FEELINGS AND PHENOMENAL EXPERIENCES

Norbert Schwarz
Institute for Social Research
University of Michigan
Ann Arbor, MI 48106-1248
NSchwarz@umich.edu

and
Gerald L. Clore
Department of Psychology
University of Illinois
Champaign, IL 61820
GClore@s.psych.uiuc.edu

To appear in
E. T. Higgins & A. Kruglanski (Eds.),
Social Psychology. A Handbook of Basic Principles.
New York: Guilford Press.
�PG�
TABLE OF CONTENTS
I. Introduction
II. Terminological Issues: Moods, Emotions, Bodily Sensations, and Cognitive Expe
riences
A. Affect, Moods, and Emotions
B. Subjective Experiences of Knowing
III. The Interface of Feeling and Thinking
A. The Cognitive Component of Feeling States: What Comes to Mind
B. The Experiential Component of Feeling States:
Feelings as a Source of Information
C. The Somatic Component of Feeling States:
A Hard Interface?
IV. Feelings and Judgment
A. Moods and Emotions
Moods
Specific Emotions
B. Bodily Sensations
Arousal States
Facial Feedback
C. Cognitive Experiences
Ease or Difficulty of Recall
Familiarity and Perceptual Fluency
Priming as a Misattribution Process
D. When are Feelings Used in Forming a Judgment?
V. Feelings and Strategies of Information Processing
A. Theoretical Approaches
Experiential Approaches
Cognitive Approaches
B. Empirical Evidence
Persuasion
Stereotyping and Impression Formation
Other Tasks
VI. Feelings and Memory: The Encoding and Recall of Information
A. Experiential Approaches
Non-mood related
Mood related
B. Cognitive Approaches
Associative network model
-state dependency
-mood congruency
-emotion congruency
VII. The Impact of Feelings on Behavior
VIII. Conclusions
References

FEELINGS AND PHENOMENAL EXPERIENCES

For the last two decades, social psychological theorizing has been dominated by the
meta theoretical assumptions of the information processing paradigm (see Lachman,
Lachman, & Butterfield, 1979, for an introduction), which fostered the exploration
of issues that could be framed in the context of a computer metaphor (see Ostrom,
1984, Strack, 1988, for a more detailed discussion). Initially, this orientation
resulted in a focus on "cold" cognitive processes and a neglect of "hot" processes,
including the role of moods, emotions, and related phenomena in human cognition and
behavior. Since the early 1980's, however, social psychologists have more than
compensated for this neglect, and the interplay of affect and cognition has become
one of the most productive, and controversial, areas of research in social
psychology. More recently, social and cognitive psychologists have turned to
another set of subjective experiences that are also difficult to capture with the
computer metaphor, namely the phenomenal experiences that accompany thought
process. Much as the emotional feelings elicited by thinking about emotional
issues may influence thought processes, non-emotional experiences, elicited by
thinking may also affect cognitive processing. These include feelings of
familiarity, the experienced ease or difficulty with which some material can be
brought to mind, and similar self-produced experiences. Moods, emotions, bodily
sensations, and the cognitive experiences accompanying thought all share an
experiential quality that allows us to conceptualize their operation with a single
set of assumptions, as we shall see below.

In the present chapter, we review the key assumptions of research concerning the
impact of feelings and other phenomenal experiences on human cognition and
behavior. In line with the general approach taken in this handbook, our focus is
on basic theoretical principles and empirical regularities, rather than complete
coverage of the literature. We begin by distinguishing different kinds of
experiential cues that provide information about different kinds of processes,
including affective, cognitive, and bodily processes. Next, within the affective
domain, we distinguish three broad approaches to the interface of affect and
cognition. One emphasizes the cognitive content of affect, focusing on the
thoughts brought to mind by affective states, a second approach emphasizes the
experience of affect, focusing on affective feelings as a source of information
about an organism's environment. A third approach emphasizes hard-wired processes,
focusing on the somatic components of affective states. Reflecting the current
state of social psychological research, we emphasize the first two approaches in
our review of the literature.

Following the exposition of the three approaches to affective states, we examine

theory and evidence concerning their roles, first in human judgment, then in
strategies of information processing, and finally in the encoding and recall of
material from memory. As might be expected, much of the research that has been
done on phenomenal experience concerns affect and emotion, but we also review
research on the cognitive consequences of nonaffective feelings. For each of these
domains, the theories that have been offered sometimes make contradictory
predictions. In addition, a number of ad hoc assumptions have been introduced that
are often at odds with one another. Nevertheless, agreement can be reached on many
replicable findings and their theoretical conceptualization, as our review will
indicate.

Terminological Issues:

Moods, Emotions, Bodily Sensations, and Cognitive Experiences

Three broad classes of feelings can be distinguished in terms of the information
they provide, including �UL1�bodily feelings, such as hunger or pain; affective
feelings, such as happiness, sadness or fear; and feelings associated with knowing,
such as feelings of familiarity, confusion, or amazement. While each of these
kinds of self-produced experiences are different, they appear to function similarly
by providing information about their respective domains. Bodily feelings indicate
the state of various bodily systems, and affective feelings reflect appraisals of
situations with respect to one's goals and concerns. In a similar way, certain
cognitive experiences reflect one's state of knowledge, as in the experience of
insight and surprise.
In information processing terms, each of these kinds of feelings provides
information with motivational implications or feedback properties (Carver &
Scheier, 1990). As discussed in later sections, the information serves as input to
judgment (e.g., Schwarz & Clore, 1983) and decision making processes (e.g., Isen &
Means, 1973), and influences processing priorities (e.g., Simon, 1967) and
strategies (e.g., Schwarz, 1990).

Affect, Moods, and Emotions

Various terminological distinctions between affect, emotion, and mood have

been offered (see Morris, 1989, for a review), making it important that we clarify
our use of these terms. Affect is sometimes used as a synonym for emotion but can
also refer simply to valence -- the positive and negative aspect of things. All
emotions are affective, but not all affective things are emotions. Preferences and
attitudes, for example, may be thought of as affective dispositions, while emotions
are affective states. Clore, Ortony, and Foss (1987) characterize emotions as
"internal, mental states focused on affect." However, that characterization does
not distinguish emotions from moods, which are also affective states.

A variety of distinctions between mood and emotion have been proposed. For
example, Nowlis and Nowlis (1956), argued that moods involve the entire person and
tend to endure for some time, while emotions are more temporary. They indicate
that moods are often experienced as less intense than emotions and may have causes
that are more obscure. Others have suggested that emotions concern appraisals of
external circumstances, while moods concern the state of internal resources
(Morris, 1992), or that emotions concern the present, while moods concern
anticipation of the future (Batson, Shaw, & Oleson, 1992). However, we shall
emphasize a distinction made by Averill (1980) that emotions require an object
while moods do not. Thus, "mood" generally refers to the feeling state itself,
while "emotion" refers both to the feelings and what the feelings are about. These
differences are apparent in ordinary language when we say that we are afraid "of"
something or angry "about" something, but that we are "in" a bad mood.

Within this view, cognitive causes may be seen as essential for emotion but
not for mood (Clore, Ortony, Dienes, & Fujita, 1993). Thus, the absence of
sunlight during long periods of cloudy weather is said to cause a depressed mood in
certain individuals. If so, the resulting mood might not reflect negative
appraisals of the person's situation of the kind found in emotions but merely a
diminished level of a particular hormone. In a similar way, anti-depressant drugs
presumably change the mood of depressed individuals directly by altering their
neurochemistry. These changes may also increase the likelihood of particular
emotions by biasing attention or by amplifying particular appraisals. The basic
cognitive model is that "emotion" refers to the consequences of ongoing, implicit
appraisals of situations with respect to positive or negative implications for
one's goals and concerns (e.g., Arnold, 1960), while "mood" refers to feeling
states themselves, when the object or cause is not a focus.

In summary, the prototypical emotion is assumed to have an identifiable

referent (what the emotion is "about"), a sharp rise time, high intensity, and
limited duration. The prototypical mood, on the other hand, lacks a clear
referent, may come about gradually, has a low intensity, and may last for an
extended period of time. In many cases, the experience of a positive or negative
emotion may also leave us in a positive or negative mood after the emotion
dissipates (Bollnow, 1956). As we shall see below, it is the unfocused nature of
moods that accounts for their pervasive influence on cognitive processes.

While considering emotions, some mention should be made of their presumed

function in the overall design of the organism, and in particular the function
served by their experiential aspect. From an information processing perspective,
Simon (1967) suggested that emotions serve as cognitive interrupts. Since
processing resources are limited, it is important the ongoing processing is
interrupted when more urgent situations arise. This appears to be accomplished
through the attention-grabbing and motivating properties of emotional feeling.

Cognitive resources are generally allocated to whatever one attends to. In

both higher and lower organisms, attention is responsive to sensory stimulation, so
that high intensity and rapidly changing stimuli in the environment, such as bright
lights, loud noises, and movement draw one's attention. However, since loudness
and brightness are not always the most important attributes of situations, this
arrangement is not fully adequate. The problem has been partially solved through
the evolution of a system of internally generated stimulation in the form of
affective feeling (Clore, 1994). Presumably using the same mechanisms that allow
attention to be guided by external stimuli, the development of emotional experience
enabled attention deployment to be partially freed from external control. As with
stimuli from the environment, the more intense these internal cues, the more
completely they command attention and redirect processing. The adaptiveness of
this system is suggested by evidence that the intensity of emotional feeling
generally reflects the importance of the situations to which they are reactions
(Frijda, Ortony, Sonnemans, & Clore, 1992). The appraisal of situational
importance, is, in turn, controlled by cognitive processes that calculate the
relevance of situations to personal goals and standards (Ortony, Clore, & Collins,
1988). In addition, evolved symbolic capacity allows past situations to be
mentally replayed, and future or hypothetical situations mentally simulated. As a
result, the internal emotional stimuli that guide attention can be still further
removed from immediate external situation.

The subjective experiences associated with affect and emotion concern the
evaluation of situations with respect to their goodness or badness. We turn next
to feelings that reflect cognitive rather than affective states, the subjective
experiences that concern one's state of knowledge, understanding, ignorance, and
confusion.

Subjective Experiences of Knowing

The least studied kind of subjective experiences are those that signal one's
state of knowledge. William F. Brewer (1992) has recently surveyed the status of
phenomenal experience in theories of memory. He concludes that contemporary
cognitive psychology shows a "neglect of conscious experience" (Tulving, 1989, p.
4), but that at the beginning of the century cognitive skills such as recognition
memory were often seen as based on unique phenomenological feelings (Calkins, 1905,
p. 257; Strong, 1913; Titchener, 1910, p. 407). In modern memory theory, Mandler
(1980) has offered a two stage model of recognition memory in which the first stage
concerns the degree to which an item is experienced as familiar. But a focus on
such experiential qualities as feelings of familiarity, feelings of confidence in
recognition, feelings of unfamiliarity, and related feelings is more characteristic
of earlier than of recent accounts (e.g., Woodworth, 1938; Moore, 1939).

In a related vein, the French physicist, Henri Poincare (1913, pp. 390-392)
Thanks to Jerry Parrott for the Poincare quote
�, refers to the feelings that accompany mathematical insight as follows:
�LM8�I have spoken of the feeling of absolute certitude accompanying the
inspiration, [....] often this feeling deceives us without it being any the less
vivid [....]. When a sudden illumination seizes upon the mind of the
mathematician, it usually happens that it does not deceive him, but [when it does]
we almost always notice that this false idea, had it been true, would have
gratified our natural feeling for mathematical elegance.�LM0�
Poincare suggests that feelings concerned with knowing often lead us to make
accurate metacognitive judgments about our insights, but that they are equally
compelling when they are false.
One reason for the paucity of research on subjective experiences of knowing,
or cognitive feelings (Clore & Parrott, 1991), may be that they so rarely occur
apart from the mental content that elicits them. Sometimes, however, they do occur
without relevant mental content, as in the tip-of-the-tongue phenomenon or in
feelings of knowing. Some relevant work has been done by developmental
psychologists interested in metamemory and in reading. In these areas one can
encounter studies of such topics as "the illusion of knowing" (Glenberg, Wilkinson,
& Epstein, 1982), "subjective certainty" (DeLoache & Brown, 1984), "realizing that
you don't know" (Markman, 1977), and "feeling of knowing experiences" (Wellman,
1977). For example, Harris studied the reactions of eight- and eleven-year-old
readers to anomalous sentences in a brief story (Harris, Kruithof, Meerum Terwogt,
& Visser, 1981). Although only the older group indicated that the anomalous lines
did not fit the story, both groups read them more slowly, suggesting that at both
ages children generate "internal signals of comprehension failure," but that only
the older children can use such signals as information to locate the obstacle to
their comprehension.
In this section we have described two of the kinds of feelings with which we
are concerned -- affective feelings and feelings concerned with knowing. In the
next we focus on cognitive, experiential, and somatic aspects of affective feeling
states.

The Interface of Feeling and Thinking

As Zajonc and Markus (1984) observed, "The interface of affect and cognition
presents a considerable theoretical and experimental challenge, because it is not
clear just how and where the affect-cognition interface should best be studied" (p.
73). In the present chapter, we focus on only one aspect of this interface, namely
the influence of emotion on cognitive processes. We do not, however, address the
impact of cognition on emotion (for a review of this literature see Clore, Schwarz,
& Conway, 1994). Below, we summarize the key assumptions of three general
approaches to the impact of feelings on thought processes. These approaches
emphasize the cognitive, experiential, and somatic components of affect
respectively. For each one, we identify the core propositions, necessarily
glossing over differences in the details of related models. Some of these
differences will be addressed as we review the available data in subsequent
sections.
The Cognitive Component of Feeling States:
What Comes to Mind
By far the largest amount of recent social psychological research into
affective influences on human cognition and behavior has been guided by the
assumption that our feelings influence the content of our thoughts. Whereas
earlier discussions of affective influences often employed perceptual metaphors,
referring, for example, to "rose colored glasses" and the like (e.g., Bollnow,
1956), research since the mid 1970's has been dominated by a focus on recall
mediated processes, in line with the general focus of information processing
models. In social psychology, this focus is reflected in extensive research into
the encoding, organization, storage, and retrieval of information about others,
often referred to as person memory research (see Wyer & Srull, 1989, for a review).
As social cognition researchers became interested in the impact of affective states
on social judgments, they observed that evaluative judgments are typically more
positive under elated than under depressed moods. Across a wide range of content
domains, from satisfaction with consumer goods (e.g., Isen, Shalker, Clark, & Karp,
1978) or past life-events (e.g., D. M. Clark & Teasdale, 1982) to reports of
happiness and satisfaction with one's life as a whole (e.g., Schwarz & Clore,
1983), subjects in elated moods report more positive evaluations than subjects in
depressed moods (see Clore et al., 1994; Forgas, 1992; Schwarz, 1990, for recent
reviews).
Given the assumptions of general judgment models, researchers naturally
assumed that these mood effects on evaluative judgments and behavioral decisions
are mediated by emotional influences on the retrieval of information from memory.
This was first suggested by Isen and colleagues (e.g., Isen et al., 1978) and later
elaborated by Bower (1981; 1991), who conceptualized the operation of affective
states in the framework of a general network model of human memory (Anderson &
Bower, 1973). Specifically, Bower (1981) proposed an associative network model in
which emotions are represented by emotion nodes (see also Bower, 1991; Bower &
Cohen, 1982; Gilligan & Bower, 1984). According to this model, emotions function
as central nodes in an associative network, which are linked to related ideas,
events of corresponding valence, autonomic activity, and muscular and expressive
patterns. When new material is learned, it is associated with the nodes that are
active at the time of learning. Accordingly, material that is learned while in a
particular affective state is linked to the respective emotion node. When an
emotion node is stimulated, activation spreads along the pathways, increasing the
activation of other nodes connected to the emotion node. Activation of a node
above a certain threshold brings the represented material into consciousness.
This and related models (e.g., Isen et al., 1978;, Spies & Hesse, 1986; Wyer
& Srull, 1989) generate four key predictions. The first prediction pertains to
state dependent learning and recall and holds that material learned in one
affective state is more likely to be recalled in the same state than in another
state. Whereas this prediction pertains to matching states at learning and recall,
a mood congruency hypothesis pertains to matching valences of the affective state
and the to-be-recalled material. It holds that positively valenced material is
more likely to be recalled in positive moods, whereas negatively valenced material
is more likely to be recalled in negative moods. Although usually referred to as
the mood congruency hypothesis, this prediction is equally applicable to specific
emotions. This congruency effect of mood at recall is presumably independent of
the influence of mood at encoding, and hence conceptually distinct from state
dependency effects. In specific cases, however, they are difficult to distinguish
(see Morris, 1989). Mood congruent recall has been most reliably observed in the
domain of autobiographical memory, with happy events being more likely to be
recalled under happy than sad moods (see Blaney, 1986; Clore et al., 1994, for
reviews). However, happy events are likely to have put one into a happy mood at
the time they occurred. Accordingly, mood congruency in autobiographical recall
may reflect the operation of state dependency as well as mood congruency. This
suggests that mood induced selective recall may be most likely to occur when both
processes operate in conjunction, so that the mood at recall matches the mood at
learning as well as matching the valence of the to-be-recalled material.
If moods affect the recall of valenced material, they are also likely to
affect both the encoding of new material and judgments formed at the time.
Accordingly, a mood congruent encoding hypothesis suggests that ambiguous material
will be encoded in terms of concepts of the same valence as one's mood. Moreover,
associations that come to mind subsequently (regardless of the ambiguity of the
material) may be mood congruent, resulting in mood congruent elaboration. By the
same token, mood at the time of judgment should also facilitate the recall of mood
congruent information about the target, resulting in mood congruent judgments, that
reflect a selective data-base.
In summary, models that emphasize the cognitive component of affect assume
that affective influences on encoding, elaboration, and judgment are mediated by
selective recall of mood congruent information stored in memory. The most
prominent of these models (Bower, 1981) traces increases in the accessibility of
mood congruent information to the activation of mood nodes connected to that
material.
These assumptions can also be used to predict affective influences on styles
of information processing. If a given affective state brings a large amount of
congruent information into consciousness, it may limit the cognitive resources
required for working on a task to which this material is irrelevant, thus
interfering with cognitive performance. However, researchers disagree with regard
to which mood is most likely to constrain our cognitive resources.
In their resource allocation model, Ellis and Ashbrook (1988) suggested that
negative affective states are likely to reduce the resources that can be allocated
to a given task. In fact, several studies indicated that people in an induced or
chronic negative mood have difficulty suppressing mood-congruent material when
instructed to do so (e.g., Howell & Conway, 1992; Wenzlaff, Wegner, & Roper, 1988).
As a result of intruding thoughts and ruminations, negative affective states may
therefore interfere with information processing that requires more than minimal
amounts of attentional resources. Other researchers, however, have argued that
positive affective states can limit attentional resources (e.g., Isen, 1987; Mackie
& Worth, 1989). Isen, for example, hypothesized that positive material in memory
"is more extensive and at the same time better integrated, so that positive affect
is able to cue a wide range of thoughts" (Isen, 1987, p. 217). If so, being in a
positive mood may limit cognitive resources due to intruding positive thoughts.
In either case, the (negative or positive) mood congruent material intruding
into consciousness would limit the resources that could be allocated to a given
task, resulting in impaired performance. However, the logic of the underlying
process assumptions renders it difficult to predict positive or negative moods are
more likely to impair performance. As we shall see below, being in a good mood
facilitates the recall of positive material more reliably than being in a bad mood
facilitates the recall of negative material, suggesting that being in a good mood
may indeed bring more intruding material to mind. On the other hand, the events
that elicit a bad mood are more likely to require attention and to trigger causal
explanation (e.g., Bohner, Bless, Schwarz, & Strack, 1988), as well as extensive
rumination (e.g., Martin & Tesser, 1989), which should also disrupt performance on
other tasks. We return to these issues below.
Finally, we note that models of the type addressed here do not provide an
easy way to conceptualize the role of cognitive experiences. Although they can
account for the differential activation of stored information, incorporating
experiential variables, such as the ease with which information comes to mind, is
difficult without making assumptions extraneous to associative network models.
Moreover, their focus on what comes to mind sometimes leads to predictions that are
incompatible with results showing that the subjective experience of ease or
difficulty of recall also has an impact, as we shall see below. Similarly,
differences in the perceived informational value of one's immediate experience are
difficult to conceptualize within recall based models. Next, we turn to an account
that emphasizes the role of the experiential components of feelings in judgment and
information processing.

