Académique Documents
Professionnel Documents
Culture Documents
2004
Dedicated to
Rashmi, Kathy, and Natalie
and
The Social Cognition Group
Contents
Preface
xi
INTRODUCTION
Types of Social Knowledge 6
The Representation of Knowledge in Memory 10
Situation Models 16
Generalized Event Representations 17
Generalized Entity Representations 22
2.
3.
26
47
vii
viii
CONTENTS
79
5.
106
6.
124
7.
151
CONTENTS
8.
ix
189
231
10.
267
CONTENTS
11.
306
12.
365
374
380
AUTHOR INDEX
SUBJECT INDEX
405
415
Preface
Authors often become disenchanted with the books they have written
almost as soon as they are published. Perhaps this is inevitable. Although
a book reports the culmination of ones theoretical and empirical work at
the time it goes to press, it nevertheless contains the seeds of discontent.
That is, the author in writing the book becomes painfully aware of the
limitations of his research and the conceptualization underlying it, and
consequently is already motivated to move beyond the confines of this
conceptualization by the time it appears in print.
At least, this has been true of me. A book I published in 1974, entitled
Cognitive Organization and Change, provided a fairly rigorous analysis of
cognitive consistency theories of the way beliefs and attitudes were
organized in memory, individual and situational differences in the
cognitive responses to new information, and the way in which different
features of information were combined to form judgments. The book
assumed that humans were analogous to computers, generating outputs
based on operations that were specified in various software routines
that were activated by the user (in the case of humans, the demands of
the social environment).
By the time the book was published, however, I had already begun to
appreciate the need for a more general conceptualization of mental
representation. A second book (Social Inference and Attribution, with Don
Carlston, published in 1979) proposed a rather complex associative
network model of social memory that took into account the role of
knowledge accessibility in judgments and decisions. In this book, we
recognized the impact that script-like representations of events could have
on social judgment. However, we did not explore this possibility with any
degree of rigor. More generally, the cumbersome nature of the network
representation, coupled with the fact that it provided little insight into
how the information represented was actually used, convinced me of the
futility of using this approach to account for many phenomena that were
being identified in research on person impression formation and the use
of judgmental heuristics.
xi
xii
PREFACE
The third book (Memory and Cognition in Its Social Context, with Thom
Srull) appeared in 1989. This book remedied several deficiencies of the
1979 volume, providing a comprehensive description of social information
processing that specified the activities at several different stages: the
interpretation, organization, and storage of information in memory, its
later retrieval and integration with other relevant knowledge to make a
subjective judgment or decision, and the translation of this subjective
response into an overt one. This rather grandiose formulation not only
generated specific predictions that received empirical support but had the
flexibility to integrate the implications of more specific theories of research
that had been developed.
By the time the book was published, however, I had become aware of
several obvious limitations. For one thing, the title of the book was a
misnomer. The theory I had proposed purported to describe the
comprehension and use of information of the sort that one encounters in
daily life, In fact, however, it was applied almost exclusively to the
processing of linguistic information of the sort people receive in laboratory
situations but are unlikely to encounter elsewhere (e.g., lists of unrelated
behaviors of a fictitious person, verbal descriptions of routine sequences
of events of little intrinsic interest, etc.) In short, although the book
purported to be about cognition in its social context, the social context
was typically absent in much of the research to which the theory was
applied.
Stimulated by our dissatisfaction with the applicability of the theory
to phenomena outside the laboratory, my students and I began to examine
the processing of information of the sort people actually encounter in the
course of their daily lives. This research took many directions. For example,
we investigated the impressions people formed on the basis of a persons
opinions about social policies with which they agreed or disagreed rather
than abstract behaviors that had few implications for matters of personal
concern to the recipients. Moreover, we studied the impressions people
formed on the basis of information conveyed in conversations rather than
written lists of unrelated behaviors described out of context. Moreover,
we explored the way individuals respond to information that violated
social norms of communication and that was conveyed for the purpose of
eliciting amusement rather than to convey an impression of its referent or
to espouse a particular point of view. In doing so, we began to take more
seriously the fact that information conveyed in social situations was
frequently in the form of a narrative, or story, and was often accompanied
by visual images that were either based on pictures or direct observation
or generated by the recipients themselves. Finally, we explored the role of
affect in information processing.
PREFACE
xiii
xiv
PREFACE
PREFACE
xv
Candy Fong, who worked with me at the Hong Kong University of Science
and Technology, whose work is also represented.
I owe a special note of appreciation to Rashmi Adaval. Both as a
collaborator and as a wife, she has contributed in both tangible and
intangible ways both to this book in particular and to my personal and
professional life more generally. Her research on both the role of narratives
in memory and judgment, and on affect and information processing, is
prominent throughout this book. However, her personal support and
encouragement has been equally invaluable. It is extremely difficult for
anyone to be both a colleague and a spouse. Her willingness to play these
two often conflicting roles, her tolerance for my idiosyncratic work habits,
and her love and support throughout the past decade, has been
inexpressibly gratifying.
Finally, I want to acknowledge my deep appreciation to Larry Erlbaum
and his staff at Lawrence Erlbaum Associates. The support he has given
me over the past 30 years has been a constant source of gratitude. Larrys
commitment to excellence, and his sincere dedication to the enhancement
of basic research in psychology, have won him the respect and admiration
of all psychologists who have had the opportunity to work with him. I
can only hope that this volume confirms his confidence in me and
approximates the standards of excellence for which he strives.
Robert S. Wyer, Jr.
P A R T
I
INTRODUCTION AND
BASIC CONCEPTS
C H A P T E R
1
Introduction
Schank and Abelson (1995) argued that virtually all the meaningful social
knowledge we acquire is in the form of stories. This claim is probably overstated (Brewer, 1995; Rubin, 1995). Nevertheless, the role of narratives in
the acquisition and transmission of information in daily life is undeniable.
The content of everyday conversations almost invariably includes stories
we tell about ourselves and personal acquaintances, descriptions of books
we have read or movies we have seen, and jokes that we tell to one another. Narratives are also invoked in the course of understanding the
causes or likely consequences of real and hypothetical social events (e.g.,
the outcome of a presidential election, or the IndiaPakistan dispute over
Kashmir). Persuasive messages and television commercials often stimulate
us to imagine the sequence of events that might result from taking a particular course of action. Life itself is in the form of a narrative, consisting of a
sequence of temporally related events that we experience as either participants or observers. In short, narratives are fundamental to an understanding of ourselves and of the world in which we live.
The importance of narrative forms of knowledge is recognized in virtually every area of psychology. Research and theory on prose comprehen3
CHAPTER 1
sion and learning have long been concerned with the comprehension and
mental representation of stories and temporally ordered sequences of
events (Graesser, 1981; Mandler & Johnson, 1977; Stein & Glenn, 1979). Developmental psychologists (e.g., Miller, 1994; Nelson, 1993) have identified
the role of stories in parentchild interaction and socialization more generally. Narrative forms of knowledge can often constitute implicit theories
that people use both to explain their own and others past experiences and
to predict the future (Dweck & Leggett, 1988; Ross, 1989). The influence of
these theories on behavior and judgments has been identified in research
on topics as diverse as personal achievement (Dweck, 1975), predictions of
future life events (Oishi, Wyer, & Colcombe, 2000), marital relations
(Holmes & Murray, 1995; Murray & Holmes, 1996), personality development
(McAdams, 1988) and communityindividual relations (Mankowski &
Rappaport, 1995).
Despite the consideration of narratives in these specific areas of research, few attempts have been made to develop a broader theoretical formulation of social information processing that permits their influence to be
conceptualized in relation to that of other types of knowledge that people
encounter in their daily lives. Several types of knowledge representations
are likely to exist (Pennington & Hastie, 1993), and their role in information
processing can differ. One must understand the factors that govern the retrieval of this knowledge, how it is used to comprehend and construe the
implications of new information to which it is relevant, and the nature of its
influence on judgments and behavioral decisions about its referents.
This book attempts to provide this understanding. In particular, it is concerned with the comprehension and use of information of the sort people
encounter in the course of their daily lives. This information can be obtained through direct experience, from movies and television, from reading
newspapers, or from conversations with friends or strangers. The information can be transmitted in writing but also visually or acoustically. Moreover, it can be conveyed in several modalities simultaneously. Finally, although the information can sometimes concern hypothetical people and
situations, it more often refers to actual persons and events about which recipients already have substantial knowledge. Finally, the information is often conveyed in a social context, and can elicit affective or emotional reactions toward either its referent or, in some cases, the communicator. Much
of this information, particularly that which is acquired through direct experience, is necessarily transmitted in the form of a narrative, that is, a temporally related sequence of events.
The conceptualization we propose addresses the way this information is
comprehended and represented in memory, and how it is later recalled and
used in making judgments and decisions. Our focus on real-world information processing does not imply that the research we bring to bear on it
INTRODUCTION
CHAPTER 1
INTRODUCTION
tines involved in driving a car (e.g., putting in the clutch, turning on the ignition, putting the car in gear, gradually releasing the clutch while stepping
on the gas, etc.) initially require conscious thought. However, they ultimately come to be performed without consultation of ones declarative
knowledge about the sequence of steps involved, and require few if any
cognitive resources (Schneider & Shiffrin, 1977). As Bargh (1997) asserted, a
very large amount of our social behavior is likely to involve the use of these
automatically activated productions.
Procedural knowledge plays a very important role in social information
processing, as I indicate frequently throughout this volume. For the present, however, I will concentrate my discussion on the content and structure
of declarative knowledge. I first consider the referents of this knowledge,
and the modality in which it is acquired. I then turn briefly to the different
ways in which this knowledge is represented in memory, and discuss the
need to understand these matters in conceptualizing the processing of information of the sort we receive in the course of daily life.
Referents of Social Knowledge
The stimulus information that impinges on us in a given social situation can
often far exceed our capacity to assimilate its implications. In a conversation, for example, several persons typically communicate to one another
both verbally and nonverbally, and in doing so, are likely to convey information about not only the topic under discussion but also about themselves. Moreover, this information is often transmitted in the context of
many other situational stimuli that are largely irrelevant to either the topic
at hand or the individuals who are discussing it. We are usually not passive
recipients of this information. Rather, we actively try to determine its meaning. Furthermore, we might construe its implications for the attainment of a
particular goal that we have in mind. In doing so, however, we necessarily
pay more attention to some features than to others. The features that receive the greatest attention depend in part on our perception of the referent (i.e., what the information is supposed to be about).
The referent of a piece of information can be either an entity (a person or
object, a place, oneself, etc.) or an event. However, most events involve persons and objects and occur in a specific (although perhaps unspecified) location. Furthermore, ones knowledge about an entity often concerns
events in which the entity has been involved. As a consequence, the referent of information is often subjective, depending on not only the context in
which the information is received but also the interests and objectives of
the recipient. For example, consider the statement, Muhammad Ali
knocked out George Foreman in Zaire. This statement could be about Ali,
Foreman, or the event itself. In some cases, of course, perceptions of the
CHAPTER 1
referent of information are determined by the context in which the information is received. Thus, someone who hears the aforementioned statement
during a discussion of Muhammad Ali is likely to identify Ali as its referent,
whereas someone who reads the statement in a magazine article about boxing might consider the event itself to be the referent. In addition, however,
individual differences in prior knowledge and motivation (e.g., whether one
is a fan of Muhammad Ali or George Foreman) can also influence perceptions of a statements referent. These situational and individual differences
in perception of the referent can have an impact on how the knowledge is
represented in memory and the likelihood that the knowledge comes to
mind later. For example, a person who learns that Ali knocked out Foreman
in a conversation about Ali, and consequently perceives Ali to be the referent, may not recall the information later when information about George
Foreman is being discussed. (For further discussion of this possibility, see
Wyer & Srull, 1989.)
Furthermore, when the information conveyed in a particular situation is
detailed and complex, recipients are likely to think more extensively about
features that pertain to its referent (as they perceive it) than other features.
Thus, the referent-related features are likely to be retained, whereas more
peripheral features seem to be forgotten. For example, a man who learns
that V. S. Naipaul won the Nobel Prize for Literature might remember the
specific context in which he acquired the information a short time after it
was conveyed. Over time, however, the information becomes part of his
general knowledge and the conditions that led to its acquisition are not
recalled.1
Affect and Emotion. Although the referents of knowledge are usually
external to the organism, internally generated stimulation can also be a
source of information. (See Strack and Deutsch, 2002, for a theoretical analysis of these alternative sources of information.) For example, external
stimuli can elicit affective or emotional reactions, and these reactions can
be used as bases for both thoughts and overt responses to the stimuli
(Schwarz & Clore, 1983, 1996; Wyer & Carlston, 1979). This does not mean
1
The dissociation of knowledge from the context in which it is learned does not mean that
contextual features are erased from memory. In the previous example, I learned about Naipauls
award from my wife, who stopped into my office to tell me about it one day at work, and I wrote
the preceding paragraph shortly afterwards. I would normally not be able to recall where and
when I acquired knowledge of the award. However, being reminded of the example while going
over this manuscript several months later was sufficient to stimulate a relatively detailed mental image of the event, including my own and my wifes physical locations in the room. Thus, although these contextual features became lost in the interim, they continued to exist in memory and could be retrieved by a subset of cues that identified the event and its use as an example
in this book. The theory to be proposed in this volume attempts to account for this and related
memory phenomena.
INTRODUCTION
that affective reactions per se are part of stored knowledge. Rather, they
are conditioned or unconditioned responses to cognitions about their referent, and can sometimes be the referents of cognitions, but they are not
themselves part of the cognitive system (Wyer, Clore, & Isbell, 1999). This
view, which contrasts with that proposed by Wyer and Srull (1989; see also
Bower, 1981), is discussed more fully in chapter 11.
Relational Knowledge. Much of the knowledge we require about entities and events is relational. These relations may be physical or conceptual.
For example, we know not only that a nose and eyes are parts of a face but
also their positions in relation to one another. Furthermore, we know that
our bedroom is down the hall from the living room, that Chicago is west of
New York, that lions are larger than rabbits, and that the earth goes around
the sun. We also know that people graduate from high school before going
to college, that smoking is a cause of lung cancer, and that people have to
order a meal in a restaurant before they eat it. This knowledge requires not
only understanding of the entities being related but the relations themselves (larger than, causes, above or below, etc.). Note that in many
instances, the mental representation of these relations requires visual imagery.
Modalities of Knowledge
Information is conveyed in many different modalities and formats. For example, it can be transmitted verbally in the form of propositions. It can also
be conveyed visually as in a picture or movie, or acoustically, as in music.
Our knowledge can likewise be in different modalities. That is, our knowledge of a person could consist in part of verbal descriptions of the persons
traits, and our knowledge of restaurants could include abstract characterizations of the events that usually take place there. However, we also have
mental images of people we meet and places we visit, and might have
acoustic representations of a persons voice, or of the Beatles rendition of
Hey Jude. Social experiences are often a configuration of verbal, visual,
and acoustic stimulation that combines to form our knowledge of what occurred.
There has been some controversy about whether information is actually
coded into memory in the modality in which it is conveyed or is represented in a common format that can be decoded into different modalities,
depending on situational demands (J. Anderson, 1978; Kosslyn, 1980; Pylyshyn, 1973). Certainly we can easily translate many visual stimuli into linguistic terms, and can form visual images of people and events on the basis
of verbal descriptions of them. Images can be acoustic as well as visual.
(This is apparent from the fact that people who have seen a foreign movie
10
CHAPTER 1
General Considerations
As the preceding discussion testifies, it is virtually impossible to discuss
the type of knowledge we acquire without making implicit assumptions
about the representation of this knowledge in memory. However, the mental representation of information can differ in several ways from the information itself. Some features of the information we receive are not stored in
memory at all, whereas other features that were not specified in the information may be added (Barclay, 1973; Loftus, 2000). Furthermore, detailed
descriptions of an event can often be interpreted in terms of abstract (e.g.,
attribute) concepts that are stored in memory independently of these descriptions (Carlston, 1980; Higgins, 1996; Wyer & Srull, 1989). In addition, information might be mentally reorganized into a form that is relevant to the
goal one happens to be pursuing (cf. Hamilton, Katz, & Leirer, 1980; Wyer &
Bodenhausen, 1985). Moreover, information that is transmitted visually may
be coded into memory linguistically, and verbal information may sometimes elicit visual images.
It might nevertheless be argued that people extract similar implications
from information regardless of the mental representation they form of it. If
this were so, an understanding of the mental representations that are
formed of information would be of little importance in predicting its impact.
We would simply need to characterize clearly the stimulus information to
which people are exposed, and to obtain a reliable indication of which as-
INTRODUCTION
11
pects of this information they retain (as evidenced by the aspects of the information they can recall). This would be sufficient to determine its impact.
In fact, however, research in social and cognitive psychology since the
1970s suggests this contention is nave. That is, a conceptualization of the
cognitive operations that are performed on information, and of the mental
representations that result from these operations, is essential for understanding its influence. Because the comprehension and mental representation of experiential knowledge is a central focus of this volume, some empirical support for this conclusion may be worth summarizing.
1. Communication and Persuasion
One of the first demonstrations in social psychology of the need to take
into account the cognitive activities that people perform in response to information was constructed by Greenwald (1968). He showed that the influence of a persuasive message was often unrelated to the implications of
those aspects of the message that participants could recall. Rather, its impact was governed by the implications of the thoughts they had about the
message at the time it was received. These cognitive responses, which included elaborations of the message content, counterarguments, and
thoughts that were unrelated to the message itself, presumably reflected recipients comprehension of the message and their evaluation of its validity
as well as the implications of other previously acquired knowledge that
they brought to bear on the position advocated.
In a study by Loken (1984), participants read a list of information items
pertaining to a college professor with instructions either to form an impression of the individual or to decide whether they wanted to take a course
from him. The information consisted of either favorable or unfavorable trait
descriptions, and either favorable or unfavorable consequences of taking
the course. In some cases, individual pieces of information had implications
for both attitudes toward taking the course and attitudes toward the professor (e.g., Dolan is interesting, Taking Dolans course is time consuming,
etc.). In other cases, they had implications for only one of these attitudes
(e.g., Dolan is clumsy, Taking Dolans class fulfills a graduation requirement, etc.). Later, participants reported their attitudes toward both the
person and the behavior, and then recalled the information they had received. The type of information that participants recalled was influenced by
the task objectives they were given. Nevertheless, the effects of experimental variables on attitude judgments did not mirror their effects on recall.
Moreover, correlations between attitudes and the evaluative implications
of attitude-relevant information that participants recalled were very low,
ranging from .05 to .24. Loken concluded that participants spontaneously
formed attitudes toward both the professor and taking the course at the
12
CHAPTER 1
time the information was received, and later reported these attitudes independently of the information they could remember at the time of judgment.
2. Person Impression Formation
In a study by Carlston (1980), participants read descriptions of a target
persons behavior that exemplified two traits with evaluatively different implications. In one condition, for example, behaviors (e.g., told his girlfriend
that her hair-do was ugly) could be interpreted as both honest and unkind.
Participants judged the target with respect to one of the two traits and
then, either a few minutes or several days later, evaluated him with respect
to the other trait. Participants second judgments were evaluatively biased
by the first. In other words, participants who initially evaluated the targets
honesty (and thus judged him to be honest) later reported him as being
more kind than participants who had not made this initial judgment. Correspondingly, participants who made kindness judgments at the outset (and
thus rated the target as unkind) judged him to be less honest than other
participants did. Moreover, these effects increased over time. Participants
apparently interpreted the information in a way that was relevant to the
first judgment they made and formed a mental representation of the target
as the type of person who had the trait being judged. Then, having done so,
they based their second judgment on the implications of this representation rather than reconsidering the original information they had received.
In a series of studies by Srull and Wyer (1983), some participants read descriptions of a persons behavior that were either predominantly favorable
or predominantly unfavorable with instructions to form an impression of the
actor. Then, after either a short or long delay (5 min or 45 min), they were
asked to report their liking for him. Other participants, who were asked to recall the behaviors they had learned, remembered fewer behaviors after a
long delay than after a short delay. However, participants rated the target
more extremely in the long delay condition (e.g., as more likable or more
dislikable, depending on the predominant implications of the behaviors). Apparently, these participants formed a generally favorable or generally unfavorable impression of the target on the basis of the predominant theme of
the behaviors they had read. After a delay, their recall of the specific behaviors diminished, but their overall impression remained. Therefore, their judgment, based on this impression alone, was more polarized.
3. Assimilation and Contrast
In a study by Higgins and Lurie (1983), subjects rated the harshness of
sentence imposed on a criminal by a target judge (Jones) along with the
sentences recommended by three context judges. Participants rated
Joness sentence as more lenient when sentences recommended by the
INTRODUCTION
13
context judges were long than when they were short. The critical conditions of the study, however, occurred a week later. At this time, participants
were given a sample of the sentences imposed by a different group of context judges that were again either long or short. Then, after receiving the information, they were asked to recall the actual sentence that Jones had imposed, as conveyed in the first session. Instead of recalling this sentence
directly, however, participants apparently recalled their rating of him as
harsh or lenient, and then reinterpreted the implications of this rating with
reference to the sentence imposed by the context judges in Session 2.
Therefore, they recalled the targets sentence to be longer when the sentences imposed by the second set of judges were long than when they were
short, and this was true regardless of the sentences imposed by the context
judges in Session 1.
4. Impact of Information That One
Is Told to Disregard
In a study by Wyer and Budesheim (1987; see also Wyer & Unverzagt,
1985), participants read a series of behaviors of a target person with instructions to form an impression of him. In some cases, the first behaviors
were favorable and the last behaviors were unfavorable. In other conditions, the order was reversed. On the pretense that an error had been
made, some subjects were told after reading the first set of behaviors that
these behaviors actually pertained to a different person than the one they
were supposed to consider and to disregard them. Other participants read
the entire set of behaviors and then were told to disregard the last set they
had considered. Finally, participants were asked both to judge the target
with respect to the traits to which the information pertained and to recall
the information they received. When participants had been told to disregard the first behaviors in the series, their judgments were not influenced
by these behaviors although they could recall the behaviors quite well.
When participants had been told to disregard the last behaviors in the series, these behaviors had an impact on their judgments although their recall of the behaviors was poor. Thus, instructions to disregard the information had opposite effects on recall and judgments in the two conditions.
These studies suggest that the mental representations that participants
form of information can exert a major influence on their judgments independently of the information itself, and that the magnitude of this influence
can increase over time. Greenwald and Banaji (1995) described several additional situations in which people are influenced by information they cannot recall and, in some cases, are unaware that they ever received (see also
Bargh, 1997). Thus, neither the implications of the information that people
receive nor the implications of the information they can remember provide
14
CHAPTER 1
15
FIG. 1.1. Relations among various components of social judgment. Observable (independent
and dependent) variables are enclosed by solid lines, and mediating variables by dashed lines.
Rectangles denote states and ovals denote processes.
16
CHAPTER 1
ume more specifically. That is, the mental representation that is formed
from information is a hypothetical construct and cannot be observed directly. However, given a specific set of assumptions about the retrieval operations that occur when people are asked to recall the information, and a
metaphor for describing the associations among features of the information
that exist in memory, one can use the amount, type, and order of information to conceptualize the content and structure of the mental representations that was formed at the time the information was received. Then, once
this is established, one can hypothesize and investigate empirically the way
this mental representation is used to attain specific objectives to which it is
relevant. Much of the research and theory summarized in this volume is
based, implicitly or explicitly, on this general approach.
The social knowledge we acquire can be represented in memory in several ways. The nature of these representations, when they are constructed,
and how they are used, are a major focus of this book. Two types of mental
representations that occupy our greatest attention are worth noting briefly.
These representations, to be denoted situation models and generalized representations, are distinguished in terms of both their content and structure
and the point at which they typically come into play in social information
processing. Although we consider both types of representations in detail in
subsequent chapters, a general overview of the representations is perhaps
desirable.
SITUATION MODELS
We assume that people who receive information about situation-specific
events, either verbally or through direct experience, form mental simulations of these events in the course of comprehending them that indicate not
only the people and objects that are involved in the events but their spatial
and temporal relatedness. These mental simulations, or situation models
(Johnson-Laird, 1983, 1989; Kintsch, 1998) are often represented in memory
in terms of mental images of the events and, in some cases, propositionally
as well. To the latter extent, the representations are somewhat analogous
to a picture plus caption. Moreover, a temporally and thematically related
sequence of events might be represented by a multiple-segment representation somewhat analogous to a comic strip. (For a similar metaphor in
conceptualizing the mental representation of event sequences, see
Abelson, 1976.)
We assume that situation models are often formed spontaneously in the
course of comprehending information (see chap. 4 for a discussion of these
processes). This information could be either obtained through direct experience or conveyed verbally. However, a necessary condition for the con-
INTRODUCTION
17
struction of these models is that the events or states to which they refer are
temporally and spatially specific. That is, the phenomena must occur at a
specific (although not necessarily specified) place and point in time. This
constraint is implicit in the assumption that a mental image is formed.
Thus, the statements, The boy kicked the ball and The book is on the table describe events or states of affairs that existed at a particular time and
place, and a mental picture can be formed of them. In contrast, the boy
owns a ball, or the book is interesting, are not spatially or temporally
constrained, and a mental picture of the situations is not constructed
(Radvansky, Wyer, Curiel, & Lutz, 1997). Rather, a generalized representation is formed on the basis of these statements that does not have an image
component.
18
CHAPTER 1
INTRODUCTION
19
20
CHAPTER 1
a point or exemplifying a principle in another context. It would be conceptually confusing to define the same narrative differently, depending on the
conditions in which it is conveyed.
It therefore seems more reasonable to define a story as simply a narrative that (a) describes events involving specific individuals or groups that
are localized in time and place, (b) contains at least one event that deviates
from expectations based on prior knowledge of the type of individuals and
situations involved, and (c) whose end state has emotional significance for
one or more of the protagonists.
3. Histories
A history, like a story, consists of a sequence of events that are specific
to a particular person or group and are localized in time and place. Unlike
stories, however, a history does not necessarily contain unique or unexpected events. Nor does it necessarily have emotional significance for the
protagonists. Histories can concern real-world events in which one is not involved. However, it can also refer to oneself or an acquaintance. The mental
representation of a personal history is clearly not a single sequence of
events that occurs over the course of a lifetime. Rather, an individual is
likely to construct several histories, each of which is restricted to life
events that are relevant to the purpose for which it is formed (communicating about oneself to another, comprehending ones own or anothers personal experience, etc.). Once constructed, these sets of segments are likely
to be stored independently of one another in memory.
As these observations imply, histories can contain overlapping segments. Moreover, a history does not usually cover all the events that occur
during the time period to which it pertains. Thus, one might construct different histories that cover the same or overlapping time period. For example, a personal history of how one became a psychologist could temporally
overlap a history of ones romance with an ex-call girl. However, the events
that compose these mental constructions might be totally different.
According to this conceptualization, stories and histories are not mutually exclusive constructs. A history would be a story to the extent it contains unexpected events and is emotionally significant for one or more of
the parties involved. However, not all histories are stories, and not all stories are histories.
4. Implicit Theories and Implicational Molecules
A theory, like a script, is composed of a prototypic sequence of events.
However, the events are causally as well as temporally related. Moreover,
the events are often defined more broadly. A theory is not localized in time
and, therefore, can be applied in many different circumstances. It might
INTRODUCTION
21
nevertheless concern a particular individual (e.g., oneself) or general category of individuals (African Americans, fraternity members, etc.) as well as
people in general. Similarly, it could be specific to a particular type of situation (living in public housing, studentfaculty interactions) or could apply
more generally. In all cases, however, the events that compose a theory are
likely to recur at different times.
For example, one might have an implicit theory that (a) people with alcoholic fathers feel insecure, (b) these feelings lead them to pursue unrealistically high goals, (c) their failure to attain these goals leads to depression
and ultimately to drinking, and (d) they wind up becoming alcoholic like
their fathers. Each segment encompasses a large number of more specific
events and functions as a general concept that can be used to interpret and
organize new information. To this extent, a theory can function as an event
stereotype.
Theories play a very important role in processing new information about
the world in which we live and making inferences on the basis of it. They
are used both to explain events that have occurred in the past and to predict the future. Moreover, they can be used to fill in gaps in a sequence of
events about which one has incomplete information.
In this regard, some theories can consist of only two or three segments
that, in combination, constitute a general principle. In these instances, the
principle could function as a linguistically coded title, or header of the
theory to which it refers. For example, the principle people get what they
deserve could be the title of a two-segment theory composed of the events
A person does something bad and The person gets punished. These theories can be conceptualized as implicational molecules of the sort suggested
by Abelson and Reich (1969; see also Wyer & Carlston, 1979, 1994). That is,
they consist of a set of general propositions that have become associated
through learning. For example, the propositions associated with the principle that people get what they deserve could compose the molecule
[P does something good (bad); Something good (bad) will occur to P].
The use of an implicational molecule is theoretically governed by a completion principle. That is, if the information one receives about a specific situation instantiates all but one of the propositions that compose a molecule,
an instantiation of the remaining proposition will be inferred. Thus, people
who have formed the implicational molecule described in the preceding example might predict that a particular person who has done a bad thing will
be punished. However, they might also infer that someone who has experienced misfortune has done a bad thing or for other reasons deserves his or
her fate. Several other examples of the use of implicational molecules, and
implicit theories more generally, are discussed in detail in chapter 10 of this
volume.
22
CHAPTER 1
INTRODUCTION
23
FIG. 1.2. (a) Metaphorical trait-based representations formed as a result of interpreting behaviors in terms of trait concepts. (b) Metaphorical representation of a person, John, described by behaviors and traits. Thicker lines denote stronger associations.
The representations of the sort shown in Fig. 1.2 are theoretically the result of conscious cognitive activity (but see Chartrand & Bargh, 1996, 2002,
for a different conclusion). Moreover, the associations shown in the figure
reflect the implications of specific assumptions about the nature of this activity. Given assumptions about the processes that underlie the extraction
of specific features from the network, these implications can be evaluated
24
CHAPTER 1
INTRODUCTION
25
C H A P T E R
2
The Architecture of the
Information-Processing System
27
tivity is initiated. Wyer and Srull simply assumed that input information is
spontaneously encoded in terms of low level semantic concepts that are
applied independently of the modality in which the information is received.
They provided no indication of how this is done. Furthermore, their assumption that all incoming information is encoded semantically constrains
the models ability to address the comprehension of information that pertains to people and events about which one already has substantial knowledge. The model assumed that referent-specific knowledge does not come
into play until a later, goal-directed stage of processing.
One implication of this assumption is that the validity of information
about well-known people and events is not assessed until some time after
the information is comprehended, and may not occur at all unless people
have a specific goal to which this assessment is relevant. This seems intuitively unlikely. Most Americans would almost immediately recognize that
the assertion Abraham Lincoln was President of the United States is true
in the course of comprehending it. By the same token, they are equally
quick to recognize that Mao Tse Tung was President of the United States
is false. These validity assessments appear to occur spontaneously in the
absence of any specific goal that requires them.
Not all information is spontaneously verified at the time it is comprehended, however. The statement Abraham Lincoln visited Boston is likely
to be comprehended without any conscious recognition that it is either true
or false. Statements that refer to unknown persons may be similarly comprehended without spontaneously validating them. The conditions in which
spontaneous validity assessments do and do not occur in the course of
comprehending information is not specified by either the Wyer and Srull
(1989) formulation or others.
Information is likely to be processed quite differently when it is comprehended with reference to previously acquired knowledge about its referents than when it is not. For example, suppose people receive information
that a person is kind, followed by information that he both (a) bought dinner for a street person who did not have enough money to eat and (b) berated his secretary for coming 10 minutes late to work. Recipients are likely
to think more extensively about the second behavior, which is ostensibly
inconsistent with the trait description than about the first one, leading this
behavior to be better remembered later (Hastie & Kumar, 1979; Srull, 1981;
Wyer & Gordon, 1982). However, suppose people receive information that a
conservative Republican has both organized a sit-in at a nuclear power
plant and picketed an abortion clinic. In this case, recipients are less likely
to think about the behaviors in terms of their consistency with the general
characteristics of the actor. Rather, they think about them with reference to
their knowledge about the social implications of the behavior and their perception that these implications are desirable. Consequently, they have
28
CHAPTER 2
better recall of the behaviors that reflect policies they consider to be incorrect or undesirable (Wnke & Wyer, 1996; Wyer, Lambert, Budesheim, &
Gruenfeld, 1992). In short, information that has little relevance to ones
prior knowledge about its referents is likely to be thought about in terms of
the semantic consistency of its implications. However, information about
persons and situations about which one has prior knowledge is usually
thought about with reference to this knowledge, and so its semantic consistency has less impact.
The aforementioned studies provide only few examples of the role of referent-specific knowledge in comprehending and validating new information. The effects of spontaneously recognizing information as true or false
in the course of comprehending it, and the reactions to the implications of
these assessments, are central to an understanding of humor elicitation
(Wyer & Collins, 1992), emotional communication (Gaelick, Bodenhausen, &
Wyer, 1985; Scott, Fuhrman, & Wyer, 1991) and persuasion (Gruenfeld &
Wyer, 1992; Wyer & Gruenfeld, 1995). This becomes clear in later chapters
of this volume.
29
In this regard, the modality in which information is presented is not necessarily the modality in which it is reported. Visual information can sometimes be encoded into memory linguistically. Moreover, linguistically coded
information can elicit mental images (Bransford, Barclay, & Franks, 1972;
Garnham, 1981; Radvansky, Gerard, Zacks, & Hasher, 1990; Reyes, Thompson, & Bower, 1980; Wyer, Adaval, & Colcombe, 2002). These processes,
however, are not articulated in the theory proposed by Wyer and Srull
(1989).
