and psychological theories of re-
sponse organization (Domjan, 1994; Timberlake, 1983, 1994).
Feeding Response Organization
Appetitive and Consummatory Behavior
The last two sections have dealt with initiation of feeding
and selection of food. Another important aspect of feeding
motivation concerns the topography and organization of
behaviors used to obtain food. The most influential view of
feeding response organization is based on Craig’s (1918) dis-
tinction between appetitive and consummatory behavior.
Consummatory behavior has typically been viewed as stereo-
typed responses that served as the endpoints of motivated se-
quences of behavior and could be defined by their quieting
effect on the behaving animal. In contrast, appetitive behav-
ior was conceived of as a sequence of variable but non-
random behavior that served to increase the likelihood of the
animal being able to perform the consummatory behavior by
increasing the likelihood of interaction with the goal stimulus
(Craig, 1918). Under this framework, specific examples of
feeding consummatory behavior would include acts like
chewing, swallowing, and stereotyped killing behavior such
as the throat bite used by large cats. Appetitive behavior
would include the typical behaviors of foraging such as motor
search. These concepts were further refined by Lorenz’s
(1937) analysis that redefined consummatory behavior as the
fixed action pattern of an instinct and suggested that it was
motivated by the buildup of action-specific energy. Appetitive
behavior remained undirected behavior whose function was
to increase the likelihood of the animal’s being able to per-
form the fixed action pattern by bringing it into contact with
the releasing stimulus.
Parallels between the concept of the consummatory act
and the reflex (Sherrington, 1906) and unconditioned re-
sponse (Pavlov, 1927) led to the importation of the appetitive/
consummatory distinction from ethological theorizing into the
realm of learning theory (e.g., Konorski, 1967). Whereas
ethologists distinguished between consummatory and appeti-
tive behaviors on the basis of response stereotypy, learning
theorists distinguished them procedurally. Consummatory be-
havior was investigated in Pavlovian conditioning proce-
dures, following Pavlov’s lead in examining the stimulus
control of consummatory reflexes. Appetitive behavior was
investigated in operant conditioning procedures that empha-
sized the flexibility of appetitive behavior by concentrating
on arbitrary responses and arbitrary stimuli to control perfor-
mance (Timberlake & Silva, 1995).
Although consummatory acts have been considered proto-
typically instinctive (Lorenz, 1937), careful research has
demonstrated a role for learning in the development of con-
summatory behavior. The best demonstration of this influ-
ence comes from the work of Hogan (1973a, 1973b, 1977)
on the development of feeding behavior in the Burmese red
junglefowl, a close relative of the domestic chicken. Hogan
(1973a) demonstrated that pecking behavior in newly
hatched chicks did not discriminate between food and sand
but that by 3 days of age, pecks were directed primarily at
food. At that age, ingestion of food facilitated pecking, but