The Experiential Component of Feeling States:

Feelings as a Source of Information
An experiential approach to the influence of moods and emotions on human
thought and behavior is well in line with traditional theorizing in the domain of
emotions (see Frijda, 1988, for a review). Cognitively oriented emotion
researchers generally assume that "emotions exist for the sake of signaling states
of the world that have to be responded to, or that no longer need response and
action" (Frijda, 1988, p. 354). Similarly, mood researchers generally assume that
moods reflect the general state of the organism (Nowlis & Nowlis, 1956), an
assumption that prompted Jacobsen (1957) to refer to moods as "barometers of the
ego." Despite the common acceptance of these assumptions, the informational
implications of immediate experience have often been neglected in social
psychological research. According to the models of information processing that
provided the conceptual framework for social cognition research, judgments are
based on information retrieved from memory. Useful as that approach has been in
many domains, it seems to miss something essential, as the following example may
illustrate.
Suppose you are asked how you like your lunch. As Clore (1992) noted,
traditional judgment models would have you proceed by identifying the ingredients,
looking up their stored liking values in memory, and integrating these values to
form a judgment. While in a good mood, the ingredients associated with a positive
liking value may be more accessible, resulting in a more favorable evaluation of
your lunch. If you have evaluated lasagna before, on the other hand, you may not
need to go through the computational task, but may identify your lunch as lasagna,
look up the stored value for lasagna, and conclude that you like your lunch because
the stored value for lasagna is positive. "Such an account seems to be missing
something. Surely liking also reflects information from the direct experience of
tasting lasagna as well as indirect influences from the retrieval of previously
stored knowledge about lasagna" (Clore, 1992). In a similar way, the impact of
moods, emotions, bodily sensations, or cognitive experiences may often reflect the
direct use of one's immediate experiences as a source of information. The
implications of this assumption have been elaborated most explicitly for the impact
of feelings on evaluative judgments and the choice of processing strategies.
According to the view that feelings may serve informative functions,
individuals may use their apparent affective response to a target as a source of
information in evaluating the target (e.g., Schwarz & Clore, 1983, 1988; Wyer &
Carlston, 1979). This possibility is most obvious when a judgment refers, by
definition, to one's affective reaction to the stimulus. For example, when asked
how much we like a person, we may base the judgment on our feelings towards them
rather than on a review of their attributes. Also, when a judgment does not refer
directly to one's feelings about a target but rather poses a task that is
particularly complex and demanding, we may base the judgment on our feelings to
simplify the task. Rather than computing a judgment on the basis of recalled
features of the target, individuals may ask themselves, "How do I feel about it?"
(Schwarz & Clore, 1988). In doing so, however, it is difficult to distinguish any
pre-existing feelings that happen to be present at the time from feelings that
actually reflect an appraisal of the target. Hence, individuals may mistake
feelings due to a pre-existing affective state as a reaction to the target
stimulus, resulting in more positive evaluations under positive than under negative
moods.
In such situations, the impact of feelings on evaluative judgments should
depend on their perceived informational value. Most important, the informational
value of feelings for the judgment at hand should be called into question when we
attribute them to a source that is irrelevant to the evaluation of the target
stimulus. Under such conditions, the feelings appear uninformative about the
target and do not affect evaluation of it. As elaborated elsewhere (Schwarz,
1990), this proposition does not imply that mood effects on evaluative judgments
require a conscious attribution of one's feelings to the target. As we shall see
repeatedly below, the default operation is that we consider our subjective
experiences, ranging from our feelings to the thoughts that come to mind, to
reflect our response to whatever we are attending to at the time. Accordingly,
although discrediting these experiences often involves conscious attributions,
reliance on the experiences generally does not involve conscious attribution.
The assumption that affective states inform the organism about the nature of
its current environment may also be used to explain the influence of affect on
information processing strategies. Several lines of research, to be reviewed
below, suggest that positive moods foster heuristic information processing, while
negative moods foster more effortful, systematic processing. Approaches based on
mood congruent recall trace such findings to the impact of intruding thoughts, but
an experiential approach traces them to the motivational properties of the
information provided by affective states.
Presumably, "emotions arise in response to the meaning structures of given
situations, (and) different emotions arise in response to different meaning
structures" (Frijda, 1988, p. 349). In general, "events that satisfy the
individual's goals, or promise to do so, yield positive emotions; events that harm
or threaten the individual's concerns lead to negative emotions" (p. 349). The
relationship between emotions and the meaning structures that constitute a
psychological situation is presumably bi-directional so that specific emotions not
only are caused by particular psychological situation, but also provide information
to the experiencer about that situation (Schwarz, 1990). Thus positive affective
states can inform the individual that a situation is safe, that it does not
threaten the person's current goals. Negative affective states, on the other hand,
can inform the individual that the current situation is problematic, that is
characterized by a lack of positive outcomes or by a threat of negative ones. If
so, one's affective state could serve as a simple but highly salient indicator of
the nature of the situation one is in.
To the extent that individuals are motivated to obtain positive outcomes and
to avoid negative ones, negative emotions would therefore inform the individual
that some action may need to be taken. Appropriate action, however, initially
requires a careful assessment of the features of the current situation, an analysis
of their causal links, detailed explorations of possible mechanisms of change, as
well as anticipation of the potential outcomes of any action that might be
initiated. Moreover, individuals may be unlikely to take risks in a situation that
is already considered problematic, and may therefore avoid simple heuristics as
well as novel solutions. Accordingly, their thought processes may reflect a
detail-oriented systematic processing style.
Positive emotions, on the other hand, may not signal a particular action
requirement and may inform the individual that his or her personal world is
currently a safe place. Hence, the individual may see little need to engage in
cognitive effort, unless this is required by other currently active goals or the
task is inherently enjoyable. In pursuing these goals, the individual may also be
willing to explore new solutions, given that the general situation is considered
benign. Thus, heuristic strategies may be preferred to more effortful, detail
oriented judgmental strategies; new procedures and possibilities may be explored;
and unusual, creative associations may be elaborated. Accordingly, the spontaneous
thought processes of individuals in a positive affective state may be characterized
by a more heuristic processing style, which is less detail-oriented and systematic,
but which allows for greater flexibility and creativity.
An approach that focuses on the informational properties of subjective
experience offers a parsimonious account of the role of feelings in human judgment
and allows for the conceptualization of moods, emotions, bodily sensations and
cognitive experiences within a unified framework. Moreover, its predictions
regarding the impact of feelings on judgment and processing style are sufficiently
distinct from the predictions generated by mood congruent recall models to allow
empirical testing, as we shall see below. On the other hand, this approach is
relatively silent with regard to emotional influences on the recall of information
from memory. Next, we turn to the somatic component of feelings.
The Somatic Component of Feeling States:
A Hard Interface?
The somatic component of feelings has been addressed in two different ways.
As Zajonc and Markus (1984) noted, most theories of emotion that address the role
of somatic processes postulate some form of experiential mediation between those
somatic processes and emotional judgments or other outcomes (e.g., Izard, 1977;
Leventhal, 1982; Schachter & Singer, 1962; Tomkins, 1962). These theories focus on
the role of the subjective experience of bodily changes, facial muscle changes, or
autonomic arousal as a crucial link in emotional process. Consistent with this
perspective, we treat somatic processes in the context of experiential approaches
in the present chapter, referring to somatic experiences as bodily feelings, as
outlined in our discussion of terminological issues.
As an alternative approach, Zajonc and Markus (1984) suggested that the
impact of somatic processes may not be experientially mediated, but may reflect
hard-wired processes. For example, Zajonc and collaborators emphasized possible
representational functions of the motor system (see Adelman & Zajonc, 1989; Zajonc
& Markus, 1984) and suggested a crucial role for the vascular system of the head in
emotion regulation (see Zajonc, Murphy, & Inglehart, 1989). Much of this work
focuses on somatic processes as determinants of emotion and little is known about
the likely role of somatic processes in mediating the consequences of affective
states. Reflecting the paucity of research that bears on the mediational issue,
and the limits of our own expertise, we will not address this perspective in the
present chapter, which focuses on feelings, that is subjective experiences (but see
Cacioppo, this volume).

Feelings and Judgment

In this section, we review research on the impact of different feelings on
social judgment. Reflecting the bulk of the available work, we first review
research on moods and emotions; subsequently, we explore the information conveyed
by bodily sensations and cognitive experiences. Where applicable, we contrast the
predictions generated by approaches discussed above concerning the activation of
affectively congruent content in memory versus the use of immediate subjective
experience.

Moods and Emotions

According to the affect-as-information view (Schwarz & Clore, 1983; Wyer &
Carlston, 1979) our affective response to a target is a useful source of
information when evaluating the target. In this process, we essentially ask
ourselves, "How do I feel about this?" (see Schwarz & Clore, 1988, for a more
detailed discussion). As we have only one window on our immediate experience,
however, we may also mistake feelings due to a pre-existing state as part of the
reaction to the target stimulus, resulting in more positive evaluations under
positive than under negative moods. As indicated earlier, the impact of our
feelings on evaluative judgments should depend on their perceived informational
value. Accordingly, we should discount our feelings as a source of relevant
information when there is reason to assume that they do not reflect our reaction to
the target. On the other hand, our feelings should seem particularly informative
when our apparent reaction to the target contradicts the plausible impact of other
influences.
Such discounting and augmentation effects (Kelley, 1972) cannot be derived
from the assumption that moods or emotions affect judgment through the selective
recall of information from memory (e.g., Bower, 1981) or by hard-wired processes
(e.g., Zajonc & Markus, 1984). These assumptions predict main effects of affective
states, while the feelings-as-information hypothesis predicts an interaction
between affective states and the perception of their likely causes (Schwarz &
Clore, 1983).
�MDBU�Moods
In line with this interaction prediction, Schwarz and Clore (1983) observed
that the impact of mood on judgments of life satisfaction was eliminated when
subjects attributed their current feelings either correctly (Experiment 2) or
incorrectly (Experiment 1) to a transient source. For example, subjects reported
higher life satisfaction, and a more elated current mood, in telephone interviews
when called on sunny rather than rainy days. This difference was eliminated,
however, when the interviewer mentioned the weather as part of a private aside,
thus directing subjects' attention to this source of their elated or depressed
feelings (Experiment 2). Similarly, recalling a sad life event did not influence
subjects' judgments of life satisfaction when they could misattribute the resulting
sad feelings to the alleged impact of the experimental room (Experiment 1). In
addition, current mood, as assessed at the end of the experiment, was more strongly
correlated with judgments of life satisfaction when subjects' attention was not
directed to a transient source of their feelings than when it was. Conceptual
replications of these findings have been reported by Keltner, Locke, and Audrain
(1993), Schwarz, Servay, and Kumpf (1985), and Siemer and Reisenzein (1994), among
others.
In combination, these findings indicate that individuals may use their
current feelings as a basis of judgment unless the diagnostic value of their
feelings for the judgment at hand is called into question. As noted above, this
discounting effect (Kelley, 1972) is incompatible with predictions based on mood
congruent recall, because the attributional manipulations discredit only the
implications of one's current feelings, not the implications of valenced
information about one's life recalled from memory. Similarly, it is incompatible
with models that trace affective influences to hard-wired processes (e.g., Zajonc &
Markus, 1984), without providing a role for inferential processes.
Moreover, the content by which mood was induced proved largely irrelevant for
the type of judgment affected, in contrast to what models of mood congruent recall
would predict. Suppose that a depressed mood is induced by thoughts about a
serious illness. If so, the valence of the mood and the content of the mood
induction should facilitate the recall of illness related material from memory,
according to network models (e.g., Bower, 1981). Hence, illness related judgments
should be more likely to show mood effects than judgments pertaining to other
content domains. Empirically, however, this is not the case (e.g., Clore, Schwarz,
& Kirsch, 1985; Johnson & Tversky, 1983; Mayer, Gaschke, Braverman, & Evans, 1992;
see Schwarz & Clore, 1988, for a review). For example, Johnson and Tversky (1983)
observed that reading descriptions of negative events, which presumably induced a
depressed and slightly anxious mood, increased judgments of risk across a wide set
of targets. The impact of mood was independent of the object of judgment or the
content by which the mood was induced. Reading about cancer, for example, affected
judgments of the risk of cancer, but had equally strong effects on judgments of the
risk of accidents and divorce. Such generalized effects, undiminished over
dissimilar content domains, are incompatible with models of mood congruent recall.
However, they are consistent with the assumption that individuals often make
judgments by consulting the feelings that the stimulus apparently elicits. If they
feel depressed and anxious, they may conclude that a risky event is indeed
depressing and threatening and evaluate the risk as more severe than they would
under a more positive mood. Accordingly, the impact of mood is independent of the
content by which it is induced, unless its informational value is discredited. In
addition, other researchers have observed mood effects on evaluative judgments in
the absence of any evidence for mood effects on the recall of relevant information
from memory (e.g., Fiedler, Pampe, & Scherf, 1986), further indicating that mood
effects on evaluative judgments may be independent of mood congruent recall.
�MDBU�Specific Emotions
The same basic logic applies to the use of specific emotions as a source of
information. For example, Schwarz et al. (1985) observed that the impact of a fear
arousing communication on subjects' attitudes was eliminated when subjects
attributed their subjective experience to the arousing side-effects of a pill, but
was enhanced when subjects assumed the pill would have tranquilizing effects.
However, the informational value of specific emotions is more restricted than the
informational value of global moods. As discussed above, emotions are specific
reactions to specific events (see Clore et al., 1994, for a review of emotion
theories), whereas moods are of a diffuse and unfocused nature (see Morris, 1989,
for a more detailed discussion). As a result of this undifferentiated and
unfocused nature, mood states may be used as information in making a wide variety
of different judgments. In fact, when subjects are induced to attribute their mood
to specific causes -- as in the Schwarz and Clore (1983) experiments reviewed above
-- its impact on judgments that are unrelated to that source vanishes. In
contrast, the source of a specific emotion is more likely to be in the focus of
attention, thus rendering the emotion uninformative for judgments that are
unrelated to the emotion inducing event. This suggests that the informational
value of specific emotions is more restricted than the informational value of
global moods.
This hypothesis is supported by research by Keltner et al. (1993), who
induced a sad mood by having subjects vividly imagine a negative life event.
Subsequently, some subjects were asked to describe "what emotions" they felt
currently, whereas others indicated where and when the negative event took place.
Having to label their current feelings with specific emotion terms should induce
subjects to identify specific causes for their current feelings, thus undermining
their informational value for judging a target unrelated to these specific causes.
Consistent with this reasoning, subjects who labeled their current feelings with
specific emotion terms reported significantly higher life satisfaction than
subjects who identified the time and location of the event, despite being in a
similarly depressed mood. In fact, describing one's current specific emotions was
as effective in reducing the impact of a sad mood as misattributing one's sad
feelings to the experimental room.
If specific emotions imply the identification of a specific cause, they may
be unlikely to affect unrelated judgments shortly after their onset, when the event
that elicited them is still salient. Once, the emotion dissipates, however, it may
leave the individual in a diffuse mood state (as described by Bollnow, 1956), which
would be likely to affect a wide range of judgments. If so, the impact of an
emotion eliciting event may be quite limited shortly after the onset of the
emotional reaction, but may generalize to unrelated targets after a sufficient
delay. To our knowledge, data bearing on this possibility are not yet available.
In addition to being more restricted with regard to the range of applicable
targets, the informational value of specific emotions is also more specific than
that of diffuse moods. On theoretical grounds, we may expect that the information
conveyed by specific emotions reflects the implications of the appraisal pattern
that underlies the respective emotion. In line with this hypothesis, Gallagher and
Clore (1985) observed that feelings of fear affected judgments of risk but not of
blame, whereas feelings of anger affected judgments of blame but not of risk.
Similarly, Keltner, Ellsworth, and Edwards (1993) observed in several experiments
that angry subjects assigned more responsibility to human agents than to impersonal
circumstances, whereas sad subjects assigned more responsibility to impersonal
circumstances than to human agents, reflecting the differential appraisal patterns
of anger and sadness.
In combination, the reviewed findings illustrate that, unlike global mood
states, specific emotions have very localized effects on judgment. Much as in the
case of moods, however, these effects are eliminated when the informational value
of the emotion for the judgment at hand is called into question (e.g., Keltner at
al., 1993; Schwarz et al., 1985). As we noted in our discussion of moods, the
discounting and augmentation effects obtained in this domain are difficult to
reconcile with the assumption that the impact of affective states is mediated by
selective recall or reflects hard-wired processes, each of which would predict main
effects of the affective state.
�MDBU�Affective Reactions Without Experience?
Some researchers maintain that affective influences on judgment may occur in
the absence of a phenomenal experience of affect (e.g., Damasio, 1994; LeDoux,
1987; Zajonc, 1980). In our reading, the data bearing on this assumption are
mixed. For example, Murphy and Zajonc (1993) observed that subjects evaluated
unknown Chinese ideographs more positively when they were preceded by the
subliminal presentation of a smiling rather than frowning face. In terms of the
feelings-as-information framework, this suggests that subjects relied on their
apparent gut reaction to the ideographs when the subliminal presentation mode
precluded the insight that their gut reaction pertained to a smiling or frowning
face, rather than the to-be-evaluated ideograph. Consistent with this explanation,
the effect was only obtained when the faces were presented subliminally, but not
when they were presented supraliminally. Apparently, supraliminal presentation
conditions allowed subjects to identify the actual source of their feelings, thus
rendering them nondiagnostic for the evaluation of the ideographs (but see Murphy &
Zajonc, 1993, for another account).
Whereas this difference between subliminal and supraliminal presentation
conditions is consistent with a feelings-as-information account, explicit attempts
to manipulate the perceived diagnosticity of subjects' affective reactions to
subliminally presented faces failed to undermine the observed effect. In an
extended replication of the Murphy and Zajonc (1993) study, Winkielman, Zajonc, and
Schwarz (1995) informed subjects that a smiling (or frowning, respectively) face
would precede each ideograph (Experiment 1) or exposed subjects to music said to
elicit positive or negative feelings (Experiment 2). Neither of these
misattribution manipulations resulted in the expected augmentation or discounting
effects. Moreover, subjects did not report experiencing any "gut" reactions during
the course of the experiment.
Although it is difficult to draw strong conclusions from the null result of
misattribution manipulations, these failures -- in combination with the apparent
absence of conscious affective experiences -- cast doubt on a feelings-as-
information account of subliminal affective priming studies. In the absence of
experienced feelings, affective priming studies may indeed be better conceptualized
as reflecting automatic evaluation processes (see Bargh, 1994, this volume, for
reviews), which have been observed with materials unlikely to elicit any feelings
(e.g., Bargh, Litt, Pratto, & Spielman, 1989), rather than feeling-based
inferences.
Bodily Sensations
The research on affective feelings and judgment reviewed above indicates that
incidental feelings affect judgment only when the nature of the feelings, the
salience of their causes, or other aspects of the situation allow them to be
experienced as reactions to the object of judgment. Research reviewed in the next
section suggests that this conclusion also applies to the impact of bodily
sensations on judgment.
�MDBU�Arousal States
All of the research on internal cues using attributional manipulations stems
in part from Schachter and Singer's (1962) original demonstration that the effect
of arousal on emotion-relevant judgment and behavior depends on attributions about
its source. The research was stimulated by the belief that undifferentiated states
of autonomic arousal could be transformed into emotions as different as anger and
euphoria simply by arranging for subjects to attribute the arousal as a reaction to
angering or amusing aspects of the experimental situation. Although the idea had
enormous influence among social psychologists at the time, the experimental
situation they so cleverly devised did not yield strong data, and over time other
more elaborate accounts of emotional variation have superceded it. Despite this
fact, Schachter and Singer's original paper does give a good account of how
implicit attributions for one's ongoing reactions in a situation are involved in
the construction of situational meaning. Moreover, their idea that the effects of
arousal on emotional reactions are not necessarily automatic as everyone had
assumed has proved enormously productive. In agreement with their assertions, the
research reviewed here suggests that many of the important effects of autonomic
arousal are mediated by its experiential aspects and depend on perceptions of its
source and significance.
For example, exploring the impact of heightened excitation levels, Zillman
and his colleagues (e.g., Zillman, Johnson, & Day, 1972; see Zillman, 1978 for a
review) had subjects engage in various forms of exercise. Shortly after the
exercise, no impact of increased excitation level was observed, presumably
reflecting that subjects were still aware of its source. After some delay,
however, subsequent judgments were affected by the residual arousal. Apparently,
subjects misinterpreted their subjective experience of arousal as a reaction to the
target, once the temporal distance of the exercise rendered this alternative source
less accessible and plausible.
Similarly, Martin, Harlow, and Strack (1992) observed that induced facial
expressions affected the interpretation of an ambiguous social situation, and that
these effects were more pronounced when subjects were aroused. However, the impact
of arousal and of facial expression was largely eliminated when subjects
(correctly) attributed their arousal to physical exercise, thus rendering their
subjective experience of excitement uninformative for the judgment at hand.
Whereas Zillman and his colleagues attributed findings of this type to a process of
"excitation transfer", it is important to note that similar effects have been
obtained in the absence of any actual excitation, based on false feedback (e.g.,
Valins, 1967, 1974). This suggests that the perception of arousal does not need to
be veridical, in contrast to the assumption that some actual excitation needs to be
present to be "transferred". The apparent information value of the perceived
arousal appears to be the critical factor.
In another well-known program of research, Zanna and his colleagues (see
Zanna & Cooper, 1976, for a review) observed that cognitive dissonance effects were
eliminated when subjects could attribute the resulting arousal state to some other
source. In combination, these lines of research indicate that individuals may draw
on their perceived arousal state as a source of information, unless its
informational value is called into question, much as we have seen above for other
feelings.