Other, more specific problems with the original theory will be noted
presently in the context in which they become relevant. The conceptualization we propose in this and later chapters is intended to address the earlier
deficiencies while at the same time preserving the general architecture of
the information-processing system of the original model and, therefore,
many of its implications. In the next section, we outline the general structure of the proposed system, noting the assumptions that differ from those
of the original model. We then summarize the assumptions we now make
concerning information storage and retrieval, which differ significantly
from those of the original formulation.
30
FIG. 2.1. Schematic representation of the information processing system. Circles and ovals refer to processing units. Rectangles denote storage units. Arrows indicate the direction of transmission of cognitive material between these units.
31
can simultaneously impinge on the processing system and can serve as input to the Comprehender (see Fig. 2.1). However, its importance for the issues of concern here is limited, and the unit will generally not be discussed
further.
1. The Work Space
The structure and function of the Work Space is similar to that of the
original model. This unit functions as a temporary store of all information
that has been encoded by the Comprehender along with other material
that is or has recently been involved in higher level goal-directed processing. As shown in Fig. 2.1, input to the Work Space can include material that
has been interpreted and transmitted to it by the Comprehender, concepts and knowledge representations that have been retrieved from Permanent Storage for use in attaining specific processing objectives, and the
results of this processing. The capacity of the Work Space is large, and as
long as its contents are relevant to the attainment of a processing objective that is active, it can be retained indefinitely. However, if a processing
goal has been attained, or if information has not been used for some time,
the Work Space may be cleared to permit easier access to material that is
relevant to other, more immediate objectives. This may be done either automatically, as new material is deposited, or volitionally (for evidence, see
Srull & Wyer, 1983). Phenomenally, this simply means that people can intentionally stop thinking about material of relevance to one objective and
concentrate on a second.2
These considerations become important in the context of an additional
assumption of both the Wyer and Srull model and the present one. That is,
only the output of higher order goal-directed processing is transmitted to
Permanent Storage. Therefore, once the Work Space is cleared, any input
information from the Work Space that has not been involved in goaldirected cognitive activity is irretrievably lost.
2. Permanent Storage
A major modification of the present model in relation to that proposed
by Wyer and Srull surrounds the conceptualization of long-term memory or,
in other words, Permanent Storage. We first describe the assumptions of the
original model, and then discuss the modifications to be made.
2
On the surface, this implication appears inconsistent with the result of research on thought
suppression (Bodenhausen & Macrae, 1998; Wegner, 1994), which suggests that conscious attempts not to think about a concept increases its accessibility in memory. However, this inconsistency is illusory, as we point out later in this volume.
32
CHAPTER 2
33
34
CHAPTER 2
The other three processing units come into play in the course of specific
goal-directed activity, and are activated on the basis of instructions in a
goal schema that is used by the Executor to monitor this activity. The characteristics of these units are not appreciably different from those described
by Wyer and Srull (1989). The Encoder/Organizer interprets stimulus information transmitted to it from the Comprehender in terms of more abstract
concepts that are relevant to the goal being pursued, or organizes the information with reference to a pre-existing configuration of features that are
specified in the information. The Inference Maker generates specific subjective inferences on the basis of new or pre-existing information using logical,
algebraic or other combinatorial rules, or, in some cases, heuristics. The
Response Selector then translates subjective inferences into an overt response, which could be a behavioral decision, a judgment along a category
response scale, or a verbal utterance.
Each processing unit is equipped with a library of routines (conceptually
equivalent to cognitive productions, as mentioned in chap. 1) that it draws
on to perform its functions. The nature of these productions can depend on
both the goal being pursued and the type of information being processed.
Thus, for example, the procedures performed by the Inference Maker depend on whether the goal is to form a general impression or to make a comparative judgment. On the other hand, they can also depend on whether
the information is conveyed visually or verbally, or whether it is conveyed
in a list or a narrative (cf. Adaval & Wyer, 1998).
The Executor
The Executor monitors the flow of cognitive material between memory storage units and special purpose processing units. It has no mind of its own,
but instead, takes instructions from a pre-existing goal schema that is active in the Goal Specification Box. These schemas presumably provide stepby-step descriptions of the sequence of steps that must be performed in order to attain the objective at hand.
For example, suppose people receive descriptions of a person for the
purpose of forming an impression. The goal schema that guides the attainment of this objective might indicate that the individual behaviors should
be interpreted in terms of trait concepts, that the evaluative implications of
these traits should be assessed, and the implications of these assessments
should then be combined to form an inference of the persons general likeableness, and that this evaluation should be communicated to another verbally. The Executor would presumably recognize the first two (trait and
evaluative) encoding tasks to be the responsibility of the Encoder/Organizer and would direct the information to this unit with instructions to perform the encoding. Once the results of the evaluative encoding are avail-
35
able, the Executor would transmit these encodings to the Inference Maker
with instructions to combine their implications into an overall inference.
Then, once this inference is computed, the Executor would transmit the inference to the Response Selector with a request to generate an overt response in a language that was applicable (a verbal utterance or, if the inference is to be reported in an experiment, a numerical value along a response
scale).
Note that the routines contained in the libraries of the various processing units, as well as the goal schemata used by the Executor to govern the
flow of information in pursuit of specific processing objectives, are generally not specified by the conceptualization we propose. These routines and
procedures are the subject of research on specific types of cognitive activity of the sort described by Wyer and Srull (1989) as well as later chapters
of the present volume. However, many social psychological theories derived from this research can incorporated into the more general conceptualization we have proposed. For example, the model of person impression
formation and memory proposed by Srull and Wyer (1989; for earlier versions, see Hastie & Kumar, 1979; Srull, 1981), which specifies the cognitive
processes involved in forming an impression and the cognitive representations that are formed as a result of these processes, can be viewed as a
specification of an impression formation goal schema along with a specification of the organizational processes performed by the Encoder/Organizer
and stored in the Encoder/Organizers library. The algebraic integration
processes postulated by Norman Anderson (1971; 1981), several of the cognitive heuristics identified by Tversky and Kahneman (1974), and the attribution processes postulated by Bem (1972) and Jones and Davis (1965), can
be viewed as theories about the routines performed by the Inference Maker
and the conditions in which they are applied. The impact of contextual
stimuli on the use of category response scales (Lynch, Chakravarti, & Mitra,
1991; Ostrom & Upshaw, 1968; Parducci, 1965; Wyer, 1974) and opinion survey responses (Schwarz, 1994; Strack, 1994) are essentially theories about
the processes performed by the Response Selector. These and other theoretical formulations can in principle be incorporated into the general framework we propose.
Effects of Multiple Goal Schemata
Two more general aspects of the conceptualization are worth noting at this
point. First, more than one goal schema might be relevant to the attainment
of a given objective. An alternative to the sequence described in our example might involve a comparison of the configuration of trait encodings to
the pre-existing knowledge representation of a known person, or alternatively, a prototype, and, if there is sufficient match, an inference of the per-
36
CHAPTER 2
37
the woman was similar to the professor. We consider the results of automatic and deliberative processing later in this volume. For the time being, it
is sufficient to note that the theory can account in principle for differences
in conscious and nonconscious processing and can specify the conditions
in which each type of processing may operate.
Storage
Our assumptions about the storage of information can be stated simply in
terms of a postulate:
Postulate 2.1. Each knowledge representation that is formed, either in the
course of comprehension or in the course of higher level, goal-directed processing, constitutes a separate unit of knowledge that is stored independently
of other knowledge units.
38
CHAPTER 2
39
n(pI k )
,
n(k )
40
CHAPTER 2
ulus set. If the resonance reaches a threshold value, T (k), the representation
is activated.
One implication of Postulate 2.2 is worth noting at this point. That is, when
features of the probe set are contained in more than one representation,
the resonance at the representation with the fewest number of noncommon
features will reach threshold faster. Thus, the probe football game is
more likely to activate an abstract representation of football games than a
specific representation of the 2002 Super Bowl.
Postulate 2.3. (a) If two or more representations resonate with a given set of
probe cues, the one that reaches threshold most quickly is identified. (b) If
several representations reach threshold simultaneously, a composite of the
features is activated. The weight of each feature in the composite is a positive
function of the proportion of times it occurs in the set of representations involved.
41
In effect, Postulate 2.4 permits the theory to account for the effect of recency of activation on concept and knowledge accessibility that is consistently observed in research on priming effects (Higgins, Bargh, & Lombardi,
1985; Srull & Wyer, 1979). Postulate 2.5 permits the theory to account for the
effect of frequency of activation that is also observed in these studies. One
implication of these postulates is worth noting. Specifically, the effect of recency of activation, which is a function of the residual resonance that exists
at a memory location after the knowledge representation at the location
has been deactivated, is transitory. That is, the effect decreases over time
as the residual resonance dissipates. In contrast, the effect of activation frequency, which is localized in the level of the activation threshold, is relatively permanent. These implications of the theory are discussed more fully
in chapter 3.
42
CHAPTER 2
43
In comparing the retrieval processes described here with the processes postulated by
Wyer and Srull (1989), one further implication should be noted. The earlier model assumed that
in the pursuit of goal-relevant information, the Work Space was searched first, before information is retrieved from Permanent Storage. In the present conceptualization, this assumption is
an implicit by-product of Postulate 2.5. That is, a set of probe cues is likely to resonate to the
greatest extent with existing knowledge representations that have been activated and used
most recently. Because the knowledge representations contained in the Work Space have typically been used recently in the pursuit of processing objectives, they would normally be reactivated more quickly than representations in Permanent Storage for use in attaining goals to
which they are relevant. To this extent, the two conceptualizations are quite compatible. On the
other hand, the content of the Work Space typically consists of representations that have either
been (a) recently recalled from Permanent Storage, or (b) recently formed and transmitted to
Permanent Storage by the Executor. These representations would be activated in Permanent
Storage according to criteria assumed by Postulate 2.2. To this extent, the implications of the assumption that the Work Space is searched first are largely redundant with those of the retrieval
postulates that we have already stated.
44
45
first two stages of processing are similar to Steps 1 and 2 of the process assumed to underlie goal-directed retrieval. That is, the Executor identifies a
set of features that happen to exist in the Work Space (either transmitted
from the Comprehender or resulting from prior cognitive activity). These
features in combination activate a representation in Permanent Storage, a
copy of which is sent to the Work Space. The Executor scans this representation for a goal specification and, if none is found, repeats the process just
described, recompiling a set of probe cues and using them to activate another representation in Permanent Storage. This continues until a goal
specification is identified as a result of the content of the material retrieved
from Permanent Storage or transmitted from the Comprehender.
Phenomenologically, this simply means that when people do not have a
particular goal in mind, the configuration of cognitions they happen to be
thinking about cues the retrieval of a previously formed representation.
Features of this representation, once retrieved, may combine with other activated concepts to cue the retrieval of yet another representation, and so
on until a goal is identified.
Thus, for example, suppose a student overhears a conversation in which
one person comments that John gave Peter an answer during the exam. A
representation of the statement that is constructed in the course of comprehending it is transmitted to the Work Space. Features of this representation (e.g., exam) along with other features that happen to be in the Work
Space at the time (e.g., psychology) are sampled, and a set of features is
compiled (e.g., psychology, exam). These features may resonate with a
representation in Permanent Storage that contains them (e.g., the thought
that the final exam in psychology is only 3 days before Christmas). As this
knowledge does not have goal implications, a new set of probe features is
compiled which might now include Christmas. These features then resonate with representations in Permanent Storage that contain this feature,
perhaps leading a representation to be retrieved that concerns buying a
Christmas present for ones boyfriend. In this case, however, the Executor
recognizes this as a goal. Consequently, the system leaves the feedback
loop. That is, the Executor identifies a goal schema pertaining to the objective and deposits it in the Goal Specification Box, and then uses the content
of the schema to direct processing relevant to attainment of the goal (e.g.,
the construction of a plan about what to buy and where to purchase it).
CONCLUDING REMARKS
The information-processing model outlined in this chapter provides a
framework for conceptualizing much if not all of the theory and research
discussed in the remainder of this volume. This will be particularly clear in
46
CHAPTER 2
our analysis of the comprehension of personal experiences, and of information about actual persons and events of the sort we encounter in the course
of daily life. Several important implications of the conceptualization are localized in Postulates 2.4 and 2.5, which govern the factors that influence the
accessibility of knowledge in memory and, therefore, the likelihood that
this knowledge is brought to bear on the processing of new information. Indeed, knowledge accessibility is a primary consideration in all phases of information processing from the initial comprehension of information to the
generation of an overt response. Before embarking on a more detailed analysis of comprehension and judgment, therefore, a more general discussion
of the determinants and effects of knowledge accessibility will be useful.
The next chapter provides this discussion.
C H A P T E R
3
Determinants and Effects
of Knowledge Accessibility
GENERAL CONSIDERATIONS
The general effects of concept and knowledge accessibility are intuitively
obvious. That is, people who are called upon to make a judgment or decision seldom consider the implications of all of the knowledge they have acquired that bears upon it. Instead, they use the first relevant judgment or
decision criterion that comes to mind, ignoring other criteria that might be
equally or more applicable. Taylor and Fiske (1978) were among the first to
47
48
CHAPTER 3
recognize this possibility, arguing that many judgments and decisions are
made off the top of the head. A more formal statement of this phenomenon was proposed by Chaiken (1980, 1987). She argued that people who
have to make a judgment or decision first apply the criterion that comes to
mind most quickly and easily and then assess their confidence that the implications of using this criterion are sufficient to attain the goal at hand.
Only if this confidence is below some minimal confidence threshold do
they search for additional criteria. Thus, they may search further if they are
highly motivated to make a correct decision, or if for some reason they
believe that the criterion they initially applied may be biased. Normally,
however, they base their judgment or decision on the first relevant criterion they happen to think about.
The implications of Chaikens conceptualization can be formalized in the
following, sufficiency postulate:
Postulate 3.1. People retrieve and use only the amount of information that
they consider to be sufficient for attaining the processing objective they are
pursuing at the time. The information they use is likely to be the first goalrelevant information that comes to mind or, alternatively, permits their objective to be attained most quickly and easily.
Chaikens formulation has been applied primarily in conceptualizing responses to persuasive communications (Chaiken, 1987; Chaiken, Wood, &
Eagly, 1996). However, it has more general applicability to information processing (Chaiken, Liberman, & Eagly, 1989), and many of the phenomena
reviewed later in this volume can be viewed from the perspective she suggests. The discussion in this chapter is restricted to representative research that bears on three main points to which the formulation outlined in
chapter 2 is particularly relevant:
1. The primary determinants of knowledge accessibility are (a) the recency with which knowledge has been acquired and/or used in the past and
(b) the frequency with which it has been applied. In the latter regard, knowledge that has been employed in many situations over a long period of time
can become chronically accessible.
2. The accessibility of knowledge representations for use in attaining a
particular processing objective can be activated by extraneous factors that
are objectively irrelevant to this objective and of which persons may not be
consciously aware. In fact, increasing awareness of these extraneous factors
is likely to decrease their influence.
3. The effects of differences in knowledge accessibility are apparent at
several stages of processing, including the spontaneous comprehension of
information at the time it is first received, the more deliberative interpreta-
49
tion of information that occurs in the course of attaining a particular goal, inference processes, and behavioral decision making. Differences in knowledge accessibility can also influence the goals that individuals pursue under
conditions in which more than one objective is potentially relevant, and the
procedures that are spontaneously activated and used to attain a particular
goal (e.g., productions of the sort postulated by Anderson, 1983).
50
CHAPTER 3
been exposed in the color-naming task, and these interpretations influenced their evaluations of the target. That is, they evaluated the target
more favorably when they had been exposed to favorable trait adjectives
than when they had been exposed to unfavorable ones.
In a similar series of studies (Srull & Wyer, 1979, 1980), participants performed a sentence construction task that required the use of concepts associated with hostility. Later, they were asked to form impressions of a target
person whose behaviors were ambiguous with respect to this trait (e.g., refused to pay the rent until the landlord painted his apartment, demanded
his money back from a salesperson, etc.). Both the number of hostilityrelated items presented in the sentence-construction task and the time interval between this task and the impression formation task (immediately, 1
hour or 1 day) were manipulated. Both of these factors influenced the likelihood that participants interpreted the targets behavior as hostile, indicating that both frequency and recency played a role. Moreover, a supplementary study indicated that the effects only occurred when participants were
exposed to trait concepts before they read the target information (Srull &
Wyer, 1980). When the concepts were activated after the information was
presented, they had no effect at all. This indicates that the effects of primed
concepts on judgments were mediated by their impact on the interpretation of the trait information at the time the information was received and
did not have a direct influence on judgments independently of this information.
The Relative Persistence of Frequency
and Recency Effects
Although the frequency and recency of prior activation both have an impact on the use of trait concepts to interpret information, the underlying basis for these effects theoretically differ. As noted earlier, the effects of recency of activation implied by Postulate 2.4 are transitory, whereas the
effects of activation frequency, which impacts on the activation threshold
(Postulate 2.5) are more enduring.
This was demonstrated empirically by Higgins, Bargh, and Lombardi
(1985). Participants were initially exposed to adjectives exemplifying both
favorable and unfavorable trait concepts (adventurous vs. reckless)
that could each be used to interpret the same behavior (e.g., crossing the
Atlantic in a sailboat). However, the relative frequency and recency of exposure to the adjectives were varied. When participants were asked to interpret the behaviors a very short time after exposure to the trait adjective,
they tended to use the more recency primed trait concept rather than the
frequently activated one. When a longer period of time had elapsed, however, the effects of activation frequency predominated over recency, as im-
51
After reading the story, participants were asked what the protagonists had
commented about.
The story is normally interpreted as the description of a card game.
However, it could also describe the rehearsal of a woodwind ensemble.
Some participants, who were students in a weight-lifting class, typically indicated that the protagonists had commented on how well they were playing cards. In contrast, students who were planning a career in music education indicated that the protagonists had commented on the sound of their
music. Thus, concepts and knowledge that were chronically accessible to
participants as a result of their curriculum and interests apparently affected their interpretation of the story and this interpretation influenced
their answers to questions about it.
The Additivity of Situationally Induced
and Chronic Knowledge Accessibility
A further implication of Postulates 2.4 and 2.5 is of particular interest in this
regard. That is, the effect of activation frequency (which are mediated by
changes in activation threshold) and the recency of activation (which are
mediated by the amount of residual resonance that remains at a knowledge
representation once it is no longer activated) are independent. This independence was demonstrated by Bargh, Bond, Lombardi, and Tota (1986).
52
CHAPTER 3
They showed that the chronic accessibility of a trait concept (inferred from
individual differences in the a priori likelihood of using the trait to describe
acquaintances in a previous experiment) and situationally induced accessibility of the trait (resulting from exposure to the trait in a laboratory task)
had independent and additive effects on the use of the concept to describe
a target person on the basis of information about the targets behavior.
Although the independence of these interpolated effects are implied by
Postulates 2.4 and 2.5, they are difficult to explain on the basis of the bin
conceptualization proposed by Wyer and Srull (1989). According to the latter conceptualization, priming trait concepts should make them equally accessible (either in the Work Space or in the semantic bin) regardless of
the frequency of their prior activation, and consequently should lead these
concepts to be used to an equal extent regardless of their chronic accessibility. In other words, recency should override the effects of chronic accessibility. Therefore, Bargh et al.s (1986) findings are embarrassing to the
Wyer and Srull model. On the other hand, they are quite consistent with the
retrieval postulates of the present conceptualization.
Generalizability of Concept Activation
Over Attribute Dimensions
In the study by Higgins et al. (1977), activating favorable and unfavorable
trait concepts that were descriptively inapplicable for interpreting the targets behavior had no influence on participants evaluations of him. This
finding suggests that the effects of activating trait concepts do not have a
direct impact on judgments of the target, but rather, influence judgments
only through their mediating influence on how the information about the
persons behaviors is interpreted. As we noted earlier, however, impression
formation is a two-stage process. Once a target persons behaviors are interpreted and a general impression is formed of him, the person may be inferred to have other attributes that are characteristic of the sort of individual implied by this impression. Thus, as Srull and Wyer (1979, 1980) found,
activating concepts associated with hostility led the target person not only
to be described as hostile but also to have other attributes that are typical
of a hostile individual. (For a more recent explication of these two processing stages, see Trope, 1986.)
The conclusion that activated trait concepts influence the interpretation
of only those behaviors to which they are descriptively applicable may nevertheless need to be qualified. Stapel and Koomen (2000) found that when
stimulus attributes with extreme evaluative implications (malevolent,
warm, etc.) are primed, they do appear to influence the interpretation of
behaviors that are descriptively unrelated to them. Apparently, adjectives
with extreme implications spontaneously activate more general evaluative
53
concepts (good or bad) that, once activated, influence the interpretation of information along evaluative as well as descriptive dimensions.
EFFECTS OF AWARENESS
Many effects of previously acquired concepts and knowledge on the interpretations of information can occur without awareness of the factors that
led them to become accessible in memory. This possibility is implicit in the
conceptualization we propose. That is, consciousness theoretically resides
in the Executor. The processes performed by the Comprehender, which are
not controlled by the Executor, occur without awareness. Correspondingly,
the knowledge representations that are involved in these spontaneous
comprehension processes can later be activated and applied without consciousness of why they came to mind or, for that matter, without awareness
that these representations had recently been used. For similar reasons, the
representations that are later identified by the Executor for use in goaldirected processing are unlikely to contain cognitions about the recency or
frequency with which the representations have been formed or used in the
past, as these cognitions are typically irrelevant to the objectives for which
the representations were formed. Therefore, these representations are
likely to be used as bases for judgments and decisions without considering
the reasons for their accessibility. Consequently, as studies cited in the preceding section testify, factors that fortuitously make knowledge accessible
in memory can increase the use of this knowledge. Moreover, this can occur without awareness of the factors that influence its accessibility.
Nonconscious Effects of Knowledge Accessibility
The first compelling demonstration of the nonconscious influence of concepts and knowledge was reported by Bargh and Pietromonaco (1982). In
this research, participants were exposed to a computerized text of perceptual vigilance that required them to press a button as soon as a light appeared on the computer screen. In actuality, the points of light were words
that were presented subliminally. In some conditions, a number of words
conveyed hostility (hostile, kill, etc.). After performing the task, participants as part of a different experiment were asked to form an impression of
a person on the basis of behaviors that were ambiguous in terms of the hostility they conveyed. Participants were unable to distinguish between
words that were presented and those that were not presented in a recognition memory task after the experiment. Nevertheless, they judged the target to be more hostile if they had been subliminally exposed to hostilityrelated words than if they had not.
54
CHAPTER 3
55
56
CHAPTER 3
liams, 1980). That is, they used the primed concepts as a basis for interpreting the information without bothering to seek alternatives.
The Effect of Thought Suppression
on Concept Accessibility
It therefore seems reasonable to conclude that when people are aware
that the concepts and knowledge that come to mind at the time they make
a judgment, they often actively avoid the biasing influence of these cognitions and, in doing so, are actually less likely to use them than they might
otherwise be. However, the effects of this cognitive activity can be somewhat ironic. Suppose individuals consciously avoid the use of a concept or
knowledge representation in responding to a stimulus. This conscious suppression takes cognitive effort. Moreover, it is impossible to consciously
avoid using a concept without thinking about the concept one is trying
to avoid. Consequently, active attempts to avoid using a concept can actually increase its accessibility in memory, making the concept more likely to
be used later under conditions in which attempts to suppress it are no longer made. Thus, as Wegner (1994) noted, instructing people to avoid thinking about a white bear may stimulate conscious thoughts about the bear,
thus making it more accessible in memory than it would have been in the
absence of this cognitive activity and, therefore, more likely to come to
mind later.4
An intriguing series of studies by Bodenhausen, Macrae, and their colleagues demonstrated the effects of conscious thought suppression in the
use of stereotypes (for a review, see Bodenhausen & Macrae, 1998). In a representative set of studies (Macrae, Bodenhausen, Milne, & Jetten, 1994),
participants were given information about a skinhead with instructions to
form an impression of him. However, some participants were told not to
rely on stereotypes in arriving at their impression, whereas others were not
given this instruction. The former participants, relative to those who had
not been told to avoid using the stereotype, (a) responded more quickly to
stereotype-related words in a subsequent lexical decision task, (b) described a second skinhead more stereotypically once the request to avoid
using stereotypes was no longer imposed, and (c) sat further away from a
skinhead while waiting to participate in a later part of the experiment. Thus,
4
A game we used to play as children anticipated Wegners insight. That is, we would tell
nave playmates to try not to think of a white horse and then, after ensuring that they were doing
so, would ask them to report the direction in which the horse was facing. They invariably found
this question quite meaningful and responded to it quickly. Attempts to suppress thoughts
about the horse typically elicited an image of the horse that individuals found easy to describe
despite their assurances that they were trying not to think about it.
57
the attempt to suppress the use of the stereotype in the initial task increased rather than decreased the accessibility of concepts associated with
the stereotype. Consequently, it increased the use of the stereotype as a basis for judgments and behavioral decisions in situations in which the sanctions against using it were removed.
58
CHAPTER 3
FIG. 3.1. Hypothetical associative network showing the relations among behaviors, bipolar trait concepts (honest, dishonest, kind, and unkind), and the
priming stimuli that activate these concepts (Ph, Pd, Pk, and Pu, respectively).
59
Primed Concept
Inapplicable for
Describing Target
Honest
Unkind
Dishonest
Kind
.280
.160
.115
.308
.407
.074
.231
.231
(M = .137). However, this was also true when the bipolar opposite (inapplicable) concepts were primed (.315 vs. .152, respectively).5
Finally, these interpretations influenced participants overall evaluations of the target. That is, participants evaluated the target more favorably when they had been primed with a trait pertaining to hostility (M =
4.55) than when they had been primed with a trait pertaining to kindness (M
= 3.76), and this was true regardless of whether the primed trait was directly applicable for interpreting the behavior (4.08 vs. 3.58 when honest
and unkind were primed, respectively) or when the bipolar opposites
were primed (5.02 vs. 3.94, when dishonest and kind were primed, respectively).
The behavioral information that participants received in the preceding
study pertained to two different dimensions, but its implications along each
dimension were very clear. An additional experiment (Park et al., 2001, Experiment 2) showed that when the implications of information along a dimension are ambiguous, activating concepts can influence not only the dimension along which the information is evaluated but also the value it is
assigned along this dimension. In this study, participants first performed a
task that required the use of concepts associated with either good health,
bad health, good taste, or bad taste. Later, they were given the description of a milk product as having more artificial sweeteners. Although
this description clearly implied that the product was bad for the health, its
implications for tastiness were ambiguous. That is, it could imply that the
5
This latter effect appears more pronounced when dishonest was primed than when kind
was primed. However, this difference results from a more general tendency to judge the target
as honest rather than as unkind, which inflates the effect of priming dishonest but decreases
the effect of priming kind. Pooling over the two trait replications, the overall effect of priming
was similar in magnitude regardless of whether the primed concept was applicable for describing the target or its bipolar opposite.
60
CHAPTER 3
product was either good tasting (sweet) or bad tasting (too sweet). Participants who had been exposed to health-related concepts described the
product as unhealthy regardless of whether good health or bad health
was primed. In contrast, participants who were exposed to taste-related
concepts described it as having either good taste or bad taste, depending
on which taste-related concept had been activated. Thus, primed concepts
influenced not only the dimension along which the information was interpreted but also the value it was assigned along this dimension.
Inference Processes: Belief Formation and Change
People who are asked to report their belief that a statement concerning it is
true are unlikely to search memory for all of the belief-related knowledge
they have acquired. Rather, they are apt to rely on the first belief-relevant
knowledge that comes to mind (Postulate 3.1). That is, they search memory
for a possible reason why the statement might or might not be true and, if
they find such a reason, use its implications as a basis for their judgment
without further ado. To this extent, the reason they identify and, therefore,
the belief they report, may depend on the recency with which the reason
had been thought about. We discuss this possibility in more detail in chapter 10 of this volume. However, two examples are worth mentioning in the
present context.
Participants in a study by Henninger and Wyer (1976) were asked to report their beliefs in a series of propositions. Some propositions were syllogistically related. That is, one, informational proposition, A (e.g., The
army is recruiting people of below average intelligence), had implications
for the validity of a second, target proposition, B (e.g., The quality of the
peace-time army is deteriorating). In some cases, beliefs in the informational proposition were reported early in the questionnaire, and the target
proposition was not considered until later. In other conditions, the target
proposition was considered at the outset. Participants use of their belief in
the informational proposition as a basis for inferring the validity of the target was inferred on the basis of a quantitative model of syllogistic inference
proposed by Wyer and Goldberg (1970; see also McGuire, 1960, 1981).6 Par6
If beliefs are defined in units of probability (along a scale from 0 to 1), the model asserts that
P(B) = P(A)*P(B/A) + P(~A)*P(B/~A)
Where, P(B) is the belief that Proposition B is true, P(A) and P(~A) are the beliefs that A is and is
not true, and P(B/A) and P(B/~A) are beliefs that B is true if A is and is not true, respectively.
This model provides an accurate quantitative description of the relations among these beliefs,
and also the effect of changing beliefs in A on beliefs in B (Wyer, 1970, 1974; for summaries of evidence, see Wyer, 1974; Wyer & Hartwick, 1980). The implications of this model are discussed
more fully in chapter 10.
61
ticipants based their beliefs in the target on their beliefs in the informational proposition when they had evaluated this proposition earlier in the
questionnaire. When they evaluated the target proposition at the outset,
however, they typically identified a different subset of previously acquired
knowledge to use as a basis for reporting their belief, and so this belief was
inconsistent with their belief in the informational proposition they considered subsequently.
Interestingly, when participants reported their beliefs again a week later,
their beliefs in the target were consistent with their beliefs in the informational proposition in both order conditions. Apparently, the informational
proposition, having been reported in Session 1, was accessible to both
groups of participants in Session 2, and so both groups of participants used
it as a basis for inferring the validity of the target.
This interpretation obviously assumes that no other relevant information bearing on the validity of the target was received or thought about during the 1-week interval between sessions. The target propositions pertained
issues that participants were unlikely to think about spontaneously, and so
this assumption seems reasonable. In an unpublished study by Wyer and
Hartwick (see Wyer & Srull, 1989), however, some participants were asked
to report their beliefs that drinking coffee was desirable after having been
exposed to a proposition that either implied that this was true (Coffee
keeps you alert) or implied that it was false (Caffeine destroys nerve
cells). Others were asked to report their beliefs that student use of the university health center would increase after exposure to a proposition suggesting that the event was likely to occur (Dental services are scheduled to
be provided at the health center) or unlikely (Some doctors at the health
center are about to lose their licenses). In both cases, participants based
their beliefs in the target proposition on the implications of the particular
information proposition to which they had been exposed. However,
whereas the effects of informational propositions on participants beliefs
about the use of the health center were still evident in a second experimental session 1 week later, their effects on beliefs in the desirability of drinking
coffee were not. The reason for this difference is rather obvious. That is,
participants in the 1-week interval between experimental sessions were
likely to have numerous coffee-drinking experiences with implications for
the desirability of this activity, and these more salient implications overrode the effects of the knowledge they had selectively activated and used in
the first experimental session. In contrast, participants had few if any experiences with implications for the use of the health center during the 1-week
interim, and so the knowledge they had activated in the first session continued to have an impact.
The studies described in this section demonstrate the effects of activating prior knowledge about an issue on beliefs and opinions. However, this is
62
CHAPTER 3
not the only type of knowledge that people bring to bear on these judgments. An equally important source of information is provided by ones
own past behavior toward the persons and objects to which the issue pertains. Bem (1972), for example, noted that people often use their past behavior (e.g., a decision to advocate a position publicly) to infer their belief
that the position is valid (see also Albarracin & Wyer, 2000). These effects
are likely to be mediated by an implicit theory that people have acquired
about the causal relatedness of their beliefs and their behavior under the
particular situational conditions at hand. We explore this possibility more
fully in chapter 10.
Response Processes: The Use of Activated Concepts
as Standards of Comparison
The effects of accessible concepts on the interpretation of information typically occur at the time the information is first received. At the time a judgment is made, the same concepts can have other effects. In particular, judgments are typically made with reference to a standard of comparison.
Moreover, these standards may be applied spontaneously, without awareness. Thus, we may judge a baby to be big and a house to be small without feeling any inconsistency in these judgments at all.
The effects of knowledge on standards of comparison can sometimes offset or override the effects that occur at the comprehension stage. A study
by Herr (1986) is of interest in this context. Participants were initially exposed to the names of persons who were known to be either moderately
hostile (e.g., Muhammad Ali or Howard Stern) or extremely hostile (e.g.,
Adolf Hitler or Genghis Khan). Later, they were asked to form impressions
of a person whose behaviors were ambiguous with respect to this trait. It
seems reasonable to suppose that the priming stimuli employed in this
study had two effects. First, they activated trait concepts that were associated with the individuals being primed, leading participants to interpret the
targets behaviors in terms of these concepts. At the same time, they activated concepts of the individuals themselves that participants later used as
standards of comparison in evaluating the target. These effects could potentially offset one another.
Herrs results were consistent with this possibility. Participants who
were exposed to moderate exemplars of hostility judged the target as more
hostile than participants in a control condition, confirming the findings obtained by Srull and Wyer (1979) and by Bargh and Pietromonaco (1982).
However, participants who were exposed to extreme exemplars judged the
target to be less hostile than control participants did. The exemplars to
which participants were exposed may have been used as standards of comparison in both conditions. When the exemplars were only moderately ex-
63
treme, however, the effect of using them as standards was not sufficient to
override the effect of the activated trait concepts on interpretation of the
information at the time it was received. When the exemplars provided a
more extreme standard of comparison, the effect of using them in this capacity predominated, producing a contrast effect. In this regard, standards
of comparison for evaluating a person are generally less likely to be activated by trait concepts per se than by the exposure to persons who possess the traits. Thus, as Stapel, Koomen, and van der Pligt (1997) found (see
also Moskowitz & Skurnik, 1999), contrast effects of priming are more likely
to be evident in the latter case than the former.
