Feedback Loops and Creation of Affect 193always involved. Thus, this view incorporates clear links be-
tween behavior and affect.
Comparison with Biological Models of Bases of Affect
It is useful to compare this model with the group of biolog-
ically focused theories mentioned earlier in the chapter. As
indicated earlier, those theories assume that two separate
systems regulate approach and avoidance behavior. Many as-
sume further that the two systems also underlie affect. Given
cues of impending reward, the activity of the approach sys-
tem creates positive feelings. Given cues of impending pun-
ishment, the avoidance system creates feelings of anxiety.
Data from a variety of sources fit this picture. Of particular
relevance is work by Davidson and collaborators involving
electroencephalography (EEG) recordings assessing changes
in cortical activation in response to affective inducing stimuli.
Among the findings are these: Subjects exposed to films in-
ducing fear and disgust (Davidson, Ekman, Saron, Senulis, &
Friesen, 1990) and confronted with possible punishment
(Sobotka, Davidson, & Senulis, 1992) show elevations in
rightfrontal activation. In contrast, subjects with a chance to
obtain reward (Sobota et al., 1992), subjects presented with
positive emotional adjectives (Cacioppo & Petty, 1980),
and smiling 10-month olds viewing their approaching moth-
ers (Fox & Davidson, 1988) show elevations inleftfrontal
activation. From findings such as these, Davidson (1992a,
1992b) concluded that neural substrates for approach and
withdrawal systems (and thus positive and negative affect)
are located in the left and right frontal areas of the cortex,
respectively.
Thus far the logic of the biological models resembles the
logic of our model. At this point, however, there is a diver-
gence. The key question is what regulatory processes are
involved in—and what affects result from—failure to attain
reward and failure to receive punishment. Gray (1987b,
1990) holds that the avoidance system is engaged by cues of
punishment and cues of frustrative nonreward. It thus is re-
sponsible for negative feelings in response to either of these
types of cues. Similarly, Gray holds that the approach system
is engaged by cues of reward or cues of escape from (or
avoidance of) punishment. It thus is responsible for positive
feelings in response to either of these types of cues. In his
view, then, each system creates affect of one hedonic tone
(positive in one case, negative in the other), regardless of its
source. This view is consistent with a picture of two unipolar
affective dimensions, each linked to a distinct behavioral sys-
tem. Others have taken a similar position (see Cacioppo,
Gardner, & Berntson, 1999; Lang, 1995; Lang, Bradley, &
Cuthbert, 1990; Watson, Wiese, Vaidya, & Tellegen, 1999).
Our position is different. We argue that both approach
and avoidance systems can create affects of both hedonic
tones because affect is a product of doing well or doing
poorly. We think that the frustration and eventual depression
that result from failure to attain desired goals involve the ap-
proach system (for similar predictions see Clark, Watson, &
Mineka, 1994, p. 107; Cloninger, 1988, p. 103; Henriques &
Davidson, 1991). A parallel line of reasoning suggests that
relief, contentment, tranquility, and serenity relate to the
avoidance system rather than to the approach system (see
Carver, 2001).
Less information exists about the bases of these affects
than about anxiety and happiness. Consider first relief-
tranquility. We know of two sources of evidence, both some-
what indirect. The first is a study in which people worked at
a laboratory task and experienced either goal attainment or
lack of attainment (Higgins, Shah, & Friedman, 1997, Study
4). Participants first were given either an approach orienta-
tion to the task (to try to attain success) or an avoidance ori-
entation (to try to avoid failing). After the task outcome
(which was manipulated), several feeling qualities were as-
sessed. Among persons given an avoidance orientation, suc-
cess caused an elevation in calmness, and failure caused an
elevation in anxiety. These effects on calmness and anxiety
did not occur, however, among those who had an approach
orientation. This pattern suggests that calmness is linked to
doing well at avoidance, rather than doing well at approach.
Another source is data reported many years ago by Watson
and Tellegen (1985). In their analysis of multiple samples
of mood data, they reported “calm” to be one of the 10 best
markers (inversely) of negative affect (which was defined
mostly by anxiety) in the majority of the data sets they exam-
ined. In contrast, “calm” never emerged as one of the top
markers of positive affect in those data sets. This suggests
that these feelings are linked to the functioning of a system of
avoidance.
The same sources also provide information on the mo-
mentary experience of sadness. In the study by Higgins et al.
(1997), failure elevated sadness and success elevated cheer-
fulness among persons with an approach orientation. These
effects did not occur, however, among participants who had
an avoidance orientation. The pattern suggests that sadness is
linked to doing poorly at approach, rather than doing poorly
at avoidance. Similarly, Watson and Tellegen (1985) reported
“sad” to be one of the 10 best markers (inversely) of the fac-
tor that they called positive affect in the majority of the data
sets they examined. In contrast, “sad” never emerged as one
of the top markers of negative affect in those data sets. This
pattern suggests that sad feelings are linked to the functioning
of a system of approach.