�MDBU�Facial Feedback
Darwin (1872/1965) observed that "Most of our emotions are so closely
connected with their expression that they hardly exist if the body remains passive"
(p. 257). This assertion implies that expression is an important, perhaps even a
necessary part of emotion. James (1890) took a more extreme view, maintaining that
emotions were nothing more than the awareness of such expressive and bodily
changes. According to James, "We feel sorry because we cry, angry because we
strike, and afraid because we tremble" (p. 243). The belief that facial, postural,
and behavioral expressions contribute to emotional experience is still popular
today (but see Mandler, 1975, for a dissenting view). Tomkins (1962), for example,
has argued that the quality of an emotional experience is governed by motoric
feedback (e.g., from postures and facial expressions), while the intensity of the
experience is governed by physiological arousal. Laird (1974, p. 476) illustrates
that idea in the following way:
�LM8�I am angry rather than euphoric or frightened because I am frowning,
clenching my fists, and gritting my teeth, and I am angry rather than just annoyed
because my heart is pounding, I have butterflies in my stomach, and I feel
flushed.�LM0�
Evidence consistent with the hypothesis that facial expressions might result in
distinctive emotional experiences has been reported by Ekman, Levenson, and Friesen
(1983). They found that posing different facial expressions of emotion (e.g.,
surprise, disgust, sadness, fear, anger, and happiness) triggered different
patterns of autonomic activity (e.g., changes in heart rate, skin temperature, and
skin conductance).
These findings imply that expression might affect emotion by creating
patterns of autonomic activity that in turn contribute to emotional experience.
However, it is also possible that the effects depend on the sensory nervous system
rather than (or in addition to) the autonomic nervous system. That is, expression
may contribute to emotion not only indirectly, as a result of changes in heart
rate, skin conductance, or other autonomic changes, but directly as a result of
experiencing the proprioceptive or muscular cues themselves.
For example, Strack, Martin, and Stepper (1988) had subjects hold a pen in
their mouths in such a manner that a smile was either facilitated or inhibited.
While doing this, they were also asked to rate the funniness of cartoons. Holding
a pen between one's teeth requires that one contract the same muscles that are used
when one smiles, while holding a pen between one's lips results in pursing the
lips, rendering a smile impossible. Although subjects were apparently not aware of
the meaning of these patterns of muscle contractions, they reported greater
amusement at the cartoons when the muscle contractions resembled a smile than when
they did not.
In a subsequent study, Stepper and Strack (1993) arranged for subjects to
contract their forehead muscle (corrugator) while trying to recall experiences in
their past exemplifying their self-assurance. Such muscular cues generally
indicate mental effort, and accordingly, subjects whose recall attempts were
accompanied by cues of mental effort rated themselves as less self-assured, a
finding to which we return in the context of cognitive experiences. Stepper and
Strack (1993) also showed that such processes are not unique to facial feedback.
They found that subjects rated themselves as more proud after success on an
achievement task if they received their performance feedback in an upright posture
rather than in a slumped posture.
Two decades ago, Bem (1972) proposed that we have little access to our own
inner states, and that we determine how we feel just as others do, primarily by
observing our behavior and expression. However, the experiments reviewed above
suggest that however important such self-observation may be, we do attend to inner
cues. Indeed, the research reported by Strack and colleagues (Strack et al., 1988;
Stepper & Strack, 1993) is especially interesting because the method makes it
implausible that the results are mediated either by experimenter demand (e.g., "The
experimenter wants me to smile and hence to rate these cartoons as funny") or by
self-observations of behavior (e.g., "Since I am smiling, I must think these jokes
are funny"). Rather it seems likely that relevant proprioceptive feedback is part
of the mix of cues that makes up the phenomenal experience of such states as humor
and pride (along the other elements of the situation), and that these are all
experienced more or less holistically.
Of course many emotional episodes do not appear to involve smiles, grimaces,
postural changes, or any other readily observable muscular contractions. However,
research by Cacioppo, Bush, and Tassinary (1992; see Cacioppo, this volume)
suggests that emotions often involve invisible, subthreshold activation of
particular facial muscles. Their data suggest that such patterns are detectable
even when the emotional reaction lasts only a few hundred milliseconds. It seems
plausible that such changes in muscle potential are important contributors to
affective experiences, providing informational feedback to the individual
regardless of whether full blown expressions occur that would provide such
information to others.
To date, we are unaware of studies examining whether misattributions of
information from facial or postural muscles would eliminate their effects on
emotional experience. However, Olson and Roese (in press) observed that
misattribution manipulations may undermine the impact of emotional expression on
judgments of the emotion eliciting quality of stimuli. Specifically, they exposed
subjects to a comedic monologue and instructed them either before listening to
deliberately inhibit overt displays of mirth, or informed them after listening that
the laboratory setting inhibits overt mirth. Compared to a control group without
specific instructions, subjects instructed to inhibit overt mirth showed fewer
smiles and laughs, but gave ratings of the funniness of the monologue that were
similar to those of the control group. Subjects who were informed that their mirth
was inhibited by the laboratory setting, on the other hand, showed levels of overt
mirth that were similar to the control group, but rated the monologue to be funnier
than did control group subjects, reflecting an augmentation effect. Unfortunately,
the manipulation used in this study was too global to identify the contributions of
specific muscular feedback. Also the dependent variable pertained to the emotion
eliciting quality of the monologue rather than to subjects' emotional experience
per se. Nevertheless, these findings suggest that misattribution procedures may be
useful for exploring the role of bodily sensations in emotional experience.
Cognitive Experiences
The same experiential theme is reflected in research into the role on non-
emotional experiences that may accompany the thought process, such as the
experienced ease or difficulty of recall, perceptual fluency, or feelings of
familiarity (see Clore, 1992; Jacoby & Kelley, 1987, for more extended reviews).
Again, these experiences, occasionally referred to as "cognitive feelings" (Clore,
1992), affect individuals' judgments only if their informational value for the
judgment at hand is not called into question, as some selected examples may
illustrate.
�MDBU�Ease or Difficulty of Recall
Tversky and Kahneman's (1973) availability heuristic holds that individuals
estimate the frequency of an event, or the likelihood of its occurrence, "by the
ease with which instances or associations come to mind" (Tversky & Kahneman, 1973,
p. 208; see Sherman & Corty, 1984, for a review). Unfortunately, the assumed
crucial role of the subjective experience of ease of recall remained ambiguous in
most studies bearing on this heuristic. For example, in the most frequently cited
study (Tversky & Kahneman, 1973, Experiment 8) subjects were read two lists of
names, one presenting 19 famous men and 20 less famous women, and the other
presenting 19 famous women and 20 less famous men. When asked, subjects reported
that there were more men than women in the first list, but more women than men in
the second list, although the opposite was the case (by a difference of one).
Given that subjects were able to recall about 50 percent more of the famous than of
the non-famous names, however, it remains unclear, what drives the overestimate:
Are subjects' judgments based on the experience of the "ease" or "difficulty" with
which they could bring the famous and non-famous names to mind, or on the content
of their recall, with famous names being overrepresented in the recalled sample?
Similar ambiguities apply to other studies. Throughout, the manipulations designed
to increase the experienced ease were also likely to affect recalled content (see
Schwarz, Bless, Klumpp, Rittenauer-Schatka, & Simons, 1991; Taylor, 1982, for more
detailed discussions).
To disentangle this confound, Schwarz et al. (1991) asked subjects to recall
either six or twelve examples of situations in which they either behaved
assertively and felt at ease, or behaved unassertively and felt insecure.
Recalling six examples was experienced as easy, while recalling twelve examples was
experienced as difficult, although all subjects could complete the task. If
subjects relied solely on the content of recall, they should report higher
assertiveness after recalling twelve rather than six examples of assertive
behavior, and lower assertiveness after recalling twelve rather than six examples
of unassertive behavior. However, the reverse was true. Specifically, self-
ratings of assertiveness showed that subjects rated themselves as less assertive
after recalling twelve rather than six examples of assertive behavior, and as more
assertive after recalling twelve rather than six examples of unassertive behavior,
in contrast to what the implications of the recalled examples would suggest. In
fact, subjects rated themselves as more assertive after recalling twelve
unassertive rather than twelve assertive behaviors (Experiment 1). Apparently, the
experience that it was difficult to recall twelve examples suggested to subjects
that they must not be very unassertive (or assertive). Similar results have been
reported by Stepper and Strack (1993), who induced the subjective experience of
ease or difficulty of recall through differential facial feedback, as reviewed
above.
Consistent with an experiential interpretation, inducing subjects to
misattribute the experienced ease or difficulty of recall to the alleged impact of
meditation music played to them (Schwarz et al., 1991, Experiment 4), reversed the
otherwise obtained pattern. Specifically, subjects relied on the content of their
recall in forming a judgment when the misattribution manipulation discredited the
informational value of the subjective experience. In this case, they reported
higher assertiveness after recalling twelve rather than six examples of assertive
behavior, and lower assertiveness after recalling twelve rather than six examples
of unassertive behavior. In contrast, the pattern described above replicated when
the informational value of the experienced ease or difficulty of recall was not
called into question. A conceptual replication of these misattribution findings
has been reported by Rothman and Hardin (1994).
As these and related findings (e.g., W�nke, Schwarz, & Bless, in press)
indicate, the impact of the experienced ease of recall depends on its perceived
informational value, as we have seen for affective experiences in the research
reviewed above.
�MDBU�Familiarity and Perceptual Fluency
Feelings of familiarity, which may result from the experienced ease of
processing has also been found to affect a wide range of judgments. For example,
Jacoby, Kelley, Brown, and Jasechko (1989) had subjects read a list of names,
informing them that they pertained to nonfamous individuals. Immediately
afterwards or one day later, subjects were presented a second list of names,
including some of the previously presented ones as well as other nonfamous and
famous names. Asked to identify the names of famous individuals, subjects
erroneously identified previously presented nonfamous names as famous. This
effect, however, was more pronounced after a one day delay. This presumably
reflects that subjects tested immediately after the initial exposure were likely to
attribute their experience of familiarity correctly to the previous presentation,
whereas subjects tested a day later attributed this experience to the name being
famous. Similar phenomena have been observed in recognition studies (e.g., Jacoby
& Dallas, 1981), which we address in the memory section of this chapter.
The subjective experience of familiarity or perceptual fluency has also been
used to provide a process account for the classic mere exposure effect, first
identified by Zajonc (1968). In a typical mere exposure experiment, subjects are
exposed to nonsense words or Chinese ideographs at different exposure frequencies
and are subsequently asked to rate the stimuli along some evaluative dimension,
such as liking or connotation of goodness. In most studies, subjects were found to
attribute more positive connotations to the more frequently seen stimuli and to
prefer them over less frequently seen ones (see Bornstein, 1989, for a meta-
analysis). Recent research suggests that these effects are, at least in part,
mediated by the subjective experience of familiarity. For example, Bonanno and
Stillings (1986) exposed subjects to subliminally presented polygons and assessed
their feeling of familiarity for previously seen and unseen polygons, in addition
to preference and recognition judgments. They obtained higher ratings of
preference and familiarity for previously seen polygons than for new polygons,
whereas subjects' recognition performance was at a chance level. If feelings of
familiarity contribute to higher preference judgments, the mere exposure effect
should be particularly pronounced when subjects are not aware that their
differential feelings in response to the stimuli may simply reflect different
exposure frequencies. Confirming this hypothesis, Bornstein and D'Agostino (1992)
showed that the same stimuli produced a stronger mere exposure effect when they
were presented under subliminal rather than supraliminal exposure conditions, thus
precluding subjects' awareness of exposure frequency. Complementing these results,
Bornstein and D'Agostino (1994) manipulated subjects' attribution of their
subjective experiences. In one study, they observed reduced liking judgments when
subjects could attribute the experience of familiarity to the subliminal exposure
procedure, reflecting a discounting effect. In a second study, they observed
increased liking judgments when subjects were discouraged from attributing the
experience of familiarity to the impact of a supraliminal exposure procedure. In
combination, these findings document the role of experienced familiarity in the
operation of the classic mere exposure effect, consistent with an experiential
approach to judgmental processes (see Bornstein & D'Agostino, 1994, for a more
detailed discussion).
�MDBU�Priming as a Misattribution Process
Much more common than studies of the experience of retrieval are studies of
the content of retrieval. In the past twenty years, much of the research on social
cognition has focused on the effects of priming on what comes to mind. Priming is
generally seen as an automatic process, which can influence subsequent thoughts
even if one is unaware of exposure to the priming stimulus (see Bargh, 1994;
Higgins, 1989, this volume, for reviews). But whether priming has an influence on
subsequent processing and judgment again seems to depend on implicit attributions
about the source of the primed thoughts and its relevance to the task at hand
(Clore, 1992; Clore & Parrott, 1991). As several studies indicate, we may only
draw on the thoughts that come to mind when we do not attribute their emergence in
consciousness to the impact of an irrelevant source.
Following Higgins, Rholes, and Jones' (1977) initial demonstration of priming
effects in impression formation, social cognition researchers initially assumed
that increasing the accessibility of a trait concept would always result in
assimilation effects, provided that the behavior is ambiguous and the trait concept
applicable. All studies that demonstrated assimilation effects, however, were
based on very subtle priming manipulations. For example, the terms used as primes
are sometimes exposed subliminally so that subjects are not fully aware of having
seen them (Bargh & Pietromonaco, 1982). They have also been presented as
background stimuli in a Stroop task (Higgins, Rholes, & Jones, 1977) or embedded in
other apparently irrelevant tasks (Srull & Wyer, 1979). However, studies that used
more blatant priming manipulations (e.g., Martin, 1986), or that reminded subjects
of the priming episode prior to asking for a judgment (e.g., Strack, Schwarz,
K�bler, & W�nke, 1993), have obtained contrast rather than assimilation effects.
Moreover, Lombardi, Higgins, and Bargh (1987) observed that assimilation effects
were obtained only for subjects who were unaware of the primes, but not for
subjects who could recall some of the primes.
These findings suggest that awareness of the source of one's thoughts plays a
crucial role in obtaining priming effects. When a trait concept is primed in a
subtle manner, subjects are likely to be unaware that it has occurred to them
because of the experimental procedures. As information processors, we operate on
the default assumption that thoughts coming to mind as we observe a behavior
reflect our own reaction to that behavior. Hence, we use the trait that comes to
mind to characterize the target person, resulting in assimilation effects. Not so,
however, when we have reason to assume that the trait has come to mind because of
the preceding task. When the concept does not seem to reflect our reaction to the
behavior, it tends not to be included in one's characterization of the target. As
Martin and his colleagues (e.g., Martin, 1986; Martin, Seta, & Crelia, 1990)
suggested, we are usually motivated to form an independent and unbiased impression
of other persons. If we suspect that our impression may be unduly influenced by an
irrelevant source, we are likely to correct it, as has first been demonstrated by
Martin (1986). This correction usually results in an impression that is biased in
the direction opposite to the primed concept, resulting in contrast effects (see
Lombardi et al., 1987; Martin, 1986; Martin et al., 1990; Strack et al., 1993 for
examples, and Martin, 1986; Schwarz & Bless, 1992; Strack, 1992 for somewhat
different accounts).
In combination, these findings indicate that we rely on the primed
information only when we assume that the thoughts that come to mind reflect our
reaction to the target. If the apparent diagnosticity of the thoughts is
undermined through awareness of an irrelevant source, then priming effects are not
obtained. One implication of this experiential view of priming effects is that the
mechanisms generally appealed to in social cognitive studies of priming are, by
themselves, insufficient to explain priming effects. Priming explanations must be
supplemented with what might be called a "thoughts-as-information" corollary (see
Higgins, this volume, for a step in this direction). From this perspective, one
can see that, like mood effects, standard priming effects on judgment also depend
on implicit misattributions for one's momentary emotional and mental experience
(see also Clore, 1992). That is, subjects must take the thoughts that occur to
them as reactions to the stimulus to be judged rather than as a consequence of the
priming manipulation or some other irrelevant influence.
�MDBU�Summary
The reviewed examples bear on the informational value of various internal
experiences including the ease or difficulty of recall, feelings of familiarity,
and trains of thought. They indicate that nonemotional experiences that accompany
the reasoning process may themselves serve as a basis of judgment, and may qualify
the implications of recalled content (see Clore, 1992, for a review of other non-
emotional experiences). These processes seem to underlie many of the findings in
the implicit memory literature (see Greenwald & Banaji, 1995, for a review). As in
the case of affective experiences, however, we do only draw on these experiences
when their information value is not called into question.

When Are Feelings Used in Forming a Judgment?

We have suggested that self-produced affective, bodily, and cognitive
experiences may be more prevalent and may play a more important role in cognitive
functioning than has generally been recognized. Important questions concern not
only how feelings affect information processing but also when or under what
conditions this takes place. In order for feelings to influence information
processing, one must attend to them. In the case of momentary, intense emotions,
feelings forcefully intrude upon consciousness, but in the case of milder or more
sustained feelings, such as moods or felt preferences, attending to experiential
data is optional. This raises questions concerning the conditions under which we
rely on our feelings rather than on other sources of information (see Strack, 1992,
for a related discussion).
At present, the empirical evidence bearing on this issue is largely
restricted to the impact of moods. As reviewed in more detail elsewhere (Clore et
al., 1994), this evidence suggests that reliance on one's mood as a source of
information is particularly likely under the following four conditions: First, when
the judgment at hand is affective in nature (e.g., liking for another person);
second, when little other information is available; third, when the judgment is
overly complex and cumbersome to make on the basis of a piecemeal information
processing strategy; and, fourth, when time constraints or competing task demands
limit the attentional resources that may be devoted to forming a judgment. In
addition, there may be individual differences in people's inclination to attend to
their feelings. These conditions may be differentially likely to hold for
phenomenal experiences other than affective states, as we note below.
As any other source of information, subjective experiences are used in
forming judgments only when they seem to bear on the particular kind of judgment to
be made. Hence, it is not surprising to learn that mood is especially likely to
influence judgments when one is asked explicitly how one feels about an object of
judgment. For example, judgments of liking and preference have been found to be
strongly influenced by respondents' moods (e.g., Clore & Byrne, 1974; Zajonc,
1980). Specific emotions, on the other hand, appear to affect judgments only along
dimensions that bear on the emotion's appraisal pattern, as reviewed above (e.g.,
Gallagher & Clore, 1985; Keltner et al., 1993). Which class of judgments is
affected by which nonemotional experience, on the other hand, is less well
understood. With regard to the two most frequently investigated experiences, we
may conclude that experiences of ease of recall affect judgments of frequency,
likelihood, and typicality, as predicted by Tversky & Kahneman's (1973)
availability heuristic (see Sherman & Corty, 1984, for a review). However, the
informational value of feelings of familiarity or of perceptual fluency may be more
general than the available evidence bearing on recognition (e.g., Jacoby & Dallas,
1981), judgments of fame (e.g., Jacoby et al., 1989), truth value (e.g., Begg,
Armour, & Kerr, 1985) or preference (e.g., Bornstein & D'Agostino, 1994) may
suggest.
Also not surprisingly, reliance on one's feelings is likely when little other
information is available. When subjects are asked to evaluate unknown future risks
(e.g., Johnson & Tversky, 1983), or the fame of a name that brings no substantive
association to mind (e.g., Jacoby et al., 1989), they may have little choice but to
rely on their perceived gut reaction.
Leading to a related conclusion, some studies indicate that the impact of
feelings decreases as the amount or salience of competing information increases.
For example, Srull (1983, 1984) reported that subjects' moods influenced their
evaluations of unfamiliar, but not of familiar products. In addition, Strack,
Schwarz, and Gschneidinger (1985, Experiments 2 and 3) observed that subjects who
provided short, non-emotional reports of a past life event used this event as a
standard of comparison, resulting in contrast effects on judgments of current life
satisfaction. Subjects asked to report a past life event in an emotionally
involving style, on the other hand, relied on the elicited mood state in evaluating
their current life satisfaction, resulting in assimilation effects (see Clark &
Collins, 1993; Clark, Collins, & Henry, 1994, for conceptual replications).
Findings of this type suggest that individuals may rely either on thought content
or on the feelings elicited, depending on which source of information is more
salient. In general, as the salience or intensity of the affective state increases
other sources of relevant information may be increasingly ignored.
Individuals may also differ in their tendency to focus on experiential as
opposed to descriptive information. For example, Carver and Scheier (1981) showed
that individuals high in Private Self-Consciousness made more extreme affective
judgments than those who were low on this dimension. The dimension of Private
Self-Consciousness is intended to tap enduring differences in the tendency to focus
on inner thoughts and feelings. Experiential focus can also be manipulated. For
example, recent research (Clore, Wong, Isbell, & Gasper, 1994) shows that after
rating a series of odors (an experiential focus), subjects are more likely to base
their judgments on feelings and are more susceptible to moods than after solving
syllogisms (an analytic focus).
Whereas the preceding findings indicate that individuals may consult their
feelings due to a lack of other relevant information, they may also do so because
too much information is available or because limited attentional resources do not
allow for the systematic use of available information. In either case, asking
oneself how one feels about the target may provide an efficient heuristic that
greatly simplifies the judgmental task and limits the demands on attentional
resources. Consistent with this assumption, Borg (1987), Schwarz, Strack, Kommer,
and Wagner (1987, Experiment 1) and Levine, Wyer, and Schwarz (1994) observed
pronounced mood effects on global judgments, that would require the consideration
of a large amount of information and the integration of its implications, but did
not obtain mood effects on more specific judgments that could easily be computed in
a piecemeal fashion. Moreover, Siemer and Reisenzein (1994) observed that mood
effects on judgments of life satisfaction were more pronounced when subjects were
under time pressure, or worked on a secondary task, while forming the judgment,
thus restricting their attentional resources.
In combination, these findings suggest that the use of one's feelings as a
basis of judgment often reflects a simplifying heuristic strategy. Most of the
currently available evidence pertains to the impact of global moods, but there is
little reason to assume that the underlying logic should not hold for other
phenomenal experiences. If individuals engage in a more effortful, piecemeal
processing strategy, on the other hand, reliance on one's feelings as a source of
information should be reduced. Forgas (1992) suggested that mood effects on
evaluative judgment may be mediated by mood congruent recall under these
conditions. However, in the absence of (mis)attribution manipulations that
discredit the informational value of subjects' current feelings, it is difficult to
determine whether the use of feelings as information or mood congruent recall
drives any particular instance of the moods on evaluative judgments, rendering many
studies ambiguous with regard to the underlying process.
Summary
The reviewed findings indicate that individuals may use both affective and
non-affective phenomenal experiences as sources of information in making judgments.
However, they will rely on these experiences only if they perceive them as
reactions to the object of judgment. If the experiences are attributed, either
correctly or incorrectly, to some other source, their informational value for the
judgment at hand is called into question, eliminating the otherwise observed
effect.

�NJ�Feelings and Strategies of Information Processing

During recent years, numerous findings have suggested that affective states may
influence individuals' spontaneous adoption of heuristic or systematic strategies
of information processing. Whereas this proposal seems uncontroversial, there is
little agreement as to which affective state may be associated with which type of
processing strategy and which of a number of possible underlying mechanisms are
involved. Affect induced differences in processing strategy have been attributed
to differences in attentional resources, differences in motivation, or some
combination of these factors. Below, we review the key proposals and findings.