The activation and use of stimuli as standards of comparison can occur
without awareness. This was demonstrated in a series of studies by Adaval
and Monroe (2002). In some of their studies, participants were exposed subliminally to either high or low numbers before being asked to judge a target
product whose price fell between the two sets of priming stimuli. Participants judgments of the target product were lower when high numbers had
been subliminally primed than when low numbers had been primed. That
is, the target was judged as less expensive, but also less desirable, in the
former condition than in the latter. Thus, priming stimuli appeared to induce a disposition to use either a high or low value as a standard of comparison regardless of the dimension along which judgments were made.
64
CHAPTER 3
they are asked to judge. Once this impression was formed, it may have influenced their later judgments independently of the behavioral information
they had received earlier. Thus, in our example, participants who were
asked to judge the targets honesty judged him as honest. Having done so,
they later judged the target as more kind than they would have done if they
had not made the initial judgment.
A quite different series of studies by Thomas Mussweiler and Fritz Strack
(for a review, see Mussweiler, 2003) provide a further example of these effects and confirm assumptions underlying their interpretation. In a typical
study (Mussweiler & Strack, 2000b), participants were first asked to indicate
whether Mt. Everest was greater or less than a specific anchor value (e.g.,
Is Mt. Everest higher or lower than
feet?). The anchor value was either high or low. In all cases, these values were sufficiently extreme that
participants would normally consider the target to fall in between them. In
some cases, however, the values were fairly plausible (e.g., 45,000 ft. vs.
2,000 ft.), whereas in other cases, the values were totally implausible (e.g.,
300,000 ft. vs. 65 ft.). Participants after making the comparative judgment
were asked to estimate the actual height of Mt. Everest. Participants typically judged the actual height of Mt. Everest to be greater when they had
compared it to a high value than when they had compared it to a low one.
Moreover, this tendency was even greater when the anchor values were implausible (115,728 ft. vs. 9,271 ft., respectively) than when they were plausible (36,106 ft. vs. 27,783 ft., respectively).
On the surface, these results appear to resemble the anchoring-andadjustment phenomenon identified by Tversky and Kahneman (1974). That
is, people in judging a stimulus arbitrarily identify a high or low anchor
value along the judgment dimension and then adjust upward or downward
in relation to this standard. However, they do not adjust enough, with the
result that their judgment is displaced toward the anchor they have arbitrarily used (for examples of this effect in social judgment, see Schwarz &
Wyer, 1985). Mussweiler and Strack (1999a) raise an alternative possibility,
however. That is, people who estimate a stimulus in relation to a standard
may activate a body of general knowledge that is consistent with it. Although they may ultimately reject the anchor value as implausible, the subset of knowledge and its implications, having been made accessible, is then
used as a basis for their own estimates.
To distinguish between the two alternative interpretations, Mussweiler
and Strack (1999a) asked participants after making their comparative judgments to perform a lexical decision task requiring them to identify words
that were semantically associated with either high anchor values, low anchor values, or irrelevant. Participants were quicker to identify words that
were associated with the anchor to which they had been exposed than the
opposite anchor, confirming the assumption that exposure to the anchor in-
65
66
CHAPTER 3
Shiv and Fedorikhin (1999), for example, found that people were inclined to
choose to eat chocolate cake rather than fruit salad when they did not have
the opportunity to think about their decision, whereas the reverse was true
when they were able to deliberate. Presumably, the visual stimulus features
of the cake led hedonic goals to be activated very quickly, whereas the
competing goal of good health took more time to be activated.
In many instances, however, people may acquire more general motivational orientations that, once activated, can influence their judgments and
behavior in a variety of specific situations. Such general motivational dispositions are often activated by features of a stimulus situation to which they
are relevant. However, their accessibility can also be influenced by extraneous factors that are not directly relevant to the situation at hand. Two quite
different bodies of literature provide evidence of this possibility. One concerns the desire to maintain beliefs in a just world (Lerner, Miller, &
Holmes, 1976). A second surrounds the relative emphasis placed on positive versus negative consequences in making a behavioral decision (cf. Higgins, 1997).
The Motivation to Believe in a Just World:
Implications for Reactions to Rape
Motivational influences on attitudes and opinions, which were called to the
attention of social psychologists by Festinger (1957), are well established.
The nature of these effects obviously depends on the particular motive involved (e.g., see Arkin, Gleason, & Johnson, 1976; Baumeister & Newman,
1994; Giner-Sorolla & Chaiken, 1987; Psyzczynski & Greenberg, 1987; Zanna
& Cooper, 1976). One motivational determinant of judgments, proposed by
Lerner et al. (1976; Lerner & Miller, 1978), is the desire to believe that the
world is just. Lerner et al. assume that people have a need to maintain beliefs in a just world in order to reassure themselves that they (who are
righteous) will personally not fall victim to misfortune for reasons beyond
their control. One manifestation of this motivational disposition is that people are not only disposed to believe that individuals who do something
wrong will be punished but also, that people who encounter misfortune are
responsible for their fate. In other words, individuals not only get what they
deserve but deserve what they get.
Lerner et al. (1976) provided several examples of this tendency. Walster
(1966), for example, found that people judged victims of an automobile accident more negatively if the victims had been seriously injured than if they
had not. Similarly, Lerner and Simmons (1966) found that observers of a
learning experiment in which a confederate was ostensibly shocked for
making errors disparaged the learner more when the shocks were severe
than when they were mild.
67
Beliefs in a just world are not the only possible basis for judgments and
decisions, of course. To this extent, the use of this criterion should be increased by situational factors that increase the accessibility of these beliefs
and the motive to maintain them. This possibility was demonstrated by
Wyer, Bodenhausen, and Gorman (1985) in a study of the cognitive mediators of reactions to rape. We were interested in whether activating aggression-related concepts in one context might influence peoples reactions to
rape situations that they encountered in a different, ostensibly unrelated
context. To do so, we asked participants to engage in two ostensibly unrelated experiments. The first study was ostensibly concerned with the things
shown in the media that college students find objectionable. On this pretense, we exposed participants to slides of 12 pictures. Nine of these pictures showed objects and events that participants were unlikely to consider offensive. The other three pictures varied. In one condition, these
slides showed aggressive acts of the sort that occurred frequently (e.g., police subduing a criminal, a boxing match, etc.) and, therefore, activated concepts that aggression is normal and socially sanctioned. In a second condition, the pictures portrayed severely negative outcomes of aggression that
activated the concept that human beings were cruel and inhumane (e.g., a
lynching episode, a dead soldier with a hole in his head, etc.). The third,
control set portrayed stimuli that might be considered unpleasant but
were unrelated to aggression (e.g., deformed babies, a smoking advertisement, etc.).
Participants rated each of the 12 pictures in terms of how objectionable
it was. Then, they were told that the experiment (which took about 10 min)
was over, but that because there was time remaining, we would like them to
help out another faculty member who was conducting a study in a different
room down the hall. The new study was introduced as an investigation of
the factors that people consider important in judging criminal cases. On
this pretense, participants were asked to read descriptions of four rape
cases, and in each case, to report several reactions. Two questions concerned (a) whether the defendant should be convicted and (b) whether he
actually was convicted. Three others concerned the victims responsibility
for the incident (e.g., whether she provoked the rape, whether she could
have avoided it, etc.).
We assumed that the pictures presented in the first experiment would
activate concepts that participants would use to construe the implications
of the rape scenarios they encountered in the second experiment. On a priori grounds, several hypotheses seemed plausible. First, if pictures of socially sanctioned acts of aggression activated concepts that aggression is
normal, they might decrease beliefs that the defendant should be convicted. On the other hand, if pictures of extremely negative consequences
of aggression activate concepts that people are cruel and inhumane, they
68
CHAPTER 3
Control
Socially
Sanctioned
Aggressive Acts
Severe
Outcomes of
Aggression
8.97
3.95
2.97
9.62
3.47
3.07
8.70
5.10
4.20
Note. Judgments are reported along a scale from 0 (not at all likely) to 10 (very likely). Based
on data from Wyer, Bodenhausen, and Gorman (1985).
69
70
CHAPTER 3
71
Endorsement of equality-oriented
proverbs
Individual participation
Group participation
M
Proportion of participants choosing candy of different types
Individual participation
Group participation
M
No Instructions
(Experiments
1 and 2)
Promotion
Focus
(Experiment 3)
Prevention
Focus
(Experiment 3)
5.07
5.60
5.34
5.08
5.07
5.07
5.49
5.60
5.55
.45
.73
.61
.50
.55
.43
.59
.81
.71
72
CHAPTER 3
equality-related proverbs, and also more likely to choose candies of different types, under prevention-focus conditions than under promotion-focus
conditions.
Note also that under promotion-focus conditions, proverb endorsement
and candy choice were each very similar to that observed under individualparticipation conditions of Experiments 1 and 2, whereas under preventionfocus conditions, they were similar in magnitude to that observed under
group-participation conditions of the first two experiments. Thus, these
data provide strong confirmation of the assumption that inducing feelings
of group membership activated a prevention focus, and therefore had the
same impact as a more direct manipulation of this motivational orientation.
Feelings of Cultural Identity. A second series of studies (Briley & Wyer,
2002, Experiments 5 and 6) confirmed these conclusions using a quite different means of inducing feelings of group membership, and also different indices of prevention focus. Specifically, participants were both European
American and Hong Kong Chinese college students whose feelings of group
membership were activated by making them aware of their cultural identity. This was done by exposing representatives of each culture to symbols
of their own or a different culture as part of an ostensibly unrelated test of
general knowledge. (Specifically, European Americans were exposed to pictures of Marilyn Monroe, a Dixieland jazz band, the American flag, etc. Chinese were exposed to pictures of the Great Wall, a Chinese musical instrument, people doing calligraphy, etc.).
Following this task, some participants were asked to state their preferences for different combinations of outcomes to themselves and another
in a hypothetical resource allocation task. Others were administered the
product choice task employed by Briley et al. (2000) and described earlier.
Exposing participants to symbols of their own culture not only increased
their preferences for equality in the resource allocation task but also increased their likelihood of choosing the compromise alternative in the
product choice task, and this was true of both European Americans and
Hong Kong Chinese. Thus, making participants aware of their cultural
identity activated a disposition to avoid negative outcomes that overrode
the chronic cultural differences in prevention focus that Briley et al. (2000)
observed.
73
74
CHAPTER 3
75
Subliminal Prime
Overt Prime
420
452
512
502
446
524
Note. Scores are converted to percentiles based on GRE norms. Adapted from Colcombe and
Wyer (2001).
P A R T
II
COMPREHENSION PROCESSES
C H A P T E R
4
The Comprehension and
Validation of Information
About Familiar People and Events:
The Role of Situation Models
The comprehension of an event can occur in two stages. The first stage is
spontaneous and automatic, and occurs whenever the event can be easily
understood in terms of concepts and knowledge that comes easily to mind
at the time. It is difficult to observe a boy pounding a nail into a piece of
wood, or to see a book lying on a table, without comprehending this experience. It is equally difficult to read descriptions of these events without comprehending them. For example, try not to understand The boy pounded a
nail into a piece of wood, or the book is on the table. For a native English
speaker, the comprehension of such statements is spontaneous and virtually uncontrollable.
In many cases, however, comprehension requires deliberative cognitive
activity. Just as we comprehend some verbal statements spontaneously, we
immediately recognize that others cannot be easily understood. For example, The nail pounded a tree into the boy is structurally identical to The
boy pounded a nail into a tree. To understand the first statement, however, one must consciously construct a cartoon-like mental image of an animated nail driving the tree into a screaming juvenile. This construction is
deliberative and requires cognitive effort.
Moreover, the comprehension of some events requires an understanding of other events that precede or follow them. This information is particularly necessary when the event is inconsistent with a preexisting conception of the things that typically occur in the situation at hand. For example,
a nuclear physics professor might have difficulty computing a 15% tip at a
restaurant. Or, a persons comment at a party might violate a conversational norm to be polite, informative, or truthful (Grice, 1975). These events
79
80
CHAPTER 4
are likely to stimulate conscious cognitive activity in an attempt to understand why they occurred.
However, this deliberative cognitive activity is only likely to occur when
the comprehension processes that are spontaneously activated by exposure to information are insufficient. Before discussing this activity and the
conditions that give rise to it, one must first understand the processes that
underlie the spontaneous comprehension of information. The present chapter focuses on these processes. We first describe the mental representations that are formed in the course of comprehending events of the sort we
often encounter or read about in daily life. Then, we discuss the processes
that underlie the construction of these representations. Finally, we consider the implications of our conceptualization for the spontaneous recognition of statements about known persons and events as either true or
false. These latter implications are of considerable importance in understanding responses to information outside the laboratory, as our discussion
in later chapters testifies.
81
(Johnson-Laird, 1983; Kintsch, 1998; Zwaan & Radvansky, 1998), spatial memory (Taylor & Tversky, 1992) and logical reasoning (Schaeken, JohnsonLaird, & dYdewalle, 1996). Other research has examined their role in both
comprehending observed events (e.g., a movie; Magliano, Dijkstra, &
Zwaan, 1996) and imagining experiences that have not yet occurred (cf.
Glenberg, 1970; Glenberg, Kruley, & Langston, 1994; Graesser, Singer, &
Trabasso, 1994). Perhaps surprisingly, however, their role in social information processing is not widely recognized.
The mental simulation of a situation typically requires assumptions
about the relations among the entities (persons and objects) involved in it.
These relations are most easily represented in a mental image. An image
is often coded visually, and thus can be likened to a picture.7 However, it
should not be equated with a photograph. For one thing, not all features
that are found in a photograph are likely to be mentioned. (Thus, the mental image that is formed in response to the statement the boy kicked the
ball might specify the nature of the ball and portray the boy in the act of
kicking it, but might not include a specification of the boys hair color or
what he was wearing.) On the other hand, a mental image, as we conceptualize it, may convey information in other modalities. Thus, a representation
of The man shouted, Get out of my way! may contain an acoustically
coded representation of the mans tone of voice as well as the thing he said.
Two types of situation models are of particular relevance in conceptualizing social experiences. Event models are simulations of specific actions
and, in isolation, are similar to states of affairs. However, a number of temporally and thematically related event models might be combined to form a
multiple-segment episode model that conveys the temporal and causal relatedness of the events as well as the events themselves.
As the preceding discussion implies, all event and episode models are
assumed to have an image component. This is true regardless of whether
the information that leads the model to be constructed is conveyed
nonverbally, orally, or in writing. When information is conveyed verbally,
the representation of it in memory is also likely to have a propositional
component that provides a linguistic description of the event or situation
being represented. To this extent, an event model might be analogous to a
picture plus caption, and an episode model to a mental comic strip.
Note, however, that although the verbal description of an event may spontaneously lead to the construction of a mental image in the course of comprehending it, an observation or a picture of the event can often be compre7
The representation of visual images in memory has been somewhat controversial (cf.
Kosslyn, 1980; Pylyshyn, 1973; Shepard & Metzler, 1971). However, we do not intend to enter into
this debate. The construct of a visual image as used in this volume, like mental representations
more generally, should be evaluated in terms of their utility as metaphors and not in terms of
their validity as depictions of the physiology of the brain.
82
CHAPTER 4
hended without recoding it linguistically. In this case, the recoding may not
normally be performed (see Postulate 3.1).
The preceding discussion can be summarized in two postulates:
Postulate 4.1. Event and episode models are mental simulations of an event,
or sequence of events, that are temporally and spatially constrained. These
models are constructed spontaneously in the course of comprehending
events that occur at a specifiable (although not necessarily specified) time
and place.
Postulate 4.2. An event model can consist of both a mental image and a linguistically coded description of the event or state of affairs to which it refers.
Although the image component of the model is obligatory, the linguistic component is optional.
83
uation than at a party, and differently with friends than with casual acquaintances. These behavioral dispositions are activated and applied without
conscious analyses of the situational factors that give rise to them.
Headers
Episode models are often assigned a header, or title, that constitutes a generalization of the sequences of events to which the model refers and that
essentially tells what the model is about. The header of a model may be
coded propositionally. Thus, the fiasco at Lydias dinner party might be
the title of an episode model whose segments depict my stumbling into a table and spilling a bloody Mary on a guests new dress, causing her to drop a
dish of guacamole on the hostesss white shag rug. Features of a models
header can serve as retrieval cues for the model itself, activating the model
as a result of the processes described in chapter 2. In this regard, many
general statements that convey attitudes and opinions could be headers of
more specific episode models that exemplify them. For example, Religious
fanaticism is dangerous could be the header of an episode model of the
events surrounding the attack on the World Trade Center.
Not all propositions are model headers, of course. On the other hand, a
given proposition might be the header of more than one model. It is interesting to speculate that the strength of ones belief in a given proposition is
a positive function of the number of episode models for which it serves as a
header. Because episode models are constructed from specific experiences, this proposition would be consistent with Fazios (1990) proposal
that the strength of an attitude (or opinion) is typically greater when it is
derived from personal experience than when it is not.
In this and the following chapter, we describe the construction of event
and episode models and their role in the comprehension of social information. Before doing so, however, it will be useful to discuss more generally a
fundamental assumption that underlies the postulation of these models,
namely, that verbal event descriptions are often spontaneously encoded in
terms of mental images.
84
CHAPTER 4
85
prehending the sentences, they should be more likely to confuse the first
pair of sentences than to confuse the second pair, as reflected in their recognition responses. This was in fact the case.
A study by Glenberg, Meyer, and Lindem (1987) is particularly compelling. Participants read a story in which a sweatshirt either was associated
with the protagonist at the outset (John put on his sweatshirt and went
jogging) or was separated from him (e.g., John took off his sweatshirt and
went jogging). Their recognition of specific features mentioned in the passage (including sweatshirt) was then assessed. The time that elapsed between the mention of the feature in the story and its occurrence in the
recognition list was the same in all cases. Nevertheless, participants recognized the target object more quickly in the first condition than in the second. Participants in the first condition apparently formed a mental image of
John wearing the sweatshirt that persisted throughout the mental representation they formed of the events that occurred later, and this image was salient at the time of recognition. In the second condition, however, this was
not the case.
Finally, a study by Black, Turner, and Bower (1979) suggested that when
people form mental images on the basis of event descriptions, they construct these images from a specific visual perspective. Participants read
passages describing pairs of events that were either likely to be imagined
from the same perspective (e.g., Mary was reading a book in her room.
John came in to talk to her.) or from different perspectives (Mary was
reading a book in her room. John went in to talk to her.). Participants typically took longer to read and comprehend the second type of passage than
the first. In the first case, readers presumably formed an image of Mary
from someone in the room that was maintained while comprehending the
second. In the second case, however, the second sentence stimulated a
mental image from the perspective of someone outside the room, and so
readers were required to shift their perspective in order to comprehend it
in the context of the first sentence. As a consequence, reading and comprehension time increased.
Differences in the Comprehension of Picture
and Text Information
Using quite different methodologies, all four studies described earlier converge on the conclusion that mental images are constructed spontaneously
in the course of comprehending information about situation-specific events.
In all cases, however, the evidence is circumstantial. More direct support
for Postulate 4.2, requires a comparison of the comprehension of verbal descriptions of events with comprehension of the same events when they are
86
CHAPTER 4
conveyed in pictures. Specifically, suppose participants read a verbal description of an event. If this description spontaneously stimulates a mental
image of the event, the event should later be identified just as easily on the
basis of a picture as it would be if a picture of the event had been conveyed
in the first place.
Wyer, Adaval, and Colcombe (2002) examined this possibility. In doing
so, we also evaluated the assumption that although verbal descriptions of
an event are spontaneously encoded in terms of mental images, pictures of
events are not spontaneously encoded linguistically. We constructed two
scenarios of about 14 events each. One scenario, which concerned a day at
school, described a male student getting up in the morning, getting
dressed, eating breakfast, leaving his apartment, finding that his bicycle
had a flat, walking to school, getting to class late, leaving, meeting a friend,
going to the library, and falling asleep at the table. The second, a dating scenario, portrayed the student getting ready for a date, meeting the date, buying tickets for a movie, buying some popcorn, eating it while watching the
show, leaving, going to a pub, observing his date flirt with a friend, sulking
while walking home, but kissing his date goodnight at the door. Four sets of
stimuli were constructed for each scenario, two composed of pictures
and two composed of captions. The two picture scenarios were composed of photographs of the events, with one set taken from different perspectives than the other. The two caption scenarios were also constructed composed of verbal descriptions that were similar in meaning but
different in wording (e.g., discovered his bicycle tire was flat vs. saw that
his bike had a flat tire).
Participants were shown the events composing one of the two scenarios
on a computer screen with instructions that we were interested in how people comprehend everyday life events. This was done in three conditions. In
picture-only conditions, only pictures of the events were presented. In caption-only conditions, only verbal descriptions of the event were conveyed,
and in a third, picture-plus-caption condition, both pictures and verbal descriptions were shown.
After receiving the event descriptions and a short delay, participants
were presented stimuli portraying a number of events with instructions to
decide whether the event had occurred in the sequence they had seen
earlier. Some of the events were described verbally and others were described in pictures. Moreover, some stimuli referred to the same events
that occurred in the original scenario but were from a different perspective (in the case of photographs) or were worded differently (in the case
of verbal descriptions). Other stimuli, however, described different events
entirely. Participants were told that the stimuli they would see would not
be identical to those they had seen earlier, and to base their judgments on
whether the stimulus event was similar in meaning to the one they had
87
seen earlier. In each case, the time that participants took to judge the
stimuli was recorded.
1. Results
Participants under all conditions presumably formed a multiple-segment episode model on the basis of the events they encountered regardless of the modality in which they were conveyed. According to Postulate
4.2, however, participants who received written descriptions of the events
should spontaneously form mental images of them and, therefore, should
have these images in memory at the time of judgment. Therefore, they
should later be able to identify pictures of the events as easily as they
would if pictures of the events had been presented in the first place. Data
shown in the top half of Table 4.1 support this hypothesis. That is, when
participants were asked to verify an event on the basis of a picture, they
responded just as quickly if the event had been described verbally (M =
1.90 s) as they did if the event had been portrayed in a similar but not
identical picture (M = 1.97 s).
However, Postulate 4.2 also implies that people do not spontaneously
code pictures verbally when these codings are not necessary to comprehend the events they portray. If this is the case, however, participants who
are later asked to verify an event on the basis of a verbal description of it
must recode the picture linguistically at the time of judgment in order to
perform the verification. This recoding presumably takes time. Consequently, participants should take longer to verify the event than they would
if a verbal description had initially been presented. This was also the case,
as Table 4.1 indicates. That is, when participants were told to identify an
event on the basis of a verbal description of it, they took longer to do so
when the event had initially been conveyed in a picture (M = 2.53 s) than
when it had been described in words (M = 1.71 s).
TABLE 4.1
Mean Time to Verify Pictures and Captions as a Function of Presentation
Order and the Type of Information Initially Presented
Narrative order
Response time to
Response time to
Scrambled order
Response time to
Response time to
Picture
Verbal
Description
Picture Plus
Verbal
Description
pictures
verbal descriptions
1.97 (19)
2.53
1.90 (14)
1.71
2.05 (20)
1.96
pictures
verbal descriptions
2.16 (19)
2.26
2.50 (18)
2.23
1.91 (16)
2.09
88
CHAPTER 4
2. A Qualification
The results just described are quite consistent with Postulate 4.2. There
are nevertheless some qualifications on the validity of this postulate. In
other conditions of Wyer et al.s study, the events were presented in scrambled order rather than in the order they occurred. In this condition, participants could not spontaneously form a multiple-segment episode model of
the sequence as a whole. Rather, comprehension of this sequence required
deliberative cognitive activity. Results shown in the bottom half of Table 4.1
indicate that in this condition, participants could verify an event on the basis of a verbal description of it just as quickly when the event had originally
been conveyed in a picture (M = 2.26 s) as when the event had been described verbally in the first place (M = 2.23 s). However, they took longer to
verify an event on the basis of a picture when the event had been described
verbally (M = 2.50 s) than when it had been described pictorially (M = 2.16).
Although this pattern of response times was unexpected, it has a plausible
explanation. Verbal symbols are easier to manipulate mentally than visual
images are. Consequently, participants who received pictures of the events
in scrambled order may have subjectively translated them into linguistic
terms in order to construct a mental representation of the sequence of
events as a whole. As a result, they could later verify a verbal description of
the events just as easily as participants who had been given verbal descriptions of the events in the first place. When participants received verbal descriptions of the events in scrambled order, however, they relied on linguistic codings alone to construct a representation of the sequence without
including visual images in this representation. Therefore, when they were
later asked to verify a picture, they were required to recode it linguistically in
order to compare it with the linguistic representation they had constructed
earlier, and this took time. As a consequence, they took longer to make this
comparison than participants who had received a picture at the outset.
89
90
CHAPTER 4
FIG. 4.1. Theoretical sequence of processing steps involving the Parser and
Simulator in the course of comprehending visual and verbal information.
1. The Parser
The Parser only comes into play when the input information is coded linguistically. In this case, the Parser transforms verbal information into the
equivalent of a SubjectPredicate proposition. The output of this Parser
then resonates with previously constructed mental representations in Permanent Storage. In this regard, a major difference between the conceptual-
91
92
CHAPTER 4
One implication of this postulate should be noted at the outset. That is,
the type of representations activated by the subject and predicate are likely
to differ. In particular, the subjects features are likely to activate a generalized person representations of the sort described in chapter 1 (Fig. 1.2).
That is, this representation may consist of a set of general features that are
independently associated with the individual to whom the subject refers.
Although some of these features can be specific behaviors the individual
has performed, they are also likely to include more general descriptions of
the persons personality and behavior.
On the other hand, the features of the predicate, which pertain to behaviors or actions, are likely to activate previously formed event and episode
models of these actions. The actors depicted in these representations could
93
sometimes be the same person that is specified in the subject of the statement to be comprehended. However, they could include many other individuals as well.
In each case, the speed with which a previously formed knowledge representation reaches threshold depends on factors specified in the retrieval
postulates described in chapter 2. That is, it increases with the number of
features of the representation that are also contained in the probe set (Postulate 2.2) and also the degree to which the representation is already vibrating as a result of prior cognitive activity (Postulate 2.4). Note that if a
preexisting knowledge representation exists that contains both the probe
cues pertaining to the subject and the probe cues pertaining to the predicate, it is likely to be activated much more quickly than other representations that contain only one set of cues. Moreover:
Postulate 4.4. If a memory search activates a previously formed knowledge
representation whose features include both the subject probe cues and the
predicate probe cues, the new information is spontaneously recognized as redundant with prior knowledge and a new representation is not formed.
When the two sets of probe cues activate different knowledge representations, it means that a previously formed knowledge representation of the
information does not exist. Then, the process we assume is somewhat more
complicated. In this case, the probe cues associated with the subject are
likely to activate a number of knowledge representations that contain them,
some of whose features are more likely to overlap than others. In addition,
the predicate probe cues activate all of the models that explicate this action or state of affairs. (If the stimulus is Michael Jordan sank a threepointer, for example, it would activate all preexisting situation models
whose predicate was sank a three-pointer or the equivalent.) The subjects of these situation models could vary, but they are likely to have a subset of features in common. In the composite set of subject features that are
activated, the features that are represented most often are weighted most
heavily. Then, the following postulate applies:
Postulate 4.5. If no previously formed model (i.e., knowledge unit) refers to
both the subject and predicate of a target statement, the features that are activated by the targets subject are compared to a weighted composite of the
features associated with the targets predicate. This comparison yields a global
estimate of similarity, S. A new situation model is formed if either (a) S is greater
than a minimum threshold value, TC, or (b) the targets subject contains a subset
of features that are common to the subject of all models activated by the probe.
In the latter case, the new model is constructed by combining the composite of
features activated by the subject and the composite of features activated by
the predicate. In all cases, however, only the features that are weighted heavily in
the composite are included in the new model that is constructed.
94
CHAPTER 4
This postulate obviously requires unpacking. Suppose the target statement is Michael Jordan sank a three-pointer. According to Postulate 4.3,
features of the Subject resonate with all previously stored representations
containing these cues, whereas sank a three-pointer resonates with all
previously formed event models of this activity. However, several of these
models are likely to have Michael Jordan as the subject. Then, according to
Postulate 4.4, these representations should be activated by resonance from
both the targets subject and the targets predicate, and the statement
should be regarded as redundant with prior knowledge.
However, suppose the target statement is Michael Jordan rode a motorcycle. In this case, the predicate may not activate models whose subject is
Michael Jordan. However, features that are frequently associated with motorcycle riders are big, muscular, and athletic. Because these features
are likely to be conveyed in the generalized person representations that are
activated by Michael Jordan, the similarity, S, should be above threshold
TC . Therefore, a new model is likely to be constructed.
In contrast, suppose the target statement were Woody Allen rode a motorcycle. In this case, the person representations activated by Woody Allen are less likely to include big, muscular and athletic. Consequently, the similarity index computed on the basis of a comparison
features and the composite set activated by rides a motorcycle may be
below threshold. Nevertheless, the features of Woody Allen are likely to
include those that are common to all previously formed models of motorcycle riders (e.g., human). Thus, a new model of this statement might still be
formed, but it would not be constructed as quickly as it would if statement
had referred to Mr. Jordan.
Finally, suppose the target statement was A hippopotamus rode a motorcycle. In this case, the similarity index would not only be below threshold, but the features of the subject would not be shared by all previously
formed models of rode a motorcycle. In this case, therefore, a situation
model of the statement would not be constructed spontaneously.
The implications of this analysis for comprehension become clear in the
context of the next postulate:
Postulate 4.6. A statement is comprehended spontaneously if it is redundant
with prior knowledge or if a new model is constructed of it. If neither is the
case, comprehension of the statement is not spontaneous, but requires (Executor-controlled) goal-directed processing.
95
96
CHAPTER 4
97
Validation
Time
Difference
1.35
1.64
1.55
1.86
0.20
0.22
1.55
2.07
0.52
1.68
1.88
0.20
1.84
1.99
0.15
2.17
2.23
0.06
Experiment 1
Exemplar-referent statements
Definitely true (e.g., Mozart wrote a Symphony)
Definitely false (e.g., Mozart played rock and roll)
Unknown validity (e.g., Jane Fonda read her
horoscope)
Category-referent statements with no known exemplar (e.g., a Musician did aerobics)
Experiment 2
True exemplar-referent statements (e.g., Jane Fonda
did aerobics)
Category-referent statements with known exemplar
(e.g., An actress does aerobics)
Note. Response times are in seconds. Adapted from Wyer and Radvansky (1999).
98
CHAPTER 4
99
statement Abraham Lincoln wore a beard is comprehended and immediately recognized a true. By the same token, Abraham Lincoln shot John F.
Kennedy is immediately recognized as false. These assessments occur
spontaneously, in the absence of any goal that requires them.
These phenomena and the conditions in which they occur are taken into
account in the formulation we propose. According to the theory, the recognition that a statement is true or false is often a necessary by-product of the
comprehension processes embodied in Postulates 4.4 and 4.5. This is most
obvious when peoples comprehension of a statement activates a preexisting situation model of the event to which it refers. In this case, awareness of the prior existence of the model is typically sufficient to affirm its
validity. There may be exceptions to this. Situation models can obviously
be constructed of fictitious events (e.g., Santa Claus came down the chimney on Christmas Eve). However, the representation of these events is
likely to contain a tag denoting their fictional character. In the absence of
this tag, statements that instantiate an existing situation model may be
spontaneously recognized as true in the course of comprehending them.
The following postulate formalizes this assumption.
Postulate 4.7. A statement is spontaneously recognized as true in the course
of comprehending it if (a) a situation model of the event or state described already exists in memory and (b) no tag has been attached to the model that
designates it as false.
100
CHAPTER 4
101
and the time to comprehend them were even more similar. Thus, these results are quite consistent with the formulation we propose.
Further Implications
The results we have brought to bear on the comprehension and validation
of propositions were obtained in a rather artificial research situation. However, results of research in other areas are consistent with this conceptualization. Of particular interest in this regard are the implications of Postulate 4.7. To reiterate, this postulate states that if a pre-existing situation
model of a statement exists in memory, and if no tag has been attached to
the model that denotes it a false, the statement will become spontaneously
as true. Thus, suppose participants construct a new situation model of a
statement in one context in the course of comprehending it but do not
spontaneously assess its validity. Later, however, they encounter the same
statement in a different context. To comprehend the statement, they presumably retrieve the situation model they constructed of it earlier (Postulate 4.4). However, because the statement is comprehended in terms of this
model, it should be spontaneously identified as true (Postulate 4.7). Put
more simply, the exposure to and comprehension of a statement at one
point in time may increase beliefs in the statements validity at a later time.
The study by Hasher et al. (1977) described in chapter 3 provides direct
support for this prediction. That is, people who had reported their beliefs
in statements about unfamiliar events and situations in an initial experimental session increased their beliefs in the statements when they encountered
them in a second experimental session 1 week later. Presumably, participants comprehension of the statements in the initial session led them to
construct situation models of the events described by the statements. Consequently, they retrieved and used these models to comprehend the same
statements when they encountered them later. However, the pre-existence
of these models in memory should increase inferences of the statements
validity, as Hasher et al.s findings indicate.
These phenomena may pertain not only to the verification of a single
event, but also to sequences of events of which multiple-segment episode
models are constructed. In a study by Green and Brock (2000), participants
read a story about a violent crime. They were clearly aware that the story
was factual. Nevertheless, they later estimated the incidence of story-related events in the real world to be higher than participants who had not
read the story.
In a similar vein, Shrum and his colleagues (OGuinn & Shrum, 1997;
Shrum, Wyer, & OGuinn, 1998) found that heavy television viewers were
more likely than light viewers to overestimate the incidence of crimerelated events in the real world. These quite different sets of studies con-
102
CHAPTER 4
verge on the possibility that people, either in the laboratory or when watching television, construct episode models of the events they encounter and
store these models in memory. Later, they are likely to retrieve the models
when contemplating events that are redundant with their implications, and
therefore estimate the events to be more likely to occur. Similar conclusions can be drawn from research on the effects of generating an explanation for hypothetical events on later beliefs that the event will occur (Ross,
Lepper, Strack, and Steinmetz, 1977; Sherman, Skov, Hervitz, & Stock, 1981).
We discuss these studies in more detail in chapter 9.
An Alternative Conceptualization
The conceptualization proposed in this chapter should be considered in
the context of a formulation proposed by Gilbert (1991) on the role of comprehension on beliefs. Gilbert hypothesized that belief formation is a twostep process, which first involves comprehension and only later involves
verification. Moreover, in order to comprehend a statement, one must necessarily entertain the possibility that it is true. Thus, for example, comprehension of the statement George W. Bush donated money to the Taliban
entails a tentative acceptance of this state of affairs, and only later is the
statements validity rejected. One implication of this conceptualization is
that people who have been asked to learn true and false statements under
conditions of high information load are more likely to recall false statements as true than to recall true statements as false. This is presumably because the heavy processing demands prevent the second (verification)
stage of processing from occurring. Thus, the statements are likely to be
judged as true, based on results of the first stage of processing alone. As a
consequence, objectively true statements are evaluated correctly but objectively false statements are not.
The processes that Gilbert postulates to occur at the comprehension
stage are compatible with those assumed by the conceptualization proposed here. That is, the comprehension of a statement about an event involves the construction of a mental simulation of it (i.e., a situation model).
Moreover, the storage of this representation in memory increases its likelihood that a statement asserting the event it describes will later be judged
as true (Postulate 4.8). On the other hand, this postulate distinguishes between the comprehension of a statement and the criteria that are used to
estimate its validity. Moreover, it specifies the conditions in which the latter assessment is made spontaneously rather than at a later, postcomprehension stage of processing. Our conceptualization and Gilberts are therefore compatible with one qualification. That is, according to the proposed
model, statements that differ substantially from the subset of previously
formed models of the events they describe are likely to be below falsifica-
103
tion threshold (TF). These statements should be judged as invalid spontaneously at the time the statements are comprehended and should later be
reported as false regardless of information processing load. Thus, for example, the proposed theory predicts that the statement Abraham Lincoln
landed on the moon would be spontaneously judged as false regardless of
processing load. Gilberts formulation does not seem to make this prediction. With this exception, however, the two formulations are consistent.
104
CHAPTER 4
meaningful event are all likely to be sufficiently similar to those of a previously formed event model to activate it. Thus, for example, the features extracted from the observation of someone taking off his coat may all resonate with a static event model of this behavior, leading the model to be
accessed and used in much the same manner it is used to comprehend a
verbal description of the event.
However, the features of an observed event sequence also include aspects of the situational context in which the sequence occurred. These features can also resonate with situational features of preexisting event and
episode models in Permanent Storage. The subset of situational features
that are most typical of these models could constitute a situational appraisal of the sort described earlier in this chapter. However, because situational features are common to all of the specific events that are observed,
these features, along with others that occur frequently in the models that
have been activated, may be transmitted to the Work Space as a separate
representation that is independent of the more specific event models that
are constructed. This situational representation could operate in several
ways. First, if the events that are specified in the activated models are relatively homogenous, the situational appraisal defined by these features
could give rise to expectations for the sorts of events that are likely to occur in the situation at hand. Second, the appraisal, once it is transmitted to
the Work Space, might be included among the eliciting conditions of a production of the sort described in chapter 1, and thus could stimulate the observers own behavior in the situation. This possibility, which could help to
account for the different effects of concept activation on behaviors described in chapter 3 (cf. Bargh et al., 1996; Colcombe & Wyer, 2001), is considered further in chapter 7, where we discuss the situational norms that
govern expectations for the communications that are exchanged in social
interaction and the impact of deviations from these expectations on the interpretation of these communications.
CONCLUDING REMARKS
To summarize, the proposed formulation specifies the conditions in which
a statement about a situation-specific event or state of affairs is spontaneously identified as true or false in the course of comprehending it. Our analysis has largely been restricted to statements about events and states that
are temporally and situationally constrained (i.e., those that are comprehended by constructing a situation model). However, similar processes
should guide the comprehension and verification of other types of statements as well. Be that as it may, the implication that many statements are
verified spontaneously in the course of comprehending them has further
105
implications. As implied by Postulate 4.9, statements that are spontaneously verified as true or false are transmitted to the Work Space along with
a designation of their validity, and this spontaneous validity assessment
may be subject to further processing. For example, communications that
are obviously true or obviously false may violate normative principles of
conversation, and therefore may stimulate a conscious effort to understand
why the communication was uttered. This additional processing may sometimes lead to a reconstrual of the statements meaning that differs from the
meaning implied by the situation model that was constructed in the course
of comprehending it. We elaborate this possibility and its implications in
chapter 6.
However, the discussion of comprehension processes in this chapter has
focused on the representations formed of statements about single events
or, alternatively, on event models. Equally important considerations surround the way in which event models are combined to form a multiplesegment episode model of a thematically related sequence of events. The
next chapter is devoted to these matters.
C H A P T E R
5
The Construction of Episode
Models and Generalized
Narrative Representations
107
novel or see in a movie. Rather, they are likely to construct several different
episode models, each pertaining to a given sequence of events that are thematically as well as temporally related. Thus, the statements John packed
his bags. John left for the airport. may be represented in a two-segment episode model because they refer to the same person and are both related to
going on a trip. In contrast, the sentences John packed his bags. Mary watered the flowers. are likely to be represented in different models because
they appear to have little to do with one another. These considerations are
captured by a more general postulate:
Postulate 5.1. Once a situation model has been formed, subsequent information is integrated into the model if and only if this information concerns the
same situation as the model or a thematically related one.
108
CHAPTER 5
To the naive reader, the events described in the passage appear to have little to do with one another. However, a reader who has prior knowledge that
the woman is trying to get rid of her noisy neighbors is much better able to
understand the causal and thematic relatedness of the events.
The question is how to determine empirically whether episode models
are formed of information and, if so, how many. A procedure for determining this was developed by Radvansky and Zacks (1991) in investigating the
109
110
CHAPTER 5
111
FIG. 5.1. Mean time to verify a statement as a function of (a) the number of persons associated with a given object (denoted by squares) or (b) the number of objects associated with a given person (denoted by circles).
The two top panels show responses to items pertaining to buying objects that could be found in either a single
location (a drugstore) or different locations. The two bottom panels show responses to items pertaining to
owning these objects. RT = response time. Based on data reported by Radvansky et al. (1997).
112
CHAPTER 5
that took place in the course of the party. One episode consisted of events
that were causally related and occurred in the same situation. Specifically,
someone bumped a guests arm as he was reaching for an hors doeuvre,
causing him to spill his drink on a womans dress. The woman thereupon
called the offender a disparaging name and stalked off to the bathroom to
clean up. In contrast, the events composing a second episode were causally
related but occurred in different physical locations. That is, it concerned a
guests telling a story about Willa, who learned that her father was dying,
packed her bags and got on a plane to San Francisco, had three drinks on
the plane to calm herself, but felt tipsy by the time the plane landed. Consequently, she forgot the name of the hospital where her father was staying
and wound up crying on the streets of San Francisco.
One of the episodes was conveyed nearer the end of the party scenario
than the other. However, the order in which the events composing each episode were described was varied, sometimes being in chronological order
and in other cases being in the reverse order (e.g., Willa had cried on the
streets of San Francisco. She had forgotten the name of the hospital where
her father was staying. She was feeling tipsy as a result of having three
drinks on the plane . . .). Moreover, the events were all mentioned in sequence or were separated by other, unrelated events that occurred at the
party and temporarily distracted the visitor from those involved in the episode being described.
Participants after reading the entire cocktail party narrative and a short
delay were asked to recall what they had read. Participants typically recalled the events composing each target scenario in chronological order regardless of the order in which they had been presented and regardless of
whether the descriptions of them were separated by unrelated events. This
suggests that they integrated the events in each scenario into a multiplesegment episode model that they later used as a basis for recalling what
had occurred, and this was true regardless of whether the events occurred
in the same physical situation or in different situations. However, they recalled the last story mentioned in the narrative before recalling the one
that occurred earlier. Thus, participants formed separate episode models
of the two event sequences despite their occurrence in the same general
situation (the party), perhaps assigning them a common header (e.g., the
cocktail party). Thus, the residual resonance of the more recently constructed episode model led it to be the first one retrieved and reported
(Postulate 2.4).
Temporal Coding of Episode Models
The conclusion that thematically unrelated sequences of events are organized into different episode models and are stored separately in memory is
quite plausible. But if this is the case, how can the temporal order of the se-
113
114
CHAPTER 5
115
events is conveyed in isolation, the different situations appear more distinct, and so different models may be constructed.
Effects of Familiarity
An additional factor that might determine the extent to which a sequence of
events is divided into different episode models is the similarity of the sequence to ones that participants have been exposed in the past and are already represented in memory. For example, it seems intuitively reasonable
to suppose that familiar sequences are encoded into broader conceptual
units than novel ones (Newtson, 1973, 1976). Thus, multiple-segment episode
models may be formed more frequently in the first case than the second.
Stan Colcombe and I (reported in Wyer et al., 2002) examined this possibility using procedures similar to those employed by Radvansky and his
colleagues and described earlier. We constructed four stories, each composed of events that occurred in three related situations. The events in two
stories were fairly novel; one described Willas ill-fated trip to visit her sick
father and a second described a womans encounter with a bear while
camping. The events composing the other two stories were routine, describing a visit to a restaurant (with scenes pertaining to entering, eating,
and paying) and going to a movie (with scenes pertaining to buying tickets,
eating some popcorn, and seeing the show).
Three versions of each story were prepared. Each version described six
events, with the number of events taken from each scene (1, 2, or 3) varying
in a counterbalanced design. (Thus, one version of the restaurant story described one event that occurred while entering, two events that occurred
while ordering, and three events that occurred while paying; another described three events that occurred while entering, one while eating, and
two while paying, etc.) Participants read the stories under one of two conditions. In other-referent conditions, the protagonist in the story was a hypothetical person. In self-referent conditions, participants were told to imagine
themselves experiencing the situations described, and the protagonist was
referred to as you.
After learning the stories to criterion, participants were asked to verify
the events that were mentioned in each story. We assumed that the extent
to which verification time increased with the number of events described in
each scene of the stories indicated the extent to which participants formed
a single episode model of the scene or separate models. Pooled over otherreferent and self-referent conditions, the time to verify novel events increased with the number presented in the scene to which they pertained
(1.23 s vs. 1.35 s when 1 vs. 3 events were presented, respectively), suggesting that a different model was formed of each scene. In contrast, the time to
verify familiar events did not depend on the number presented (1.48 s vs.
116
CHAPTER 5
1.40 s, respectively). These data therefore confirm the expectation that participants are more likely to construct a single episode model from a series
of familiar events than from a sequence of novel ones.
In addition, the effects of set size were contingent on the nature of the
protagonist. Specifically, verification times increased with the number of
events presented in a scene when the protagonist was described as you
(1.32 s vs. 1.41 s, when 1 vs. 3 events were presented, respectively). This
was not the case, however, when hypothetical others were involved (1.40 s
vs. 1.33 s, respectively). Thus, participants were apparently less likely to
construct a single multiple-segment episode model of events that they
imagined themselves experiencing than events that they imagined an unfamiliar other experiencing, and this was true even when the situation described was very familiar. These results suggest that the criteria that people employ in comprehending self-referent events may differ from those
they apply when comprehending information about unfamiliar others. More
direct evidence of this difference (Colcombe & Wyer, 2002) is described in
chapter 6.
117
event prototypes (Colcombe & Wyer, 2002) or event schemas (Graesser &
Nakamura, 1982; Wyer & Srull, 1989), can function as implicit theories about
the events that occur in the type of situation at hand. As we note in the next
chapter, however, the abstraction of generalized representations from specific exemplars may occur less frequently than is often assumed. Colcombe
and Wyer (2002) showed that people typically do not form generalized narrative representations of events on the basis of exemplars unless they are
exposed to several different exemplars of the generalized narrative in temporal proximity. Even then, the abstraction may not occur unless people
are stimulated to think about the exemplars in relation to one another.
Thus, although generalized event representations undoubtedly exist in
memory, they may be constructed in other ways.
In fact, many generalized event representations might not be based on
personal experience at all. Rather, people might acquire a general understanding of the causal relations among event concepts through social learning, without directly experiencing the individual events that exemplify
these concepts. Thus, a child may be told that people who study hard get
good grades in school, and that these grades help them to get a good job
upon graduation. This sequence of events could constitute an implicit theory that the child accepts without first-hand knowledge of any specific instance in which the theory holds.
Finally, people may construct generalized event representations in the
course of making a judgment to which it is relevant. For example, people often wish to explain why a past event occurred, or to predict a future event.
Alternatively, they might wish to assess the desirability of a particular
course of action. These events could either be situation specific or more
general. For example, one might wish to explain either why a particular
friend got AIDs or, alternatively, why the incidence of AIDs in underdeveloped countries is increasing. Because the processes of constructing these
generalized representations are somewhat less obvious than the process of
constructing other types, these processes and their implications are worth
discussing in somewhat more detail.
General Considerations
To explain or predict a specific event, people might attempt to identify a
pre-existing episode model in which the event is embedded and use segments that precede or follow this event as bases for their inference. When
such a model does not exist, however, a two-stage process may be required. First, people identify a generalized propositional representation of
the event to be explained (i.e., people get AIDs). Then, having done so,
they search memory for propositions that describe possible antecedents of
it and, if they find one, they base their explanation on this proposition.
118
CHAPTER 5
119
Thirty-two of the propositions comprised 16 pairs of assertions that occupied the positions of A and B in a causal statement of the form, If A, then
B. Moreover, the likelihood that A was true, and the clarity of its implications for B, were systematically varied over the 16 pairs. (Thus, for example, a pair in which A was likely to be true and that had clear implications
for B consisted of the statements, Trucks conveying heavy cargo destroy
highways, and The weight limit on truck cargo will be decreased. In contrast, a pair in which A was likely to be true and had unclear implications
for B was composed of the statements, There will be no national borders
by the end of the year 2000 and War will be a thing of the past by the end
of the [20th] century.) This task presumably increased the accessibility of
the propositions in memory and, therefore, increased the likelihood that
participants would later identify and use them for purposes for which they
were applicable.
Participants then completed a two-part questionnaire in which they were
asked to report their beliefs in half of the propositions they had encountered in the comprehension task (i.e., the likelihood that the events described would occur) and to estimate the desirability of the events described in the remaining ones. This was done in such a way that for a given
AB pair, participants reported either (a) their beliefs in both A and B, (b)
the desirability of both propositions, (c) their belief in A and the desirability of B, or (d) their belief in B and the desirability of A. They were then dismissed but returned for a later session 1 week later, at which time they
were asked to recall as many of these propositions as they could.
We expected that participants who were asked to report their belief in B
would search memory for an antecedent of the event. Moreover, they
should use the first relevant antecedent that comes to mind (Postulate 3.1;
see Henninger & Wyer, 1976). Therefore, if A had been encountered in the
comprehension task and had clear implications for B, participants should
identify and use it as a basis for their judgment, thereby establishing an association between A and B that might otherwise not have existed. Similarly,
participants who were asked to report the desirability of A should search
for a consequence and identify B, and this should also establish an association between the two propositions. In contrast, participants who are asked
to report their belief in A should not identify B. Correspondingly, people
who are asked to estimate the desirability of B should not consider A. In
these conditions, therefore, an association between A and B should not be
formed.
The strength of association that participants formed between A and B in
the first session of the experiment should be reflected in their recall of the
propositions later. Specifically, suppose participants had established an association between A and B. Then, if they recall B during the second session,
they should think of A as well. In other words, their likelihood of recalling A
120
CHAPTER 5
TABLE 5.1
Strength of Association Between Syllogistically Related Propositions as a Function
of the Plausibility of the Antecedent (A), the Clarity of its Implications for the
Consequence (B) and the Type of Ratings Made of A and B
Rating Made
of Consequence (B)
Rating Made
of Antecedent (A)
Likelihood
Desirability
Likelihood
Desirability
.734
.462
.598
.505
.156
.331
.442
.274
.358
.798
.345
.572
.320
.375
.347
.439
.402
.420
.366
.482
.424
.393
.295
.344
Note. Numbers refer to the difference between the probability of recalling A given that B was recalled and the probability of recalling A given that B was not recalled. Adapted from Wyer and Hartwick
(1984).
if they recall B should be greater if they have recalled B than if they have
not. The difference between these two conditional probabilities therefore
provided an index of the strength of association that people had formed between A and B in the first session of the experiment.
This index is shown in Table 5.1 as a function of (a) the a priori likelihood
that A was true, (b) the clarity of As implications for B, and (c) the type of
rating that participants had made of the propositions in the first session of
the experiment (likelihood vs. desirability). When A had clear implications
for B, it was more strongly associated with B when participants had reported either the likelihood that B was true (.598) or the desirability of A
(.572) than when they had reported either the likelihood that A was true
(.358) or the desirability of B (.331). When As implications for B were unclear, however, these contingencies were not apparent. Thus, when As implications for B were not obvious, participants were less likely to recognize
their relatedness during the first session of the experiment, and so they
based their judgments on other criteria. As a result, an association between
the two propositions was not formed, and so their recall of B had no impact
on the likelihood of recalling A.
Other Determinants of Association
Wyer and Hartwicks (1984) findings obviously did not provide a complete
picture of the associations that are formed among general propositions of
the sort that compose an implicit theory or implicational molecule. McGuire
121
and McGuire (1991) identified several other factors that can exert an influence on these associations. For example, they postulated that people are
likely to think about antecedents of events over which they feel they have
control, but are more inclined to think about consequences of events over
which they have little influence. Therefore, because people feel they have
more control over the events in their personal lives than over events in
the society at large, they tend to think about antecedents of personal life
events (e.g., having a paper rejected for publication) but are more likely
to think about consequences of social events (e.g., a terrorist bombing).
For possibly similar reasons, people are disposed to think about the antecedents of events that are likely to occur in the distant future but to think
about the consequences of immediately impending events over which they
have less control.
Other motivational factors could also influence the type of associations
that people form. For example, people may be motivated to think about desirable rather than undesirable consequences of events that are likely to
occur, and to think about factors that facilitate rather than inhibit the occurrence of events they consider desirable. Although the effects of these
factors were not consistently supported by the McGuires data, their potential importance should not be entirely dismissed.
Indeed, a study by Albarracin and Wyer (2001) provided evidence of
these effects. Briefly, participants who had been induced to feel either
happy or unhappy by writing about a past experience read a persuasive
message containing either strong or weak arguments in favor of performing
a particular behavior. This was done in either the presence or absence of
situational distraction. After reading the message, participants reported
their attitudes toward the behavior being advocated, followed by estimates
of the likelihood and desirability of specific behavioral consequences. Participants who were not distracted based their beliefs and evaluations of the
consequences on the content of the message, and these beliefs and evaluations determined their overall attitude toward the behavior. In contrast,
participants who were distracted from thinking about the communication
based their attitudes on the extraneous affect they were experiencing and
misattributed to their feelings about the behavior. Once these attitudes
were formed, however, they had reciprocal effects on outcome-related beliefs and evaluations. For example, participants with favorable attitudes reported stronger beliefs that desirable consequences would occur, and reported that likely-to-occur outcomes were more desirable.
The Construction of Implicational Molecules
Propositions that become associated as a result of the processes described
in this chapter constitute an implicit theory about the causal relatedness of
the events they describe that can be used to comprehend and make infer-
122
CHAPTER 5
ences about more specific experiences. A conceptualization of these inference processes was mentioned briefly in chapter 1. To reiterate, sets of general propositions that become associatively linked in memory can form
implicational molecules that are stored in memory as conceptual units and,
therefore, can be brought to bear on the comprehension of new information that exemplifies them (Abelson & Reich, 1969; Wyer & Carlston, 1979,
1994). In combination, the propositions that compose a molecule can exemplify a general principle that one uses to comprehend and make inferences
about the world in which one lives. Many types of molecules can potentially
exist.10 The molecules of primary concern in the present context consist of
two or more verbally coded propositions, each describing a different event
or state of affairs. In combination, the segments of a molecule exemplify a
generalization about the world in which we live. For example, the generalization that smoking causes lung cancer might be represented in the twosegment molecule:
[P smokes; P has (will get) lung cancer].
Other generalizations may have moral overtones. For example, the generalization that people get what they deserve might be conveyed in a justdeserts molecule; i.e.:
[P does a bad thing; P encounters misfortune]
or, alternatively,
[P does a good thing; P has good fortune].
Not all implicational molecules pertain to events. Some, implied by cognitive balance theory (Heider, 1957), may concern the sentiment relations among a group of individuals (Picek,
Sherman, & Shiffrin, 1975; Sentis & Burnstein, 1979). Other, syllogism molecules may specify
the logical relations among propositions (Loken & Wyer, 1983). These possibilities, and empirical evidence bearing on them, have been summarized elsewhere (cf. Wyer & Carlston, 1979,
1994).
10
123
done something wrong. A more formal statement of these inference processes is provided in chapter 10, where we discuss the impact of implicit theories on judgments and decisions in some detail.
The associations that give rise to the construction of an implicational
molecule can often be formed as a result of the cooccurrence of its constituents in the course of comprehending information and making judgments.
However, they can also be motivationally determined. As Kunda (1990)
pointed out, people can often be motivated to believe that they have desirable attributes and that these attributes will allow them to have a happy
and successful life. Therefore, they may often construct implicit theories
and implicational molecules that confirm this belief. In doing so, they may
selectively attend to new information, and selectively seek and retrieve previously acquired knowledge, that confirms their desire to maintain a favorable view of themselves and their world. This possibility is elaborated in
chapter 10, where we consider the motivational factors that underlie the
construction and use of implicit theories in more detail.
C H A P T E R
6
The Impact of Generalized
Event Representations on
Comprehension and Memory
124
125
126
CHAPTER 6
Preliminary Considerations
The assumption that people construct event prototypes is quite common in
research and theory on prose comprehension (e.g., Bower, Black, & Turner,
1979; Graesser, 1981; Graesser, Gordon, & Sawyer, 1979; Schank & Abelson,
1977). These representations are unlikely to be formed spontaneously, however. Rather, they are constructed in the pursuit of specific processing objectives that require them. This might be done in a manner similar to that
suggested by Hintzmans (1986) conceptualization of the construction of
schematic representations from exemplars. Once such a representation is
formed, however, people store it in memory, and can later recall and use it
to comprehend new experiences to which it is applicable.
Several theories have been proposed to address these matters (Bower et
al., 1979; Graesser et al., 1979; Graesser & Nakamura, 1982; Trafimow &
Wyer, 1993). Most of these theories employ a principle of cognitive economy similar to that implied by Postulate 4.4. That is, they assume that when
people encounter a sequence of events that is redundant with the events
contained in an event prototype, they usually do not construct a detailed
representation of these events. Rather, they retain only enough features
that permit them to reconstruct the events that occurred on the basis of
the prototype. However, the theories make different assumptions about the
nature of the features that are retained.
For example, Graesser et al. (1979; see also Bower et al., 1979) assume
that when a sequence of events can be predicted on the basis of a more
general event prototype, people typically do not store a representation of
these events in memory at all. Rather, they store only a pointer to the prototype (i.e., a set of retrieval cues that identify it) along with a set of translation rules that instantiate features of the prototype in terms of specific features of the exemplar. Thus, suppose people read that Joe went to Doms
Patio Villa for dinner, ordered lasagna and a bottle of Chianti, and paid
$15.95 for the meal. They might not construct an episode model of the
events. Instead, they might store a pointer to a prototypic restaurant scenario along with the translation rules, customer = Joe, restaurant =
Doms, meal = lasagna, drink = Chianti, etc. These translation rules,
along with the content of the prototype, would be sufficient to reconstruct
what had occurred. They could also be used to infer the occurrence of
other events that were not specified in the information but typically occur
in restaurants (e.g., that John looked at the menu or actually ate the meal).
Note that these inferred events cannot be distinguished in the representation from the events that were actually mentioned. As a result, errors may
occur in recalling the information later (Graesser et al., 1979).
A related conceptualization was proposed by Graesser and Nakamura
(1982). They suggest that when people read descriptions of events the
events that compose a prototype, they comprehend these events in terms
127
of the prototype and copy them into the representation they form, instantiating the generic features of the prototype in terms of specific ones
that are mentioned in the information. However, they do not include segments of the prototype to which the information does not refer. Thus, the
events that were specified in the information and those that were not mentioned can be more easily distinguished.
A limitation of both conceptualizations, however, surrounds the representation of events that are not instantiations of the prototype. The occurrence of these events in daily life is less often the exception than the rule.
For example, a man who visits a restaurant is likely to look at the menu, order the meal, eat and pay the bill. However, he might also ponder a painting
on the wall, say hello to a colleague who is seated at a nearby table, and
spill some of his wine on the tablecloth. The way in which these latter
events are integrated into the representation of the experience as a whole
is not clearly specified. In fact, the theories simply assume that the events
are appended to the representation as tags. To this extent, there would
be no way to reconstruct the point in the sequence that the events occurred. In many cases, however, this reconstruction is obviously possible.
A conceptualization developed by David Trafimow (Trafimow & Wyer,
1993) eliminates this ambiguity. Trafimow, like Graesser et al. (1979), assumes that people who encounter a sequence of events that exemplify a
prototypic sequence do not normally retain these events in the representation they form. However, there is an important exception to this rule. That
is, when a nonprototypic event occurs, people retain a prototypic event
that occurred in temporal proximity to it in order to localize its position in
the sequence. Thus, in our restaurant example, people who learn about
Johns visit to Doms might not normally retain the events that occurred.
However, suppose that while John was ordering his meal, he said hello to a
colleague at the next table. In this case, the person may retain a depiction
of Johns ordering the meal in the representation, thus permitting the point
at which the nonprototypic event took place to be reconstructed.
1. An Empirical Test
Trafimow designed an intriguing series of experiments to evaluate these
possibilities. Suppose that when participants interpret a sequence of events
in terms of a prototype, they only retain prototype-relevant events in the
representation they construct if these events are necessary to localize
nonprototypic ones. Then, the addition of nonprototypic events to a sequence should increase the number of prototypic ones that are retained
and, therefore, should increase the ability to recall these events at a later
point in time. On the other hand, suppose participants do not comprehend
the sequence in terms of a prototype, but rather, construct an episode
128
CHAPTER 6
model of the events. In this case, additional event descriptions are likely to
increase comprehension difficulty and, therefore, to decrease the likelihood of retaining the presented events in the representation that is formed.
In other words, the addition of the nonprototypic events might decrease the
recall of the other, prototype-related ones. (Evidence that the likelihood of
recalling a given event decreases as the total number of events increases is
consistent with the set size effects observed in other research paradigms;
see Anderson & Bower, 1973; Srull & Brand, 1983.)
To evaluate this possibility, Trafimow and Wyer (1993) identified several
situations with which college students were familiar (e.g., cashing a check,
photocopying a piece of paper, making tea, etc.) and constructed six events
that typically occurred in each. The photocopying sequence, for example,
included found the article, got some change, found a machine,
aligned the copy, put in the coins, and pressed the button. However,
several of the events composing each sequence were not unique to the situation at hand, but could occur in other situations as well. Therefore, without a prior indication of the situation being described, the nature of the situation was unlikely to be immediately identified. In addition to the
prototypic events, six unrelated events were identified that could plausibly
occur in the situation of concern (e.g., remembered he had to return a
phone call, ate a piece of candy, saw a person he knew, etc.).
Stimulus stories were then constructed by combining the prototyperelated events and the prototype-unrelated ones. In one experiment (Trafimow & Wyer, 1993, Experiment 2), the stories contained four prototyperelated events and either two or six unrelated ones. In some cases, the
story was introduced with a statement that identified the situation (e.g.,
John needed to photocopy a paper) and in other cases, the situation was
not indicated until the end (After he had photocopied the paper, John
left). Participants read four stories that differed in terms of (a) the number
of prototype-unrelated events presented, and (b) whether the prototype
was identified at the outset or not until the end. Then, after a short delay,
they were asked to recall all of the events they had read about.
Results were quite consistent with expectations. When the situation was
not identified at the outset, participants were unlikely to identify and use a
prototype to comprehend the information. In this case, increasing the number of nonprototypic events presented decreased the proportion of prototypic events that were recalled (from .218 to .195, when 6 vs. 2 unrelated
events were presented, respectively). When the situation was identified at
the outset, however (thus stimulating participants to comprehend in the information in terms of a prototype), increasing the number of nonprototypic
events presented increased the proportion of prototypic events that participants recalled (from .179 to .279 when 2 vs. 6 nonprototypic events were
presented).
129
More subtle implications of the theory were also confirmed. For example, if participants retain prototypic events in the representation in order to
localize the nonprototypic events, the associations they form between
these events should lead one event to cue the recall of the other. To this extent, participants recall of a prototypic event should be preceded by the recall of the nonprototypic event that came before it in the sequence. When
the prototypic situation was identified at the outset, this occurred with a
greater probability than would be expected by chance. When the situation
was not specified until the end of the story, however, this was not the case.
A more general prediction of the theory results from its implication that
when no prototypic-unrelated events are included in the sequence, no prototypic events at all should be retained in the representation that is formed.
Under these conditions, participants theoretically have no way of distinguishing between the prototypic events that were presented and those that
were not. Consequently, participants should not only have poor recall of
the presented events but also should tend to recall events as presented
when they actually were not mentioned. On the other hand, suppose several unrelated events are added to the representation, leading the presented prototype-related events to be retained. Then, these events can
later be distinguished from the prototypic events that were not presented,
and so intrusion errors should decrease. An additional study (Trafimow &
Wyer, 1993, Experiment 3) confirmed these predictions. Participants read
stories composed of four prototype-related events and either four unrelated events or no such events. When the situation was identified at the outset, participants recalled a greater proportion of prototypic events when
four unrelated events had been presented than when none had been presented (.21 vs. .13), confirming the results of the earlier study. At the same
time, the number of intrusion errors was significantly lower in the former
condition than the latter (0.25 vs. 0.96).
2. Methodological Implications
Trafimow and Wyers (1993) findings, which were replicated by Colcombe and Wyer (2002, Experiment 1), provide evidence of the impact of using event prototypes to comprehend situation-specific sequences of events.
However, the conditions in which this occurs may be more limited than one
might expect. For one thing, prototypic event representations are usually
not formed spontaneously. Rather, they are only constructed in order to attain a specific objective that requires them. Moreover, even if these representations have been formed and stored in memory, they may not always
be retrieved and used. As we assumed in chapter 4, people who encounter
a situation-specific sequence of events are likely to comprehend it with reference to a pre-existing episode model of the person and/or type of event in
130
CHAPTER 6
131
tensity. After carefully positioning the wave generators in a circle around the
bulb, she simultaneously discharged all of them. This short blast from many
different sources at low intensity separated the filaments and preserved the
glass bulb. Thus, the light bulb was repaired, and the womans physics experiments continued successfully.
Three other stories concerned (a) putting out the fire in a burning woodshed, (b) taking over a corporation, and (c) extinguishing an oil well fire.
Participants were exposed to either one or five of these context stories
followed by a target story. The target, which concerned an attack on the
stronghold of an evil dictator, had four versions. Two versions described
four theme-related events and either two or six unrelated ones. Two other
versions described four theme-unrelated events and either two or six related ones. For example, a story containing six theme-related events and
four theme-unrelated events was:
A small country was under the rule of an evil dictator who was despised by
the people. The dictator lived in a fortress with many roads leading to it like
spokes in a wheel. The dictator was cruel. Rebels realized that they had enough
forces to overrun the fortress if they could attack all at once. The rebel general
noted that the weather had been turning colder lately. A rebel spy reported back
that the dictator had planted mines that would allow only a few men to pass
at once along the many roads leading to the fortress. Given that all of the rebel forces must attack the fortress at once, it seemed as though the rebels
plans were foiled. Birds flocked in a nearby tree. However, the rebel general
was very smart and instructed his men to split up into several groups, each
taking a different road to the fortress. A dog howled in the distance. In these
small groups, the men could pass over the mines without setting them off and
132
CHAPTER 6
meet up again at the fortress. Thus, the rebels overthrew the dictator, and the
people in the kingdom lived happily ever after.
133
134
CHAPTER 6
Self
Different Person
.63
.49
.51
.45
.70
.57
.55
.45
135
themselves (.47 vs. .57) or a parent (.51 vs. .62), and regardless of whether
the stories were based on routines that participants themselves performed
(.53 vs. .67) or those that another had engaged in (.45 vs. .53). Thus, these
data suggest that participants did not use a generalized representation to
interpret either stories about themselves or stories about their parent.
2. The Use of Previously Formed Prototypes
to Comprehend Anothers Behavior
There are two possible reasons why participants did not use prototypic
representations in the preceding study. First, as noted earlier, people might
not form prototypic representations of their own and well-known others
behaviors. Another possibility, however, is that these generalized representations may exist in memory, but people may simply not draw upon them
for use in comprehending new experiences involving themselves and familiar others. A second series of studies support the latter possibility. Participants in these studies read stories similar to those employed by Trafimow
and Wyer (1993). That is, each story consisted of four behaviors that exemplified a prototypic script (cashing a check, changing a tire, etc.) and either
two or six prototype-unrelated behaviors. In one study (Colcombe & Wyer,
2002, Experiment 1), participants were told that the protagonist in the stories was a hypothetical person (Joe), as in Trafimow and Wyers original
studies. In a second experiment, however (Colcombe & Wyer, 2002, Experiment 4), they were told to imagine that the story was about their roommate.