Theoretical Approaches
�MDBU�Experiential Approaches
Several authors have proposed models that entail an experiential approach to
the impact of moods on processing strategies, although the models differ with
regard to the specific information that they assume moods provide. They either
assume that affective states inform us about the state of the environment (Schwarz,
1990; Weary, Marsh, Gleicher, & Edwards, 1993); the contingency of hedonically
relevant rewards (Wegener, Petty, & Smith, 1994); or that they serve as input into
specific decisions (Martin, Ward, Achee, & Wyer, 1993). All of these models assume
that the choice of heuristic or systematic processing strategies reflects
differences in motivation, rather than, for example, differences in the cognitive
capacity of happy and sad subjects.
Mood and the state of the environment. Two related approaches have
emphasized that affective states may influence processing motivation by signalling
different states of the environment. Exploring the impact of chronic states of
mild depression, Weary and her colleagues (see Weary et al. 1993, for a review)
have proposed that depression is characterized by uncertainty about one's ability
to understand, predict, and control one's social environment and to produce desired
outcomes, as suggested by Seligman's (1975) learned helplessness model. At least
at mild and moderate levels of depression, this uncertainty is supposed to give
rise to control motivation, which leads to the adoption of an accuracy goal. This
accuracy goal, in turn, presumably results in a systematic style of information
processing that can be characterized as highly resource-dependent, vigilant, and
complex, with more extensive processing of and search for relevant information.
Consistent with this control motivation assumption, other research has shown that
temporary exposure to uncontrollable situations may elicit increased attributional
activity, more careful and deliberate processing of available information, and
increased information search (e.g., Pittman & D'Agostino, 1985, 1989; Pittman &
Pittman, 1980).
This proposal emphasizes dispositional differences between depressed and non-
depressed individuals, and it traces depression and the accompanying processing
style to experiences of uncontrollability. A similar proposal has been offered to
account for the impact of temporary moods on processing strategy. Emphasizing the
experiential component of feelings, Schwarz and his colleagues (Schwarz, 1990;
Schwarz & Bless, 1991; Schwarz, Bless, & Bohner, 1991) suggested that our feelings
inform us about the nature of our current psychological situation, as discussed in
our initial review of theoretical approaches. Accordingly, negative affective
states would inform individuals that their current situation is problematic, and is
either characterized by a lack of positive or a threat of negative outcomes. This
would presumably increase individuals' temporary control motivation, much as Weary
et al. (1993) assume for chronic depression. In contrast, positive affective
states would inform individuals that their situation is safe and does not threaten
current goals; positive outcomes are not lacking, and there is no threat of
negative outcomes. Thus, being in a good mood would signal that the current
situation is safe, presumably decreasing individuals' temporary control motivation.
Assuming that our cognitive processes are tuned to meet the requirements signalled
by our feelings, we may expect to see the following differences in the processing
strategies that are spontaneously adopted by individuals in a happy or sad mood:
First, sad individuals may be more willing to invest cognitive effort than happy
individuals, who may only do so when required by some other currently active goal
or when the activity is inherently enjoyable. Second, sad individuals may be more
likely to process information at a lower level of abstraction, paralleling the
shift in focus of attention that has been observed when things are not going as
expected (Wegner & Vallacher, 1986). In contrast, happy individuals may be more
likely to rely on pre-existing global knowledge structures, which have served them
well in the past. If so, the processing style of sad individuals may be more data-
driven than the processing style of happy individuals, as has recently been
suggested by Bless (1994). Finally, situations characterized as benign may afford
greater opportunities for the exploration of novel ideas and new solutions than
situations characterized as problematic. As a result, happy individuals may be
more likely than sad individuals to engage in playful creativity and the
exploration of new options (see Isen, 1987, for supporting data and a different
account, to be addressed below).
In summary, the assumption that our feelings inform us about the nature of
our current situation (Schwarz, 1990) leads to predictions about the impact of
temporary sad moods that are compatible with Weary et al.'s (1993) assumptions
about chronically depressed individuals. However, it also predicts differences as
a function of being in a happy mood that cannot be derived from the analysis of
depressive thought. Finally, both models predict less systematic information
processing under severe chronic depression, although for different reasons.
According to Weary et al., severely depressed individuals have expectations of
extreme uncontrollability, which undermine the elicitation of control motivation.
From the feelings-as-information perspective, it is noteworthy that
phenomenological studies of severe depression suggest that the experience of
"sadness" or of "being in a bad mood" may not be part of the melancholic state of
severe depression (see T�lle, 1982, p. 232 ff. for a review). Hence, the
information conveyed by chronic melancholy may differ from that conveyed by milder
dysphoric states. Moreover, any feeling that extends over prolonged periods of
time may eventually loose its informational value for specific situations.
Mood and reward contingencies. Whereas the above takes on the issue assume
that our moods inform us about the nature of our current psychological situation,
Wegener, Petty, and Smith (1994) suggested that different moods may signal
different reward contingencies, with potential implications for mood management.
Although couched in terms of learning theory, their approach shares the basic
assumption that moods serve informative functions. According to their reasoning,
"one might consider a person's mood as a discriminative stimulus that signals what
kind of hedonic reinforcement contingencies are operating" (p. 9). When one's
current mood is positive, almost any activity is likely to strip one of one's
current good mood, unless the activity is at least as pleasant as the mood itself.
As a result, happy people are expected to scrutinize the hedonic implications of
possible activities. Hence, they may only engage in systematic information
processing if the material that they may process is pleasant, but not if it is
unpleasant. Conversely, when one's current mood is negative, almost any activity
may be more pleasant than the mood one is in. Hence, one may engage in activities
without scrutinizing their hedonic implications. As a result, "happy people should
manage mood by paying attention to the hedonic consequences of future action to a
greater extent than sad people" (p. 9), a hypothesis that the authors refer to as
the hedonic contingency hypothesis (see also Wegener & Petty, 1994).
With regard to processing strategies, this hypothesis predicts that happy
people engage in systematic processing strategies if they expect the outcome to be
rewarding, but not if they expect it to be unpleasant. Sad people, however, are
supposed to engage in systematic processing irrespective of its hedonic
implications, although we suppose this should only hold when they don't expect the
outcome to be worse than they feel anyway. We return to this issue below, when we
review the relevant findings. For the time being, we note that the hedonic view
underlying this and related conceptualizations of mood regulation has not gone
unchallenged. As Erber (in press) noted, people's striving towards positive
feeling states may be less prevalent than commonly expected and people are more
than willing to forego shortterm positive experiences in the interest of longterm
gains.
Mood as input for specific decisions. Whereas the above approaches assume
that our feelings inform us about the state of the environment (Schwarz, 1990,
Weary et al., 1993) or the contingency pattern of rewards (Wegener et al., 1994),
Martin et al., (1993) suggested that our feelings serve as input for more specific
decisions about the task or one's task performance. According to their variation
of the feelings-as-information theme, individuals may evaluate their performance by
asking themselves, "How do I feel about it?" (Schwarz & Clore, 1988). As reviewed
in the preceding section of this chapter, such reliance on one's apparent affective
response would result in more positive performance evaluations when in a happy
rather than sad mood. As a result, happy individuals may infer high performance
satisfaction and may hence terminate their efforts. In contrast, sad individuals
may infer low performance satisfaction and may hence expand more effort. Empirical
findings support this conjecture. Specifically, Martin et al. (1993) asked
subjects to generate a list of birds from memory and instructed some subjects to
ask themselves how satisfied they are with their performance and to stop when they
feel that this "is a good time to stop". Under this performance-oriented decision
rule, sad subjects spent more time listing birds, and listed more birds, than happy
subjects.
If mood serves as input for specific decisions, however, its impact should
depend on the decision at hand. Suppose that subjects are not trying to evaluate
how well they are doing but how much they enjoy the task. In that case, being in a
good mood should indicate high enjoyment, whereas being in a bad mood should
indicate low enjoyment. Martin et al. (1993) tested this possibility by
instructing other subjects to stop when they "no longer enjoy the task." Under this
enjoyment-related decision rule, sad subjects spent less time on the task, and
listed fewer birds, than happy subjects, presumably reflecting that they were more
likely to infer that they no longer enjoyed the task. Finally, subjects who were
not given an explicit decision rule also spent more time on the task, and listed
more birds, when they were in a sad rather than in a happy mood, suggesting that
they spontaneously adopted a task-oriented decision rule. A conceptual
replications of these findings have been reported by Hirt, Melton, McDonald, and
Harackiewicz (1995) and Levine, Hirt, McDonald, Melton, and Martin (1995). In the
latter study, Levine et al. also observed that the otherwise obtained impact of
moods was eliminated when their informational value was discredited, as predicted
by the feelings-as-information framework.
As these results indicate, the same affective state may result in different
performance behavior, depending on the decision rule for which it serves as input:
Whereas sad moods result in more effort than happy moods under a performance-
oriented rule, the reverse holds true under an enjoyment rule. Depending on the
nature of the task and situational circumstances, individuals may spontaneously
adopt one or the other rule, resulting in opposite effects of mood on effort
expenditure. We may assume, however, that a performance orientation is likely to
predominate in most performance settings, hence facilitating effort expenditure
under sad mood in most cases. Moreover, one's mood may itself influence which
decision rule one is likely to adopt. To the extent that being in a sad mood
signals a problematic situation, a performance-oriented decision rule may seem more
appropriate than an enjoyment-oriented rule. In contrast, the reverse may hold if
an elated mood signals that the environment poses no particular problem, thus
allowing one to pay attention to enjoyment considerations (see Schwarz & Bohner, in
press, for a more detailed discussion).
Following the review of relevant findings, we will present an outline that
incorporates each of these experiential approaches. Before doing so, however, we
need to address cognitive approaches to the impact of moods on processing strategy.
�MDBU�Cognitive Approaches
Other researchers have traced the impact of affective states on strategies of
information processing to the cognitive component of feelings. To the extent that
affective states bring affect related material to mind, this material may limit the
cognitive resources available. As noted earlier, this has been proposed for elated
as well as depressed moods. On the one hand, Ellis and Ashbrook (1988) suggested
that dysphoric affective states limit attentional resources,resulting in the use of
less information and simpler judgmental procedures under depressed moods.
Qualifying this assumption, Hertel and Hardin (1990) suggested that depressed mood
may only lead to a deficit in initiative, which results in poorer performance on
tasks that require the generation of complex hypotheses or of organizational
schemes, but may not affect performance on well-structured tasks.
Other researchers suggested that elated affective states may limit cognitive
resources (Isen, 1987; Mackie & Worth, 1989). Isen, for example, hypothesized that
positive material in memory "is more extensive and at the same time better
integrated, so that positive affect is able to cue a wide range of thoughts" (Isen,
1987, p. 217). Isen further suggested that "this larger amount of material in mind
might also result in a defocusing of attention" (Isen, 1987, p. 237). This
defocusing, in turn, is hypothesized to increase reliance on heuristic processing
strategies under elated affect (e.g, Isen, Means, Patrick, & Nowicki, 1982; see
Isen, 1987, for a review).

Empirical Evidence
Numerous studies found a pronounced impact of affective states on subjects'
performance on a wide variety of tasks (see Clore et al., 1994, for a review).
Unfortunately, it is difficult to draw strong conclusions from many of the observed
results, due to several complications. First, the cognitive processes underlying
performance on a given task are often not well understood, rendering performance on
many tasks thoroughly nondiagnostic. Second, even on tasks that are well
understood, several different processes may result in the same outcome, rendering
it difficult to infer any particular process in the absence of additional data.
Accordingly, we restrict our discussion to tasks that are sufficiently well
understood to allow at least limited inferences. In the domain of social
psychology, these include the processing of persuasive messages, person perception
and stereotyping, as well as decision making (for a more extended review, covering
additional areas of research, see Clore et al., 1994).
�MDBU�Persuasion
The cognitive dynamics underlying the processing of persuasive communications
are among the better understood phenomena in social psychology (see Eagly &
Chaiken, 1993; Petty & Cacioppo, 1986, for reviews). In general, a message that
presents strong arguments is more persuasive than a message that presents weak
arguments, provided that recipients are motivated and able to process the content
of the message. If recipients do not engage in elaborative processing of message
content, the advantage of strong over weak arguments is eliminated. Accordingly,
one may explore the impact of mood states on processing strategies by testing the
relative impact of argument strength under different mood states. Several
researchers followed this strategy (see Mackie, Asuncion, & Rosselli, 1992;
Schwarz, Bless, & Bohner, 1991, for extensive reviews).
Processing capacity vs. cognitive tuning predictions. With an exception to
be addressed below, the available research has consistently shown that individuals
in an elated mood are less likely to engage in systematic elaboration of a
counterattitudinal message than individuals in a non-manipulated or a depressed
mood (e.g., Bless, Bohner, Schwarz, & Strack, 1990; Bless, Mackie, & Schwarz, 1992;
Bohner, Crow, Erb, & Schwarz, 1992; Mackie & Worth, 1989; Worth & Mackie, 1987).
In these studies, happy recipients were moderately and equally persuaded by strong
as well as weak arguments. Moreover, their cognitive responses showed no
differences as a function of argument strength, and they reported similar
proportions of agreeing or disagreeing thoughts in response to strong or weak
arguments. In contrast, sad recipients were strongly persuaded by strong
arguments, but not by weak arguments. In addition, sad recipients reported more
disagreeing thoughts in response to weak, and more agreeing thoughts in response to
strong messages. In combination, these findings indicate that recipients in a sad
mood are more likely to spontaneously engage in systematic message elaboration than
recipients in a happy mood.
These findings are inconsistent with the assumption that depressed moods
limit cognitive capacity (e.g., Ellis & Ashbrook, 1988), which would predict
decreased message elaboration under depressed mood. However, they are compatible
with the assumption that elated moods may limit cognitive capacity as well as with
the assumption that our cognitive processes are tuned to meet the requirements
signalled by different affective states. This leaves us with the question of which
of these processes drives the observed effects. One issue on which these process
assumptions differ is the ease or difficulty of overriding the effects of mood. If
elated moods limit cognitive capacity due to intruding thoughts, their impact
should be independent of the perceived informational value of one's mood and should
be difficult to override. In contrast, the cognitive tuning assumption holds that
our feelings inform us about the processing requirements of the current situation.
If so, the impact of moods on processing style should be reduced when their
informational value is discredited, much as we have seen for the impact of moods on
evaluative judgments (e.g., Schwarz & Clore, 1983). In addition, the cognitive
tuning assumption suggests that our moods influence which processing strategy is
spontaneously adopted, but that this spontaneous preference can be overridden by
current processing goals. With regard to both issues, the currently available
evidence favors the cognitive tuning assumption.
Two studies bear on whether the impact of happy moods can be overridden by
explicit processing instructions. As part of a study described above, Bless et al.
(1990, Experiment 1) explicitly instructed some of their subjects to pay attention
to the quality of the tape recorded arguments presented to them at a fixed speed.
Under this instruction, happy subjects differentiated between strong and weak
arguments, suggesting that happy individuals have sufficient attentional resources
to engage in systematic message processing. In another study, Mackie and Worth
(1989) manipulated the time available to subjects, a assuming that having more time
to process a written message would compensate for the hypothesized reduction in
cognitive resources. Specifically, they told some of their subjects that they may
take all the time they wanted to read, and re-read, the persuasive message. Under
this condition, happy subjects differentiated between strong and weak arguments.
Mackie and Worth (1989) interpreted this to reflect that increased processing time
can compensate for reduced resources under elated mood. It is conceivable,
however, that informing subjects that they may take their time, and may reread the
message, may convey to them that a carefully considered response to the message is
being asked for. If so, Mackie and Worth's findings would parallel the impact of
processing instructions obtained in the Bless et al. (1990) study. On the other
hand, Bless et al.'s instruction to pay attention to message quality may have
provided subjects a more focused task and may hence have reduced resource demands,
rendering their findings compatible with a reduced resources account. To
distinguish between these two accounts, future research will need to manipulate
subjects' processing capacity or processing goals in ways that are not open to
reinterpretation in terms of the respective other concept. We return to this issue
below, in the context of stereotyping research. In any case, however, happy
subjects' ability to process message content on instruction suggests to us that any
constraints on processing capacity under elated moods are unlikely to be very
severe. Consistent with this view, happy subjects did not show less recall of the
arguments presented to them than sad subjects under either processing instruction
(Bless et al., 1990), in contrast to what a reduced resources account might
predict.
More importantly, a recent study by Sinclair, Mark, and Clore (1994)
indicates that the impact of mood states may be eliminated when their informational
implications are called into question. Students were approached on early spring
days when the weather was sunny and pleasant or cloudy and unpleasant and were
presented with strong or weak persuasive messages. To discredit the informational
implications of subjects' mood, Sinclair et al. did or did not draw their
attention to the weather, following a procedure previously used by Schwarz and
Clore (1983, Experiment 2). When subjects' attention was not drawn to the weather,
the previously obtained interactive effects of mood and message quality were
observed. In this case, sad respondents were persuaded by strong but not weak
messages, while happy respondents were moderately but equally persuaded by both.
However, when the weather was made salient as a potential cause of subjects'
momentary feelings, mood no longer played a role and a main effect of message
quality emerged.
 While the findings of Sinclair et al. (1994) are perfectly consistent with
the assumption that our feelings inform us about processing requirements, one may
potentially argue that knowing the source of one's mood reduces the likelihood of
intruding thoughts, thus rendering the findings compatible with a capacity
constraints account. In our reading, this seems more plausible for depressed
moods, which are supposed to limit processing capacity due to rumination about the
mood inducing event and its implications (e.g., Ellis & Ashbrook, 1988). In this
case, knowing the source of one's mood may reduce the likelihood of rumination.
However, processing impairments under depressed mood have not been observed in
persuasion studies. A similar account seems less plausible, however, for elated
moods, which are supposed to limit processing capacity by a process of automatic
priming of extensive mood congruent material stored in memory (e.g., Isen, 1987).
Hence, to provide a capacity-based account of the results of Sinclair et al.
(1994), one would have to assume that awareness of the source of one's mood limits
the presumably automatic process of mood congruent recall. Although this
possibility does not seem very likely, it illustrates the difficulties encountered
in distinguishing between different accounts in this domain.
Mood management: The hedonic contingency hypothesis. Challenging the
conclusions of the above research, Wegener et al. (1994) demonstrated that being in
a happy mood may increase rather than decrease systematic message elaboration under
specific conditions. According to their hedonic contingency hypothesis (summarized
above), happy individuals scrutinize the hedonic implications of activities before
engaging in them because any activity that is less pleasant than their current mood
would interfere with their good feelings. From this point of view, the previously
reviewed studies found decreased message elaboration under happy mood conditions
because they presented counterattitudinal messages, which carried the potential of
interfering with subjects' elated mood. To test this assumption, Wegener et al.
(1994, Experiment 2) exposed happy and sad subjects to largely identical messages
that were framed as either pro- or counterattitudinal and presented strong or weak
arguments. Moreover, the pro-attitudinal message was introduced by a statement
that informed subjects that the "primary quality" of the message is that it "makes
people feel HAPPY if they think carefully about the information in the article"
(emphasis in the original). Conversely, the counterattitudinal message was
introduced as making "people feel SAD if they think carefully about the information
in the article" (emphasis in the original).
As predicted, when the message was counterattitudinal and subjects expected
that it would make them feel sad, happy subjects did not engage in systematic
message processing, as reflected in the absence of an argument strength effect.
Not so, however, when the message was pro-attitudinal and subjects expected that
thinking about it would make them feel happy. In this case, happy subjects engaged
in message elaboration, resulting in higher persuasion under strong than weak
argument conditions. In combination, these findings indicate that being in a happy
mood may increase as well as decrease systematic processing, depending on the
expected hedonic consequences.
Wegener et al. (1994) further hypothesized that sad subjects' message
processing should be less affected by the expected hedonic consequences. According
to their hedonic contingency analysis, nearly any activity is likely to be more
pleasant than one's current sad mood and is hence likely to improve it. Sad
individuals may therefore engage in message elaboration without scrutinizing its
possible hedonic implications. Although this may be the case when the hedonic
implications are not obvious, it seems unlikely to us that sad subjects would have
missed the blatant information about the alleged affective impact of the message
(see instructions above).
Nevertheless, sad subjects engaged in systematic message elaboration when the
message was said to make them happy as well as when the message was said to make
them sad. In fact, they tended to elaborate the counterattitudinal message, which
was said to make them feel sad, more than the pro-attitudinal message, which was
said to make them feel happy. In our reading, this renders it unlikely that
message elaboration was primarily driven by mood management considerations under
sad mood conditions, in contrast to what Wegener et al. (1994) concluded. �1B
Independent of this concern, to which we return below, Wegener et al.'s study
demonstrates that happy moods may increase as well as decrease message elaboration,
in contrast to what previous research suggested. Moreover, the observed increase
in message elaboration under happy mood conditions is again incompatible with a
capacity constraints account, further supporting the conclusions drawn from the
previously reviewed studies (Bless et al., 1990; Sinclair et al., 1994). Next, we
offer a theoretical integration of these apparently contradictory findings.
Towards an integration. We propose that the mood management processes
addressed by Wegener et al. (1994) need to be conceptualized in the context of the
information that our feelings convey about the nature of the current situation.
Suppose that being in a good mood signals that the current situation poses no
problem (Schwarz, 1990). If so, there is no external task that urgently needs
attention (unless another goal is active; Bless et al., 1990). Hence, we may
attend to the likely hedonic consequences of an activity, basically using an
enjoyment rule (Martin et al., 1993). As a result, one obtains the pattern
reported by Wegener et al. (1994), namely, systematic message elaboration under
happy moods if the message has positive hedonic consequences, but not if it has
negative hedonic consequences. Of course, this contingency depends on happy
subjects realizing that thinking about the message may make them feel good. If so,
they are likely to engage in systematic message processing. If this hedonic
implication is not apparent, being in a good mood is likely to result in decreased
message elaboration, as has been observed in previous studies.
Attending to short-term hedonic consequences, however, may be a luxury that
we are only willing to afford in situations that do not seem to pose any particular
problems. Suppose that a negative mood signals a problematic situation (Schwarz,
1990). If so, it would be maladaptive to ignore this signal and to base one's
processing decisions on short-term hedonic implications. Hence, adoption of an
enjoyment rule may be unlikely (unless the experimenter asks subjects to use one,
as in Martin et al., 1993). Accordingly, the message would be processed
systematically even if it is likely to make one feel bad in the short run. This
may be particularly likely if the implications of the message are indeed negative,
as was the case for the counterattitudinal message used by Wegener et al. (1994,
Experiment 2), thus confirming one's concern that one may have to deal with some
problem. As a result, being in a sad mood would foster systematic processing
independent of short-term hedonic considerations. In fact, negative material may
actually receive more processing attention than positive material, in contrast to
what a straightforward mood management approach would predict.
From this perspective, the hedonic consequences of an activity play a more
important role under happy than under sad mood because the basic "there's no
problem that needs attention" signal sets us free to attend to short-term hedonic
consequences. Conversely, the "better watch out" message conveyed by negative
moods renders processing decisions based on short-term hedonic consequences less
likely, as ignoring a potential problem would be maladaptive. These considerations
parallel previous discussions of the likely impact of alternative goals (Schwarz,
1990), except that the goals considered here are mood management goals.