(In this study, the protagonist was referred to as he or she, depending
on whether participants were male or female, respectively). In all cases, the
situation to which the sequence pertained was identified at the outset, thus
ensuring that participants aware of the applicability of the prototype for
comprehending it.
Results of the study are summarized in Table 6.2. When the stories referred to a hypothetical person, Trafimow and Wyers (1993) findings were
TABLE 6.2
Mean Proportion of Prototype-Related and Prototype-Unrelated
Events Recalled as a Function of the Number of Unrelated
Events Mentioned in the Story and the Type of Protagonist
Protagonist
Unfamiliar Other
Roommate
.30
.42
.50
.39
136
CHAPTER 6
replicated. That is, participants recalled a greater proportion of prototyperelated events when six unrelated events were presented than when two
were presented (.42 vs. .30, respectively), confirming the fact that participants used a prototype to comprehend the story. When participants imagined the events occurring to their roommate, however, their recall of the
prototype-related events was poorer when six unrelated events were presented than when two were presented (.39 vs. .50, respectively). In this
case, therefore, a prototypic event representation was not employed.
Concluding Remarks
Colcombe and Wyers findings suggest that even when a generalized event
representation exists in memory, people do not always use it to comprehend information about themselves or familiar others. This conclusion is
consistent with the retrieval postulates described earlier, and the results
reported by Wyer and Radvansky (1999). That is, people who receive information about the behavior of a well-known person can often construct a situation-specific event model involving this person very quickly. This is particularly true if a previously formed event model of the persons behavior
already exists in memory. In this case, they are likely to use this model as a
basis for comprehending the information rather than relying on more general knowledge that comes less quickly to mind.
In retrospect, this conclusion is not surprising. Unless people have a reason to think about different instances of their own or familiar others behavior in relation to one another, they are unlikely to engage in this cognitive
activity. In making behavioral decisions, a previously formed episode
model of the behavior and its consequences can often serve just as well as
an abstract prototype. Moreover, as noted earlier, routinized behavior may
be governed by productions and performed with minimal cognitive mediation. The influence of these productions is particularly evident when one is
distracted by thinking about something else. (For example, a person who is
thinking about how to revise a paper while driving to the grocery store may
suddenly find himself about to enter his office parking lot. Similarly, a person may often forget whether he has locked the door to the house before
leaving.) As Bargh (1997) contended, a very large proportion of our behavior may be governed by automatic processes, and there is little need to use
a generalized representation of this behavior that we have stored as part of
declarative knowledge.
Circumstances can nevertheless arise in which a prototypic event for example representation is personally useful. When a persons behavior has
undesirable consequences, the individual might be motivated to avoid
these consequences in the future and, therefore, may attempt to explain it.
To generate this explanation, the person might retrieve a number of epi-
137
138
CHAPTER 6
tially relevant to this decision. Rather, they retrieve only a subset of this information and do not search further unless their confidence in the validity of
the information or its implications is below threshold (Postulate 3.1; see also
Chaiken, 1987). This sufficiency postulate was applied earlier in conceptualizing the effects of knowledge accessibility on judgments. However, it has implications for memory as well. For example, if a generalized representation of
information has recently been formed in the course of attaining a specific
goal, this representation is likely to be used as a basis for not only making
judgments but also reconstructing the events that occurred. That is, the original representation may only be consulted if the implications of the more recently formed goal-specific representation are below confidence threshold.
A classic study by Loftus and Palmer (1974) provides an example. Some
participants who had been shown a picture of an automobile accident were
later asked to indicate how fast they thought the car was going when it hit
the tree. Others were asked how fast it was going when it crashed into the
tree. Participants estimated that the car was going faster in the second condition than the first. Moreover, when later asked to remember details of the
picture they had seen, participants in the second condition were also more
likely to report (incorrectly) seeing broken glass at the scene of the accident. Participants apparently reconstructed the scene of the accident in a
way that was consistent with implications of the question they had been
asked, forming a new mental representation that contained features in addition to those that were conveyed in the picture. Their later memory was
then based on this newly formed image independently of the one they had
constructed at the time they saw the picture.
In Loftus and Palmers study, however, participants processing objective stimulated them to add features to the representation they had formed
initially. When people wish to describe an experience to others, the representation they form is more likely to omit features that are of peripheral relevance to the description they are providing. This could also decrease the
accuracy of recalling information under conditions in which this representation is consulted. In a study by Schooler and Engstler-Schooler (1990),
participants were shown a series of human faces and asked to describe
some of these events verbally. Later, they were given a memory test for the
faces. One might expect that the increased amount of processing involved
in describing the picture would make it more memorable. In fact, however,
participants were less accurate in identifying the faces they had described
than faces they had seen but not described. Participants who wrote an abstract description of a face apparently formed a new representation of it in
which certain specific features that were useful in distinguishing it from
other faces were omitted, and then used this new representation, which
was less detailed than the original representation, as a basis for their later
recognition responses.
139
140
CHAPTER 6
141
they formed in the course of this activity to be sufficient for evaluating protagonists statements but not their nonverbal behaviors. Therefore, they
should resort to the use of episode models to verify these behaviors. Consequently, although their recognition of protagonists statements should be
adversely affected by the goal-specific representation they formed, their
recognition of the protagonists behaviors should not be.
3. Results
The results of both studies were very similar and were generally consistent with expectations. Table 6.3 shows the difference between recognition
accuracy under each communication objective condition and accuracy under comprehension-only conditions.11 Writing a description of the sequence
of events that occurred decreased recognition of both protagonists statements and their nonverbal behaviors, and this was true regardless of
whether participants had watched the movie with this objective in mind or
were not told of the objective until afterwards. In contrast, communicating
an impression of the protagonists only decreased participants recognition
of things the protagonists said and did not influence their memory for
things the protagonists did. Moreover, this decrease was only evident when
participants were not given an impression-formation objective until after
they had seen the movie.
The effects of objectives that participants had in mind at the time they
watched the movie deserve further attention. All of the statements and behavior that participants consider interesting enough to communicate to
others are likely to be of interest to themselves as well and, therefore, to be
represented in the episode models they would normally form spontaneously while watching the movie. Therefore, the content of the goal-specific
representation formed by participants with an event-description objective
should be similar regardless of whether they were aware of this objective at
the time they watched the movie or not. To this extent, the decrement in accuracy that results from their use of this representation as a basis for recognition should not depend on when this objective was imposed. This was
the case, as Table 6.3 indicates.
11
The index of recognition accuracy used in Experiment 1, which corrects for guessing, is
given by the equation:
11
P(Acc) =
where P(Hit) and P(False Alarm) are the probabilities of identifying an item as having been presented when it was or was not presented, respectively (Hilgard, 1951). This index could not be
applied to Experiment 2, as the number of nonpresented items was too small to obtain a reliable
index of P(False Alarm).
142
CHAPTER 6
TABLE 6.3
Effects of Communication Objectives, and the Point at Which
These Objectives Were Induced on Recognition of Protagonists
Statements and Nonverbal Behaviors
Experiment 1
Event-description objectives
Induced before watching movie
Induced after watching movie
Impression-description objectives
Induced before watching movie
Induced after watching movie
Experiment 2
Statements
Behaviors
Statements
Behaviors
.189a
.065
.360
.297
.070
.044
.046
.095
.130
.108
.000
.021
.045
.022
.046
.046
In contrast, participants who are told at the outset that they would be
asked to describe their impressions may include features in the representation they form while watching the movie that are relevant to their impressions but are not depicted in the episode models they would form if they
were only trying to comprehend what is going on. The implications of these
additional features may then be included in the communication they generate later, and their recognition may benefit, as results also suggest.
Effects of Reminders on Recognition
Suppose people after seeing a movie or observing an interaction are later
reminded of the events that occurred. They are likely to comprehend this
reminder with reference to a previously formed representation of the original experience. In the course of doing so, they presumably construct yet another representation of what went on. According to the conceptualization
we propose, the content of this new representation, and its consequent effects on memory, may depend on the nature of the previously formed representation that was used to interpret the reminder.
Adaval and Wyer (2003) investigated these possibilities in a study that
provides further support for the implications of Experiments 1 and 2. In this
study, participants who had watched the movie with instructions to comprehend it were then asked to describe their impressions of the protagonists. After doing so, however, they were dismissed with instructions to return the next day for a second session of the experiment. In this session,
143
144
CHAPTER 6
145
ries are stored in memory. These stories may then become the basis for
their self-perceptions. This means that the self-perceptions that people develop are likely to depend on the type of situations in which they are typically called upon to tell stories about themselves and the expectations
that others have for the content of these stories. Two bodies of research
and theory bear on these processes. One concerns the impact of parentchild interaction in the construction of self-stories. A second area of
research provides insight into cultural and situational influences on individuals personal and social identity.
146
CHAPTER 6
147
However, American parentchild co-narratives were more often characterized by a self-favorability bias (Miller et al., 2001). These stories focused much less frequently on a childs past transgression, and when they
did, the transgression was treated humorously, portraying the child in a
positive light despite his or her misdeed. Thus, for example, one story concerned a child who had written on the wall and lied about it. Although the
mother acknowledged the seriousness of the childs negative act, she at the
same time made light of it in helping the child communicate about the act
to others. Another example:
Tommy and his older brother remembered being punished for some misdeed
committed a few days earlier, but none of the participantsTommys brother
or mothercould remember what they had done wrong, a baffling if not inconceivable state of affairs from a Taiwanese perspective. This practice of down
playing transgressions in the narrative medium seemed to be part of a wider
set of practices that [European American] caregivers used to protect their
childrens self esteem. (p. 168)
An equally or more important difference in Eastern and Asian story telling was evident in the stories that parents told about themselves. Taipai
mothers self-stories typically provided positive examples for the child to
emulate. As Miller et al. (2001) suggested:
. . . Taipai mothers goal seems to be to encourage their children to live up to
their own high standards by citing concrete examples of admirable behavior
from their own or other family members lives . . . [T]eaching by example is
one of several traditional Chinese socializing strategies . . . (p. 170)
148
CHAPTER 6
stories, which emphasized positive outcomes while minimizing the seriousness of misdeeds, seem likely to induce a general promotion focus (Higgins, 1997), that is, a tendency to attach importance to positive outcomes
with little concern about negative consequences of behavior.
Unlike Americans, Chinese parents encouraged their children to focus
their stories on their misdeeds and to acknowledge responsibility for them.
Moreover, parents stories about themselves established a high standard of
excellence that their children were encouraged to emulate. In combination,
these socialization practices seem likely to induce a prevention focus,
that is, a concern with negative consequences of behavior. These general
differences seem likely to stimulate the development of different implicit
theories about oneself as responsible for negative outcomes and, moreover, in the desire to avoid these outcomes. In fact, this difference exists between European and Asian adults, and can have a profound impact on judgments and behavior (cf. Aaker & Lee, 2001; Briley et al., 2000; see chap. 3).
Cultural Influences on Socialization
The narrative representations that individuals acquire about themselves
and others are transmitted directly through parentchild interaction. As
Miller et al.s research testifies, the character of these individual narratives
can vary over cultures, and also over communities and subcultures within a
society. However, cultural differences can also exist in more general narrative representations that individuals construct. Mankowski and Rappaport
(1995) noted that the norms and values that pervade societies are governed
by a shared cultural narrative that not only serves to guide behavior in specific situations but influences individuals feelings of personal and social
identity. As they indicated:
. . . the phenomena of identity development and change may be understood in
terms of the appropriation of shared narratives into ones personal life story
on the one hand, and the creation of new narratives or modification of existing narratives on the other. (p. 213)
149
gious figure can also provide the basis for shared norms and values in societies in which the religion predominates.
The suggestion that individuals in a society share a dominant cultural
narrative does not imply that the entire narrative exists as such in the mind
of a single individual. Moreover, the influence of such a narrative on the
construction of self-narratives may not be completely conscious. Just as
features of the content of childrens self-stories are modified to conform to
the values of their parents or others to whom they are communicating, a
cultural narrative may imply general norms and values that individuals
consider socially desirable and, therefore, influence the content of the stories that individuals construct about themselves, leading some things to be
emphasized and others to be omitted. This may be done without a clear understanding of the roots of these norms and values. Moreover, the functional autonomy of cultural norms and values from their original content is
well established in Boorstins (1973) sociological analysis of the development of American thought. (Thus, e.g., the conviction that individuals have
the right to own firearms is a remnant of a period in the development of the
American West in which law enforcement was either nonexistent or ineffective, and individuals were required to take law into their own hands in order to survive.)
It seems reasonable to suppose that once self-narratives become tailored to cultural norms and values, they are later retrieved without a conscious consideration of the personal experiences on which they are based.
Thus, they ultimately become regarded as true by the storytellers themselves. As a result, they become part of individuals self-identity (McAdams,
1988) that are told to ones children and others independently of the original narrative representations of the events that were formed at the time
they occurred.
CONCLUDING REMARKS
The influence of generalized narrative representations on memory clearly
depends on the nature of the representation and how it is used. When
event prototypes of the sort referred to elsewhere as scripts (Shank &
Abelson, 1977) or event schemas (Graesser, 1981) are brought to bear on
the comprehension of verbal descriptions of events involving fictitious individuals, they can influence the content of the representations formed of
these events and stored in memory. However, these prototypes are much
less likely to come into play in comprehending events that people imagine
themselves or familiar others experiencing.
Generalized representations of single experiences that people form in
the course of telling or writing about these experiences may have a more
150
CHAPTER 6
C H A P T E R
7
Pragmatic Influences on the
Interpretation of Statements
Made in a Social Context
152
CHAPTER 7
GENERAL CONSIDERATIONS
Peoples perceptions of the pragmatic implications of a message can often
be stimulated by their own goals in the situation at hand. A woman who
must decide whether to accept a date might want to know whether the invitation reflects a genuine desire for her company or only a desire to have
sex. More generally, people who wish to form an impression of a person or
product on the basis of a communicators description of the referent may
need to assess the likelihood that this description is biased. They may attempt to identify the pragmatic implications of the communicators message in order to make this assessment.
On the other hand, people often spontaneously recognize that a statements literal meaning is not the meaning the communicator actually intends to convey. This occurs when the statements literal meaning violates
normative expectations for the content and style of communications that
PRAGMATIC INFLUENCES
153
The expectations for communications in a social situation can pertain not only to the content of the messages that are exchanged but the style in which they are transmitted. The extent
to which parties to a conversation elaborate answers to one anothers questions, or ask questions in return, may influence perceptions of the communicators attitudes toward both the conversation and one another (Wyer, Swan, & Gruenfeld, 1995). Nonverbal behaviors can obviously
play a role as well, as we note later in this chapter.
12
154
CHAPTER 7
PRAGMATIC INFLUENCES
155
veyed, in a way that is more consistent with the principle that is violated
that is, whether the statement is interpreted as redundant with prior knowledge or is interpreted as false. We consider these two possibilities in turn.
156
FIG. 7.1. Theoretical flow diagram of processing steps involved in identifying the implications of
statements that are recognized as true or false. The source and recipient of the statement are
denoted S and R, respectively.
PRAGMATIC INFLUENCES
157
until they reach the age of 18, and this is certainly a good thing, or Americans are not allowed to vote until they reach the age of 18, whereas citizens
of other countries can vote much earlier.). To this extent, recipients would
not change their belief in the validity of the statement, but instead might
change their attitude toward the state of affairs that the statement describes. Note, however, that these effects should only occur under conditions in which uninformative communications are counternormative. When
the messages are not intended to convey new information, they should theoretically be taken at face value, and attempts to make them informative
should not be evident.
Effects of Uninformative Statements
on Recipients Beliefs
Gruenfeld and Wyer (1992) investigated implications of the preceding analysis. Participants read a series of statements about social events. In some
cases, the statements asserted the validity of target propositions that college students were likely on a priori grounds to believe to be false (e.g.,
The CIA is engaged in illegal drug trafficking, Lyndon Johnson was responsible for the assassination of President John F. Kennedy, etc.). In
other cases, however, the statements denied the validity of these propositions (e.g., The CIA is not engaged in illicit drug trafficking, etc.) and,
therefore, were uninformative. Finally, some participants believed that that
the statements were taken from newspapers (whose purpose is to convey
new information), whereas others believed they were taken from an encyclopedia (whose purpose is to preserve archival knowledge). Participants
first judged the importance of the information conveyed in the statements.
Then, after doing so, they estimated their beliefs in both the target propositions and related ones (e.g., The CIA engages in other illegal activities,
Lyndon Johnson was an enemy of John F. Kennedy, etc.).
Statements that affirmed the validity of the propositions constituted new
information, whereas statements that denied their validity were redundant
with participants prior knowledge about the persons and events in question. However, we expected that when participants were told that the statements came from encyclopedias, they would accept the statements without
further consideration. In contrast, participants who were told that the statements came from newspapers should assume that they were intended to be
informative and, therefore, might infer that there was some reason to suppose that the propositions being denied were in fact true. To this extent,
they might actually increase their belief in these propositions.
This was in fact the case. Table 7.1 shows the difference between beliefs
in propositions under each experimental condition (reported along a 010
scale) and beliefs reported by control participants who were not exposed
158
CHAPTER 7
TABLE 7.1
Effects of Assertions on Beliefs in Target
Propositions and Related Propositions
Assertion Type
Affirmation
Denial
1.16*
1.81*
1.06*
0.48
1.15*
1.62*
1.20*
0.91*
Note. Cell entries refer to differences between beliefs reported after exposure to an assertion and context-free beliefs reported in the absence of the assertion. Differences denoted by asterisks are significantly greater than 0, p < .05. Adapted from Gruenfeld and Wyer (1992).
to either denials or affirmations. Affirmations that were taken from encyclopedias increased participants beliefs that target propositions were true,
whereas denials had little effect. These effects confirm the assumption that
affirmations constituted new information but denials did not. News headlines that affirmed the validity of the target propositions also increased beliefs that the propositions were true. (This increase was somewhat less
than it was when the statements came from encyclopedias, suggesting that
participants perceived newspapers to be relatively less credible.) On the
other hand, denials of the propositions validity that were taken from newspapers also increased participants beliefs that the propositions were true.
In fact, this increase was just as great as the increase produced by affirmations! Thus, participants attempted to interpret denials in a way that made
them informative, leading them to infer that there might be some reason to
suppose that the propositions were true, and this led the statements to
have a boomerang effect.
Effects of Uninformative Statements on Attitudes
As noted earlier, however, some uninformative assertions cannot possibly
be false. In this case, participants attempts to make them informative could
lead them to infer that they are implicit expressions of opinion. To this extent, the assertions might influence participants estimates of the desirability of the state of affairs being denied by the proposition rather than their
beliefs in their validity. Gruenfeld and Wyer (1992) reported evidence of this
as well. For example, exposure to the news headline, Wearing seatbelts is
required by law in Illinois significantly decreased participants opinion
that wearing seatbelts should be required. Similarly, exposure to the state-
PRAGMATIC INFLUENCES
159
ment, Many food products are sold in plastic and styrofoam containers
decreased subjects beliefs that there should be restrictions on the packaging of food product, and exposure to Not all college students drive American automobiles increased beliefs that college students should drive these
automobiles.
General Implications
1. Communication and Persuasion
The conditions constructed by Gruenfeld and Wyer (1992) provide examples of innuendo effects of the sort that are used intentionally by propagandists. Their use is particularly common in the political arena. For example,
a candidate might attempt to plant a seed of doubt about an opponents integrity, not by asserting that the opponent has engaged in illegal activities
but rather, by actually denying it. For example, the assertions I personally
have no evidence that my opponent embezzled funds while serving as an
executive of General Motors, or I do not for one minute believe that my
opponent is a card-carrying member of the Ku Klux Klan, which might normally go without saying, could stimulate people to suppose that the opponent if not an embezzler or a Klan member, is nonetheless likely to have engaged in activities that give rise to these suspicions.
A perhaps more familiar instance of the intentional application of the informativeness principle occurs in advertising. Suppose an advertisement
for Brand X asserts, X contains no hydratropine. Consumers may have no
idea what hydratropine is. In an attempt to make the statement informative,
however, they may infer that there is some reason why promoters of Brand
X made the statement. In particular, they may infer that (a) hydratropine is
undesirable and (b) other products have it. Correspondingly, the equally
meaningless assertion, X contains hydratropine might stimulate them to
infer that hydratropine is desirable and that brands other than X do not
have it.
2. Communication of Emotions
Similar effects occur in interpersonal communication. For example, a
woman might comment to her spouse, Its cold in here. If she has rarely
complained about the room temperature in the past, her husband may regard this as new information and take steps to remedy the situation. However, suppose the wifes remark is one of many similar comments she has
made over a period of time. In this case, the husband would consider the
statement to convey little new information about his wifes feelings. Rather,
he might infer that the statement is a veiled criticism of his compulsion to
save energy by keeping the thermostat at 65F. As this example points out,
160
CHAPTER 7
the recognition of a statement as uninformative, and the construal of its intended meaning, often requires a substantial amount of shared knowledge.
Watzlawick, Beavin, and Jackson (1967) noted that people in close relationships often acquire a private meaning system that permits them to convey
thoughts and feelings of which other listeners are unaware. This shared
knowledge presumably has its impact through its mediating influence on
perceptions of one anothers statements as uninformative and the reinterpretation of their literal meaning that results from it.
REACTIONS TO COMMUNICATIONS
THAT ONE BELIEVES TO BE UNTRUE
Statements that are transmitted to the Work Space by the Comprehender
may often contain tags that identify them as false or, at least, inconsistent
with the implications of prior knowledge (Postulate 4.6). These statements,
like statements that are redundant with prior knowledge, are likely to stimulate cognitive activity in an attempt to understand why they were made.
These processes may be similar to those that underlie responses to violations of the informativeness principle. However, the inferences that result
from the activity are somewhat different.
The processes that seem likely to underlie reactions to untrue communications can be summarized with reference to the left half of Fig. 7.1. Specifically, individuals who hear a statement they recognize as false (S < TF , as
defined in Postulate 4.6) are likely to assess whether the communicator
knows that this is the case. If the source is likely to believe it is true, the recipients may take the statements literal implications at face value. They
might also do so if they think that the source knows the statement is false
but is unaware that they share this knowledge. In this case, they might conclude that the source is trying to deceive them. However, suppose recipients believe that the source is aware of their knowledge. Then, they are
likely to infer that the source does not mean the statement to be taken literally, but rather, intends it to be ironic.
Reactions to ironic statements can depend on both their topic and
whether their literal implications are favorable or unfavorable. In some
cases, ironic statements are intended to be funny. In other cases, however,
they can be indirect expressions of hostility. These motives are sometimes
difficult to distinguish. Suppose a woman whose husband has had three
traffic accidents in the past 2 years comments to her guests, My husband
is a wonderful driver. The comment could be a true expression of opinion,
and an attempt to defend her spouses abilities despite his poor driving record. To the extent it is intended to be sarcastic, however, it could either
convey hostility or be meant as a tease. Which interpretation the guests
PRAGMATIC INFLUENCES
161
make is likely to depend on their view of the womans more general feelings
toward her spouse. If they believe that the couples relationship is basically
harmonious, they are likely to infer that the woman is teasing. If the couple
is known to have a history of marital conflict, however, they might interpret
the statement as hostile.
These observations are rather self-evident. The conceptualization outlined in Fig. 7.1b has greater interest, however, when a statements validity
is clearly a matter of opinion. Suppose a person comments, What this
country needs is another Ronald Reagan. This comment is more likely to
be ironic if it is made by a liberal Democrat than if it is made by a conservative Republican. On the other hand, conservative Republicans are less
likely than liberal Democrats to interpret the statement as ironic. That is,
they are more likely to believe that the statement is true, and to accept its
literal meaning as the intended meaning regardless of its source.
Accuracy Violations and Amusement:
The Identification of Witticisms
To reiterate, people often interpret a statement spontaneously as false in
the course of comprehending it (Postulate 4.6). When it is obvious to recipients that the source also believes the statement to be false and, therefore
interpret the statement as ironic, they are often amused. This was demonstrated empirically by Isbell, Wyer, and Collins (2002). Participants were exposed to a number of vignettes in which one person made a comment to another. In some cases, the comment was likely to be true. In other cases, the
comment was literally false, but if interpreted as ironic, was similar in
meaning to the first statement. Finally, in some cases, the intended meaning
of the statement was more favorable than its literal meaning, and in other
cases, it was less so. For example, one vignette described a discussion between two students about where they had spent the summer. In some versions, one student made a true statement that conveyed either a favorable
description of its referent (Switzerland is certainly a great place to spend
the summerall those lakes and high mountains) or an unfavorable description (Champaign, Illinois, is certainly an awful place to spend the summerall that corn and high humidity). In other cases, however, the students statement was literally false (Switzerland is certainly an awful place
to spend the summer . . . , and Champaign, Illinois is certainly a great
place to spend the summer . . .), but its intended meaning was similar to
that of the corresponding true statement. In another scenario, two professors discussed who was the most (or least) intelligent public figure in the
20th century. In this case, the literally true statements were Einstein gets
my vote (for most intelligent) and Dan Quayle gets my vote (for least intelligent) and the corresponding false statements were Einstein gets my
162
CHAPTER 7
vote (for least intelligent) and Dan Quayle gets my vote (for most intelligent).
Participants read one version of each vignette with instructions to estimate how amused they would be if they overheard the statement being
made under the conditions described. Pooled over vignettes, participants
regarded ironic statements as more amusing than literal ones (4.9 vs. 2.7),
and this was true regardless of whether the intended meaning of the statements was favorable (4.6 vs. 1.9) or unfavorable (6.3 vs. 3.5).
It would obviously be incorrect to conclude that all ironic statements are
amusing. Sarcasm can often not be amusing at all. A person who has just
been turned in by a classmate for cheating on an exam might respond to
the other, You are certainly a fine friend, which is intended to convey
quite the opposite. This comment is unlikely to elicit amusement. The necessary conditions for humor elicitation are discussed in chapter 8 (see also
Apter, 1982; Wyer & Collins, 1992). Nevertheless, the evidence that ironic
statements are more likely to elicit amusement than their true counterparts
is worth noting.
Effects of Invalid Communications
on Information Search
The sources of the ironic statements presented in Isbell and Wyers study
obviously believed that their statements were false, and so recipients were
likely to believe that the statements were intended to amuse. In many instances, however, people are uncertain of a sources beliefs in the statements that he or she makes. This is particularly true when the statements
express opinions that are somewhat controversial. People who encounter
statements of opinion that they personally believe to be invalid may often
attempt to seek additional information that helps them to understand why
the statements were made. In addition, they may review their prior knowledge about the issues to which the statements pertain in order to confirm
their belief that the opinion expressed is incorrect.
This cognitive activity is likely to be reflected in recipients ability to recall the information later. Information that is thought about more extensively with reference to prior knowledge is generally easier to recall (Craik
& Lockhart, 1972; Wyer & Hartwick, 1980). If this is so, however, people
should be more likely to recall opinion statements they spontaneously recognize as false than either statements they consider to be true or statements whose validity they have not spontaneously evaluated.
This possibility has not been investigated under conditions in which information is conveyed in conversations. If the proposed conceptualization
is valid, however, the cognitive activities that underlie peoples recall of
opinion statements they recognize as false should occur even when the
PRAGMATIC INFLUENCES
163
statements are conveyed out of their social context. Several studies bear on
this possibility. In a study of impression formation reported by Wyer, Lambert, Budesheim, and Gruenfeld (1992), participants received information
about a target person who was characterized as either a conservative Republican or a liberal Democrat and, in addition, was described by either favorable or unfavorable trait adjectives (e.g., kind, hostile, etc.). This
general description was followed by a list of both (a) behaviors that were
both evaluatively consistent and evaluatively inconsistent with the trait descriptions (e.g., invited a charity collector for lunch, shouted and honked
at slow drivers, etc.), and (b) opinion statements that reflected either a liberal or conservative political orientation. Participants after receiving the information were asked to recall the information they had read.
Participants were more likely to recall behaviors that were consistent
with the target persons initial trait description than behaviors that were inconsistent with it, replicating the results of many other studies (Hastie &
Kumar, 1979; for a review, see Srull & Wyer, 1989). However, they were more
likely to recall opinions with which they disagreed than opinions with
which they agreed, and this was true regardless of the consistency of these
opinions with the targets political ideology. Thus, when participants encountered opinion statements about matters with which they were familiar,
they comprehended these statements with reference to their previously acquired knowledge, and devoted a particular amount of time trying to understand the basis for opinions they regarded as incorrect. Therefore, they
had better recall of these opinions regardless of the opinions consistency
with the general attitudes of the individual who expressed them.13
1. Individual Differences in the Processing
of Invalid Information
People obviously differ in the amount of cognitive effort they perform in
an attempt to reconcile information they consider to be invalid. Wnke and
Wyer (1996) speculated that differences in this motivation might vary with
participants own social and political ideology. Participants with a liberal
social and political orientation are inclined to consider the underlying
causes of a persons behavior when evaluating a person (Fletcher, Danilovics, Fernandez, Peterson, & Reeder, 1986). Conservatives, on the other
hand, are more disposed to evaluate a persons behavior in terms of its
13
Wnke and Wyer (1994) found that participants had better recall of behaviors that were
ideologically inconsistent with the liberal or conservative orientation of the actor. However, this
occurred only when the actors ideology was different from participants own (i.e., when the actor was an outgroup member). When the actors ideology was the same as participants (i.e.,
the target was an ingroup member), participants recall of the targets behavior was similar regardless of whether it was inconsistent or consistent with his ideology (.42 vs. .48, respectively).
13
164
CHAPTER 7
165
PRAGMATIC INFLUENCES
TABLE 7.2
Mean Proportion of Behaviors Recalled as a Function of Perceived
Desirability, Participants Ideology, and Instructional Conditions
Opinion-Focus Conditions
Desirable behaviors
Undesirable behaviors
Difference
Behavior-Focus Conditions
Liberal
Participants
Conservative
Participants
Liberal
Participants
Conservative
Participants
.37
.64
.27
.39
.33
.06
.46
.54
.08
.35
.66
.31
told to use a criterion that differed from the one they typically employed,
this spontaneous cognitive activity was disruptive, and so recall differences were not evident.
166
CHAPTER 7
son, 1981; Carlston & Smith, 1996; Wyer & Carlston, 1994). In this research,
participants are usually asked to form an impression of someone on the basis of trait and behavior descriptions that are conveyed out of their social
context. In doing so, they are expected to take the information at face value,
independently of its source. To this extent, the implications of this research
for person impression formation outside the laboratory may be limited.
The research described in the remainder of this chapter calls attention
to several of these limitations. To provide a perspective on this research,
we first review briefly the conclusions that have been drawn in the typical
research paradigm in which person impression formation has been investigated. We then describe the results of a number of studies in which similar
information is conveyed in the context of informal conversations under
conditions in which the pragmatic implications of the information is taken
into account. Finally, we describe the role of nonverbal and paralinguistic
features of the communications exchanged in a conversation that influence
recipients perceptions of their implications and, therefore, the impressions
they form of both the topic being discussed and the persons who are discussing it.
Impression Formation in Nonsocial Contexts
Extensive research has been conducted on the impressions that people
form of someone on the basis of information that describes the persons
traits and behaviors, and rigorous theories have been proposed to describe
the processes that underlie these impressions (Carlston & Smith, 1996;
Wyer & Srull, 1989). In a typical study (cf. Srull, 1981; Wyer, Bodenhausen, &
Srull, 1984; for a review, see Srull & Wyer, 1989), participants are asked to
form an impression of a target person on the basis of a favorable or unfavorable trait description followed by a number of behaviors. Some behaviors are descriptively and evaluatively consistent with the initial trait description of the person and others are inconsistent with this description.
The target person is typically fictitious, and the trait description is conveyed in a way that suggests that its validity is not open to question. (For
example, the person might ostensibly be a character in a novel, and the
trait adjectives might have been used by the author to describe him; see
Wyer & Martin, 1986.) The mental representation that is formed under these
conditions, which presumably exemplifies a generalized person representation of the sort described in chapter 1, has been conceptualized using an associative network metaphor (Hastie & Kumar, 1979; Srull & Wyer, 1989).
That is, the concepts formed from information are represented by nodes,
and associations between them (formed by thinking about the concepts in
relation to one another) are represented by pathways. Stronger associa-
PRAGMATIC INFLUENCES
167
14
168
CHAPTER 7
PRAGMATIC INFLUENCES
169
Budesheim, 1987; Wyer & Unverzagt, 1985), the combined influence of stereotypes and idiosyncratic trait descriptions on person impressions (Wyer
& Martin, 1986), and the different impressions that are formed of persons
and groups (Srull, 1981; Wyer & Gordon, 1982). Nevertheless, the processes
implied by the theory are restricted to conditions in which information is
presented out of its social context. Quite different considerations arise
when a persons traits and behaviors are described in the context of a conversation. For one thing, an acquaintances trait descriptions of someone in
the course of a conversation are matters of opinion and, as such, may not
be valid. Consequently, listeners may attempt to establish the validity of
these descriptions. To do this, they may construe the implications of the
descriptions for characteristics of the speakers who provide them as well
as for the individual being described.
Wyer, Budesheim, and Lambert (1990) examined this possibility. In their
studies, the trait and behavior descriptions that participants received were
very similar to those presented in more traditional impression formation
research (e.g., Wyer, Bodenhausen, & Srull, 1984). However, the information
ostensibly pertained to an actual person and was provided by acquaintances of the person in the course of a conversation. Specifically, participants were told they would listen to a tape-recorded conversation between
a male and a female student about a mutual acquaintance. The tape began
with instructions to the two speakers to select the person they wanted to
talk about and then, after doing so, to write down three general attributes
they would use to describe this person. At this point, the tape stopped, and
the listeners were given photocopies of the trait adjectives the speakers
had had ostensibly written down. The favorableness of each speakers trait
description of the target and the dimension to which this description pertained (intelligence or kindness) were varied independently.
After participants had read the target descriptions, the taped conversation was restarted, and participants heard the experimenter tell the speakers
to reminisce about things the target person had done that they had either
seen or heard about. The speakers then proceeded to exchange anecdotes,
interspersed with occasional pauses and promptings by the experimenter to
make the conversation seem natural. Over the course of the conversation,
each speaker mentioned 13 behaviors of which 6 were favorable (either kind
or intelligent), 6 were unfavorable (unkind or unintelligent) and 1 (occurring
near the beginning of the conversation) was neutral. These behaviors were
very similar to those presented in studies by Wyer et al. (1984) and Wyer and
Martin (1986), but were conveyed as they might be mentioned in a normal
conversation. An excerpt of one conversation, about a target named Don,
was as follows (for a complete transcript, see Wyer et al. (1990):
Experimenter: Okay, then, why dont we begin. Who wants to start off?
170
CHAPTER 7
M: Well, lets see . . . well, he walks a mile to school each day, even when
its freezing. I guess I thought of that because its so damn cold today.
I asked him if he wanted a ride, but he said he didnt.
F: Yeah, he never seems to get sick. It must be because he works out 3
or 4 times a weekhes in pretty good shape.
M: [Laughs] Better than me, anyway. . . . I remember he won the university chess championship last fall.
F: [Laughs] Yeah. . . . I remember we went out to dinner that night to
celebrate and Don swore at the waiter at Eddies because he didnt
get served as quickly as he wanted.
M: Yeah, I can just see him doing that . . . lets see . . . well, he cant seem
to follow simple directions people give him. Remember that time last
week at Jerrys?
F: Yeah, [Laughs] . . . Oh, by the way, do you know Jerry lost his job at
the Union?
M: No, really? Gee, thats rough.
F: Yeah. I thought of that cause he told me Don had loaned him 20
bucks for a date he had and stuff.
M: [Pause] Jeez, this is a lot harder than I thought. [To experimenter]
You want specific behaviors, right? Like I know hes always smiling
and saying hello to people when I walk down the street. But one particular instance doesnt stand out.
Experimenter: Well, try to be as specific as you can.
M: [Pause] Hmm . . . well, lets see . . . well, someone said hed flunked a
mechanics training course I took last summerI dont know why he
took it, anyway.
The other tape was identical except that the roles of the male and female
speakers were reversed.
We anticipated that people who listened to the conversation would spontaneously form impressions of the speakers as well as the target himself. To
understand the nature of these different impressions, we employed two instructional conditions. In speaker-impression conditions, participants were
explicitly told to form impressions of the speakers, and in target-impression
conditions, they were told to form an impression of the person the speakers
were discussing. In both cases, however, participants after listening to the
tape reported their liking for the target person along a -5 to +5 scale. In addition, they reported their perception of how well each of the speakers
liked the target and rated their own liking for the speakers. Finally, they recalled the behaviors they had heard the speakers mention.
PRAGMATIC INFLUENCES
1.
171
Speaker-Impression Conditions
172
CHAPTER 7
TABLE 7.3
Proportion of Behaviors Recalled as a Function of the Speaker Who
Mentioned Them and the Behaviors Evaluative Consistency
With Each Speakers Trait Description of the Target
A. Speaker-impression conditions
Behaviors mentioned by male
Consistent with males trait description
of target
Inconsistent with males trait description of target
M
Behaviors mentioned by female
Consistent with males trait description
of target
Inconsistent with males trait description of target
M
B. Target-impression conditions
Behaviors mentioned by male
Consistent with males trait description
of target
Inconsistent with males trait description of target
M
Behaviors mentioned by female
Consistent with males trait description
of target
Inconsistent with males trait description of target
M
Consistent With
Females Trait
Description of
Target
Inconsistent With
Females Trait
Description of
Target
.430
.514
.412
.403
.416
.438
.476
.421
.351
.430
.390
.527
.439
.451
.441
.489
.396
.437
.416
.422
.409
.424
.430
.423
.437
.528
.482
.409
.423
.472
.500
.441
Note. In each quadrant of the table, diagonal cells are those in which the two speakers trait
descriptions of the target are evaluatively similar, and off-diagonal cells are those in which the
descriptions are evaluatively dissimilar. Adapted from Wyer, Budesheim, and Lambert (1992).
Thus, suppose the female speaker described the target in terms of favorable traits and the male described him in terms of unfavorable traits.
Then, the mental representation that presumably results from the cognitive activity described in the preceding paragraph can be captured in an
associative network of the form shown in Fig. 7.3a, where F+ and M- refer
to concepts of the male and female speakers, respectively, and the favorable and unfavorable behaviors mentioned by the speakers are denoted
b+ and b-, respectively.
PRAGMATIC INFLUENCES
173
2. Target-Impression Conditions
In these conditions, participants were told to form an impression of the
target rather than of the speakers. Despite these instructions, the results
were very similar to those observed in speaker-impression conditions. For
example, the favorableness of the speakers initial trait descriptions of the
target influenced both participants perceptions that the speakers liked the
target (1.58 vs. -0.93, when both speakers trait descriptions were favorable
vs. unfavorable, respectively) and also participants liking for the speakers
(1.04 vs. -0.42, respectively). However, in contrast to earlier studies in
which trait descriptions of the target provided the primary basis for evaluations of him (Srull & Wyer, 1989), participants liked the target less when
both speakers described him favorably (M = -0.33) than when they did not
(M = 0.75). Thus, these data, like those obtained in speaker-impression conditions, indicate that participants used the speakers trait descriptions of
the target to form impressions of the speakers themselves, and then, having done so, used their impressions of the speakers as comparative standards in evaluating the target without considering the favorableness of the
speakers trait descriptions for characteristics of the target himself.
The recall data in this study were also similar to those observed in
speaker-impression conditions except that in this case, participants were
primarily disposed to confirm their impression of the female speaker rather
than the male. These data are summarized in the bottom half of Table 7.3.
When both speakers trait descriptions of the target were similar in favorableness, participants had somewhat better recall of behaviors that were
evaluatively inconsistent with these descriptions (M = .448) than behaviors
that were consistent with them (M = .417). When the two speakers descriptions differed in favorableness, participants were more inclined to recall behaviors that were inconsistent with the female speakers description of the
target (M = .482) than behaviors that were consistent with these descriptions (M = .416), and this was true regardless of whether the behaviors were
mentioned by the male or the female. Thus, both sets of data could suggest
that participants organized the information around an evaluative concept
of the female speaker. Thus, they thought more extensively about behaviors she had personally mentioned that were inconsistent with this concept
in an attempt to reconcile their occurrence. On the other hand, they
thought more about behaviors the other speaker mentioned that were inconsistent with this concept and, therefore, bolstered their perception that
the females trait description of the target reflected a general disposition to
judge people favorably or unfavorably independently of the targets actual
attributes. The representation that participants may have formed under
this condition, therefore, would resemble that shown in Fig. 7.3b.
174
CHAPTER 7
3. Summary
Considered in combination, the results obtained under speaker-impression and target-impression conditions suggest that participants who overhear a conversation about a person are inclined to form impressions of the
speakers, and that this is true regardless of whether they are explicitly told
to do so or are asked to form an impression of the person the speakers are
discussing. Moreover, they tend to use the speakers descriptions of the targets behavior to confirm these impressions. Then, once these impressions
are formed, participants use them as comparative standards in evaluating
the person the speakers are discussing rather than basing their impressions on the literal implications of the speakers descriptions of him.
PRAGMATIC INFLUENCES
175
1. General Considerations
Participants in informal conversations are expected not to offend the
people to whom they are communicating. They are also expected to be
modest and not to toot their own horns unnecessarily. Consequently, people who hear someone make a statement that appears to violate one of
these norms may attempt to reinterpret it in a way that is more consistent
with their normative expectations. These processes may be similar to those
that result from violations of informativeness and truthfulness norms.
For example, the interpretation of a statement as hostile or a tease may
depend in part on whether the statement violates a norm to be polite. This,
in turn, may depend on whether the target of the statement is present at
the time the statement is made. Thus, for example, a persons comment
that a colleague has trouble balancing his checkbook is likely to be seen as
a tease if the comment is directed to the colleague himself. In the colleagues absence, however, a politeness norm might not be applicable, and
so the statement is more likely to be interpreted as antagonistic.
However, the effects of politeness norms on comprehension are not independent of its effects on truthfulness. Shortly after the publication of my
1974 book on cognitive organization and change, a friend remarked, Congratulations on your book. Im looking forward to the publication of the
English language version.15 Because I interpreted the statement as a tease
rather than an expression of hostility, I found it amusing. However, this reaction was probably diluted by my lack of confidence in my writing skills
and, therefore, my belief that the statements implications might be valid. If
I had had more confidence in my writing ability, I might have construed the
implications of the statement to be definitely false. I might then have appreciated the irony much more.
A related implication of this analysis is that statements are more likely to
be interpreted as teases if their literal implications are extremely disparaging than if they are less so. This is because they are more likely to be seen
as invalid. I may be more disturbed by a mildly negative comment about my
writing ability than by a comment that my writing is slightly superior to that
of a chimpanzee. This is because I am inclined to believe that the first remark is valid and thus to take it literally, whereas I perceive that the second
statement is clearly invalid and therefore, interpret it as a tease.
15
15
176
CHAPTER 7
PRAGMATIC INFLUENCES
177
Experimenter: Okay, then, why dont we begin. Why dont you start, Don?
O: Well, lets see . . . well, you walk a mile to school each day, even when
its freezing. I guess I thought of that because its so damn cold today.
Every time I offer you a ride, you always say no.
T: Yeah, I never seem to get sick. It must be because I work out 3 or 4
times a weekIm in pretty good shape.
O: [Laughs] Better than me, anyway. . . . I remember you won the university chess championship last fall.
T: [Laughs] Yeah . . . I remember we went out to dinner that night to celebrate and I swore at the waiter at Eddies because I didnt get served
as quickly as I wanted.
O: Yeah . . . I remember after we left, and we were supposed to go to
that party at Jerrys, you were trying to follow the directions hed
given you to get to his house and you got us totally lost! . . . Oh, by
the way, do you know Jerry lost his job at the Union?
T: No, really? Gee, thats rough.
O: Yeah. I thought of that cause he told me you had loaned him 20
bucks for a date he had and stuff.
T: [Pause] Jeez, this is a lot harder than I thought. [To experimenter]
You want specific behaviors, right? Like I know Im always smiling
and saying hello to people when I walk down the street. But one particular instance doesnt stand out.
Experimenter: Well, try to be as specific as you can.
T: [Pause] Hmm . . . well, lets see . . . well, I flunked a mechanics training course I took last summerI dont know why I took it, anyway.
Participants after hearing the conversation were asked to indicate how well
they liked the two speakers, and then were asked to recall as many of the
behaviors as they could.
We expected that participants would have better recall of statements
that O made that violated a politeness norm, but would have better recall of
statements the target mentioned that violated a modesty norm. This was in
fact the case. Table 7.4 shows the mean proportion of favorable and unfavorable behaviors that participants recalled in each instructional condition
as a function of the person who mentioned them. Participants had better recall of behaviors that O mentioned if they were unfavorable than if they
were favorable (.49 vs. .39, respectively), but had better recall of behaviors
the target himself mentioned if they were favorable than if they were unfavorable (.45 vs. .38, respectively). The first difference was more pronounced
when participants were explicitly told to focus their attention on the behaviors that O mentioned, and the second was more pronounced when partici-
178
CHAPTER 7
TABLE 7.4
Proportions of Favorable and Unfavorable Behaviors Recalled as a
Function of Task Objectives and the Person Who Mentioned Them
Task Objectives
Behaviors mentioned by
other (O)
Favorable
Unfavorable
Behaviors mentioned by
target (T)
Favorable
Unfavorable
Target Impression,
Target Focus
Target Impression,
Other Focus
Other Impression,
Other Focus
.42
.45
.40
.54
.35
.49
.47
.45
.50
.54
.37
.34
PRAGMATIC INFLUENCES
179
180
CHAPTER 7
teresting information) or an encyclopedia (whose goal is to preserve archival knowledge). Regardless of the passages source, however, participants
after reading it were told to convey its content to a nave recipient in a way
that was either as interesting as possible or as accurate as possible. Thus,
in some cases, the assumed objectives of the messages original source and
those of the communicator were ostensibly similar, and other cases, these
objectives differed. After delivering the message (which was tape recorded), both communicators and nave recipients evaluated the seriousness of the disease and reported their specific beliefs about its prognosis,
cause, and treatment. These latter beliefs were coded so that more positive
values indicated stronger beliefs in the content of the original passage.
Communicators own beliefs and opinions about the disease were not affected by either the source of the passage or their own objectives in communicating it. Moreover, a content analysis of communicators speech revealed no differences in the actual content of the information they
conveyed in different conditions. Nevertheless, communicators exhibited
more characteristics of poor speaking style (e.g., a greater frequency of
uhhs and errs), and took longer to deliver the speech (an indication of
rambling), when their communication objectives differed from those of the
original message source than when they were similar. Correspondingly, recipients reported less strong beliefs and opinions about the disease and its
seriousness in the former condition than the latter.
These results are perhaps not too surprising. However, their importance
in the present context lies in their implications that communicators take
into account not only their own objectives in conveying a message but also
the pragmatic implications of the source of the information they are transmitting, and that these factors combine to influence recipients perceptions
of the validity of the information being communicated.
The Impact of Conversational Style on Liking
for the Communicator
The speech style characteristics identified by Gruenfeld and Wyer seem
likely to influence reactions to communications exchanged in an actual conversation. In this context, they might influence participants impressions of
one another as well as the topic under discussion. Additional factors are
likely to come into play, however. People who engage in an informal conversation are typically expected to converse in a way that will make it easy
for others to respond. One obvious way of accomplishing this is to ask
questions of one another. A related technique is to elaborate answers to
one anothers questions, thereby increasing the likelihood that the response will stimulate a relevant idea in the listener. These techniques may
be acquired through social learning and be applied spontaneously, without
PRAGMATIC INFLUENCES
181
much deliberation. When the techniques are not applied, however, recipients are likely to find the conversation awkward and difficult to conduct,
and these reactions could affect their liking both for the conversation itself
and for the other participants.
These effects are particularly likely to occur in get-acquainted conversations, where participants are motivated to learn about one another and to
identify areas of mutual interest. For example, suppose a participant in
such a conversation asks, Where are you from? If the other responds,
Chicago. How about yourself? the first person finds it quite easy to continue the dialogue. Alternatively, if the other responds, Well, I live in Chicago now, but I just moved here from Florida, the questioner might be
stimulated to ask about Florida and why the person moved, or might describe his or her own recent visits to the state. However, suppose the other
simply responds Chicago, without elaborating or asking a question in return. Then, the first individual is burdened with the task of finding a different topic to discuss or a different question to ask. This burden increases
the difficulty of continuing the conversation and this difficulty, in turn could
decrease the persons liking for both the conversation and the interaction
partner.
But other factors can enter into the picture as well. An individuals failure to elaborate answers to anothers questions, or the failure to ask questions in return, could also be interpreted as an indication that the individual
is not really interested in interacting with the questioner. This inference
could also decrease the questioners liking for the individual independently
of the effects of these conversational characteristics on the ease of conducting the conversation per se.
Wyer, Swan, and Gruenfeld (1995) investigated these effects in a getacquainted conversation. College students who were previously unacquainted took part in the study in same-sex pairs. However, only one member of each pair was assigned the role of the actual subject. The subjects
partner was recruited as an accomplice. Subjects and their partners arrived
at the experiment in different rooms so they would not see one another
prior to the experiment. Subjects were told we were interested in how people engage in get-acquainted conversations, and that they would be asked
to take part in a 5-minute conversation with another student much as they
might if they met the person for the first time at lunch or on a break between classes. However, we indicated that to ensure that some of the information exchanged in each conversation was similar, we would like them to
ask their partner five questions at some point during the conversation:
1. Where are you from?
2. Do you like school?
3. Where do you live on campus?
182
CHAPTER 7
PRAGMATIC INFLUENCES
183
them. Finally, they judged the partner with respect to several personality
attributes. Of these, the most relevant to his or her behavior in the conversation was aloofness.
We expected that both partners elaboration and their reciprocation of
questions would influence participants ease of conducting the conversation, and that this, in turn, would infer their liking for the conversation and,
as a result, their liking for the partner. However, we also speculated that if
subjects found that their partner did not elaborate and reciprocate their
questions, they would interpret this behavior as an indication that their
partner had little interest in the interaction and might not like them, and
that these perceptions might also influence their liking for their partner. In
fact, both possibilities were evident.
The effects of conversational style can be seen most easily from path
analyses. Figure 7.4 shows the significant paths connecting the two communication style characteristics to the participants estimates of the ease of
conducting the conversation (Ease), their enjoyment of the conversation
(Enjoy), their perceptions that their partner (O) was interested in them (O
int S) and liked them (O like S) and their liking for the partner (like O). Females perception of the ease of conducting the conversation influenced
both males and females liking for their partner through its mediating influence on both their enjoyment of the conversation and their inference that
the partner liked them. However, whereas males perception of the ease of
conducting the conversation was primarily a function of their partners reciprocation of their questions, females perceptions were influenced primarily by the degree to which their partner elaborated responses to their
questions. In addition, partners elaborations also affected subjects perceptions that the partner was interested in them, and this factor also increased
their liking for their partner. On the other hand, partners conversational
style had no impact on participants perceptions of their aloofness, nor did
these perceptions influence liking for the partners.
The gender differences obtained in the study are provocative. As Wyer
et al. (1995) point out, women are typically more motivated than men to establish a sense of connectedness to the persons with whom they interact
(Chodorow, 1979; Gilligan, 1982; Tannen, 1990).Their partners elaboration
of answers to their questions may have increased their feelings of
connectedness, as evidenced by its impact on these subjects perceptions
that their partner was interested in them and liked them. In contrast, males
placed less emphasis on these factors. Perhaps because they are less socially skilled than females, however, they found that the ease of conducting
the conversation was easier when their partners asked them questions.
In summary, partners counternormative communication style influenced both mens and womens liking for their partner through its mediating impact on their perception of ease of conducting the conversation. In
184
CHAPTER 7
FIG. 7.4. Path diagrams of the effects of the partners conversational style on
subjects liking for their partner (O) and the variables that mediate these effects. Elab = Os elaboration of responses to the subject question. Quest = Os
reciprocation of the subjects question. Ease = subjects perception of the ease
of conducting the conversation. OintS = subjects perception that O was interested in them. O like S = subjects perception that O liked them, and like O =
subjects liking for the partner.
PRAGMATIC INFLUENCES
185
186
CHAPTER 7
TABLE 7.5
Subjects Liking for Their Partner as a Function of the Partners
Conversational Style and Initial Trait Description
Initial Trait Description of Partner
ConversationIrrelevant
ConversationRelevant
None
8.56
6.63
7.50
8.78
6.78
7.78
8.95
3.88
6.56
8.76
5.89
trait descriptions of the partner only had an influence on subjects evaluations of her when her behavior in the conversation deviated from expectations and, therefore, stimulated subjects to seek other information that
might account for it. Moreover, this search led the information to have an
impact regardless of whether it was relevant or irrelevant to the explanation that subjects were seeking.
PRAGMATIC COMMUNICATION
IN CLOSE RELATIONSHIPS
The discussion in this chapter has focused largely on the pragmatic influences of communications that are exchanged among casual acquaintances.
When people know one another very well, the pragmatic implications of
their shared messages may be difficult for others to discern. As noted earlier, married couples can often convey feelings to one another through a
statement that seems quite innocuous to observers who are unaware of the
pool of shared knowledge that the partners bring to bear on its interpretation (Watzlawick et al., 1967). Thus, for example, a wife who asks her spouse
at a party what time it is might be interpreted as making a serious request
for information. However, it could be interpreted by her husband as an indirect expression of anger because of his failure to leave the party early
enough for her to see David Letterman.
On the other hand, miscommunication often arises between partners in
close relationships as well as casual acquaintances. This is particularly true
in the case of emotions. One reason for this could be that the expression of
some emotions is inherently ambiguous. However, other factors play a role
as well. A study by Gaelick, Bodenhausen, and Wyer (1985) provided insight
into these possibilities. In an initial session of their experiment, married
couples engaged in a 10-minute tape-recorded discussion of a problem they
were having in their relationship. Then, in a second session, each partner
PRAGMATIC INFLUENCES
187
viewed the tape of the conversation and identified three statements that
were made in the conversation that had had an important effect on their
feelings toward one another. Finally, in a third session, partners reviewed
the segments of the tape containing the statements that both they and their
spouse had identified. Partners rated statements they had personally made
in terms of the feelings they intended to convey, their expectations for how
the partner would interpret the statements, and how they thought the partner would respond. They rated statements their partner had made in terms
of the feelings their partner intended to convey, how their partner thought
they would respond, and their actual response. Factor analyses of these ratings revealed they fell along two independent dimensions pertaining to love
and hostility.
Several interesting results emerged. In general, partners attempted to
convey the emotion they perceived their spouse had conveyed to them.
However, they were only accurate in perceiving their spouses hostility.
Consequently, feelings of hostility were actually reciprocated, but feelings
of love were not. One implication of this is that hostility was more likely to
escalate over the course of the conversation than feelings of love.
Second, when wives communicated in a way they intended to be
affectively neutral, their husbands interpreted their statements as expressions of hostility. In contrast, when husbands communicated in a way they
intended to be neutral, their wives interpreted their statements as expressions of love. Gaelick et al. (1985) interpreted these miscommunications in
terms of the stereotyped social role expectations that men and women hold
for one another. Specifically, women are expected to be loving and affectionate. Therefore, when women responded in a way they intended to be
neutral, their husbands interpreted this counternormative comment as hostile and, as noted earlier, were likely to respond hostilely in return. Men,
however, are expected to be hostile and aggressive. Consequently, when
they intended to convey neutral affect, their wives interpreted this expectancy-deviant statement as an indication of love, and presumably reciprocated this emotion. Unfortunately, however, because expressions of love
were typically misperceived, these attempts to deescalate the conflict were
unlikely to do much good.
CONCLUDING REMARKS
Although this chapter has covered a lot of ground, it clearly does not provide a complete account of the factors that theoretically influence peoples
perceptions of the pragmatic implications of the information they receive in
social situations. For example, the extensive research on social attribution,
stimulated by theories of Jones and Davis (1965) and Kelley (1967, 1987),
188
CHAPTER 7
C H A P T E R
8
The Dynamics of Humor Elicitation:
The Effects of Informational
Context on the Interpretation
and Elaboration of Narratives
The pilot was killed when his landing gear malfunctioned. Flying airplanes
can be dangerous.
Marys husband and children were killed when a plane overshot the runway and crashed into their home near the airport. Flying airplanes can be
dangerous.
190
CHAPTER 8
In some instances, however, the concepts and knowledge that are activated and used to interpret the events that occur at the beginning of a sequence cannot be applied to the events that come later. In such cases, recipients must reinterpret the initial events in terms of different concepts in
order to construct a complete understanding of the sequence of events as a
whole. In the preceding example, suppose the statement, Flying airplanes
can be dangerous is presented first. An individual who happens to interpret this statement as equivalent in meaning to It can be dangerous to fly
an airplane might then encounter the statement Marys husband and children were killed. . . . To comprehend the information as a whole, the recipient would presumably reinterpret the first statement as referring to the
hazards of living near an airport.
The comprehension processes that are involved in this example are similar to those described in chapter 7 when a statement violates normative expectations. The only difference is that the expectations are not only activated by the social context in which it occurs. In addition, they result from
the interpretation that has been given to other information in the communication in which the statement is embedded.
Reinterpretations can also occur in the course of comprehending direct
experiences. A waiter whose outward appearance gives the impression of
elegance and sophistication but who is observed to spill soup in a customers lap may be reconceptualized as a pretentious oaf. However, reinterpretations of written or oral communications are more common. Particularly frequent examples occur in jokes and stories whose punch line
stimulates a reinterpretation of the events leading up to it. For example,
consider the following story:
A young Catholic priest is walking through town when a prostitute accosts
him. How about a quickie for twenty dollars? she asks.
The priest, puzzled, shakes her off and continues on his way, only to be
stopped by another prostitute. Twenty dollars for a quickie, she offers.
Again, he breaks free and goes on up the street.
Later, as he is nearing his home in the country, he meets a nun outside her
convent. Pardon me, sister, he asks, but whats a quickie?
Twenty dollars, she says, The same as it is in town.
The events described in the punch line of this story stimulate the reinterpretation of two features of the previous information. First, ones characterization of the nun is revised to include her service as a prostitute. Second,
the question, Whats a quickie?, which was first assumed to be equivalent
to What does a quickie mean?, is reinterpreted as equivalent to Whats
a quickie cost?
Not all reinterpretations of information elicit amusement, as our example of flying airplanes testifies. In fact, the cognitive and motivational un-
191
derpinnings of humor elicitation are not widely understood. This is surprising in light of the central role that humor plays in social communication. It is a rare conversation in which at least one participant does not
respond with amusement to another persons comments or behavior.
Jokes, witticisms, and other humorous events are commonplace in social
interaction situations and can have an impact on the quality of the interactions. Although jokes are often told for the purpose of entertaining, they
can also be used strategically to decrease tension in a heated discussion
(Kane, Suls, & Tedeschi, 1977) or to enliven a boring one. In short, the
transmission and comprehension of humor are central features of social
experience.
Theoretical and empirical analyses of social interaction processes have
largely ignored these processes. Several theories of humor elicitation have
been proposed, beginning with the work of Freud (1928, 1960). However,
most research on humor (for summaries, see Chapman & Foot, 1976, 1977;
Goldstein & McGhee, 1972; McGhee & Goldstein, 1983) has focused on its
motivational bases (cf. LaFave, Haddad, & Maeson, 1976; Zillman & Cantor,
1976) and its personality correlates (Lefcourt & Martin, 1986; Mindess et al.,
1985). Yet, only a few attempts have been made to explicate the cognitive
processes that underlie perceptions that something is funny (Apter, 1982;
Long & Graesser, 1988; Suls, 1972, 1977). And of these, only Apters conceptualization is applicable to situations in which humor is elicited spontaneously in informal social interaction.
In this chapter, we provide a theoretical analysis of humor elicitation in
the context of the comprehension processes we have described in earlier
chapters. Our theory avoids many of the deficiencies of existing formulations while at the same time providing an example of the potential applicability of our more general conceptualization of comprehension phenomena
to information processing outside the laboratory. According to this conceptualization, humor elicitation is a by-product of the comprehension processes postulated in chapter 4 and in some cases, of higher order comprehension processes similar to those discussed in chapter 7. As we have
noted, however, merely the reinterpretation of previously acquired information in the light of new information is not a sufficient condition for humor elicitation. Thus, several additional factors not discussed in earlier
chapters must be taken into account. We first circumscribe the conditions
in which humor is spontaneously elicited by information that people receive and consider the effects of more deliberative, elaborative processes
that occur once the humor-eliciting interpretation of information has been
identified. We then apply the conceptualization to a specific area in which
an understanding of humor elicitation is particularly important, namely, the
amusement elicited by jokes and stories that perpetuate a social stereotype
or are likely to be considered offensive for other reasons.
192
CHAPTER 8
HISTORICAL BACKGROUND
To provide a perspective on the conceptualization to be proposed in this
chapter, a brief review of earlier theories of humor elicitation may be helpful. These theories fall into two categories, one of which focuses on the motivational antecedents of humor elicitation and the other of which concerns
the cognitive processes that give rise to it. Considered in isolation, each
theory is insufficient to account for the numerous circumstances in which
amusement may occur. (A possible exception to this rule is Apters [1982]
theory to be discussed presently.) Nevertheless, the factors that the various theories postulate to influence humor elicitation are generally compatible. That is, the factors postulated by several theories could contribute to
humor elicitation, although not necessarily for the reasons the theories assume.
Motivational Theories of Humor
1. Arousal-Reduction Theories
Responses to humor have sometimes been conceptualized in terms of a
release of tension or reduction in arousal. According to Freud (1905/1960,
1928), for example, peoples responses to humor-eliciting stimulus events is
motivated by the need to reduce tension or arousal they are inhibited from
expressing directly. This arousal, which was often assumed to be aggression- or sex-related, might either be induced by the stimulus itself or exist
before exposure to the stimulus. For example, the humor a joke elicits
could vary with the intensity of suppressed emotions that have previously
become associated with the type of stimulus to which the joke is relevant.
Although Freuds conceptualization can account for some humoreliciting experiences, it is clearly not sufficient to explain the humor that is
elicited by all types of communications. For example, the conceptualization
appears to apply only under conditions in which features of the humoreliciting stimuli are similar to those to which the suppressed emotion is relevant. Given the wide diversity of stimuli that a given individual finds humorous, one would have to postulate a very large number of suppressed
emotions in order to argue that this is a necessary antecedent of humor
elicitation.
A conceptualization of humor elicitation by Berlyne (1969, 1971) is
broader in scope. He assumed an inverted-U relation between physiological
arousal and the experience of pleasure. That is, pleasure increases with
arousal up to a point and then decreases, ultimately reaching a level at
which it becomes aversive. Berlyne viewed a joke as a scenario that induces arousal beyond its optimal value, followed by a punch line that de-
193
194
CHAPTER 8
ten require access to a large body of knowledge about the type of persons
and events described. For example, consider a joke from the Antioch Humor Test (Mindess et al., 1985):
A blind man enters a department store, picks up his dog by its tail and begins
swinging it over his head. A clerk hurries over and says, Can I help you, sir?
No, thanks, the man replies, Im just looking around.
In this joke, two planes of thought are brought together by the punch line,
one of which pertains to the idiomatic meaning of looking around and the
other of which concerns its literal meaning. However, an appreciation of
the joke also requires knowledge that blind individuals often rely on seeingeye dogs to do their looking.
1. Suls Formulation of Humor Elicitation
A more formal statement of incongruity-resolution processes, provided
by Suls (1972, 1983), is consistent with the conceptualization of comprehension processes we proposed in chapters 4 and 5 as well as more general
theories of prose comprehension (e.g., Graesser, 1981). Suls assumes that
when people begin to read a story, the persons and events described activate a body of conceptual knowledge (e.g., event or episode models as conceptualized in this volume) that can be used to interpret it. This knowledge,
once activated, is brought to bear on the interpretation of information that
is conveyed subsequently. However, recipients may later encounter information that cannot be understood in terms of this knowledge. When this occurs, they must identify alternative concepts and knowledge structures that
can be used to comprehend the new information in the context of the old.
This process often requires a reinterpretation of the original information,
and the generation of this reinterpretation elicits humor.
Suls likens joke comprehension to a problem-solving task in which comprehension is analogous to solving the problem and amusement is analogous to the pleasure that people get from arriving at the solution. This analogy has additional implications. Most obviously, people must recognize
that a problem exists (i.e., that the information cannot be interpreted in
terms of previously acquired concepts and knowledge). More important,
the problem must be neither so easy that the pleasure derived from solving
it is minimal nor so difficult that its solution requires excessive effort. This
suggests that humor is more likely to be elicited by jokes that are moderately difficult to comprehend than by jokes that are either too easy or too
hard. We consider this possibility more fully in the conceptualization we
outline later.
195
196
CHAPTER 8
tion of a situation to elicit humor, two things must be true. First, the reinterpretation must not replace the original interpretation that was made. That
is, the perception of the experience that results from the reinterpretation
must not change ones perception of the experience that purported to exist
at the outset. Rather, both interpretations must be taken into account simultaneously. (For a similar assumption, see Koestler, 1964.) Second, the
perception of the experience that is established by the new information
must in some sense be diminished in importance or value relative to the
conditions that were first assumed.
To see the applicability of the diminishment and nonreplacement principles, reconsider the quickie joke described earlier. To reiterate, this joke
actually exemplifies two types of shifts in interpretation. One, purely semantic shift occurs in the interpretation of Whats a quickie? The second
shift occurs from the perception of the nun as a chaste and devout woman
to the perception of her as a prostitute. Note, however, that the reinterpretation of Whats a quickie? that is implied by the nuns response does not
negate the interpretation that was apparently intended by the priest. Moreover, the nun, although turning out to be a prostitute, remains a nun. In
other words, the new perception of the reality of the situation does not invalidate the appearance of the situation that existed before the reality was
revealed. Finally, note that the nuns holiness is diminished as a consequence of being a prostitute on the side. Thus, both of the conditions that
Apter (1982) postulated to be necessary for human elicitation are met.
Diminishment should not be confused with disparagement. Although disparaging reinterpretations may often be diminishing, not all diminishing reinterpretations are disparaging. In the Wizard of Oz, for example, people are
amused by the lion who purports to be ferocious but who turns out instead
to be meek. The lions true attributes are less unfavorable than his purported ones. They are nevertheless more mundane and, therefore, elicit
amusement.
3. Summary
Apters conceptualization can account for a number of humor-elicitation
phenomena that are not easily explained by other formulations. To give
some examples:
Puns. Puns purport to convey interesting information but are revealed
to be silly and, therefore, trivialize the communication itself. For example:
Two Eskimos sitting in a kayak were chilly. However, when they lit a fire in
their craft, it sank, proving once again that you cant have your kayak and
heat it too.