�MDBU�Stereotyping and Impression Formation

Paralleling the distinction between heuristic and systematic processing
strategies in the persuasion domain, recent models of person perception distinguish
between two different processing strategies involved in impression formation (M.
Brewer, 1988; Fiske & Neuberg, 1990). At the one extreme, judgments may be
primarily based on the implications of the target's category membership, with
little attention to the target's specific behaviors. At the other extreme,
judgments may be primarily based on individuating information about the target,
with little impact of information about the target's category membership.
Researchers agree that judgments based on category membership information require
less processing capacity and/or motivation to process than judgments based on
individuating information (e.g., M. Brewer, 1988; Bodenhausen, 1990, 1993; Fiske &
Neuberg, 1990; Macrae, Milne, & Bodenhausen, 1994). Hence, we may expect that
being in a good mood facilitates reliance on category membership information,
whereas being in a depressed mood facilitates reliance on individuating
information, paralleling the findings obtained in the persuasion domain.
Several studies support these predictions. Bodenhausen (1993) presented
subjects in different mood states with descriptions of an alleged student
misconduct and asked subjects to determine the target's guilt. Subjects in a happy
mood rated the offender as more guilty when he was identified as a member of a
group that is stereotypically associated with the described offense than when he
was not. In contrast, the guilt judgments of subjects in a neutral or sad mood
were not affected by the target's category membership. Similarly, Edwards and
Weary (1993) observed that non-depressed subjects were more likely to rely on
category membership information than chronically depressed subjects, who seemed to
engage in a more effortful analysis of the individuating information provided to
them. In a similar vein, Sinclair (1988) observed that subjects in a sad mood made
more use of detailed performance information. Sad subjects were also less likely
to show halo effects and more likely to be accurate in a performance appraisal task
than those in happy moods, with neutral mood subjects falling in between. In
addition, sad subjects have been found to show less pronounced primacy effects in
impression formation tasks than happy subjects (Sinclair & Mark, 1992). Finally,
Hildebrandt-Saints and Weary (1989) observed that chronically depressed individuals
sought more, and more diagnostic, information about another person than non-
depressed subjects. Moreover, depressives did so independent of whether they
expected to interact with the person in the future, whereas non-depressives'
information search increased when future interaction was expected.
In combination, these findings from the domain of impression formation
indicate that being in a depressed mood fosters a more systematic style, while
being in an elated mood fosters a less systematic processing style. Moreover,
happy subjects' reliance on category membership information can be overridden by
manipulations that increase their processing motivation, such as personal
accountability for one's judgment (Bodenhausen, Kramer, & S�sser, 1994, Experiment
4) or an anticipated interaction with the target (e.g., Hildebrandt-Saints & Weary,
1989). Overall, the mood and stereotyping findings are consistent with what we
observed in the domain of persuasion research. On the one hand, the obtained
patterns are incompatible with the assumption that being in a depressed mood limits
cognitive capacity. On the other hand, the results reviewed so far do not allow us
to differentiate between the cognitive tuning assumption and the hypothesis that
being in an elated mood limits cognitive capacity. Fortunately, however, these
assumptions lead to differential predictions under specific conditions.
Specifically, the cognitive tuning assumption holds that being in an elated
mood informs us that the current situation is not problematic, which may result in
decreased processing motivation and increased reliance on general knowledge
structures, which have served us well in the past. If required, however,
individuals in an elated mood can engage in systematic processing strategies, as
has been observed in the persuasion (e.g., Bless et al., 1990) as well as the
stereotyping domain (e.g., Bodenhausen et al., 1994). These findings are
consistent with the assumption that their ability to engage in systematic
processing is not constrained by limited resources. Accordingly, individuals in an
elated mood may consider individuating information in forming an impression if it
contradicts the implications of their assumptions based on the target's category
membership. As previous research has indicated (see Fiske & Neuberg, 1990), such
inconsistent information should receive increased attention and elaboration,
resulting in an increased impact on the impression formed. Such an increased
impact of stereotype inconsistent information would be difficult to reconcile with
the assumption that being in an elated mood limits processing capacity, as it would
reflect a considerable degree of attention to, and elaboration of, detailed
individuating information. This possibility cannot be evaluated on the basis of
the previously reviewed research, which did not include conditions under which the
implications of the individuating information contradicted the implications of
category membership information.
Introducing the relevant conditions, Bless, Schwarz, and Wieland (1994)
observed in several experiments that subjects in a happy mood paid attention to
individuating information that contradicted the implications of the stereotype.
Specifically, these subjects reported the most negative impression of the target
when the target's specific negative behaviors contradicted the implications of his
membership in a positively evaluated category. This impact of stereotype
inconsistent information was more pronounced under happy than under neutral mood
conditions, and was not obtained under sad mood conditions. This pattern of
findings suggests several important conclusions.
First, it suggests that subjects in a happy mood were more likely to process
the information presented about the target in terms of their general knowledge
structures pertaining to the target's category. As a result, they were more likely
to notice the inconsistency between the individuating information and the
stereotype than subjects in a neutral mood, resulting in a pronounced impact of the
inconsistent information. Conversely, subjects in a sad mood presumably focused on
the implications of the individuating behaviors without relying on their general
knowledge structures to begin with, consistent with the findings of the research
reviewed above (e.g., Bodenhausen, 1993; Edwards & Weary, 1993). As a result, they
did not notice the inconsistency of the individuating information and the
implications of the stereotype. Accordingly, this information did not receive
increased elaboration and was hence not more likely to influence their judgment
than other individuating information. In combination, this pattern suggests that
being in a depressed mood facilitates a systematic processing strategy that
involves attention to detailed, individuating information that is unlikely to
involve processing in terms of general knowledge structures. As a result of this
data-driven strategy, inconsistencies between the individuating information and the
stereotype are likely to go unnoticed.
Conversely, being in a good mood facilitates reliance on pre-existing
knowledge structures, which may reflect a less effortful processing strategy. This
theory-driven strategy results in judgments that reflect the implications of the
stereotype, provided that the individuating information is not blatantly
inconsistent. Accordingly, previous research obtained category-based judgments
under elated moods, and behavior-based judgments under depressed moods, reflecting
that no inconsistencies between both sources of information were introduced. If
the individuating information contradicts the implications of the stereotype,
however, the theory-driven processing style presumably facilitated by happy moods
renders these contradictions obvious. Hence, the inconsistent information is
elaborated on, resulting in a particularly pronounced impact on impression
formation.
As a second conclusion, these findings suggest that the impact of happy moods
is unlikely to reflect limited cognitive resources. If happy subjects had limited
resources, they should have relied on the category membership information presented
to them, with little attention to the detailed individuating information. Instead,
these subjects were most likely to notice the inconsistency and to elaborate on its
implications, resulting in more extreme impressions. As this elaboration
presumably requires considerable resources, it is difficult to reconcile with the
assumption that subjects' cognitive resources were limited by being in a good mood.
�MDBU�Conclusions
In summary, research in the domain of persuasion and stereotyping
consistently demonstrated that subjects in a sad mood are more likely to
spontaneously engage in systematic information processing than subjects in a happy
mood, who may only do so if required by some other goal. Whereas this general
conclusion is uncontroversial, the evidence bearing on the underlying processes is
limited and often indirect. In our reading, the observation that the impact of
happy moods can be overridden by instructions (Bless et al., 1990); that happy
subjects elaborate on stereotype inconsistent individuating information (Bless et
al., 1994); and that they engage in systematic processing when they expect it would
make them feel good (Wegener et al., 1994), renders it unlikely that the observed
effects of happy moods are due to limited processing capacity (as suggested by
Mackie & Worth, 1989, and others). Moreover, the evidence is clearly incompatible
with the assumption that being in a sad mood limits processing capacity (as
suggested by Ellis & Ashbrook, 1988, and others). Rather, the obtained findings
are most consistently accounted for by assuming that cognitive processes are tuned
to meet the processing requirements signalled by one's affective state (Schwarz,
1990). Consistent with this assumption, undermining the perceived informational
value of one's affective state has been found to eliminate the impact of moods on
processing style (Sinclair et al., 1994), much as it has been found to eliminate
the impact of moods on evaluative judgments (e.g., Schwarz & Clore, 1983).
However, the specific information provided by one's feelings is not
necessarily limited to information about the general nature of one's current
situation. As many studies on evaluative judgment demonstrated, moods may serve as
input into a variety of judgments. Accordingly, their impact may differ depending
on the judgment at hand. Thus, if we ask ourselves if we enjoy the task, feeling
good may lead us to continue an activity that seems enjoyable. However, the same
feeling may lead us to terminate our efforts if we asked ourselves if we are
satisfied with what we already achieved (see Martin et al., 1993). We conjecture
that people are less likely to spontaneously rely on an enjoyment rule in making
performance decisions if a bad mood tells them that their current situation is
problematic, much as we conjecture that they are less likely to attend to short
term hedonic consequences under these conditions. In our reading, such hedonic
considerations are more likely when a good mood informs us that our current
situation is benign. This suggests that we should see more flexibility in the
choice of processing strategies under good than under bad mood conditions, an
assumption that is compatible with the currently available evidence. Nevertheless,
if individuals are induced to bring their mood to bear on a specific judgment, its
impact may strongly differ from the effects that have been documented so far (see
Schwarz & Bohner, in press, for conjectures about different possible effects of
mood at different steps of the intention-action sequence). Throughout, however, we
can conceptualize these effects as reflecting the use of one's feelings as a basis
of judgment, as suggested by the feelings-as-information approach.
�MDBU�Other Tasks
Following this review of findings in the social psychological literature, we
hasten to add that findings from other domains of research are considerably less
consistent (see Clore et al., 1994, for an extensive review). However, it is
typically difficult to draw strong conclusions from this research, as the tasks are
often not well understood and different processing assumptions lead to similar
predictions. A few examples may illustrate this point.
Exploring the impact of moods on learning and memory, several researchers
concluded that being in a sad mood may interfere with the elaboration of to-be-
learned material. In general, words that are presented in a context that requires
more elaboration at the encoding stage are better recalled than words presented in
a context that requires less elaboration (e.g., Stein & Bransford, 1979).
Exploring the impact of induced depressed moods on this phenomenon, Ellis, Thomas,
and Rodriguez (1984) observed that being in a sad mood eliminated the otherwise
obtained elaboration effect. Similar findings have been reported for naturally
depressed individuals (Potts, Camp, & Coyne, 1989). Whereas these researchers
attributed the reduced elaboration of the presented material to capacity
constraints under depressed mood, it may also indicate that sad individuals
generally process information at a lower level of abstraction, which would also
interfere with the organization of information into larger meaningful units. This
possibility is suggested by the observation that sad individuals pay more attention
to individuating information in stereotyping studies, making less use of their
general knowledge structures, as we have seen above.
Similarly, induced negative affect has been shown to limit the spontaneous
organization of to-be-learned information into meaningful units. In perceptual
grouping tasks that involve the repeated presentation of letter strings that are
identical in sequence but differ in grouping (e.g., CA DM ET and C ADME T),
learning is typically facilitated by reorganizing the material into more meaningful
units (e.g., CAD MET). Leight and Ellis (1981) observed that induced negative
affect interfered with such reorganization and led to reduced recall. Related
evidence for less organization under negative affect has been obtained with
clinically depressed subjects. For example, Watts and Cooper (1989) asked
depressed and control subjects to memorize a story. Usually, people better
remember the central aspects than the peripheral aspects of a story (Mandler,
1984); this effect, however, relies on people organizing the presented material in
terms of the story structure. In contrast, depressed subjects did not show a
recall advantage for central material, which suggests that they did not organize
the presented material. Once again, these findings may reflect capacity
constraints under depressed moods as well as processing the stimuli at a lower
level of abstraction, interfering with the organization of new information into
meaningful units.
The impact of happy and sad moods has also been explored in the domain of
logical problem solving. This literature consistently shows that being in a
depressed mood facilitates performance on tasks that require the detection of
covariation. Specifically, research on covariation detection (see Alloy, 1988, for
a review) demonstrated that nondysphoric subjects are likely to overestimate the
degree of contingency between their actions and their outcomes. In contrast,
subjects experiencing induced or chronic negative affect are not as prone to
exhibit this "illusion of control" (e.g., Alloy & Abramson, 1979). Furthermore,
the illusion of control observed under neutral mood increases under good mood
(Alloy, Abramson, & Viscusi, 1981). These findings are consistent with the
assumption that depressed affective states elicit a more systematic and data-driven
processing strategy, which is less likely to be used under elated affect. In fact,
if being in a depressed affective state signals that one's current environment is
problematic, paying attention to the covariation of one's acts and the obtained
outcomes would be highly adaptive. If being in an elated mood signals that one's
current environment is safe, on the other hand, paying attention to this
covariation may seem less important. Thus, one may argue that covariation
detection is a task to which the assumed signalling and tuning function of
affective states is highly germane.
The available evidence is less consistent for other logical problem solving
tasks. For depressed affect, increased as well as decreased performance has been
reported. For example, Fiedler and Fladung (1986, reported in Fiedler, 1988)
observed that subjects in an induced sad mood produced fewer logical
inconsistencies in a multi-attribute decision task than subjects in a good mood.
Specifically, the latter were twice as likely to violate transitivity of preference
than the former, by producing inconsistent triads of the form A > B, and B > C, but
A < C. Other studies, however, indicated poorer performance for subjects in a sad
mood. Logical problem solving research with depressed subjects selected on the
basis of self-report measures, peer nomination (for children), or clinical criteria
repeatedly revealed performance deficits under depressed affect (Dobson & Dobson,
1981; Perkins, Meyers, & Cohen, 1988; Silberman, Weingartner, & Post, 1983).
Moreover, being in a happy mood has been shown to improve rather than impair
performance for a number of tasks. For example, children in a happy mood performed
better than neutral mood control subjects on discrimination problems that required
the identification of which stimulus in a set of three is correct (Masters, Barden,
& Ford, 1979), whereas children in a sad mood performed worst on this task.
Similarly, adults in a happy mood outperformed subjects in a neutral mood on
mathematical problem solving tasks (Kirschenbaum, Tomarken, & Humphrey, 1985).
In fact, such mixed findings are to be expected in the problem solving domain
because none of the processes hypothesized will necessarily result in improved
performance. For example, the cognitive tuning assumption generally predicts
increased systematic processing under negative affect. However, systematic
processing may not result in improved performance if subjects do not have access to
appropriate algorithms. Similarly, heuristic processing strategies may either
facilitate or impede logical problem solving, depending on whether the adopted
heuristics are applicable to the current task. Furthermore, such heuristics may be
discarded in the face of task demands or instructions, rendering it difficult to
make strong predictions. It is therefore not surprising that the most consistent
results have been observed on tasks with which subjects are most familiar and which
do not require the application of potentially unknown algorithms, namely
covariation detection, persuasion, and impression formation.

Feelings and Memory:

The Encoding and Recall of Information
As in the other domains, there are, at least in principle, both cognitive or
content-based approaches and experiential approaches that might be taken to memory
phenomena. Memory is, of course, primarily about content, and theories focus on
such questions as how thinking about one domain of memory content makes recall of
other content more likely. Very little research has been done on the role of
experiential factors in memory, although such factors were deemed important by
early memory theorists (see W. F. Brewer, 1992). Even in the domain of mood and
emotion, almost all memory research has taken a content-based approach (e.g.,
Bower, 1981), and most of our review will focus on this material. Before we turn
to this dominant approach, however, we explore the role of subjective experiences
in memory.

Experiential Approaches
�MDBU�Experiential Factors in Non-mood Related Memory
Experiential considerations have played a key role in early treatments of
memory processes, but have lost popularity and were rediscovered only recently. W.
F. Brewer (1992) concludes from his review of the role of phenomenal experience in
memory research:
�LM8�The history of the study of memory has been a dramatic roller coaster ride
from the position that the only interesting aspects of memory are those that are
consciously experienced to a position that denied or ignored any consciously
experienced aspect of memory. However, now there is a small [...] growing group of
researchers who think that the proper study of human memory includes a treatment of
memory in terms of both behavior and conscious experience (p. 36).
�LM0� A role for subjective experiences in memory arises mainly for recognition
memory. Some early memory theorists suggested that recognition occurs through a
unique phenomenological feeling of familiarity (e.g., Titchener, 1910, p. 407). As
noted earlier, some recent theorists have voiced a similar view (e.g., Mandler,
1980). For example, Jacoby and Dallas (1981) observed in a recognition experiment
that subjects could accurately identify rare words that had previously been shown
to them, but provided numerous false alarms in response to common words.
Apparently, they misattributed the sense of familiarity that resulted from their
frequent exposure to common words to the recency of exposure, erroneously
concluding that the word was presented during the preceding learning task. Such
findings parallel the previously discussed studies on the role of familiarity and
perceptual fluency in judgment.
However, as W. F. Brewer (1990) points out, not all memorial activity
involves feelings of familiarity. Engaging in habitual actions is not accompanied
by any re-experience of having done the act before (Thorndike, 1898), and the
retrieval of facts from long term memory also generally does not involve feelings
of mental experiences (Angell, 1909). Although, when experiencing difficulty
recalling a fact or name, the effort may be accompanied by experiential cues that
indicate whether one is getting warmer or colder in one's mental search. Indeed,
there may be a general relation between inner experience and how automatic the
retrieval is. As Hunter (1964) and others pointed out, in learning to type or
learning a foreign language, finding the appropriate key or word initially requires
various intervening mental activities, but with practice becomes automatic.
Feelings of familiarity may also vary with context. For example, one may
have a conscious experience of recognition when seeing an acquaintance in a foreign
city that one would not have when encountering the same person in one's home
community where no cognitive updating or mental accommodation is required. And
just as old objects may feel familiar, new ones or rearrangements of old ones may
feel unfamiliar. Patterns perceived or stored as a unit, such as a person's face,
often yield feelings of unfamiliarity when the person shaves off their beard or
replaces glasses with contact lenses.
Feelings, regardless of their source, may play similar roles in information
processing. Feelings of familiarity act like affective feelings and also like
feelings associated with cognitive states of surprise, or bodily feelings of pain.
In each case, the feeling alerts the experiencer to focus conscious attention and
processing resources on the apparent cause of the feeling. Thus, when irritated or
in pain, one focuses on the irritating or painful stimulus and ways to eliminate
it. When surprised, one focuses on the surprising aspect of the situation and ways
to revise one's conception of the situation to incorporate the new feature.
Similarly, when something seems familiar, one may focus on a search for the source
of the feeling. In the case of familiarity, this involves a memory search, an
attempt to recollect the original situation. Just as emotional feelings stimulate
the tacit hypothesis that the situation is characterized by features that
correspond to the class of stimuli that elicits such an emotion, so too feelings of
familiarity stimulate the tacit hypothesis that information about prior exposure to
a familiar-seeming stimulus already exists in memory. When this is the case, it is
generally worth retrieving the information rather than trying to recompute it on
the basis of new experience. In this way, feelings of familiarity may keep one
from constantly discovering what one already knows. Disorders of this system may
make one a victim of memory loss. Sachs (1985), for example, relates the sad story
of an amnesic patient whose father had recently died. The patient had been very
close to his father. Each morning he asked to see his father and each morning was
shocked again to learn his death. Without the ability to retrieve this episodic
information, he was doomed to re-experience this tragedy each day.
�MDBU�Mood Related Memory Phenomena
Mood related memory phenomena have not been explicitly addressed from an
experiential perspective. However, the feelings-as-information approach generates
a number of testable hypotheses. As reviewed in the preceding sections, this
approach assumes that positive affective states signal a benign situation, whereas
negative affective states signal a problematic situation (Schwarz, 1990). If so,
we may assume that negative feelings focus attention on potentially problematic
features of the situation, thus rendering it more likely that they will be recalled
later on. Moreover, cues that signal a problematic situation are likely to result
in attention to information at a lower level of abstraction, consistent with Wegner
and Vallacher's (1986) observation that failures to obtain a desired outcome
increase attention to details of one's action strategy. In addition, the
systematic style or heuristic style of information processing fostered by negative
or positive feelings, respectively, is likely to affect the organization of
information in memory. Consistent with this assumption, Bless, Hamilton, and
Mackie (1992) observed that being in a happy mood increased, whereas being in a sad
mood decreased, clustering in the recall of person information. In fact, the
impact of mood eliminated the usually obtained effect of memory or impression
formation instructions (e.g., Hamilton, Katz, & Leirer, 1980), which only affected
clustering under neutral mood conditions. This finding is consistent with research
on mood and stereotyping, as reviewed above.
But with respect to mood congruency effects in memory, the topic of the
lion's share of mood and memory research, experience-based approaches such as the
feelings-as-information model, have been silent. But if one were to approach this
issue from an experiential standpoint, one might take two different routes. The
first approach draws on the use of one's feelings as a basis of evaluative judgment
according to a "How-do-I-feel-about-it?" heuristic, whereas the second draws on the
assumption that our feelings inform us about the nature of our psychological
situation.
The first take on the issue begins with the observation that individuals may
use a judgment as a retrieval cue for reconstructing the information that
presumably provided the basis of judgment in the first place (e.g., Higgins &
Lurie, 1983; Higgins & Stangor, 1988). This raises the possibility that
individuals' mood may influence their evaluative judgments, which in turn may serve
as mood congruent retrieval cues, resulting in mood congruent recall. To use one
of Bower's (1981) examples, individuals who are asked to recall events from their
kindergarten days may first ask themselves, "Well, kindergarten days. What were
they like?". In doing so, they may form a global evaluation that is based on their
current mood as reviewed above. Facing the task to report specific episodes, they
may then use this global evaluation as a retrieval cue to guide the recall of
specific information, resulting in an increased recall -- or reconstruction -- of
mood congruent information.
This reasoning suggests that mood effects on recall should only be obtained
under conditions that give rise to mood effects on evaluative judgments in the
first place. Hence, misattribution manipulations should eliminate the impact of
moods on the recall. Moreover, the likelihood of mood congruent recall should vary
with other variables known to determine reliance on one's feelings as a basis of
judgment (as reviewed in a preceding section). However, experimental evidence
bearing on these possibilities is not available.
An alternative approach is suggested by Higgins' (1987) self-discrepancy
theory. This theory postulates that discrepancies between one's ideal self (one's
own or others' hopes, wishes or aspirations) and one's actual self give rise to
dejected feelings of sadness, which relate to concerns about the absence of
positive outcomes. In contrast, discrepancies between one's ought self (one's own
or others' beliefs about one's duties, obligations and responsibilities) and one's
actual self give rise to agitated negative feelings, which relate to concerns about
the presence or threat of negative outcomes. Whereas priming the respective
discrepancies results in the respective feelings (Higgins, 1987), we may again
assume that the respective feelings inform us about the nature of our psychological
situation. If so, dejected negative feelings would signal a lack of positive
outcomes, whereas agitated negative feelings would signal a threat of negative
outcomes. This proposal differs from the bulk of mood and memory research by
emphasizing the nature of potential outcomes, rather than their valence (see
Higgins, in press, for a more detailed discussion). In fact, Higgins and
Tykocinski (1992) observed that subjects with chronic actual:ideal discrepancies
were more sensitive to positive rather than negative outcomes in a person
description presented to them, resulting in high recall of information related to
positive outcomes, independent of whether these outcomes were obtained (resulting
in an event of positive valence) or not (resulting in an event of negative
valence). Conversely, subjects with chronic actual:ought discrepancies were more
sensitive to negative rather than positive outcomes described to them, resulting in
high recall of information related to negative outcomes, again independent of
whether these outcomes were obtained (resulting in an event of negative valence) or
not (resulting in an event of positive valence). To the extent that our feelings
signal different psychological situations, and hence elicit processes that parallel
the impact of self-discrepancies, it is therefore conceivable that they draw
attention to different classes of outcomes,independent of their valence. This
conjecture may provide an account for the inconsistent findings in the mood and
memory literature, to be reviewed below, which focused on the valence of the
recalled material rather than the type of outcome.

Cognitive Approaches
By far the largest amount of research on the interface of affect and
cognition has focused on affective influences on the recall of information from
memory. Although the key hypotheses have initially been introduced by Isen and
colleagues (see Isen et al., 1978), most research in this area has been
conceptualized in terms of Bower's (1981; see also Bower & Cohen, 1982; Gilligan &
Bower, 1984; Bower, 1991) extension of Anderson and Bower's (1973) human
associative memory model.
�MDBU�Associative Network Model
 We have already reviewed the key features of Bower's (1981) model in the
section on theoretical approaches to the interplay of affect and cognition. To
reiterate, the model assumes that emotions function as central nodes in an
associative network, which are linked to related ideas, events of corresponding
valence, autonomic activity, and muscular and expressive patterns. When new
material is learned, it is associated with the nodes that are active at the time of
learning. Accordingly, material that is learned in a particular affective state is
linked to the respective emotion node. When an emotion node is stimulated,
activation spreads along the pathways, increasing the activation of other nodes
connected to the emotion node. Activation of a node above a certain threshold
brings the represented material into consciousness.
Related models have been offered by Isen et al. (1978), Spies and Hesse
(1986), and Wyer and Srull (1989; see Clore et al, 1994, for a comparison), but our
present discussion will focus on Bower's model, in line with the bulk of the
available research. This model generates four key predictions -- pertaining to
state-dependency and mood-congruency in recall, mood congruent encoding and
elaboration, and judgmental biases -- that received considerable attention in
recent research. Given that numerous extensive reviews of this work are available
(e.g., Blaney, 1986; Clore, et al., 1994; Forgas & Bower, 1988; Singer & Salovey,
1988), however, we will only highlight the key issues and some selected findings.
State Dependency. As a first implication, the model predicts state-dependent
recall as a function of matching moods at the time of learning and at the time of
recall. As a considerable body of research indicates (see Baddeley, 1990 ;
Tulving, 1983, for reviews), any contextual factor may serve as a discriminatory
cue, facilitating memory when the context at recall matches the context at
learning. Whereas many studies have focused on the learning environment as a
discriminatory cue, subjective states of the person may serve the same function.
According to the associative network model, material that is learned while in a
certain mood, for example, is associated with the respective mood node. If this,
rather than another, mood node is activated at the time of recall, the excitation
spreading from the mood node increases the likelihood that the excitation of the
node that represents the learned material exceeds threshold, thus bringing the
material into consciousness. Note that this state-dependent recall prediction is
based on matching moods at learning and recall.
Support for state-dependent recall has been obtained in several studies that
used a "two-list interference paradigm", where, for example, List A is learned
while in a happy, and List B is learned while in a sad mood. As predicted, being
in the same mood at the time of recall facilitated recall in several studies,
whereas being in the opposite mood inhibited recall (e.g., Bower, Monteiro, &
Gilligan, 1978, Exp. 3; Schare, Lisman, & Spear, 1984, Exp. 3). However, other
studies (e.g., Bower & Mayer, 1985; Marshall-Garcia & Beck, 1985; Wetzler, 1985)
failed to replicate state dependent recall under these conditions. Of these
studies, Bower and Mayer's presents the most problematic failure to replicate
because these authors used the materials employed in the successful Bower et al.
(1978) study and induced moods of comparable intensity. Unfortunately, a
theoretical rationale that would account for the inconsistent findings seems
lacking. Indeed, Bower and Mayer (1985, p. 42) concluded that "mood-dependent
retrieval is an evanescent will-o'-the-wisp, and not the robust outcome suggested
by earlier reports."
Mood Congruency. A second prediction is based on matching valences of mood
at the time of recall and of the to-be-recalled material, independent of the mood
at time of learning. This prediction is referred to as mood congruent recall and
holds that being in a positive mood facilitates the recall of positively valenced
material, and inhibits the recall of negatively valenced material. The reverse
holds for being in a negative mood. Note that this prediction pertains to the
match of mood at recall and valence of the to-be-recalled material. In contrast to
the state-dependency prediction addressed above, this mood congruency prediction
does not involve matching moods at the time of learning. The activation spreading
from an emotion node at the time of recall is assumed to increase the activation of
related nodes above threshold, resulting in increased recall of material that
matches the valence of the current mood.
Unfortunately, the conceptually straightforward distinction between state
dependency and mood congruency is difficult to sustain in the domain that produced
the most consistent findings. Specifically, mood congruent recall has been
demonstrated most convincingly for autobiographical memories. In several studies
(for reviews see Blaney, 1986; Morris, 1989; Singer & Salovey, 1988), subjects were
more likely to recall happy memories when in a happy rather than sad mood, or sad
memories when in a sad rather than happy mood (e.g., Bower, 1981; Madigan &
Bollenbach, 1982; Mathews & Bradley, 1983; Natale & Hantas, 1982; Snyder & White,
1982; Teasdale & Taylor, 1981; Teasdale, Taylor, & Fogarty, 1980). Note, however,
that happy events are likely to induce a happy mood at the time of their
occurrence, whereas sad events are likely to induce a sad mood. If so, the valence
of the event is inherently confounded with the experiencer's mood at the time of
encoding. Accordingly, facilitative effects of mood on the recall of
autobiographical memories may reflect state-dependent recall as well as mood
congruency (see Morris, 1989, p. 72, for a related discussion).
In light of the inconsistent findings obtained in the state-dependent recall
studies reviewed above, which have typically used neutral rather than strongly
valenced material, this suggests that mood effects on recall may be most reliable
under conditions where (a) the valence of the to-be-recalled material matches the
mood at recall and (b) the material was initially learned in a matching mood state,
as Clore et al. (1994) noted. Thus, mood congruent selectivity in recall may be
most likely under conditions that simultaneously satisfy the criteria for state-
dependency and mood congruency.
However, inconsistent findings have also been reported in this domain. For
example, Parrott and Sabini (1990) obtained mood incongruent recall of
autobiographical memories in several studies. However, the obtained incongruency
was restricted to the first event recalled. More importantly, the valence of the
first event recalled depended on the specific nature of the mood induction. When
subjects were in a happy or sad mood due to naturally occurring events (return of a
graded exam in Experiment 1; sunny or cloudy weather in Experiment 2), or were not
aware that the experimental procedures were intended to change their mood
(Experiment 4), the first event they recalled was likely to be mood-incongruent.
However, when subjects were instructed to put themselves into a certain mood, and
were hence aware of the intended effect of the manipulation (Experiment 3), the
first event recalled was likely to be mood-congruent. Parrott and Sabini (1990)
therefore concluded that the standard laboratory procedures may unduly facilitate
the emergence of mood-congruency effects. Putting oneself into a certain mood, and
maintaining this mood over the course of the experiment, presumably requires the
production of mood congruent thoughts and the avoidance of mood incongruent
thoughts. When this avoidance of mood incongruent thoughts is not encouraged, on
the other hand, sad subjects may try to improve their mood by avoiding further
negative thoughts and by deliberately searching for positive material, resulting in
mood-incongruent recall.
This latter assumption is known as the "mood repair" hypothesis, introduced
by Isen and her colleagues (see Isen, 1984; Erber, in press, for discussions). At
first glance, this assumption seems to provide a compelling account for Parrott and
Sabini's (1990) findings as well as for a repeatedly observed asymmetry in the
recall of autobiographical events. Specifically, subjects in a happy mood have
been found to recall more happy, and fewer sad, memories than subjects in a neutral
mood, reflecting facilitating as well as inhibiting effects of happy moods. In
contrast, however, being in a sad mood seems to inhibit the recall of happy
memories, but it does not seem to facilitate the recall of sad ones, relative to
being in a neutral mood (e.g., Natale & Hantas, 1982; Salovey & Singer, 1985; see
Blaney, 1986; Singer & Salovey, 1988, for more detailed discussions). Thus, sad
subjects recall fewer happy events, but they do not recall more sad events, in
contrast to what the associative network model (Bower, 1981) would suggest.
According to the mood repair hypothesis, this reflects that sad subjects
deliberately attempt to improve their mood by avoiding further negative thoughts.
Note, however, that a mood repair assumption would also suggest that sad
subjects attempt to recall positive memories to improve their mood. Increased
recall of happy events under sad moods, however, has typically not been observed,
with exception of the Parrott and Sabini (1990) studies. These studies, however,
yielded a finding that is very difficult to reconcile with the assumptions
underlying the mood repair notion: Whereas sad subjects' first event recalled was
likely to be a happy one, happy subjects' first event recalled was likely to be a
sad one. Whereas the first observation is compatible with a mood repair notion, it
seems quite counterintuitive that individuals would want to regulate their happy
mood by recalling sad events (see Clore et al., 1994, and Wyer & Srull, 1989, p.
383, for more extended discussions of these controversial findings).
Other researchers suggested that the observation that happy moods facilitate
the recall of mood congruent material, whereas sad moods primarily inhibit the
recall of mood incongruent material, reflects structural differences in the storage
of positive and negative material. These authors (e.g., Cramer, 1968; Isen, 1984;
Matlin & Stang, 1979) proposed that positive material is more interconnected in
memory than negative material. If so, being in a good mood would facilitate the
recall of some positive event, which in turn would facilitate the recall of other
positive events, assuming that they are connected in memory. In contrast, being in
a bad mood would also facilitate the recall of a congruent, negative event.
However, this event would not facilitate the recall of other negative events,
assuming that negative events are less likely to be interconnected. Unfortunately,
data that would directly bear on these structural assumptions are not available.
Moreover, it seems unclear why memories of positive events should be more
interconnected in memory, given that negative events are more likely to trigger
explanatory efforts (e.g., Bohner, Bless, Schwarz, & Strack, 1988), which, in turn,
should establish pathways between negative memories. In fact, other researchers
(e.g., Higgins, Van Hook, & Dorfman, 1988) argued that negative events are more
likely to be interconnected in memory.
Finally, it is worth noting that much as it is difficult to separate the
relative contributions of state dependency and mood congruency in studies using
autobiographical recall, it is also difficult to separate the relative
contributions of mood congruent recall and mood congruent encoding. Specifically,
many studies that obtained selective recall of mood congruent material have allowed
subjects to control the length of exposure to the to-be-remembered material (e.g.,
Derry & Kuiper, 1981; Forgas & Bower, 1987), as Morris (1989) noted. Accordingly,
the obtained selectivity in recall may reflect mood congruent elaboration of the
presented material at encoding, as well as a facilitating effect of mood at the
recall stage.
Specific Emotions. Research on the impact of specific emotions on memory has
also been guided by the hypothesis that emotions should activate emotion-congruent
material, consistent with the assumptions of Bower's (1981) associative network
model. Hence, emotion congruent material should be easier to recall, more readily
perceived, and more likely to interfere with competing material when one
experiences the respective emotion. However, the available data, mostly pertaining
to anxiety, do not provide strong support for the operation of some general form of
emotion congruence (see Mathews & MacLeod, 1994, for an extensive review).
Instead, the emotion seems to elicit a focus on material that is content
relevant rather than on material that is simply feeling consistent. For example,
Mogg, Mathews, and Eysenck (1992) observed in an attentional paradigm, that the
tendency of anxious subjects to focus on emotionally threatening words was not a
general effect across all negative words. Rather, anxious subjects were faster in
responding when the activated words matched their specific major domain of worry.
Similarly, Mathews and Klug (1993), crossed the valence of a set of words with
whether the content was or was not likely to be related to the concerns of anxious
patients. Content related words interfered more than did content unrelated words
-- and they did so regardless of their valence. Mathews and MacLeod (1994)
therefore concluded after an extensive review of the literature, "It is the match
with current domain of concern, rather than emotional valence or congruence in a
general sense, that determines the information that is given processing priority"
(p. 37). We note that this conclusion is compatible with the assumption that our
feelings inform us about our current situation, thus directing our attention to
features that are likely to make us anxious.
It is tempting to conjecture that a similar focus on current concerns, rather
than general valence, may underlie the inconsistent findings regarding the
influence of mood on memory, although the diffuse nature of moods would be likely
to elicit attention to a broader class of relevant concerns than the specific
nature of emotions, such as anxiety. As noted previously, the mood and memory
literature focused on the valence of recalled events, independent of their specific
nature. Hence, a recalled positive event may reflect that one obtained a positive
outcome or successfully avoided a negative outcome, whereas a negative event may
reflect that one obtained a negative outcome or did not obtain a positive outcome.
Recall, however, that Higgins and Tykocinski (1992) observed that subjects with
chronic actual:ideal discrepancies showed better recall for positive outcomes
described to them, independent of whether these were obtained positive or avoided
negative possibilities. Conversely, subjects with chronic actual:ought discrepancy
showed better recall for negative outcomes, independent of whether they were
obtained negative or lost positive possibilities. These findings indicate that
"people who self-regulate in relation to an ideal self-guide are sensitive to
psychological situations involving positive outcomes (presence or absence), whereas
people who self-regulate in relation to an ought self-guide are sensitive to
psychological situations involving negative outcomes (presence or absence)"
(Higgins & Tykocinski, 1992, p. 528). To the extent that our current feelings
signal different psychological situations, we may conjecture that they elicit
temporary attempts at self-regulation that are tuned to match the psychological
situation they presumably signal, in a way that is similar to the impact of chronic
self-discrepancies. Hence, they may elicit attention to different classes of
outcomes, rather than to different valences (see Higgins, in press).
If so, individuals in a sad mood may recall information related to positive
outcomes, independent of whether these outcomes were obtained or not. Only the
former, however, would reflect a positive valence. On the other hand, individuals
in an anxious mood may recall information related to negative outcomes, independent
of whether they were obtained or not. Again, only the former would reflect a
negative valence. Explorations of this possibility would require a coding of the
recalled material in terms of outcomes rather than valence, as well as a
distinction between dejected and agitated negative moods, which are difficult to
distinguish on the basis of the global mood inductions used in most studies.
Summary. In summary, mood congruent recall has been found to be a rather
fragile phenomenon, that is sometimes difficult to obtain in empirical studies
(Blaney, 1986; Bower & Mayer, 1985; Morris, 1989). Most importantly, mood
congruent recall is most likely to be obtained for self-referenced material, such
as autobiographical memories, and it is "impossible or difficult to demonstrate
when stimulus exposure occurs under sets that are explicitly antithetical to self-
referencing" (Blaney, 1986, p. 232). Moreover, mood congruency may be limited to
relatively unstructured material and tends to be difficult to find when material is
presented in narrative form, such that positive and negative elements are clearly
interconnected (Mecklenbr�uker & Hager, 1984; Hasher, Rose, Zacks, Sanft, & Doren,
1985) or otherwise well organized (Fiedler, Pampe, & Scherf, 1986). Finally, the
relative contributions of mood congruency, state dependency, and mood congruent
encoding are difficult to disentangle in many studies and mood congruent
selectivity in recall may be most likely to be obtained when several of these
processes operate in combination. Moreover, congruency in recall has not been
observed for more specific emotions, most notably anxiety (Mathews & MacLeod,
1994).
Despite the considerable interest in the issue and the large body of
available research, a compelling conceptual framework that accounts for the
accumulated inconsistencies pertaining to the mood and memory link is not yet
available (but see Forgas, 1992, for a different perspective). Reflecting our own
biases, we suppose that paying closer attention to the information provided by
one's mood and its relevance to the recall task may help in addressing the relevant
conditions, as discussed in our previous conjectures, but relevant empirical tests
are missing.

The Impact of Feelings on Behavior

As reflected in the present review, most of the research has focused on the
influence of feelings on cognitive processing. Attention to the impact of feelings
on behavior has been more limited, mostly addressing mood effects on helping
behavior (see Isen, 1984; Schaller & Cialdini, 1990, for reviews). Below, we
address whether moods and emotions are likely to influence behavior directly, or
indirectly, via an impact on motivation and cognition.
Many emotion theorists (e.g., Davidson, 1994; Frijda, 1986; Leventhal, 1982)
believe that emotions affect behavior more or less directly. It is common to
assume, for example, that fear involves behavioral tendencies to escape, that anger
involves activation of aggressive responses, that shame involves tendencies to
hide, and so on. We propose, however, that these links may be more indirect than
is generally assumed. Such words as "behavior," "response," and "action," even
when qualified by such words as "tendencies," "readinesses," or "inclinations,"
imply that specific muscle groups and motor circuits are activated when one is
angry, fearful, or ashamed. Such claims suggest, rather implausibly, that one's
legs are programmed to run when afraid, one's arm is programmed to hit when one is
angry, or one's hands programmed to cover one's face when ashamed. Fear and
related emotions may redistribute blood from the viscera to the large muscles, and
overall muscular tension may change, and such effects might indeed enable more
forceful or rapid action. But such general activation is not the same thing as a
specific motor program or an action tendency for a specific behavior.
We are inclined to look for the effects of emotion in experience and
motivation rather than in behavior. It seems likely that the direct effects of
emotions are motivational, changing the salience and priority of goals rather than
changing behavior or even behavior readiness in some reflexive way. Agreement
about the likely goals of angry, fearful, or ashamed persons should be much easier
to achieve than agreement about likely behaviors. For example, fear clearly
involves a desire to avoid harm or loss, but from knowing only that they are
afraid, we cannot predict whether people will sell their stocks, listen to the
weather report, or start running. The immediate effects of emotion, therefore, may
be more mental than behavioral, emphasizing the importance of the processes that
were the focus of the present chapter.

Conclusions
In line with the theme of the present handbook, our review of different
approaches to te interplay of affect and cognition focused on basic principles and
explored different mechanisms by which feelings may influence cognitive processes.
Each of these mechanisms can account for some, but not all, of the available data.
Moreover, many of the reviewed process assumptions are -- in principle -- not
mutually exclusive and each one of them may hold under some conditions. In a
commendable effort to offer an integration, Forgas (1992, 1994, 1995) proposed a
multi-process "affect infusion model" (AIM) that attempts to specify the conditions
under which the different process assumptions identified in previous research are
likely to hold. We agree on many of the core features of this model. Yet, a
selective discussion of some of its ambiguities is helpful in identifying major
gaps in our current state of knowledge.
The model distinguishes four different processing strategies. If the target
is familiar and a previously formed judgment is accessible in memory, individuals
are assumed to rely on a direct access strategy, provided that the judgment is not
personally relevant. Mood is not assumed to play a role in this case. The
prototypical example given (Forgas, 1992, p. 239) is stereotyping, which is
presumably reflecting the recall of a previously formed impression of a group.
Yet, as we have seen above, individuals are more likely to rely on stereotypes when
they are in a good rather than bad mood (e.g., Bodenhausen et al., 1994). In line
with our previous discussion of mood and processing strategies, we assume that
affective states influence the likelihood of a direct access strategy, with
individuals in a good mood being more likely to rely on a previously formed
judgment and individuals in a bad mood being more likely to recompute a judgment
based on currently accessible details (see Bless, Mackie, & Schwarz, 1992, for
empirical evidence).
As a second possibility, the model introduces a motivated processing
strategy, which individuals are likely to employ when they want to reach a certain
conclusion (e.g., Kruglanski, 1989; Kunda, 1990). This strategy is based on a
partial search of information, as conceptualized by Kunda (1990), and may serve
mood management goals.
A third possibility pertains to a substantive processing strategy, which is
based on extensive memory search and elaboration of the recalled material. Once
individuals engage in this strategy, their judgments are predicted to reflect mood
congruent recall and elaboration (Bower, 1981). People are assumed to use this
strategy under conditions of unconstrained processing capacity and high accuracy
motivation to form judgments that are demanding (as exemplified by atypical,
unusual or complex targets) and of some importance to them. Being in a sad mood is
assumed to facilitate the adoption of this processing strategy via increased
accuracy motivation, but may impede the adoption of this strategy via decreased
cognitive capacity (Forgas, 1992, p. 239).
In contrast, a fourth possibility pertains to a heuristic processing
strategy. Once individuals engage in this strategy, their affective states are
likely to influence their judgments by serving as an informational basis (Schwarz &
Clore, 1983). People are assumed to use this strategy under conditions of limited
processing capacity and low accuracy motivation to form judgments that are simple
(as exemplified by high typicality and low complexity of the target) and/or of
limited importance to them. Being in a good mood is assumed to facilitate the
adoption of this processing strategy via decreased accuracy motivation.
Most of the social psychological literature on mood and cognition pertains to
these latter two strategies, each of which predicts mood congruent judgments.
Hence, the affect infusion model (Forgas, 1992, 1994, 1995) fares well in terms of
outcome predictions. The conditions identified as triggering different strategies,
on the other hand, are more controversial and the available evidence is often
nondiagnostic and compatible with different assumptions. For example, is it really
the case that more complex judgmental tasks are likely to trigger a substantive
processing strategy, whereas simple tasks trigger a heuristic processing strategy?
Or do people resort to a "How-do-I-feel-about-it?" heuristic to simplify complex
tasks, which is not needed for less demanding tasks? Whereas the affect infusion
model (Forgas, 1992) suggests the former, Schwarz, et al. (1987) suggested the
latter -- and none of the studies allegedly supporting each claim involved a
misattribution manipulation that would allow us to distinguish between a mood-as-
information and a mood-congruent recall account. Moreover, does reliance on one's
feelings as a source of information always reflect a heuristic processing strategy?
We assume that this is the case when the judgment is nonaffective in nature, but
when the judgment pertains to one's feelings (e.g., judgments of liking), reliance
on these feelings does not reflect a heuristic shortcut but the use of the most
diagnostic substantive information available. And what are the implications of the
assumption that being in a sad mood fosters substantive processing, which
presumably gives rise to mood-congruent recall and elaboration, while being in a
good mood fosters heuristic processing? Does this suggest that the impact of good
moods is likely to reflect a mood-as-information process, whereas the impact of sad
moods is likely to reflect a mood-congruent recall process? In contrast to this
conjecture, the impact of sad moods is more reliably eliminated by misattribution
manipulations than the impact of happy moods (e.g., Schwarz & Clore, 1983).
Moreover, happy subjects are more likely to recall positive material, but sad
subjects are not more likely to recall negative material (see Singer & Salovey,
1988), as discussed in our review of the memory literature. How can this be
reconciled with the relatively symmetric impact of happy and sad moods in studies
reviewed as evidence for mood-congruent recall effects in social judgment (Forgas,
1992)?
As yet there are not definitive answers to many of these and other relevant
questions. The proliferation of research on affect and cognition during the last
two decades has resulted in a multitude of findings, but most of them can be framed
in terms of two global approaches: An experiential approach that focuses on the
informational value of subjective experiences, which include moods, emotions,
bodily sensations, and cognitive experiences; and a cognitive approach that focuses
on the impact of affective states on the content of the thoughts that come to mind
rather than the experience of having the thoughts. Each of these approaches is
supported by a number of unique findings, whereas other findings are compatible
with both. While researchers working within each approach are able to produce many
of the core effects with considerable reliability, the interplay of the underlying
processes is awaiting an encompassing conceptualization and more rigorous testing
than has been provided in studies designed to demonstrate the operation of one or
another mechanism. We hope that our accentuation of the principles underlying each
approach will help in tackling this formidable task.