197
A group of chess enthusiasts were standing in a hotel lobby discussing their recent tournament victories. After about an hour, the manager went up to them
and asked them to disperse. But why? they asked. Because, he said, I cant
stand chess nuts boasting in an open foyer.
Mahatma Gandhi, as you know, walked barefoot most of the time, which produced an impressive set of calluses on his feet. He also ate very little, which
made him rather frail and with his odd diet, he suffered from bad breath. This
made him a super callused fragile mystic hexed by halitosis.
In each case, the humor is elicited by the realization that the communication is not intended to be informative but rather, sets up the reader for a
play on words.
Shaggy Dog Stories. These stories are characterized by a lengthy description of events that appear to lead up to something interesting or exciting but turn out to be totally mundane. In this case, the humor does not
arise from an interpretation of the stories semantic content. Rather, it results from a reinterpretation of the situational context in which the information is presented. That is, the storys length and content gives the appearance of leading to something of interest and importance, whereas the
punch line reveals it to have no interest whatsoever.
Slapstick. There are many forms of slapstick humor. A potential challenge to the diminishment assumption could be the Laurel and Hardy variety, in which amusement is elicited by protagonists repeatedly bludgeoning one another with two-by-fours. Certainly hitting someone with a stick is
not inherently funny. The diminishment in this situation presumably arises
from the realization that the protagonists are not, in fact, killed or even
maimed by one anothers actions. That is, they only appear to hurt one another. Thus, the actual situation turns out to be more mundane than the
purported one.
Expectancy Deviations. In a study by Nehrhardt (1976), blindfolded
participants were asked to judge a series of weights. After several exposures to weights that were fairly similar to one another, they encountered
a weight that was either much lighter or much heavier than the others. On
receiving this weight, subjects typically smiled or laughed. Assuming that
participants were amused rather than simply registering surprise, the question is why. It seems reasonable to assume that participants who encountered the deviant weight inferred that they were being tricked and that
the experiment was not a serious study of weight judgment after all. In
other words, participants interpreted the situation as a whole as less im-
198
CHAPTER 8
portant than they had originally interpreted, and this reinterpretation elicited amusement.
In summary, the diminishment assumption not only helps to account for
a number of humor-eliciting phenomena that are hard for other theories to
explain, but also accounts for conditions in which experiences are reinterpreted but do not elicit amusement. For example, scientific discoveries, the
attachment of meaning to a persons dreams, and the reinterpretation of
events in mystery novels all involve a reinterpretation of a situation in light
of new information. In these cases, however, the reality that is implied by
the new information is of greater importance or value than the original, and
so amusement is not experienced.
A COMPREHENSIONELABORATION THEORY
OF HUMOR ELICITATION
The conceptualization of humor elicitation that Wyer and Collins (1992) proposed has much in common with incongruity-resolution theories and borrows particularly heavily from Apters (1982) formulation. At the same time,
it is derived largely from the assumptions surrounding the comprehension
of information outlined in previous chapters. Specifically, we assume that
amusement is the by-product of comprehension processes similar to those
described in chapters 4 and 5. However, the amount of humor that a joke
elicits can depend on the difficulty of comprehending it. Moreover, it can be
increased or decreased as a result of the amount and type of cognitive elaboration that is performed after its humor-eliciting interpretation has been
identified. We first outline the basic assumptions of the formulation and
provide empirical evidence of its implications. We then turn more specifically to the role of cognitive elaboration in humor elicitation and its implications for reactions to stories that people consider offensive. In our discussion, we focus largely on the humor elicited by jokes and stories to which
event and episode models pertain. As Wyer and Collins (1992) indicated,
however, the conceptualization is potentially applicable to the humor that
is elicited by social experiences in general.
Comprehension Processes
Although the comprehension processes we assume to underlie humor elicitation have been elaborated in previous chapters, it may be worthwhile to
reiterate them in the present context. This will again be done in terms of
postulates. The first postulate is simply a recapitulation of the processes
described in chapter 4.
199
When people have no goal in mind other than to comprehend the information they receive, the concepts and knowledge they apply are typically
those they have used most frequently and recently in the past (cf. Higgins,
Bargh, & Lombardi, 1985; Srull & Wyer, 1979; see Postulates 2.4 and 2.5). If
recipients have a more specific goal in mind, however, this goal may activate concepts that are relevant to its attainment, and these concepts may
then be used instead.
The second postulate reiterates the processes that underlie the construction of episode models, and formalizes the effect of previously encountered stimulus events on the processing of information about subsequent
ones.
Postulate 8.2. Once a mental representation has been formed of an experience, this representation, along with the concepts and knowledge that were
activated in the course of constructing it, is used to comprehend thematically
related experiences that occur subsequently.
200
CHAPTER 8
Similarly, the initial reference to the nun in the quickie joke mentioned
earlier activates expectations for her attributes that make her later comment to the priest difficult to interpret. Consequently, recipients identify a
different domain of knowledge (e.g., knowledge about prostitutes) in order
to comprehend it, with the result that the nun is reconceptualized as being
a prostitute as well.
These processes are captured by a third postulate, which essentially recapitulates assumptions made by Suls (1972, 1983) and Apter (1982):
Postulate 8.3. When information about an event cannot be interpreted in
terms of concepts drawn from the same domain of knowledge that was applied to previous events, recipients attempt to identify concepts and knowledge in a different domain that is applicable to both the new event and the
preceding ones. If these concepts can be found, the previously learned events
are reinterpreted in terms of them.16
Like Apter (1982), we further assume that diminishment can occur along
many dimensions and at several levels of generality. For example, the reinterpretation of a stimulus event might paint a more mundane (or, in some
cases, less desirable) picture of a person or event than the original interpretation implied. Alternatively, the reinterpretation could render the event it16
Note that Postulate 8.3 could be applied not only to stories but also to observed experiences or single statements that people make in a social context. To this extent, this postulate
governs processes similar to those described in chapter 7.
16
201
202
CHAPTER 8
will cost you $20, just like in town. Thus, this response preserves the original meaning of the question rather than stimulating a reinterpretation of it.
The third and fourth versions were similar to the first two except that the
person the priest asked was a third prostitute rather than a nun; these versions, then, eliminated the shift in perception of the respondent. Results
shown in Table 8.1 are very clear. That is, the original story, in which both
diminishing shifts in meaning were present, was judged as amusing. However, eliminating either shift in interpretation decreased amusement, and
eliminating both reinterpretations decreased it still further.
In a second study by Collins and Wyer, participants read a story that
could be interpreted in two different ways, one of which was less likely to
be identified than the other. One story, for example, was likely to be interpreted spontaneously as a conversation about the best way to administer
harsh physical punishment to children, but could also be interpreted as a
discussion of the best way to open a jar of pickles. A second story was most
likely to be interpreted as a mans comments to a woman in the course of
making love in the shower, but could also convey his comments in the
course of washing a dog.
No indication was given at the beginning of the story about the nature of
its subordinate theme. In some versions, however, a statement was inserted near the end of the story that was anomalous when considered in
terms of concepts activated by the dominant theme but made sense in
terms of the subordinate one (specifically, But honey, you know theres
nothing tougher than getting into a jar of pickles, and Honey, bring me the
flea powder, in the two stories, respectively). In other versions, this concluding statement was omitted. Participants were told to read the story for
understanding as they would if they encountered it in a magazine or novel.
After doing so, however, they reported how amused they were by it. Participants judged the stories more amusing when the statement that activated
the subordinate, trivializing theme was present than when it was not.
TABLE 8.1
Mean Humor Elicited by Quickie Joke as a Function of the Effects
of the Punch Line on Interpretation of Story Features
Effect of Punch Line on
Perceptions of Nun
Change
No Change
5.60
3.33
3.67
2.18
Note. Judgments reported along a scale from 0 (not at all humorous) to 10 (extremely humorous). Adapted from Wyer and Collins (1992).
203
Considered in isolation, this finding could have several alternative interpretations. For example, humor might be elicited simply by any reinterpretation of the story regardless of whether or not it was diminishing. This
problem was eliminated in a study by Linda Isbell (reported in Isbell, Wyer,
& Collins, 2002). Participants read a version of the pickle story that was
similar to that employed by Collins and Wyer but varied in terms of the relative salience of the two alternative interpretations. The story was given a
title that identified either the serious theme (Disciplining the Children) or
the mundane one (Getting a Pickle). Participants read this story along
with three filler stories with instructions that (a) the stories could each be
about two quite different situations, (b) each story was given a title that
identified one of the situations, and (c) they should generate an alternative
title that reflected a different interpretation. (The three filler stories described situations that were similar in importance; e.g., a game of cards vs.
a woodwind ensemble rehearsal.) After generating an alternative title for
each story, they gave their reactions to the story along several 010 scales,
one of which pertained to the amusement they experienced. As expected,
participants were more amused when their reinterpretation of the story
was more mundane than the original title conveyed (M = 6.12) than when it
was less so (M = 4.29), and this was true regardless of which version of the
story they read.
Thus, the results of these studies converge on the conclusion that stories elicit more amusement if they are reinterpreted in a way that diminishes the importance of either their referents or the stories themselves. At
the same time, the results do not indicate that diminishment is a necessary
condition for humor elicitation. In the preceding experiment, for example,
both reinterpretations that participants generated could have elicited some
humor, differing only in magnitude. The study of peoples responses to witticisms, described in chapter 7, is worth reconsidering in this context. To
reiterate, participants read some scenarios in which a statement was either
likely to be taken literally or likely to be viewed as ironic. Moreover, the intended meaning of the ironic statement was either less favorable than its literal meaning (thereby diminishing the value of its referent) or more so. Diminishing reinterpretations elicited more amusement than enhancing ones
(6.3 vs. 4.6 along a 010 scale), consistent with Postulate 8.4. On the other
hand, both types of ironic statements were interpreted as substantially
more amusing (M = 5.5) than their true counterparts (M = 1.7).17
17
A second aspect of these data is also worth noting. That is, statements with unfavorable implications were judged more amusing than statements with favorable implications, and this was
true regardless of whether the statements were taken literally (3.5 vs. 1.9) or ironic (6.3 vs. 4.6).
This finding, which would be consistent with disparagement theories (Zillman & Cantor, 1976),
suggests that inconsistency resolution is also not a necessary condition for humor elicitation.
17
204
CHAPTER 8
The humor elicited by this joke presumably derives from the reinterpretation of Descartes reply in the context of his philosophical conclusion, I
think, therefore, I am. However, someone who is not very familiar with Descartes philosophy might find the joke very bewildering.
Even when a jokes humor-eliciting interpretation can be identified, it
might not be perceived as funny if an excessive amount of cognitive activity
is required to understand it. On the other hand, jokes that are too easy to
understand are also unlikely to elicit much amusement. Thus, as suggested
by Suls (1972, 1983) problem-solving analogy, jokes that are either too easy
or too difficult to comprehend may be less amusing than those that are
moderately difficult. To formalize:
Postulate 8.5. The amount of amusement that is potentially elicited as a result of reinterpreting a stimulus event is a nonmonotonic (inverted-U) function of the time and effort that is required to make this interpretation.
In a sense, this postulate is not new. Many years ago, McClelland, Atkinson, Clark, and Lowell (1953) also hypothesized a nonmonotonic relation
novelty of a stimulus (and, therefore, the difficulty of understanding it in
terms of previously formed concepts and knowledge) and judgments of its
pleasantness. The question is why this relationship exists. Freud (1905/
205
206
CHAPTER 8
TABLE 8.2
Humor Elicited by Stories as a Function of Allusion
to Subordinate Theme in the Title and Text
Theme Suggested in Title
Subordinate
Dominant
6.17
576
6.85
4.32
5.76
3.87
6.67
2.94
Note. Judgments are reported along a scale from 0 (not at all humorous) to 10 (extremely humorous). Adapted from Wyer and Collins (1992).
The first interpretation of Adams response, which is induced by the warning, Stand back!, connotes danger. The implications of this interpretation
are diminished by the subsequent realization that the thing is only an
erection rather than a real danger and that the warning results from
Adams sexual naivete. To understand the joke, however, one must realize
that the thing is, in fact, an erection, and must know that Adam and Eve
have no prior knowledge of sex. Identifying these concepts and knowledge
requires cognitive effort, the magnitude of which should depend on
whether the relevant information is explicitly provided.
To evaluate this possibility, we constructed eight versions of the joke
that varied in the explicitness of information bearing on three types of
knowledge that were necessary to understand it: (a) the thing referred to
an erection, (b) it was Adams first such experience, and (c) Adam did not
know its size. Thus, the punch line when all three pieces of information
were provided was Stand back! This is my first erection and I dont know
how big it gets! In contrast, the punch line when none of the three pieces of
information was provided was simply Stand back!
207
Erection mentioned
Erection not mentioned
Ignorance of Size
Not Mentioned
First Time
Mentioned
First Time
Not Mentioned
First Time
Mentioned
First Time
Not Mentioned
5.57
6.00
6.43
8.00
2.63
2.12
3.25
1.00
Note. Judgments are reported along a scale from 0 (not at all humorous) to 10 (extremely humorous). Adapted from Wyer and Collins (1992).
Humor ratings are shown in Table 8.3 as a function of the amount and
type of information conveyed in the punch line. The joke was generally
more amusing when the punch line explicitly stated that Adam did not
know the size of the erection. More important, however, is the fact that
when Adams lack of knowledge was not explicit, providing one or both of
the other pieces of information increased the ease of understanding the
joke and, therefore, increased the humor it elicited. When Adams lack of
knowledge was explicitly mentioned, however, providing the other two
pieces of information presumably reduced cognitive effort below the optimal level required to appreciate the joke. Therefore, it decreased humor
ratings.
3. Situational and Individual Differences
in Comprehension Difficulty
The ease of identifying a humor-eliciting interpretation is not only a function of the explicitness of the information itself. Situational and individual
difference factors can play a role as well. For example, situational variables
that increase the accessibility of comprehension-relevant concepts in memory should increase the humor elicited by difficult-to-comprehend stimulus
experiences but should decrease the humor elicited by easy-to-comprehend experiences. On the other hand, factors that distract recipients from
comprehending the information in terms of these concepts should have the
opposite effects.
Two studies provide indirect support for this possibility. In a study by
Schick, McGlynn, and Woolam (1972), participants were exposed to two sets
of cartoon strips. All strips were originally from the Peanuts series and so
the characters and their personalities were well known. However, half of
the strips were redrawn to make the characters unfamiliar. Participants
were exposed to several strips in succession and asked to indicate how
208
CHAPTER 8
amused they were by each. People are likely to have some difficulty comprehending strips containing unfamiliar characters as they do not have a
previously formed body of knowledge to draw upon for use in construing
the implications of the strips and their situational context. However, comprehension difficulty should decrease as the characters become more familiar, and so the humor the strips elicit should increase. On the other
hand, suppose the characters in the strips are already familiar. Then, increases in the number of strips should not have this effect. In fact, humor
judgments of the redrawn cartoon strips increased as a function of the number of exposures to the strips. In contrast, the humor elicited by the original Peanuts strips was high at the outset and did not change as a function
of the number of exposures.
A perhaps more interesting finding surrounds the combined effects of familiarity and participants chronic anxiety (as inferred from the Taylor Manifest Anxiety Scale). Chronic anxiety might function as internal noise that increases participants difficulty in comprehending the stimuli. If so, it should
increase the humor elicited by familiar-character cartoon strips, which in
the absence of this noise are below the optimal level of comprehension difficulty. However, anxiety might decrease the humor elicited by unfamiliarcharacter strips, which may be above the optimal level of difficulty in the
absence of distraction. This was in fact the case.
It can be difficult to interpret the effects of situational manipulations of
comprehension difficulty without an a priori understanding of the difficulty
level of the stimuli to be comprehended independently of these manipulations. A study by Goldstein, Suls, and Anthony (1972) provides an example.
Subjects were initially shown photographs of either aggressive stimuli or
automobiles as part of an aesthetic preference task, thereby activating concepts that were associated with the domain to which the pictures pertained. Later, in an ostensibly unrelated experiment, participants judged
the funniness of cartoons in either the same domain to which the pictures
they saw were relevant or in the other, unrelated domain. Cartoons elicited
more amusement in the former condition than the latter. This could suggest
that the concepts activated by the photographs facilitated the comprehension of the cartoons that participants encountered later and, therefore, increased the ease of identifying their humor-eliciting features. According to
Postulate 8.4, however, this increase should only occur if the cartoons are
fairly difficult to comprehend in the absence of this facilitating influence. If
the cartoons were very easy to comprehend in the absence of priming, experiences that further increase the ease of comprehending them should
have precisely the opposite effect. Thus, Goldstein et al.s finding would
only be consistent with the conceptualization we propose if the stimuli
were of the first variety.
209
210
CHAPTER 8
The second friend says, I wish I were in the bosom of my family. Immediately, he is also gone.
The Texas Aggie thinks for a moment and says, Gee, I wish my friends were
with me now.
The humor elicited by this story presumably results from the realization
that the Aggies wish essentially negates the effects of the friends wishes,
returning them to the same abysmal situation they had been in before.
Once readers of the story identify these implications, they are likely to form
a mental image of the friends reactions to finding themselves back on the
island again. To this extent, the opportunity to think about the joke is likely
to increase the amusement it elicits. In contrast, consider the following:
Q.
A.
Why did the Texas Aggie want people to save their burned-out light
bulbs?
He needed them for the darkroom he was building.
Although this joke might be mildly amusing, it has little elaboration potential and so thinking about the joke after comprehending it should have little
impact on amusement. (Moreover, note that both jokes might be seen as
disparaging the intelligence of Texas A&M University students, to whom the
name Texas Aggie is typically applied. Students from this university might
be inclined to elaborate the nonhumorous implications of both jokes, and
this could decrease their amusement.)
Effects of Repetition
Some jokes continue to elicit amusement even when they are repeated one
or more times. This can be true despite the fact that the punch line is well
remembered and, therefore, no longer elicits a new interpretation of the
stimulus events being described. Other jokes, however, are less likely to
bear repeating, even though they might have been considered funny at the
time they were first heard. Suls (1972) conjectured that repetition effects reflect a general tendency for novel stimuli to become better liked as they become more familiar (Zajonc, 1968; but see McClelland et al., 1953). If this
were so, however, repetition should have similar effects on all jokes. Thus,
this conceptualization cannot easily explain why some repeated jokes continue to elicit humor but others do not.
It seems more reasonable to interpret repetition effect in terms of Postulate 8.6. If a joke has high elaboration potential, all possible implications of
it are unlikely to be considered at the time the joke is first encountered. To
this extent, repeating the joke might stimulate a different subset of implications than it did the first time, and these new implications could elicit hu-
211
mor. Thus, the desert island joke described earlier is likely to elicit somewhat different mental images and elaborations each time it is told or
thought about. Eventually, with still further repetitions, no new elaborations are likely to come to mind and so humor is no longer elicited. (All
jokes become stale eventually.) However, jokes that have low elaboration
potential to begin with (e. g., the burned out light bulb joke) may elicit little humor when they are repeated even once.
Cognitive elaborations may be aided by external stimulation. In fact,
much of the humor that is generated by professional comedians results
from their ability to stimulate their audience to elaborate the implications
of a humor-eliciting event with which they are already familiar. Moreover,
many movies and stories concern protagonists whose humor-eliciting behavior, although initially unexpected, is repeated in one form or another
throughout. Don Quixote, for example, repeatedly behaves in foolish ways
while appearing to be distinguished and chivalrous. Inspector Clouseau, the
French detective created by Peter Sellers in the Pink Panther movies appears sophisticated but is actually a buffoon. In such cases, the audience
soon becomes aware of the protagonists actual characteristics, and so his
or her later behavior is not unexpected. Yet, the behavior continues to provide amusement. For example, the first Pink Panther movie one sees leads
the ostensibly sophisticated Inspector Clouseau to be reinterpreted as an
inept detective who only appears sophisticated. Once this is established,
new elaborations of the reinterpretation are encountered as the film presents further instances of Clouseau appearing sophisticated but being in
reality a buffoon. The repeated instances of Clouseaus humor-eliciting behavior constitute externally generated elaborations of ones initial reinterpretation of the character and situation and function in much the same way
as self-generated elaborations. These elaborations elicit humor for much
the same reasons that self-generated elaborations do.
These considerations help to conceptualize the humor elicited in a large
number of situations in which the comprehension processes we postulate
might otherwise seem irrelevant. That is, many everyday situations appear
to elicit amusement although they do not require a reinterpretation. A colleague who is already known to be incompetent but who (perhaps like Inspector Clouseau) behaves in a pompous fashion elicits humor whenever
he or she inadvertently says or does something that betrays this incompetence. Neither the new event nor the colleagues attributes are reinterpreted. However, the colleagues behavior essentially constitutes an externally generated elaboration of the implications of a past event (an earlier
instance of the colleagues buffoonery that occurred in the context of his or
her pomposity) that did stimulate a diminishing reinterpretation at the time
it occurred.
212
CHAPTER 8
Sex differences in reactions to the latter joke were less clear a priori. Females
might find the joke more embarrassing than men. On the other hand, both
men and women might consider the joke to disparage people with disabilities. Moreover, cognitive elaboration of the joke could stimulate unpleasant
olfactory images and might decrease humor for this reason as well.
Be that as it may, we reasoned that if people are encouraged to think
about the implications of these jokes, they should judge a joke to be more
amusing if they elaborate its humor-eliciting aspects than if they think
about its humor-irrelevant aspects. However, suppose participants give
their spontaneous reactions to the joke without engaging in this postcomprehension elaboration. Then, they should be equally amused by the
joke regardless of the sort of elaboration they might otherwise be disposed
to perform.
1. Method
We told participants we were interested in reactions to stories of the
sort they encounter in daily life, and that to study this, we would like them
to react to a number of stories of the sort they might hear in situations out-
213
Sexual jokes
Males
Females
Scatological joke
Males
Females
Offensiveness Ratings
Spontaneous
Reaction Conditions
Thought
Conditions
Spontaneous
Reaction Conditions
Thought
Conditions
4.94
4.64
7.08
2.62
1.87
1.12
2.07
5.44
5.16
5.63
2.87
3.00
2.62
1.62
1.29
4.78
214
CHAPTER 8
215
216
CHAPTER 8
experimenter then proceeded to convey one of two jokes. One, sexual joke,
was the Adam and Eve joke described earlier in this chapter (see Table
8.3). The other was irrelevant to sexuality.
After this aside, the experimenter indicated that before going on with Dr.
Collins experiment (s)he would like the participant to perform two unrelated task. One was a filler task that was intended to disguise the relatedness of the studies. The second was a word association task in which participants were asked to select one of three words that was most similar to a
fourth. One alternative in each set had sexual overtones. For example:
Feeling: touch tingle stimulate
Oral: eating sex communication
Excited: winning aroused horny
Female
Experimenter
Male
Experimenter
Female
Experimenter
5.40
4.86
3.00
2.00
1.89
1.92
6.28
6.96
5.13
5.52
4.42
3.96
217
when they had heard a female tell a sexual joke than when they had heard
her tell a nonsexual joke, indicating that the females joke activated sexuality-relevant concepts. Correspondingly, as shown in the bottom half of Table 8.5, they were more amused by the sexual jokes they encountered in
these conditions. In contrast, male participants sexuality-related word associations increased only slightly when a male experimenter told a sexual
joke, and female participants sexuality-related associations actually decreased in these conditions. As a result, the male experimenters telling of a
sexual joke only slightly increased males amusement in response to the
sexual jokes they encountered later, and decreased womens amusement.18
Combined Effects of Elaboration
and Comprehension Difficulty
Although we have considered the effects of comprehension difficulty and
cognitive elaboration separately, these effects are actually interdependent.
For one thing, people cannot elaborate the implications of information they
are unable to interpret. Moreover, the experience of difficulty in comprehending the information could itself stimulate cognitive elaborations. For
example, if recipients believe they have had more difficulty understanding
a joke than they should, they might be stimulated to think about themselves and their competence rather than about the humor-eliciting aspect
of the information. Such humor-irrelevant elaborations could offset any
amusement that the joke would otherwise elicit.
A third consideration arises from the fact that people are typically neither willing nor able to devote an unlimited amount of time to the processing of any given stimulus event (Sperber & Wilson, 1986; Taylor & Fiske,
1978). This is particularly true in a social interaction in which a continuous
stream of events commands attention. Suppose people are unwilling to
spend more than a given amount of time thinking about a stimulus event.
To this extent, the more time and effort they devote to comprehending a
stimulus event, the less time they will devote to a subsequent elaboration
of its implications.
If this is so, the combined effects of comprehension difficulty and cognitive elaboration on humor elicitation can be described more precisely.
Suppose the total time that people are willing to devote to (a) the comprehension of an experience and (b) the cognitive elaboration of its humor18
Our interpretation of these results of course assumes that participants cognitive elaborations of the experimenters joke activated concepts that influenced their interpretation of the
sexual jokes they encountered later, thus increasing the humor the jokes elicited. This assumption is supported by a significant correlation between the amusement they reported in response
to these jokes and the number of sexuality-related word associations that participants made (r =
.39, p < .05).
18
218
CHAPTER 8
219
tion that people perform in the two conditions. However, it could be localized at the comprehension stage as well. For example, jokes often mention
the ethnic group to which they pertain at the outset. People who do not expect to be offended by a disparaging joke about this group may spontaneously activate a body of knowledge about attributes that are stereotypically
associated with the group and use this knowledge to comprehend the rest
of the joke or story. In such cases, activating the stereotype may facilitate
comprehension of the joke and, perhaps, increase the humor that it elicits.
On the other hand, people who identify with the group may anticipate that
the joke is going to offend them, and may think about the motives of its
source. This humor-irrelevant activity could interfere with the identification
of the jokes humor-eliciting interpretation. Whether or not this interference
increases or decreases amusement, however, depends on the difficulty of
comprehending the joke more generally.
The extent to which mentioning a stereotyped group activates humorirrelevant cognitive activity can also depend on characteristics of the storyteller. A joke that pertains to Jews is more likely to stimulate this activity in a
Jewish listener than in a Ku Klux Klan member. On the other hand, it is less
likely to do so if the joke-teller is Jewish than if (s)he belongs to the KKK.
However, not all jokes that refer to members of a stereotyped group are
disparaging. If people who anticipate that a joke will perpetuate a negative
stereotype engage in cognitive activity that interferes with comprehension, it
could affect the amusement the joke elicits even if the joke turns out not to
perpetuate the stereotype at all. Consider the following joke, for example:
A bus passenger notices that the woman sitting beside him is staring at him.
Im sorry to bother you, she says, but I wonder if you would mind my asking you a personal question. Are you by any chance Jewish?
I certainly dont mind your asking, replies the man good-naturedly, but
no, I dont happen to be Jewish.
The woman continues to stare at him, however, and asks again, Are you
sure youre not Jewish?
Yes, Im sure, says the man, now somewhat irritated. My mother is Roman Catholic, and my father is Japanese. Theres really no chance that I am
Jewish.
This still doesnt satisfy her, and as the trip continues, she becomes even
more insistent. Now, I know that youre Jewish. Tell me youre Jewish. Admit
it!
Finally, after repeated denials, the man, only to keep from being pestered
further, says, Okay. Have it your way, Im Jewish.
Well, the woman responds, you certainly dont look Jewish.
Thus, the joke does not disparage Jews, but rather, disparages the stereotype. However, this does not become clear until the punch line. The allusion
220
CHAPTER 8
to the ethnic group in the material preceding the punch line could stimulate
Jewish recipients to engage in cognitive activity that interferes with their
later comprehension. To this extent, they might consider the joke to be less
funny than non-Jewish recipients would, even though the jokes implications are consistent with their a priori values. (This conjecture assumes
that the joke is already moderately difficult to understand. If it were very
easy to understand, the increases in comprehension difficulty resulting
from this activity could increase Jewish individuals appreciation of the
joke rather than decreasing it.)
Effects of Cognitive Elaboration
Although individual differences in attitudes and values undoubtedly influence the comprehension of a joke or story, their primary effects are likely
to result from their impact on the cognitive elaboration that occurs subsequently. Suppose individuals encounter the following joke:
Q.
A.
This joke is extremely easy to understand. Its humor derives from the
different meanings of tips, coupled with knowledge of the stereotype of
Jews as being tight with money. Thus, differences in the humor the joke
elicits are likely to be due largely to differences in the type of cognitive
elaboration that recipients perform subsequent to its comprehension. That
is, people who are generally offended by the perpetuation of the stereotype
may think about why the joke is told and may question the attitudes and
motives of the storyteller. However, people who personally believe the stereotype, or are not concerned with perpetuating it, might not engage in this
humor-irrelevant elaboration.
There are two qualifications on this prediction, however. First, differences in reactions to the joke should only be evident when people have an
opportunity to elaborate its implications after comprehending it. When this
opportunity does not exist, recipients may be equally amused by the joke
regardless of their feelings about the group being disparaged. (This could
explain why people often laugh at jokes they hear at a party that they
would find unamusing in a less cognitively demanding situation in which
they had more opportunity to think about the jokes implications.)
A second contingency surrounds the elaboration potential of a joke.
When a jokes humor-eliciting reinterpretation evokes visual images, or if
the situation it implies can be elaborated, individuals who are not offended
by the joke may elaborate its humor-eliciting implications and, therefore,
may be more amused by the joke than they might otherwise be. However,
221
jokes of the sort provided in our example, which are simply plays on words,
have low elaboration potential. The amusement elicited by these jokes is
unlikely to increase with the opportunity to engage in elaboration regardless of recipients attitudes toward the stereotyped group. In fact, if individuals have nothing to distract them, any additional postcomprehension
thoughts they might have about the joke could pertain to the source of the
joke, and perhaps the validity of the stereotype, and these thoughts could
occur even if these individuals do not identify with the group being disparaged. As a result, the amusement experienced by these individuals, like persons who are disposed a priori to elaborate the humor-irrelevant implications of the joke, might decrease when they have an opportunity to think
about the joke more extensively.
A second study with Linda Isbell (see Isbell et al., 2002) examined this
possibility. Participants under thought and spontaneous reaction conditions similar to those in Isbells earlier study were asked to read and give
their reactions to six stories. Four of the stories were nonhumorous, but
two others were jokes that had negative implications for the protagonists
intelligence. One of these jokes (e.g., the desert island joke described earlier) had high elaboration potential, whereas the other (e.g., the burnedout light bulb joke) had low elaboration potential. Moreover, the protagonist in one joke was described as Polish and in the other was described as a
Texas Aggie. (The type of protagonist associated with each joke was of
course counterbalanced.)
The midwestern college students who participated in this study were undoubtedly aware of the stereotype of Polish as unintelligent. However, they
were typically not of Polish extraction themselves, and were unlikely to be
offended by the jokes. (Participants reactions to the jokes in the experiment confirmed this assumption; 65% of the participants reported being
not at all offended by these jokes, or 0 along the 010 scale used to report
their judgments.) Although Texas Aggies are sometimes stereotyped as unintelligent as well, this stereotype was generally unknown in the student
population from which participants were drawn. Based on the assumptions
outlined earlier, we expected that encouraging participants to think about
when Polish jokes with high elaboration potential would stimulate them to
elaborate the humor-eliciting aspects of the jokes, leading them to be more
amused by the jokes than they would otherwise be. In contrast, encouraging participants to think about Polish jokes with low elaboration potential
should stimulate them to elaborate the jokes negative implications for the
stereotyped group to which the jokes refer, and thus should lead them to
be less amused by the jokes than they would otherwise be.
Results were consistent with these assumptions. Participants reported
being equally amused by Texas Aggie jokes regardless of whether they
were told to think about them (M = 4.16) or to give their spontaneous reac-
222
CHAPTER 8
tions (M = 4.52). In contrast, thinking about Polish jokes with high elaboration potential increased the amusement the jokes elicited from 4.04 (under
spontaneous reaction conditions) to 5.13 (under thought conditions),
whereas thinking about Polish jokes with low elaboration potential decreased amusement from 4.96 to 3.86. These differences were reflected in an
interaction of elaboration potential and instructional conditions in an analysis of reactions to Polish jokes alone, F(1,40) = 4.27, p < .05.
Thus, the results provide insight into peoples reactions to jokes that
perpetuate a negative stereotype. Thinking about a joke that perpetuates a
negative stereotype can obviously decrease the amusement experienced
by persons who are offended by the joke. However, thinking about such a
joke can also decrease the amusement experienced by persons who are not
offended by the joke, provided the jokes humor-eliciting implications cannot easily be elaborated.
The Effects of Social Context on Responses
to Sexist Humor
Humor-irrelevant cognitive elaborations are particularly likely when participants themselves belong to the stereotyped social category and, therefore,
are apt to be offended by it. However, this is not the only factor that can influence humor-irrelevant elaboration. In some cases, recipients might perceive that a joke-teller is motivated by a desire to ingratiate his or her audience, or alternatively, to convey hostility toward the person or group that
the joke disparages. They might think about the implications of these motives regardless of whether or not they are personally offended by the joke.
Consequently, they might be relatively less amused by the joke than they
would be if the joke-teller were simply trying to be funny. In other cases,
people might be aware of the motives that lead a joke to be told and still
elaborate its humor-eliciting implications. (This is particularly true if the
jokes elaboration potential is high, as suggested by the results of Isbells
experiment described earlier.)
The effects of these contextual factors may be particularly evident (at
least, among college students) when jokes exploit a negative stereotype of
men or women. Consider the following joke:
A man and a woman are in the elevator together. The woman suddenly starts
taking off her clothes and says, Make me feel like a woman. The man takes
off his clothes, throws them on the floor, and says, Here, fold these.
223
ture of these reactions may depend on both the sex of the communicator
and the social context in which the joke was originally told. The effects of
these factors may be mediated by their influence on recipients assumptions about the speakers motives and, therefore, on the amount and type
of cognitive elaboration that they perform.