�LS1�References
Adelmann, P. K., & Zajonc, R. B. (1989). Facial efference and the experience of
emotion. Annual Review of Psychology, 40, 249-280.
Alloy, L. B. (1988). Cognitive processes in depression. New York: Guilford Press.
Alloy, L. B., & Abramson, L. Y. (1979). Judgment of contingency in depressed and
non-depressed students: Sadder but wiser? Journal of Experimental Psychology:
General, 108, 441-485.
Alloy, L. B., Abramson, L. Y., & Viscusi, D. (1981). Induced mood and the illusion
of control. Journal of Personality and Social Psychology, 41, 1129-1140.
Anderson, J. R., & Bower, G. H. (1973). Human associative memory. Washington, D.C.:
Winston.
Angell, J. R. (1909). Psychology (4th ed., revised). New York: Henry Holt.
Arnold, M. B. (1960). Emotion and personality. New York: Columbia University Press.
Averill, J. R. (1980). A constructivist view of emotions. In R. Plutchik & H.
Kellerman (Eds.), Emotions: Theory, research, and experience (Vol. 1, pp. 305-339).
New York: Academic Press.
Baddeley, A. (1990). Human memory. Theory and practice. Hillsdale, NJ: Erlbaum.
Bargh, J. A. (1994). The four horseman of automaticity: Awareness, intention,
efficacy, and control in social cognition. In R. S. Wyer, & T. K. Srull (Eds.),
Handbook of social cognition (2nd ed.; Vol. 1, pp. 1-40). Hillsdale, NJ: Erlbaum.
Bargh, J. A., Litt, J., Pratto, F., & Spielman, L. A. (1989). On the preconscious
evaluation of social stimuli. In F. A. Bennett &K. M. McConkey (Eds.), Cognition
in individual and social contexts. Amsterdam, NL: Elsevier/North Holland.
Bargh, J. A., & Pietromonaco, P. (1982). Automatic information processing and
social perception: The influence of trait information presented outside of
conscious awareness on impression formation. Journal of Personality and Social
Psychology, 43, 437-449.
Batson, C. D., Shaw, L. L., & Oleson, K. C. (1992). Differentiating affect, mood,
and emotion. In M. S. Clark (Ed.), Review of Personality and Social Psychology,
Vol. 11 (pp. 294-326). Beverly Hills, CA: Sage.
Begg, I., Armour, V., & Kerr, T. (1985). On believing what we remember. Canadian
Journal of Behavioral Science, 17, 199-214.
Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in
experimental social psychology, 6, 1-62. New York: Academic Press.
Blaney, P. H. (1986). Affect and memory: A review. Psychological Bulletin, 99, 229-
246.
Bless, H. (1994). Stimmung und die Nutzung allgemeiner Wissensstrukturen: Ein
Modell zum Einfluss von Stimmungen auf Denkprozesse /Mood and the use of general
knowledge structures/. Habilitationsschrift. Heidelberg, FRG: University of
Heidelberg.
Bless, H., Bohner, G., Schwarz, N., & Strack, F. (1990). Mood and persuasion: A
cognitive response analysis. Personality and Social Psychology Bulletin, 16, 331-
345.
Bless, H. Hamilton, D.L., & Mackie, D.M. (1992). Mood effects on the organization
of person information. European Journal of Social Psychology, 22, 497-509.
Bless, H., Mackie, D. M., & Schwarz, N. (1992). Mood effects on encoding and
judgmental processes in persuasion. Journal of Personality and Social Psychology,
63, 585-595.
Bless, H., Schwarz, N., & Wieland, R. (1994). Mood and stereotyping: The impact of
category membership and individuating information. Unpublished manuscript.
Heidelberg, FRG: University of Heidelberg.
Bodenhausen, G.V. (1990). Stereotypes as judgmental heuristics. Evidence of
circadian variations in discrimination. Psychological Science, 1, 319-322.
Bodenhausen, G. V. (1993). Emotions, arousal, and stereotypic judgments: A
heuristic model of affect and stereotyping. In D. M. Mackie & D. L. Hamilton
(Eds.), Affect, cognition, and stereotyping (pp. 13-37). San Diego, CA: Academic
Press.
Bodenhausen, G.V., Kramer, G.P. & S�sser, K. (1994). Happiness and stereotypic
thinking in social judgment. Journal of Personality and Social Psychology, 66, 621-
632.
Bohner, G., Bless, H., Schwarz, N., & Strack, F. (1988). What triggers causal
attributions? The impact of valence and subjective probability. European Journal of
Social Psychology, 18, 335 - 345.
Bohner, G., Crow, K., Erb, H.-P., & Schwarz, N. (1992). Affect and persuasion: Mood
effects on the processing of message content and context cues and on subsequent
behavior. European Journal of Social Psychology, 22, 511-530.
Bollnow, O. F. (1956). Das Wesen der Stimmungen. Frankfurt: Klostermann.
Bonanno, G. A., & Stillings, N. A. (1986). Preference, familiarity, and recognition
after repeated brief exposures to random geometric shapes. American Journal of
Psychology, 99, 403-415.
Borg, I. (1987). The effect of mood on different well-being judgments. Archiv f�r
Psychologie, 139, 181-188.
Bornstein, R. F. (1989). Exposure and affect: Overview and meta-analysis of
research, 1968-1987. Psychological Bulletin, 106, 265-289.
Bornstein, R. F. (1992). Subliminal mere exposure effects. In R. F. Bornstein & T.
S. Pittman (Eds.), Perception without awareness: Cognitive, clinical, and social
perspectives (pp. 191-210). New York: Guilford.
Bornstein, R.F. & D'Agostino, P.R. (1992). Stimulus recognition and the mere
exposure effect. Journal of Personality and Social Psychology, 63, 545-552.
Bornstein, R. F., & D'Agostino, P. R. (1994). The attribution and discounting of
perceptual fluency: Preliminary tests of a perceptual fluency/attributional model
of the mere exposure effect. Social Cognition, 12, 103-128.
Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129 - 148.
Bower, G. H. (1991). Mood congruity of social judgments. In J. P. Forgas (Ed.),
Emotion and social judgments (pp. 31-53). Oxford, UK: Pergamon.
Bower, G. H., & Cohen, P. R. (1982). Emotional influences in memory and thinking:
Data and theory. In M. S. Clark & S. T. Fiske (Eds.), Affect and cognition (pp.
291-332). Hillsdale, NJ: Erlbaum.
Bower, G. H., & Mayer, J. D. (1985). Failure to replicate mood congruent retrieval.
Bulletin of the Psychonomic Society, 23, 39-42.
Bower, G. H., Monteiro, K. P., & Gilligan, S. G. (1978). Emotional mood as a
context of learning and recall. Journal of Verbal Learning and Verbal Behavior, 17,
573-585.
Brewer, M. B. (1988). A dual process model of impression formation. In T.K. Srull
& R.S. Wyer (Eds.), Advances in social cognition (Vol. 1, pp. 1-36). Hillsdale,
NJ: Lawrence Erlbaum Associates.
Brewer, W. F. (1992). Phenomenal experience in laboratory and autobiographical
memory. In M.A., Conway, D.C. Rubin, H. Spinner, & W.A Wagenaar (Eds.),
Theoretical perspectives in autobiographical memory (pp. 31-51). Netherlands:
Kluwer.
Caccioppo, J. T., Bush, L. K., & Tassinary, L. G. (1992). Microexpressive facial
actions as a function of affective stimuli: Replication and extension. Personality
and Social Psychology Bulletin, 18, 515-526.
Calkins, M. W. (1905). An introduction to psychology (2nd ed.). New York:
Macmillan.
Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control-
theory approach to human behavior. New York: Springer-Verlag.
Carver, C. S., & Scheier, M. F. (1990). Origins and functions of positive and
negative affect: a control-process view. Psychological Review, 97, 19-35.
Clark, D. M., & Teasdale, J. D. (1982). Diurnal variation in clinical depression
and accessibility of memories of positive and negative experiences. Journal of
Abnormal Psychology, 91, 87-95.
Clark, L. F., & Collins, J. E. (1993). Remembering old flames: How the past affects
assessments of the present. Personality and Social Psychology Bulletin, 19, 399-
408.
Clark, L. F., Collins, J. E., & Henry, S. M. (1994). Biasing effects of
retrospective reports on current self-assessments. In N. Schwarz & S. Sudman
(Eds.), Autobiographical memory and the validity of retrospective reports (pp. 291-
304). New York: Springer Verlag.
Clore, G. L. (1992). Cognitive phenomenology: Feelings and the construction of
judgment. In L. L. Martin & A. Tesser (Eds.), The construction of social judgment
(pp. 133-164). Hillsdale, NJ: Erlbaum.
Clore, G. L. (1994). Why emotions are felt. In P. Ekman & R. J. Davidson (Eds.),
The Nature of Emotion: Fundamental Questions (pp. 103-111). New York: Oxford
University Press.
Clore, G.L., & Byrne, D. (1974). A reinforcement affect model of attraction. In
T.L. Huston (Ed.), Foundations of interpersonal attraction (pp. 173-170). New York:
Academic Press.
Clore, G. L., Wong, G. M., Isbell, L., & Gasper, K. (1994). Odor, Experiential
Orientation, and Affect. Proceedings of the International Society for Research on
Emotion.
Clore, G. L., Ortony, A., Dienes, B., & Fujita, F. (1994). Where does anger dwell?
In T. K. Srull & R. S. Wyer (Eds.), Advances in social cognition, Vol. 5.
Hillsdale, N. J.: Erlbaum.
Clore, G. L., Ortony, A., & Foss, M. (1987). The psychological foundations of the
affective lexicon. Journal of Personality and Social Psychology, 53, 751-766.
Clore, G. L., & Parrott, W. G. (1991). Moods and their vicissitudes: Thoughts and
feelings as information. In J. Forgas (Ed.), Emotion and social judgment (pp. 107-
123). Oxford: Pergamon Press.
Clore, G. L., Schwarz, N., & Conway, M. (1994). Affective causes and consequences
of social information processing. In R. S. Wyer, & T. K. Srull (Eds.), Handbook of
social cognition (2nd ed.; Vol. 1, pp. 323-418). Hillsdale, NJ: Erlbaum.
Cramer, P. (1968). Word association. New York: Academic Press.
Damasio, A. R. (1994). Descarte's error: Emotion, reason and the human brain. New
York: Grosset/Putnam.
Darwin, C. (1872). The expression of the emotions in man and animals. London: J.
Murray.
Davidson, R. J. (1994). Complexities in the search for emotion-specific physiology.
In P. Ekman & R. J. Davidson (Eds.), The Nature of Emotion: Fundamental Questions
(pp. 237-242). New York: Oxford University Press.
DeLoache, J. S., & Brown, A. L. (1984). Where do I go next? Intelligent searching
by very young children. Developmental Psychology, 20, 37-44.
Derry, P. A., & Kuiper, N. A. (1981). Schematic processing and self-reference in
clinical depression. Journal of Abnormal Psychology, 90, 286-297.
Dobson, D. J. G., & Dobson, K. S. (1981). Problem-solving strategies in depressed
and nondepressed students. Cognitive Therapy and Research, 5, 237-249.
Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort Worth, TX:
Harcourt Brace Jovanovich.
Edwards, J. A. & Weary, G. (1993). Depression and the impression-formation
continuum: Piecemeal processing despite the availability of category information.
Journal of Personality and Social Psychology, 64, 636-645.
Ekman, P., Levenson, R. W., & Friesen, W. V. (1983). Autonomic nervous system
activity distinguishes among emotions. Science, 221, 1208-1210.
Ellis, H. C., & Ashbrook, P. W. (1988). Resource allocation model of the effects of
depressed mood states on memory. In K. Fiedler & J. Forgas (Eds.), Affect,
cognition, and social behavior. Toronto: C.J. Hogrefe.
Ellis, H. C., Thomas, R. L., & Rodriguez, I. A. (1984). Emotional mood states and
memory: Elaborative encoding, semantic processing, and cognitive effort. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 10, 470-482.
Erber, R. (in press). The self-regulation of moods. In L. L. Martin & A. Tesser
(Eds.), Striving and feeling: Interactions between goals and affect. Hillsdale, NJ:
Erlbaum.
Fiedler, K. (1988). Emotional mood, cognitive style, and behavior regulation. In K.
Fiedler & J. Forgas (Eds.), Affect, cognition, and social behavior (pp. 100-119).
Toronto: Hogrefe International.
Fiedler, K., Pampe, H., & Scherf, U. (1986). Mood an memory for tightly organized
social information. European Journal of Social Psychology, 16, 149-164.
Fiske, S. T., & Neuberg, S. L. (1990). A continuum of impression formation, from
category-based to individuating processes: Influences of information and motivation
on attention and interpretation. In M. P. Zanna (Ed.), Advances in Experimental
Social Psychology (Vol. 23, pp. 1-74). San Diego, CA: Academic Press.
Fletcher, G. J. D., Danilovics, P., Fernandez, G., Peterson, D., & Reeder, G. D.
(1986). Attributional complexity: An individual difference measure. Journal of
Personality and Social Psychology, 56, 757-764.
Forgas, J. P. (1992). Affect in social judgments and decisions: A multi-process
model. In M. P. Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 25,
pp. 227-275). San Diego: Academic Press.
Forgas, J.P. (1994). The role of emotion in social judgments: An introductory
review and an affect infusion model (AIM). European Journal of Social Psychology,
24, 1-24.
Forgas, J. P. (1995). Emotion in social judgments: Review and a new affect infusion
model (AIM). Psychological Bulletin,
Forgas, J. P. (1993). On making sense of odd couples: Mood effects on the
perception of mismatched relationships. Personality and Social Psychology Bulletin,
19, 59-71.
Forgas, J. P., & Bower, G. H. (1987). Mood effects on person perception judgments.
Journal of Personality and Social Psychology, 53, 53-60.
Forgas, J. P., & Bower, G. H. (1988). Affect in social and personal judgments. In
K. Fiedler & J. Forgas (Eds.), Affect, cognition, and social behavior (pp. 183-
207). Toronto: Hogrefe International.
Frijda, N. H. (1986). The emotions. New York: Cambridge University Press.
Frijda, N. H. (1988). The laws of emotion. American Psychologist, 43, 349-358.
Frijda, N., Ortony, A., Sonnemans, J., & Clore, G. (1992). The complexity of
intensity: issues concerning the structure of emotion intensity. In M. S. Clark
(Ed.) Review of Personality and Social Psychology, Vol. 11 (pp. 60-89). Beverly
Hills, CA: Sage.
Gallagher, D., & Clore, G. L. (1985, May). Effects of fear and anger on judgments
of risk and blame. Paper presented at the meetings of the Midwestern Psychological
Association, Chicago.
Gilligan, S. G., & Bower, G. H. (1984). Cognitive consequences of emotional
arousal. In C. Izard, J. Kagen, & R. Zajonc (Eds.), Emotions, cognition, and
behavior (pp. 547-588). Cambridge, UK: Cambridge University Press.
Glenberg, A. M., Wilkinson, A. C., & Epstein, W. (1982). The illusion of knowing :
failure in the self-assessment of comprehension. Memory & Cognition, 10, 597-602.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes,
self-esteem, and stereotypes. Psychological Review, 102, 4-27.
Hamilton, D. L., Katz, L. B., & Leirer, V. O. (1980). Cognitive representation of
personality impressions. Journal of Personality and Social Psychology, 39, 1050-
1063.
Harris, P. L., Kruithof, A., Meerum Terwogt, M., & Visser, T. (1981). Children's
detection and awareness of textual anomaly. Journal of Experimental Child
Psychology, 31, 212-230.
Hasher, L., Rose, K. C., Zacks, R. T., Sanft, H., & Doren, B. (1985). Mood, recall,
and selectivity in normal college students. Journal of Experimental Psychology:
General, 114, 104-118.
Hertel, P. T., & Hardin, T. S. (1990). Remembering with and without awareness in a
depressed mood: Evidence of deficits in initiative. Journal of Experimental
Psychology: General, 119, 45-59.
Higgins, E. T. (1987). Self-discrepancy theory: A theory relating self and affect.
Psychological Review, 94, 319-340.
Higgins, E. T. (1989). Knowledge accessibility and activation: Subjectivity and
suffering from unconscious sources. In J. S. Uleman & J. A. Bargh (Eds.),
Unintended thought (pp.75-123). New York: Guilford Press.
Higgins, E.T. (in press). Ideals, thoughts, and regulatory outcome focus: Relating
affect and motivation to distinct pains and pleasures. In P. M. Gollwitzer & J. A.
Bargh (Eds.), The psychology of action: Linking cognition and motivation to
behavior. New York: Guilford.
Higgins, E. T., & Bargh, J. A. (1987). Social cognition and social perception.
Annual Review of Psychology, 38, 369-425.
Higgins, E.T., & Lurie, L. (1983). Context, categorization, and memory: The
"change-of-standard" effect. Cognitive Psychology, 15, 525-547.
Higgins, E. T., Rholes, W. S., & Jones, C. R. (1977). Category accessibility and
impression formation. Journal of Experimental Social Psychology, 13, 141-154.
Higgins, E.T., & Stangor, C. (1988). A "change-of-standard" perspective on the
relations among context, judgment, and memory. Journal of Personality and Social
Psychology, 54, 181-192.
Higgins, E. T. & Tychocynski, O. (1992). Self-discrepancies and biographical
memory: Personality and cognition at the level of psychological situations.
Personality and Social Psychology Bulletin, 18, 527-535.
Higgins, E.T., Van Hook, & Daubman, (1988). $$. Social Cognition, 6, 177-206.
Hildebrand-Saints, L., & Weary, G. (1989). Depression and social information
gathering. Personality and Social Psychology Bulletin, 15, 150-160.
Hirt, E.R., Melton, J. R., McDonald, E. E., & Harackiewicz, J.M. (1995). Processing
goals, task interest, and the mood-performance relationship: A mediational
analysis. Unpublished manuscript; Bloomington, IN: Indiana University.
Howell, A., & Conway, M. (1992). Mood and the suppression of positive and negative
self-referent thoughts. Cognitive Therapy and Research, 16, 1-21.
Hunter, I. M. L. (1964). Memory. Baltimore, MD: Penguin Books.
Isen, A. M. (1984). Toward understanding the role of affect in cognition. In R. S.
Wyer, Jr. & T. K. Srull (Eds.), Handbook of social cognition (Vol. 3, pp. 179-236).
Hillsdale, NJ: Erlbaum.
Isen, A. M. (1987). Positive affect, cognitive processes, and social behavior. In
L. Berkowitz (Ed.), Advances in experimental social Psychology. (Vol. 20; pp. 203-
253). New York: Academic Press.
Isen, A., & Means, B. (1983). The influence of positive affect on decision-making
strategy. Social Cognition, 2, 18-31
Isen, A. M., Shalker, T. E., Clark, M. S., & Karp, L. (1978). Affect, accessibility
of material in memory, and behavior: A cognitive loop? Journal of Personality and
Social Psychology, 36, 1-12.
Izard, C. E. (1977). Human emotions. New York: Plenum Press.
Jacobsen, E. (1957). Normal and pathological moods: Their nature and function. In
R. S. Eisler, A. F. Freud, H. Hartman, & E. Kris (Eds.), The psychoanalytic study
of the child (pp. 73-113). New York: International University Press.
Jacoby, L. L. & Dallas, M. (1981). On the relationship between autobiographical
memory and perceptual learning. Journal of Experimental Psychology: General, 110,
306-340.
Jacoby, L. L. & Kelley, C. M. (1987). Unconscious influence of memory for a prior
event. Personality and Social Psychology Bulletin, 13, 314-336.
Jacoby, L. L., Kelley, C. M., Brown, J., Jasechko, J. (1989). Becoming famous
overnight: Limits on the ability to avoid unconscious influences of the past.
Journal of Personality and Social Psychology, 56, 326-338.
James, W. (1890). Principles of psychology. (Vols. 1 & 2). New York: Dover.
Johnson, E., & Tversky, A. (1983). Affect, generalization, and the perception of
risk. Journal of Personality and Social Psychology, 45, 20-31.
Kelley, H. H. (1972). Causal schemata and the attribution process. Morristown, NJ:
General Learning Press.
Keltner, D., Ellsworth, P., & Edwards, K. (1993). Beyond simple pessimism: Effects
of sadness and anger on social perception. Journal of Personality and Social
Psychology, 64, 740-752.
Keltner, D., Locke, K. D., & Audrain, P. C. (1993). The influence of attributions
on the relevance of negative feelings to satisfaction. Personality and Social
Psychology Bulletin, 19, 21-30.
Kirschenbaum, D. S., Tomarken, A. J., & Humphrey, L. L. (1985). Affect and adult
self-regulation. Journal of Personality and Social Psychology, 48, 509-523.
Kruglanski, A .W. (1980). Lay epistemo-logic -- process and contents: Another look
at attribution theory. Psychological Review, 87, 70-87.
Kruglanski, A. W. (1989). Lay epistemics and human knowledge: Cognitive and
motivational bases. New York: Plenum.
Kuhl, J. (1983). Emotion, Kognition und Motivation, II. (Emotion, cognition, and
motivation.) Sprache und Kognition, 4, 228-253.
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108,
331-350.
Lacey, J. I., & Lacey, B. C. (1970). Some autonomic nervous system relationships.
In P. Black (Ed.), Physiological correlates of emotion (pp. 205-227). New York:
Academic Press.
Lachman, R., Lachman, J.T., & Butterfield, E.C. (1979). Cognitive psychology and
information processing. Hillsdale, NJ: Erlbaum.
Laird, J.D. (1974). Self-attribution of emotion: The effects of expressive
behavior on the quality of emotional experience. Journal of Personality and Social
Psychology, 33, 475-486.
Laird, J. D., Wagener, J., Halal, M., & Szegda, M. (1982). Remembering what you
feel: Effects of emotion on memory. Journal of Personality and Social Psychology,
42, 646-657.
Lang, P. J. (1988). What are the data of emotion? In V. Hamilton, G. Bower, & N.
Frijda (Eds.), Cognitive science perspectives on emotion and motivation (367-398).
Amsterdam: Martinus Nijhoff.
Lassiter, G. D., & Koenig. (1991). Depression and the unitization of behavior.
Manuscript under review.
Lazarus, R. S., Kanner, A. D., & Folkman, S. (1980). Emotions: A cognitive-
phenomenological analysis. In R. Plutchik & H. Kellerman (Eds.), Emotion: Theory,
research, and experience: Vol. 1. Theories of emotion. New York: Academic Press,
pp. 189-217.
Lazarus, R. S. (1966). Psychological stress and the coping process. New York:
McGraw-Hill.
Lazarus, R. S., Averill, J. R., & Opton, E. M. (1970). Toward a cognitive theory of
emotion. In M. Arnold (Ed.), Feeling and emotion. New York: Academic Press.
LeDoux, D. (1987). Emotion. Handbook of physiology: The nervous system (Vol. 5, pp.
419-459). Washington, DC: APA.
Leight, K. A., & Ellis, H. C. (1981). Emotional mood states, strategies, and state-
dependency in memory. Journal of Verbal Learning and Verbal Behavior, 20, 251-266.
Leventhal, H. (1982). The integration of emotion and cognition: A view from the
perceptual-motor theory of emotion. In M. S. Clark & S. T. Fiske (Eds.), Affect and
Cognition (pp. 121-156.) Hillsdale, NJ: Erlbaum.
Levine, G. M., Hirt, E. R., McDonald, E.E., Melton, J.R., & Martin, L.L. (1995,
May). Awareness of mood source in performance decisions. Midwestern Psychological
Association, Chicago, IL.
Levine, S., Wyer, R.S. Jr., & Schwarz, N. (1994). Are you what you feel? The
affective and cognitive determinants of self-esteem. European Journal of Social
Psychology, 24, 63-77.
Liu, J. H., Karasawa, K., & Weiner, B. (1992). Inferences about the causes of
positive and negative emotions. Personality and Social Psychology Bulletin, 18,
603-615.
Lombardi, W. J., Higgins, E. T., & Bargh, J. A. (1987). The role of consciousness
in priming effects on categorization: Assimilation versus contrast as a function of
awareness of the priming task. Personality and Social Psychology Bulletin, 13, 411-
429.
Lutz, C., & White, G. M. (1986). The anthropology of emotions. Annual Review of
Anthropology. 15, 405-436.
Mackie, D. M., Asuncion, A.G., & Rosselli, F. (1992). The impact of affective
states on persuasion processes. In M. Clark (Ed.), Review of Personality and Social
Psychology (Vol. 14, pp. 247-270). Beverly Hills, CA: Sage.
Mackie, D. M., & Worth, L. T. (1989). Cognitive deficits and the mediation of
positive affect in persuasion. Journal of Personality and Social Psychology, 57,
27-40.
Macrae, C. N., Milne, A. B., & Bodenhausen, G. V. (1994). Stereotypes as energy-
saving devices: A peek inside the cognitive toolbox. Journal of Personality and
Social Psychology, 66, 37-47.
Madigan, R. J., & Bollenbach, A. K. (1982). Effects of induced mood on retrieval of
personal episodic and semantic memories. Psychological Reports, 50, 147-158.
Mandler, G. (1975). Mind and emotion. New York: Wiley.
Mandler, G. (1980). Recognizing: the judgment of previous occurrence. Psychological
Review, 87, 252-271.
Mandler, G. (1984). Mind and body. New York: Norton.
Mandler, J. M. (1984). Stories, scripts, and scenes: Aspects of schema theory.
Hillsdale,N.J.: Erlbaum.
Manstead, A. S. R., & Tetlock, P. E. (1989). Cognitive appraisals and emotional
experience: Further evidence. Cognition and Emotion, 3, 225-240.
Markman, E. M. (1977). Realizing that you don't understand: a preliminary
investigation. Child Development, 48, 986-992.
Marsh, K. L., & Weary, G. (1989). Depression and attributional complexity.
Personality and Social Psychology Bulletin, 15, 325-336.
Marshall-Garcia, K. A., & Beck, R. C. (1985). Mood and recognition memory: A
comparison of two procedures. Bulletin of the Psychonomic Society, 23, 450-452.
Martin, L. L. (1986). Set/reset: Use and disuse of concepts in impression
formation. Journal of Personality and Social Psychology, 51, 493-504.
Martin, L. L., Harlow, T. F., & Strack, F. (1992). The role of bodily sensations in
the evaluation of social events. Personality and Social Psychology Bulletin, 18,
412-419.
Martin, L. L., Seta, J. J., & Crelia, R. A. (1990). Assimilation and contrast as a
function of people's willingness and ability to expend effort in forming an
impression. Journal of Personality and Social Psychology, 59, 27-37.
Martin, L. L., Ward, D. W., Ach�e, J. W., & Wyer, R. S. (1993). Mood as input:
People have to interpret the motivational implications of their moods. Journal of
Personality and Social Psychology, 64, 317-326.
Martin, L. L., & Tesser, A. (1989). Toward a motivational and structural theory of
ruminative thought. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought (pp.
306-326). New York: Guilford Press.
Masters, J. C., Barden, R. C., & Ford, M. E. (1979). Affective states, expressive
behavior, and learning in children. Journal of Personality and Social Psychology,
37, 380-390.
Mathews, A., & Bradley, B. (1983). Mood and the self-reference bias in recall.
Behavior Research and Therapy, 21, 233-239.
Mathews, A. M. & Klug, F. (1993). Emotionality and interference with color-naming
in anxiety. Behavior Research and Therapy, 31, 57-62.
Mathews, A. & MacLeod, C. (1994). Cognitive approaches to emotion and emotional
disorders. Annual Review of Psychology, 45, 25-50.
Matlin, M. W., & Stang, D. (1979). The Pollyanna Principle: Selectivity in
language, memory, and thought. Cambridge, MA: Shenkman.
Mayer, J. D., Gaschke, Y. N., Braverman, D. L., & Evans, T. W. (1992). Mood-
congruent recall is a general effect. Journal of Personality and Social Psychology,
63, 119-132.
Mecklenbr�uker, S., & Hager, W. (1984). Effects of mood on memory: Experimental
tests of a mood-state-dependent retrieval hypothesis and of a mood-congruity
hypothesis. Psychological Research, 46, 335-376.
Mogg, K., Mathews, A. M., & Eysenck, M. (1992). Attentional bias to threat in
clinical anxiety. Cognition and Emotion, 6, 149-159.
Moore, T. V. (1939). Cognitive psychology. Chicago: J. B. Lippincott.
Morris, W. N. (1989). Mood: The frame of mind. New York: Springer Verlag.
Morris, W. N. (1992). A functional analysis of the role of mood in affective
systems. Review of Personality and Social Psychology, Vol. 11 (pp. 256-293).
Beverly Hills, CA: Sage.
Murphy, S. T., & Zajonc, R. B. (1993). Affect, cognition, and awareness: Priming
with optimal and suboptimal stimulus exposures. Journal of Personality and Social
Psychology, 64, 723-739.
Natale, M., & Hantas, M. (1982). Effect of temporary mood states on selective
memory about the self. Journal of Personality and Social Psychology, 42, 927-934.
Nowlis, V., & Nowlis, H. H. (1956). The description and analysis of mood. Annals of
the New York Academy of Sciences, 65, 345-355.
Olson, J. M., & Roese, N. J. (in press). The perceived funniness of humorous
stimuli. Personality and Social Psychology Bulletin.
Ortony, A., Clore, G. L., & Collins, A. (1988). The cognitive structure of
emotions. New York: Cambridge University Press.
Ostrom, T.M. (1984). The sovereignity of social cognition. In R.S. Wyer & T.K.
Srull, Eds., Handbook of social cognition. (Vol. 1, pp. 1-38). Hillsdale: Erlbaum.
Parrott, W. G., & Sabini, J. (1990). Mood and memory under natural conditions:
Evidence for mood incongruent recall. Journal of Personality and Social Psychology,
59, 321-336.
Perkins, S. C., Meyers, A. W., & Cohen, R. (1988). Problem-solving ability and
response to feedback in peer-nominated mildly depressed children. Cognitive Therapy
and Research, 12, 89-102.
Petty, R. & Cacioppo, J. (1986). Communication and persuasion: Central and
peripheral routes to attitude change. New York: Springer-Verlag.
Pittman, N. L., & Pittman, T. S. (1980). Deprivation of control and the attribution
process. Journal of Personality and Social Psychology, 39, 377-389.
Pittman, T. S., & D'Agostino, P. R. (1985). Motivation and attribution: The effects
of control deprivation on subsequent information processing. In J. H. Harvey & G.
Weary (Eds.), Attribution: Basic and applied issues (pp. 117-142). San Diego, CA:
Academic Press.
Pittman, T. S., & D'Agostino, P. R. (1989). Motivation and cognition: Control
deprivation and the nature of subsequent information processing. Journal of
Experimental Social Psychology, 25, 465-480.
Poincare, H. (1913). The foundations of science (G. B. Halsted, trans.). New York:
The Science Press.
Potts, R., Camp, C., & Coyne, C. (1989). The relationship between naturally
occurring dysphoric moods, elaborative encoding, and recall performance. Cognition
and Emotion, 3, 197-205.
Sachs, O. (1985). The man who mistook his wife for a hat. London: Gerald Duckworth
& Co.
Salovey, P., & Singer, J. A. (1985, August). The effects of mood on the recall of
childhood and recent autobiographical memories. Paper presented at annual
convention of the American Psychological Association, Los Angeles, CA.
Schachter, S., & Singer, J. E. (1962). Cognitive, social, a physiological
determinants of emotional state. Psychological Review, 69, 379-399.
Schaller, M. & Cialdini, R. (1990). Happiness, sadness, and helping: A
motivational integration. In R. Sorrentino & E. T. Higgins (Eds.), Handbook of
motivation and cognition: Foundatiions of social behavior. (Vol. 2, pp. 265-296).
New York: Guilford Press.
Schare, M. L., Lisman, S. A., & Spear, N. E. (1984). The effects of mood variation
on state-dependent retention. Cognitive Therapy and Research, 8, 387-408.
Schwarz, N. (1990). Feelings as information: Informational and motivational
functions of affective states. In E.T. Higgins & R. Sorrentino (Eds.), Handbook of
motivation and cognition: Foundations of social behavior (Vol. 2, pp. 527-561). New
York: Guilford Press.
Schwarz, N., & Bless, B. (1991). Happy and mindless, but sad and smart? The impact
of affective states on analytic reasoning. In J. Forgas (Ed.), Emotion and social
judgment (pp. 55-71). Oxford: Pergamon.
Schwarz, N., & Bless, H. (1992). Constructing reality and its alternatives:
Assimilation and contrast effects in social judgment. In L.L. Martin & A. Tesser
(Eds.), The construction of social judgment (pp. 217-245). Hillsdale, NJ: Erlbaum.
Schwarz, N., Bless, H., & Bohner, G. (1991). Mood and persuasion: Affective states
influence the processing of persuasive communications. In M. Zanna (Ed.), Advances
in Experimental Social Psychology (Vol. 24, pp. 161-199). San Diego, CA: Academic
Press.
Schwarz, N., Bless, H., Strack, F., Klumpp, G., Rittenauer-Schatka, H., & Simons,
A. (1991). Ease of retrieval as information: Another look at the availability
heuristic. Journal of Personality and Social Psychology, 61, 195-202.
Schwarz, N., & Bohner, G. (in press). Feelings and their motivational implications:
Moods and the action sequence. In J. A. Bargh & P. Gollwitzer (Eds.), The
psychology of action: Linking thought and motivation to behavior. New York:
Guilford.
Schwarz, N., & Clore, G. L. (1983). Mood, misattribution, and judgments of well-
being: Informative and directive functions of affective states. Journal of
Personality and Social Psychology, 45, 513-523.
Schwarz, N., & Clore, G. L. (1988). How do I feel about it? Informative functions
of affective states. In K. Fiedler & J. Forgas (Eds.), Affect, cognition, and
social behavior (pp. 44-62). Toronto: Hogrefe International.
Schwarz, N., Servay, W., & Kumpf, M. (1985). Attribution of arousal as a mediator
of the effectiveness of fear-arousing communications. Journal of Applied Social
Psychology, 15, 74-78.
Schwarz, N., Strack, F., Kommer, D., & Wagner, D. (1987). Soccer, rooms and the
quality of your life: Mood effects on judgments of satisfaction with life in
general and with specific life-domains. European Journal of Social Psychology, 17,
69-79.
Seligman, M. E. P. (1975). Helplessness: On depression, development, and death. San
Francisco, CA: Freeman.
Sherman, S. J., & Corty, E. (1984). Cognitive heuristics. In R.S. Wyer & T. K.
Srull (Eds.), Handbook of social cognition (Vol. 1, pp. 189-286). Hillsdale:
Erlbaum.
Siemer, M., & Reisenzein, R. (1994). Effects of mood on evaluative judgments:
Influence of reduced processing capacity and mood salience. Manuscript under
review.
Silberman, E. K., Weingartner, H., & Post, R. M. (1983). Thinking disorder in
depression. Archives of General Psychiatry, 40, 775-780.
Simon, H. (1967). Motivational and emotional controls of cognition. Psychological
Review, 74, 29-39.
Sinclair, R. C. (1988). Mood, categorization breadth, and performance appraisal:
the effects of order of information acquisition and affective state on halo,
accuracy, information retrieval, and evaluations. Organizational Behavior and Human
Decision Processes, 42, 22-46.
Sinclair, R. C., & Marks, M. M. (1992). The influence of mood state on judgment and
action: Effects on persuasion, categorization, social justice, person perception,
and judgmental accuracy. In L. L. Martin & A. Tesser (Eds.), The construction of
social judgment (pp. 165-193). Hillsdale, NJ: Erlbaum.
Sinclair, R. C., Mark, M. M., & Clore, G. L. (1994). Mood-related persuasion
depends on misattributions. Social Cognition, 12, 309-326.
Singer, J. A., & Salovey, P. (1988). Mood and memory: Evaluating the network theory
of affect. Clinical Psychology Review, 8, 211-251.
Snyder, M., & White, P. (1982). Moods and memories: Elation, depression, and
remembering the events of one's life. Journal of Personality, 50, 139-167.
Spies, K., & Hesse, F. W. (1986). Interaktion von Emotion und Kognition.
(Interaction of emotion and cognition.) Psychologische Rundschau, 37, 75-90.
Srull, T. K. (1983). Affect and memory: The impact of affective reactions in
advertising on the representation of product information in memory. In R. Bagozzi &
A. Tybout (Eds.), Advances in consumer research (Vol. 10). Ann Arbor, MI:
Association for Consumer Research.
Srull, T. K. (1984). The effects of subjective affective states on memory and
judgment. In T. Kinnear (Ed.), Advances in consumer research (Vol. 11; pp. 530-
533). Provo, UT: Association for Consumer Research.
Srull, T. K., & Wyer, R. S., Jr. (1979). The role of category accessibility in the
interpretation of information about persons: Some determinants and implications.
Journal of Personality and Social Psychology, 37, 1660-1672.
Stein, B. S., & Bransford, J. D. (1979). Constraints on effective elaboration:
Effects of precision and subject generation. Journal of Verbal Learning and Verbal
Behavior, 18, 769-777.
Stepper, S., & Strack, F. (1993). Proprioceptive determinants of emotional and
nonemotional feelings. Journal of Personality and Social Psychology, 64, 211-220.
Strack, F. (1988). Social Cognition: Sozialpsychologie innerhalb des Paradigmas der
Informationsverarbeitung. [Social cognition: Social psychology in the paradigm of
information processing.] Psychologische Rundschau, 39, 72-82.
Strack, F. (1992). The different routes to social judgments: Experiential versus
informational strategies. In L. L. Martin & A. Tesser (Eds.), The construction of
social judgments (pp. 249-276). Hillsdale, NJ: Erlbaum.
Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating
conditions of the human smile: A non-obtrusive test of the facial feedback
hypothesis. Journal of Personality and Social Psychology, 53, 768-777.
Strack, F., Schwarz, N., & Gschneidinger, E. (1985). Happiness and reminiscing: The
role of time perspective, mood, and mode of thinking. Journal of Personality and
Social Psychology, 49, 1460-1469.
Strack, F., Schwarz, N., Bless, H., K�bler, A., & W�nke, M. (1993). Awareness of
the influence as a determinant of assimilation versus contrast. European Journal of
Social Psychology, 23, 53-62.
Strong, E. K. J. (1913). The effect of time-interval upon recognition memory.
Psychological Review, 20, 339-372.
Strongman, K. T. (1987). The psychology of emotion. New York: Wiley.
Sullivan, M. J. L., & Conway, M. (1989). Negative affect leads to low-effort
cognition: Attributional processing for observed social behavior. Social Cognition,
7, 315-337.
Sweeney, P. D., Anderson, K., & Bailey, S. (1986). Attributional style in
depression: A meta-analytic review. Journal of Personality and Social Psychology,
50, 974-991.
Taylor, S. E. (1982). The availability bias in social perception and interaction.
In D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty:
Heuristics and biases (pp. 190-200). Cambridge: Cambridge University Press.
Teasdale, J. D., & Taylor, R. (1981). Induced mood and accessibility of memories:
An effect of mood state or induction procedure? British Journal of Clinical
Psychology, 20, 39-48.
Teasdale, J. D., Taylor, R., & Fogarty, S. J. (1980). Effects of induced elation-
depression on the accessibility of memories of happy and unhappy experiences.
Behavior Research and Therapy, 18, 339-346.
Thondike, E. L. (1898). Animal intelligence. Psychological Review, 2,(whole No. 8).
Titchener, E. B. (1910). A text-book of psychology. New York: Macmillan.
Tomkins, S. S. (1962). Affect, imagery, and consciousness (Vol. 1). New York:
Springer-Verlag.
T�lle, R. (1982). Psychiatrie (6th edition). Heidelberg: Springer Verlag.
Tulving, E. (1983). Elements of episodic memory. Oxford, UK: Oxford University
Press.
Tulving, E. (1989). Memory: Performance, knowledge, and experience. European
Journal of Cognitive Psychology, 1, 3-26.
Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency
and probability. Cognitive Psychology, 5, 207-232.
Valins, S. (1967). Emotionality and information concerning internal reactions.
Journal of Personality and Social Psychology, 6, 458-463.
Valins, S. (1974). Persistent effects of information about intenal reactions:
Ineffectiveness of debriefing. In H. London & R. E. Nisbett (Eds.), Thought and
feeling: Cognitive alteration of feeling states (pp. 116-126). Chicago: Aldine.
W�nke, M., Schwarz, N., & Bless, H. (in press). The availability heuristic
revisited: Experienced ease of retrieval in mundane frequency estimates. Acta
Psychologica.
Watts, F. N., & Cooper, Z. (1989). The effects of depression on structural aspects
of the recall of prose. Journal of Abnormal Psychology, 98, 150-153.
Weary, G., Marsh, K. L., Gleicher, F., & Edwards, J. A. (1993). Social-cognitive
consequences of depression. In G. Weary, F. Gleicher, & K. L. Marsh (Eds.), Control
motivation and social cognition (pp. 255-287). New York: Springer Verlag.
Wegener, D.T., & Petty, R. E. (1994). Mood management across affective states: The
hedonic contingency hypothesis. Journal of Personality and Social Psychology, 66,
1034-1048.
Wegener, D. T., Petty, R. E., & Smith, S. M. (1994). Positive mood can increase or
decrease message scrutiny: The hedonic contingency view of mood and message
elaboration. Unpublished manuscript. Columbus, OH: Ohio State University.
Wegner, D.M. & Vallacher, R.R. (1986). Action identification. In R.M. Sorrentino &
E.T. Higgins (Eds.), Handbook of motiva~tion and cognition: Foundations of social
behavior. New York: Guilford.
Wellman, H. M. (1977). Tip of the tongue and feeling of knowing experiences: a
developmental study of memory monitoring. Child Development, 48, 13-21.
Wenzlaff, R. M., Wegner, D. M., & Roper, D. (1988). Depression and mental control:
The resurgence of unwanted negative thoughts. Journal of Personality and Social
Psychology, 55, 882-892.
Wetzler, S. (1985). Mood state-dependent retrieval: A failure to replicate.
Psychological Reports, 56, 759-765.
Winkielman, P., Zajonc, R. B., & Schwarz, N. (1995, May). Subliminal affective
priming is impervious to attributional manipulations. Midwestern Psychological
Association, Chicago, IL.
Woodworth, R. S. (1938). Experimental psychology. New York: Henry Holt.
Worth, L. T., & Mackie, D. M. (1987). Cognitive mediation of positive mood in
persuasion. Social Cognition, 5, 76-94.
Wyer, R. S., & Carlston, D. (1979). Social cognition, inference, and attribution.
Hillsdale, NJ: Erlbaum.
Wyer, R. S., & Srull, T. K. (1989). Memory and cognition in its social context.
Hillsdale, N.J.: Erlbaum.
Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality
and Social Psychology Monograph, 9(2), 1-27.
Zajonc, R. B. (1980). Feeling and thinking. Preferences need no inferences.
American Psychologist, 35, 151-175.
Zajonc, R.B., & Markus, H. (1984). Affect and cognition: The hard interface. In C.
Izard, J. Kagan, & R. B. Zajonc (Eds.), Emotions, cognition and behavior (pp. 73-
102). Cambridge: Cambridge University Press.
Zajonc, R.B., & Murphy, S. T., & Inglehart, M. (1989). Feeling and facial
efference: Implications of the vascular theory of emotion. Psychological Review,
96, 395-416.
Zanna, M. P., & Cooper, J. (1976). Dissonance and the attribution process. In J. H.
Harvey, W. J. Ickes, & R. F. Kidd (Eds.), New directions in attribution research
(Vol. 1, pp. 199-217). Hillsdale, NJ: Erlbaum.
Zillman, D. (1978). Attribution and misattribution of excitatory reactions. In J.
H. Harvey, W. I. Ickes, & R. F. Kidd (Eds.), New directions in attribu~tion
research (Vol. 2, pp. 335-368). Hillsdale, NJ: Erlbaum.
Zillman, D., Johnson, R. C., & Day, K. D. (1972). Attribution of apparent arousal
and proficiency of recovery from sympathetic activation affecting excitation
transfer to aggressive behavior. Journal of Experimental Social Psychology, 10,
503-515.