1. Perception of the Motives for Telling Sexist Jokes
The motives that people perceive to underlie the communication of sexist jokes were investigated by Isbell et al. (2002). Male and female college
students were introduced to the study with instructions that we were particularly interested in the motives that people attribute to individuals who
tell jokes that convey negative stereotypes of men and women. With this
preamble, we gave participants a short questionnaire containing four (female-bashing) jokes that perpetuated a negative stereotype of women and
four other (male-bashing) jokes that conveyed a negative stereotype of men.
Each joke was preceded by an indication of the sex of both the person who
told the joke and the individual to whom it was told. (Each joke was associated an equal proportion of times with each of the four speakeraudience
configurations.)
Participants were asked to imagine that each joke had been told in the
situational context described and to estimate the likelihood that the joketeller was (a) just trying to be funny, (b) trying to gain social approval, or
(c) being intentionally hostile or aggressive. They also estimated the likelihood that the joke-teller personally believed in the stereotype being conveyed. These estimates, reported along scales from 0 (not at all) to 10 (extremely), are shown in Table 8.6. as a function of the joke-tellers sex, the sex
of the audience, and the type of joke that was told. Between-cell comparisons, shown in the table, suggest the following conclusions:
1. Participants in all cases inferred that the joke-tellers primary motive
was to be funny (M = 7.30). However, they perceived this motive to be particularly strong (M = 8.08) when someone told a joke that disparaged the opposite sex to a member of his or her own sex (i.e., when a man told a femalebashing joke to another man, or a woman told a male-bashing joke to another
woman).
2. Joke-tellers were perceived to have less strong personal beliefs in the
stereotype implied by a joke when the joke disparaged members of their
own sex (M = 3.16) than when it disparaged members of the opposite sex (M =
5.10). However, this difference was evident both when men told a malebashing joke (2.28 vs. 4.97) and when women told a female-bashing joke (4.09
vs. 5.23).
224
CHAPTER 8
Female Joke-Teller
Female
Audience
Male
Audience
Female
Audience
7.67ab
2.72b
3.69b
1.66b
7.47ab
1.84b
6.75a
1.09b
6.59b
5.53a
3.06b
4.47a
8.12a
4.41a
4.75b
3.44a
7.46
3.62
4.56
2.66
8.04a
4.58ab
6.46a
2.79
7.00ab
5.88a
2.62b
3.75
7.08ab
4.17b
5.29a
2.71
6.46b
3.92b
3.38b
3.38
7.14
4.63
4.43
3.16
Note. Means in each row with unlike superscripts differ at p < .05.
225
226
CHAPTER 8
Female
Audience
Female Joke-Teller
Male
Audience
Female
Audience
Amusement Ratings
Male participants
Male-bashing joke
Female-bashing joke
Female participants
Male-bashing jokes
Female-bashing jokes
4.21
4.18
4.03
4.72
2.43
3.51
4.33
5.07
6.17
3.37
6.80
3.77
5.96
4.04
6.89
3.39
Offensiveness Ratings
Male participants
Male-bashing jokes
Female-bashing jokes
Female participants
Male-bashing jokes
Female-bashing jokes
1.46
1.00
2.50
1.98
3.00
1.93
1.63
1.43
0.90
2.43
0.87
3.70
1.57
3.54
0.39
3.23
227
CONCLUDING REMARKS
Humor is a fundamental ingredient of social communication in everyday
life. A specification of the conditions in which it is elicited is therefore a necessary component of any theory of real-world information processing. We
have argued that the experience of amusement is partly a result of the ba-
228
CHAPTER 8
P A R T
III
INFERENCE PROCESSES
C H A P T E R
9
Crimes, Vacations, and Political
Candidates: The Construction and
Use of Narratives in Social Judgments
232
CHAPTER 9
233
their verdict were given a recognition memory test containing some statements that had actually been made in the course of the testimony and others that had not. Participants were more likely to report a statement as having been mentioned in the story if its implications were consistent with the
narrative-based causal scenario on which they based their verdict than if
its implications were inconsistent with this scenario.
A second series of studies (Pennington & Hastie, 1988) examined these
inference processes more directly. Participants listened to the transcript of
an actual murder trial containing testimony for both the prosecution and
the defense. In witness-order conditions, the evidence favoring a given verdict was organized according to the witness who provided it, as in the original trial. In story-order conditions, the evidence for a given side was presented in the order it became relevant in the sequence of events that had
allegedly occurred. (That is, testimony about events that preceded the incident was presented first, followed by testimony about the incident itself,
the arrest, the autopsy, etc.). This was done in a 2 2 design so that the order of presenting the prosecution testimony and the order of presenting
the defense testimony were varied independently. Participants after receiving the testimony reported both their verdict and their confidence that
their decision was correct.
When the defense and prosecution testimonies were presented in different orders, 73% of the participants recommended the verdict implied by the
evidence that was conveyed in story order. When the evidence favoring
both verdicts was presented in the same way, the percentage of participants who favored a guilty verdict did not depend on how the evidence was
ordered. However, participants were much more confident of their judgments when the evidence was in story order. In short, evidence that was
conveyed in an order that facilitated the construction of a narrative account of the crime was apparently easier to comprehend, and so recipients
were more confident of its implications. Consequently, they were more
likely to base their judgments on this evidence than on evidence conveyed
in other ways.
This conclusion has further implications. For example, if people have
formed a narrative of the sequence of events that occur, they may be more
receptive to additional information that is consistent with the implications
of the narrative as a whole than to information that conflicts with these implications. Moreover, this could be true regardless of the credibility of the
informations source. In contrast, if people are unable to construct a coherent narrative representation of the information, they may construe the implications of each piece separately. In this case, the credibility of source of
this additional information should be given more weight.
234
CHAPTER 9
A COMPARISON OF NARRATIVE-BASED
VERSUS NONNARRATIVE-BASED
INFORMATION PROCESSING
General Considerations
A narrative account of a crime is of particular relevance in assessing the
likelihood that the defendant was motivated or able to commit it. Participants in Pennington and Hasties studies may conceivably have tried to
construct a narrative account of the events that occurred regardless of how
the information was presented, but simply found it easier to attain this objective when the information was conveyed in temporal order. As these authors point out, however (Hastie & Pennington, 1995), many of the decisions
we are called upon to make in daily life are not typically based on narratives. Rather, they are based on spatial, functional, hierarchical, and argumentative criteria. The mental representations that are formed in the
course of applying these criteria may be quite different. Thus, for example,
computer commands might be cognitively organized on the basis of functional considerations, and the arguments bearing on a particular issue
might be organized according to the alternative points of view they support. Goal-related differences in the representations that people form of information, which were discussed in chapter 6, are considered more generally by Pennington and Hastie (1993).
Even within a knowledge domain, a number of alternative strategies
might be used to compute a judgment or decision. For example, the evaluation of a consumer product could be based on either an imagined sequence
of events that surround the use of the product and its consequences or a
piecemeal analysis of the products individual features (cf. Fiske &
Pavelchak, 1986). In such cases, the strategy that people use, and the mental representation they construct in the course of applying it, may depend
on situational or informational factors that exist at the time the information
is received. Two factorsthe format in which the information is presented
and the presence or absence of picturesare of particular interest.
1. Effects of Presentation Format
on Processing Strategies
When information describes a temporally ordered sequence of events
that are situationally constrained, people may spontaneously construct an
episode model of this sequence in the course of comprehending it (Postulate 4.1). Under these conditions, people are likely to base their judgments
on the implications of the model as a whole without analyzing its individual
features. In contrast, suppose individuals receive the same information in
an ostensibly unordered list. In this case, they may not attempt to assess
235
236
CHAPTER 9
237
were conveyed in an ostensibly unordered list. In narrative-format conditions, for example, the India vacation brochure began:
On your vacation, you will start out from the capital of India, Delhi, and move
on to see the Taj Mahal. Later, you will go west and see the palaces and temples in the colorful deserts of Rajasthan . . . before heading south. Further
south, you will visit the beaches of Goa, the tropical forests and backwaters of
Kerala, and . . . complete your trip at the southernmost tip of India.
This was followed by written narrative descriptions of the places to be visited. For example:
Only a short trip from Delhi is Agra, the home of the Taj Mahal. A mausoleum
built by Shah Jahan for his empress, the Taj is widely regarded as the most
beautiful man-made structure in the world. Remarkable at all times of the day,
you will visit as the sun rises above the early morning mists and return on a
moon-lit night when the Taj seems to float unattached above the blue-green
pools in front of it.
In contrast, the India brochure under list-format conditions described aspects of the vacation without indicating the order in which the events described would occur. This brochure began with a simple listing of highlights:
Some features of your vacation experience are:
a visit to the capital, Delhi
the cool mystery of the forests
the forest tribes
the Taj Mahal at Agra
palaces and temples in Rajasthan, etc.
Moreover, the individual places and situations were described in a similar manner:
Agra, home to one of the most beautiful man-made structures in the world.
the Taj Mahala mausoleum built by Shah Jahan for his empress.
a beautiful spectacle both when the sun rises above the early morning mists
and on moonlit nights when the Taj seems to float above the blue-green
pools in front of it.
238
CHAPTER 9
1. Experiment 1
In the first study, an attractive picture accompanied the verbal description of each event. However, the relative dominance of the pictures and verbal descriptions varied. In some cases, the pictures were large and the verbal descriptions were in small (10-point) font. In other cases, the pictures
were relatively small and the verbal descriptions were in large (18-point)
font. To ensure that format and picture-word dominance were not confounded with information content, the particular vacation representing
each combination of these variables was counterbalanced. Participants after reading both brochures evaluated the extent to which the brochure pertaining to each vacation made them want to go there. Then, they reported
the difficulty they had had in imagining each of the two vacations and finally, recalled the events contained in each brochure.
Processing Efficiency. The ease of processing the information under
each presentation format condition was inferred from both participants
self-reported ease of imagining each vacation and their memory for events
that would occur in it. Data pertaining to these indices are shown in the
first two sections of Table 9.1. When the text information was dominant,
participants found it fairly easy to imagine the vacation regardless of format, and had relatively better recall of the places they would visit when the
information was conveyed in a list. When the information was conveyed in
a narrative, however, increasing the dominance of the picture decreased
the difficulty of imagining the vacation and increased recall of the places to
be visited. When the same information was presented in a list, on the other
hand, increasing the dominance of the pictures had the opposite effects. In
other words, pictures appeared to facilitate the processing of information
that was presented in a narrative. However, they interfered with the semantic, piecemeal processing of this information when it was conveyed in a list.
Effects on Judgments. The combined effects of format and pictures in
evaluations of the vacations are consistent with the processing differences
just described. The bottom section of Table 9.1 shows evaluations of the
first vacation that participants considered.19 When the text was dominant,
participants rated the vacation somewhat less favorably when the information about it was conveyed in a narrative than when it was described in a
list. However, increasing the dominance of pictures increased evaluations
19
Participants appeared to use the same criteria to judge the second vacation they considered that they used to judge the first one, even though the information was conveyed in a different format. Thus, the effects of format on judgments of the second vacation were the mirror image of its effects on judgments of the first. (This was true in the second experiment to be
reported as well.)
19
239
Narrative Format
List Format
3.2
4.5
6.3
4.7
.52
.48
7.3
6.5
.37
.55
4.1
7.9
of the vacation in the first case and decreased it in the second. As a result,
evaluations of a vacation when pictures were dominant were much more favorable when the vacation was described in a narrative than when it was
described in a list.
2. Experiment 2
A second experiment confirmed these conclusions. In this study, the
text-dominant condition was replaced by one in which participants received only text information and no pictures were presented at all. In addition, some participants were explicitly told to imagine themselves actually
having the experience described on each page of the brochure. We reasoned that if the effects of picture results from the mental images they stimulate explicitly telling participants to form their own mental images of the
events might have similar effects.
Results confirmed this reasoning. Participants evaluations of the first vacation they considered are shown in Table 9.2 as a function of presentation
format, the presence of pictures and imagination instructions (see Footnote
20). When the vacation was described in a narrative, participants evaluated
it more favorably when both pictures were presented and they were told to
imagine the events described (M = 8.1) than they did under other conditions (M = 6.4). When the vacation was described in a list, however, participants evaluated it somewhat less favorably in the first condition than in
other conditions (5.8 vs. 6.1). Thus, both studies converge on the conclusion that people form mental images of events that are conveyed in a narrative, and that situational conditions that encourage the construction of
these images increase the extremity of their judgments. On the other hand,
these same conditions interfere with the piecemeal processing that under-
240
CHAPTER 9
TABLE 9.2
Evaluations of Vacations as a Function of Presentation Format,
Presence of Pictures and Imagination Instructions
Narrative Format
List Format
8.1
6.3
5.8
6.8
6.4
6.5
6.1
5.5
Imagination instructions
Pictures presented
No pictures
No imagination instructions
Pictures presented
No pictures
Note. Adapted from Adaval and Wyer (1998).
241
Wyers studies, it is unclear which was the case. It was also unclear whether
the facilitating effects of pictures are restricted to conditions in which episode models are likely to be formed from the information presented or
whether their effects generalize to conditions in which the narrative information is unlikely to activate these models (e.g., abstract descriptions of
events that are not temporally or situationally constrained).
Adaval, Isbell, and Wyer (2003) provided evidence of this generalizability. Some of the more important judgments that people are called upon
to make outside the laboratory are their evaluations of candidates for public office. These judgments are presumably based in part on the candidates
stands on social and political issues. Ever since the 1960 KennedyNixon debates, however, it has been widely recognized that voters evaluations of
politicians are influenced at least as much by the global images of the candidates that are largely unrelated to the candidates political ideology or
specific issue positions (Englis, 1994). These images are undoubtedly stimulated in part by a candidates physical appearance, as conveyed in a magazine or on television. However, they can also be influenced by descriptions
of the individuals activities over the course of his or her lifetime that have
implications for intelligence, morality, or steadfastness. Thus, information
that a candidate had smoked marijuana in college, or performed a heroic
deed in World War II, can have an impact on inferences about the candidates character and, therefore, judgments of his or her suitability for office
at a much later point in time.
Adaval et al. (2003) examined the processes that underlie the construction of the images that people form of politicians using procedures similar
to those employed in the vacation studies. The first two studies confirmed
Adaval and Wyers earlier findings and, in doing so, provided evidence that
the processing strategies that participants employed under different presentation format conditions were activated and applied automatically, with
little awareness. Two other studies used recognition response times to understand more specifically the cognitive representations that were formed
of information under conditions in which the strategies were employed. A
final study explored the role of both pictures and mental imagery in political information processing and clarified further the processes that contributed to judgments in the earlier experiments.
Participants in an initial study were told we were interested in the impressions people form of famous personalities. On this pretense, they
read brochures about two political figures who had been well known during the period they held office. Each brochure began with an overview of
the major events in the politicians life, followed by 12 additional pages
each describing an event or situation in more detail. As in the vacation
studies, the information in one brochure was conveyed in a narrative and
the information in the other was conveyed in the form of a list. In narra-
242
CHAPTER 9
tive-format conditions, for example, the brochure describing one politician (Thomas Winters) began:
Thomas Winters was a well-known political figure between 1950 and 1975. He
was a veteran of World War II and served as an executive of General Motors
before becoming Governor of Michigan. He then served two years as a U. S.
Senator, and ended his career as a special envoy to China.
Other activities included urging the government to halt the bombing in Vietnam, donating his summer home for use by a charitable organization, hosting the Pope during his visit to America, and helping to revise the state budget to provide support for crime prevention.
In contrast, the brochure under list-format conditions described the
events in the politicians life in bullet form and did not indicate their temporal relatedness. Thus, the brochure pertaining to Winters began:
Thomas Winters was a well-known political figure between 1950 and 1970. He
was:
a member of the U. S. Senate
a World War II veteran
a General Motors executive
Governor of Michigan
special envoy to China
Although the individual events were conveyed in the same order they were
presented in narrative-format conditions, they were also conveyed in bullets that had no temporal implications:
Was sensitive to the interests of the public while Governor of Michigan.
Went on television to oppose the construction of a nuclear waste processing
plant that would contaminate the citys water supply.
243
Impressions of politician
Pictures
No pictures
Number of items recalled
Pictures
No pictures
Narrative Format
List Format
3.98
3.58
3.63
3.95
5.79
5.04
4.92
4.98
244
CHAPTER 9
facilitated memory for the events in the former case but not the latter. In
contrast to Adaval and Wyers studies (see Footnote 20), these effects were
evident for both the first politician that participants considered and the
second.
Deliberative Versus Spontaneous Influences of Format
on Information Processing Strategies
If participants intentionally use a holistic or piecemeal strategy when they
encounter information in a narrative or a list, respectively, explicit instructions to employ the strategies should override the effects of format. On the
other hand, suppose the strategies have the status of cognitive productions
that are activated and applied automatically when information to which
they are applicable is encountered. Then, they may have an influence regardless of participants conscious attempts to employ other criteria.
To evaluate these possibilities, participants in a second experiment were
exposed to stimulus materials identical to those in the first study. In this
case, however, we told participants explicitly which processing strategy we
would like them to use. Under piecemeal-instruction conditions, we indicated:
Sometimes people form impressions of a person by focusing on specific
events in a persons life (e.g., winning a lottery). At other times, however, they
try to think about the persons life as a whole and whether it was admirable,
successful, happy, etc. In this study, we would like you to use the first strategy. That is, when reading about the political figures described in the brochures,
try to imagine the specific events that occurred in each politicians life and their
implications for your impression of the politician. Use these individual events as a
basis for your impression.
245
Schematic instructions
Pictures
No pictures
Piecemeal instructions
Pictures
No pictures
Mean
Pictures
No pictures
Narrative Format
List Format
3.94
3.46
3.61
4.03
3.76
4.00
3.12
4.02
3.85
3.73
3.37
4.03
adding pictures slightly increased evaluations of politicians who were described in a narrative and decreased evaluations of politicians based on information conveyed in a list. As a result, the advantage of a narrative over a
list format was significantly greater when pictures were presented (3.85 vs.
3.37) than when they were not (3.73 vs. 4.02), F(1,56) = 7.49, p < .01. These effects did not significantly depend on the instructions that participants were
given concerning the strategy they should use (p > .10).
The effects of task demands cannot be ruled out entirely, however. Note
that pictures facilitated impressions to the greatest extent when the information was conveyed in a narrative and participants were told to apply a schematic criterion (3.94 vs. 3.46 under picture and no-picture conditions, respectively). Correspondingly, they interfered with impressions to the greatest
extent when participants were told to base their judgments on specific
events and the information as conveyed in a list (3.12 vs. 4.02, respectively).
These data therefore suggest that participants did attempt to apply the criteria they were told to use, and these conscious efforts added slightly to the
spontaneous effects that were induced by the format of the information.
However, neither the effect of task demands nor its interaction with pictures
was statistically significant. Thus, the impact of these demands was very
small in relation to the effects of presentation format per se.
Recall data also confirmed the results of the first experiment. That is,
participants recalled a significantly greater number of events when the information was conveyed in a narrative than when it was conveyed in a list
(4.91 vs. 4.22), and this effect was greater when pictures were presented
(4.99 vs. 3.85) than when they were not (4.82 vs. 4.58). However, these effects also did not depend on instructional set.
246
CHAPTER 9
2. Self-Report Data
The failure for participants judgments to be appreciably influenced by
explicit instructions concerning the criteria they should use suggests that
the strategies that are elicited by different presentation formats were applied unintentionally. In fact, the use of these strategies, and the impact of
pictures on their effectiveness, may occur without awareness. Several aspects of the supplementary data we collected suggest this conclusion. For
one thing, participants self-reports indicated they attempted to apply the
same strategy in evaluating the second politician they saw that they had applied in judging the first one, even though the information about the second
politician was conveyed in a different format. In fact, however, their actual
judgments of the two politicians were quite dependent on format in much
the same way.
Second, the effect of format on participants actual evaluations of the
politicians depended on whether pictures were presented but did not depend significantly on task demands (see Table 9.4). In contrast, its effect on
the criteria that participants reported using as a basis for these evaluations
depended on task demands but not on the presence of pictures. Moreover,
analyses of participants estimates of the extent to which pictures facilitated the impressions they formed did not significantly depend on either
task demands or presentation format.
These results therefore argue against the possibility that the effects of
presentation format reflect a conscious attempt to comply with implicit format-based expectations concerning the criteria to be used in making judgments. Rather, different presentation formats appear to activate different
processing strategies that participants apply automatically, and that have
effects independently of the strategies they actively try to use. The presence of pictures may nevertheless facilitate or interfere, depending on the
particular strategy that participants actually employ.
The Role of Pictures in Narrative-Based Processing:
A Reconceptualization
Although the interactive effects of pictures and presentation format in the
preceding studies were very similar to those observed by Adaval and Wyer
(1998), they nevertheless raise questions about the content and structure of
the mental representations that underlie these effects. The assumption that
the event descriptions conveyed under narrative-format conditions were
stored in memory as a sequence of temporally related events was confirmed by Adaval et al. (2003) in a study of recognition memory. Participants were exposed to stimuli under conditions comparable to those constructed in the first experiment by Adaval et al. (2003). Rather than judging
the politicians, however, they were given a timed recognition memory test
247
for the verbal descriptions of the events presented. The order in which the
recognition items were presented was systematically varied. We assumed
that if two presentation items were associated in memory, exposure to one
of these items would spontaneously activate the second, and so the speed
with which the second item could be identified would be increased. When
the original information had been conveyed in a narrative, participants
were quicker to identify an event if it was preceded by a description of the
event that had come before it in the presentation sequence (M = 3.68 s) than
if it was preceded by a different item (M = 4.16 s). When the items had been
conveyed in a list, however, this difference was negligible (4.21 s vs. 4.18 s).
These data suggest that the events were organized in memory in a temporal sequence when they were conveyed in a narrative, but that this was not
the case when the events were presented in a list.
The influence of pictures on the representations that participants
formed is less clear. We assumed that when information was conveyed in a
narrative, participants would form mental images of the events and that
these images would become part of the representations of these events
that they stored in memory. Unlike Adaval and Wyers (1998) studies, however, many of the pictures used in the present experiment were only indirectly connected to the event descriptions they accompanied. (E.g., the verbal description of Harrison as displaying courage under enemy gunfire in
World War II was accompanied by a picture that showed him in an army
uniform but did not portray him as actually engaged in combat. Similarly,
the verbal description of him as showing sensitivity to the needs of the
poor and disabled, and as initiating legislation to provide assistance to
homeless children, was accompanied by a picture of him seated a desk
rather than as actively involved in this behavior.) Thus, although the pictures may have facilitated the construction of mental images of the events
presented, their specific content may not itself have been part of the representation that was formed and stored in memory.
A second recognition memory study confirmed this speculation. In this
study, both pictures and verbal event descriptions were used as test stimuli. Participants did not recognize an event description any more quickly if
it had been preceded in the recognition series by the picture that had accompanied it in the original stimulus materials than if it had been preceded
by a novel picture, and this was true regardless of whether the events had
been conveyed in a narrative (1.84 s vs. 1.81 s) or in a list (1.66 s vs. 1.60 s).20
In other words, the specific picture that accompanied the event description
20
The generally shorter recognition times in this study than the preceding one is attributable
to the fact that American students participated in this study and Hong Kong Chinese students in
the earlier one. The latter students generally took longer to read the English phrases than Americans did.
20
248
CHAPTER 9
was not associated with the description in memory under either format
condition.
These data therefore suggest that pictures played a somewhat different
role in the representations formed under these conditions than they played
in Adaval and Wyers (1998) research. The nature of this role is suggested
by a conceptualization of visual information processing by Barsalou (1993).
That is, a photograph of a person might stimulate the formation of a perceptual symbol that people mentally manipulate along with other symbols
to construct a visual image of the persons activities in not only the situation to which the picture pertains but other situations as well. (This could
occur in much the same way that a previously formed mental picture of a
colleagues physical appearance stimulates an image of his or her actions in
a story that someone tells.) Moreover, this perceptual symbol may be applied across situations, establishing a stronger connection between the
events that occur in a narrative than would otherwise exist. Pictures of a
person could create such a symbol and, therefore, could facilitate the construction of a visual image of events in which the individual is involved
even if they are not themselves directly related to these events.
If this is the case, however, pictures of the person should facilitate the
construction of a coherent image-based narrative representation of the persons life even when the pictures do not become associated with the specific events that accompany them. For that matter, pictures of the politician
should have similar effects even if they do not accompany the written event
descriptions at all, but are presented at the outset, before any of these descriptions are provided.
These speculations were confirmed in an additional study described in
the next section. In this study, pictures did not accompany the verbal descriptions of the politicians life events. In some cases, however, four photographs of the politician were provided at the outset, before the event descriptions were presented. Each picture provided a clear image of the
politicians physical appearance but contained few other indications of the
context in which the pictures were taken.
Relative to conditions in which no pictures at all were presented, introducing pictures at the outset increased the favorableness of participants
evaluations of the politicians, and this was true both when the information
was conveyed in a narrative (3.56 vs. 2.88) and when it was listed (3.35 vs.
2.56). Thus, pictures that were presented at the outset had the same influence under narrative-format conditions that they had when they were conveyed in the context of the event descriptions. Furthermore, presenting pictures before the event descriptions in list-format conditions eliminated the
interference effects that occurred when the pictures accompanied these descriptions, leading the pictures to have a positive influence on judgments
under these conditions as well.
249
250
CHAPTER 9
Pictures
Narrative format
List format
Mean
No pictures
Narrative format
List format
Mean
Nonvisualizers
Imagination
Instructions
No Imagination
Instructions
Imagination
Instructions
No Imagination
Instructions
3.33
3.50
3.42
3.08
2.83
2.96
2.73
3.07
2.90
4.04
3.87
3.96
3.08
3.71
3.40
3.10
2.67
2.88
3.83
3.54
3.68
2.67
2.46
2.56
251
curred in his life than when they were not, and this was true regardless of
the format in which the information was conveyed. Moreover, pictures had
no effect whatsoever on these participants evaluations. That is, visualizers
apparently formed clear images of the events even in the absence of pictures, and so making the pictures available had little additional effect.
In contrast, pictures had a positive influence on nonvisualizers evaluations when these participants were not explicitly told to imagine the politicians life events. Furthermore, imagination instructions had a positive effect provided no pictures were presented. When nonvisualizers were asked
to imagine the politicians life events and pictures were presented, however, the two sources of mental images appeared to interfere with one another, thus eliminating the positive effect that each factor had in isolation.
As noted earlier, the verbal information we presented typically described abstract events that were unlikely to elicit visual images spontaneously. Nevertheless, chronic visualizers found it easy to construct images
when they were asked to do so, and the presence of pictures neither helped
nor hindered them. Therefore, the images that these participants formed increased the extremity of their judgments regardless of whether or not pictures were presented. Nonvisualizers, who do not normally form visual images on the basis of verbal information, were also benefited by doing so
provided no pictures were presented. When pictures were presented, however, these individuals apparently found it difficult to integrate their implications into the images they formed, and this led them to form less clear impressions of the politicians than they might otherwise have done.
Effects of Image-Based Representations
on the Processing of New Information
The preceding series of studies focused on how visual and verbal information about a politician combined to affect his image. Once this image is
formed, however, it can have an impact on the way other information is processed and consequently on the implications drawn from this information. A
study by Wyer, Budesheim, Shavitt, Riggle, Melton, and Kuklinski (1991) is of
interest in this regard. Nonacademic employees were recruited for a study of
the way people make judgments of political candidates on the basis of information of the sort they receive during an election campaign. On this pretense, they received two types of information about a member of the U.S.
House of Representatives who had recently run for the Senate in a neighboring state. First, participants were shown a videotaped nonpolitical speech of
the candidates remarks at a bicentennial celebration at which he was asked
to present an award to a local dignitary. The speech, delivered by an accomplished character actor, was identical in content in all conditions but was delivered in either a forceful, articulate manner that conveyed a favorable im-
252
CHAPTER 9
253
Delayed-Information
Conditions
Delayed-Judgment
Conditions
53.8
44.6
47.7
42.9
64.5
30.9
48.1
40.1
48.1
47.3
57.3
31.0
about his issue stands were presented (i.e., under delayed-information conditions), participants based their evaluations of the candidate on their agreement with his issue positions, and the candidates similarity to them in ideology had virtually no effect. However, when participants learned of the
candidates issue stands immediately after they viewed his image-inducing
speech (i.e., under no-delay and delayed-judgment conditions), they based
their evaluations on the candidates general ideology, and their agreement
with him on specific issues had little influence. This was true under both nodelay and delayed-judgment conditions. Thus, unlike the direct effects of the
candidates image on judgments, the indirect effects of image on the processing of issue information were not a function of its salience at the time of judgment. Rather, they depended on the salience of the candidates image at the
time the issue stand information was conveyed.
In short, the salience of the candidates image at the time his issue
stands were learned altered the way in which the implications of these issue stands were construed. When a global image of the candidate was not
salient to them, participants assessed their agreement with his stands on
specific issues and based their judgments on this criterion independently of
the ideological implications of the candidates positions. When the candidates image was salient to participants at the time they heard about his issue positions, however, they applied a global criterion in assessing the implications of his issue positions as well. Consequently, their agreement with
the candidate on specific issues had relatively little effect.21
21
An alternative interpretation of these results might be that participants experienced overload when the candidates videotaped speech and his issue stands were conveyed in temporal
proximity and, therefore, they devoted less cognitive effort to an assessment of the candidates
21
254
CHAPTER 9
issue positions. If this were so, however, they would presumably have been inclined to use the
candidates image as a heuristic, leading it to have greater effect on judgments than it otherwise
would. In fact, however, the candidates image had no greater effect under delayed-judgment
conditions (when the two types of information were presented together) than under delayedinformation conditions. Therefore, this alternative interpretation does not seem viable.
255
6. Instructions to imagine the events that are described verbally can influence the impact of these events in much the same way as pictures. However,
when pictures and imagination instructions are both provided, people who
are not normally disposed to form visual images may have difficulty integrating pictures into the images they form when they are explicitly asked to do
so. In this case, therefore, pictures can decrease the impact of the information relative to conditions in which the pictures are not presented.
7. Narrative representations that are formed on the basis of observed
events can induce a holistic processing strategy that, once activated, influences the processing of verbal information that is conveyed subsequently. In
particular, it stimulates the use of more global criteria for judgments, and
correspondingly decreases the use of criteria that require more analytic,
piecemeal processing.
These conclusions may have more general implications. For example, information is likely to have greater impact on judgments and decisions if it
stimulates individuals to form a narrative-based account of an experience,
and to form mental images of the events that occurred, than if it does not.
This possibility, which was suggested by results of the study by Reyes et al.
(1980), was also recognized by Nisbett and Ross (1980). To borrow their example, a person who is considering the purchase of a particular brand of
dishwasher may be less influenced by statistical evidence of its superior
maintenance and repair record than by a neighbors vivid description of her
unique experience with the machine, which broke down 2 weeks after she
purchased it, flooded her kitchen and dining room, and ruined a valuable
rug. As noted earlier, narrative-based information processing may not always
generate more extreme evaluations than other well-learned processing strategies. Nevertheless, as this anecdotal example suggests, when narrativebased representations are easy to construct and are likely to elicit visual images, they are generally more likely to have an impact than abstract information whose implications are more difficult to assimilate and interpret.
256
CHAPTER 9
of delineating the role of narratives in judgment. Nevertheless, a conceptualization of the research from the perspective proposed in this chapter has
implications for the situational and informational factors that determine
which of several alternative narrative representations are likely to be constructed and used.
The Impact of Constructing Explanatory Narratives
on the Prediction of Future Events
People are likely to predict that an event is likely to occur if they can construct a plausible narrative of the events that might lead up to it. This implies that increasing the accessibility in memory of a plausible narrative account of the event should increase beliefs that the event will occur. Several
studies bear on this possibility. Ross, Lepper, Strack, and Steinmetz (1977)
asked some participants to read a clinical case study with instructions to
explain why the individual might have committed suicide after leaving therapy. Others read the same passage with instructions to explain why the
protagonist might have donated a sizable sum of money to the Peace
Corps. Participants after generating their explanations were asked to predict the likelihood that the protagonist had engaged in a number of behaviors after leaving therapy, one of which was the event they had explained.
Participants were told that the case study had been taken from a textbook
and that neither the experimenters nor anyone else had any knowledge of
what had actually happened to the individual. Nevertheless, participants
typically predicted that the event they had explained was more likely to
have occurred than the events they had not explained. Participants who
were asked to explain an occurrence presumably constructed a plausible
narrative of the sequence of events that led up to it and then, having done
so, used this narrative as a basis for their later predictions.
Results of other studies are consistent with this interpretation. Sherman,
Skov, Hervitz, and Stock (1981), for example, arbitrarily asked participants
to explain why they might either succeed or fail on an anagrams task and
then, after generating their explanation, asked them to predict their actual
performance on the task. Participants predicted the occurrence of the outcome they had arbitrarily been asked to explain. Moreover, their actual
task performance confirmed their prophecy. (That is, participants who had
been asked to explain why they might succeed, and predicted success, performed better than those who had been asked to explain why they might
fail.) Participants apparently used the implications of their narrative-based
prediction as a standard at the time they actually performed the task, motivating them to attain the performance level it implied.
Narrative-based explanations are likely to be generated spontaneously
when individuals receive information that is inconsistent with expectations.
257
258
CHAPTER 9
reaching the terminal, Al learns that his fight left on time. However, Bob
finds that his flight was delayed by 25 minutes and left only 5 minutes before he got there. Who is more upset? Ninety-six percent of the participants
who are asked this question agree that Bob would be more upset than Al.
Presumably, the reason is that Bob can easily construct a scenario whereby
he could have saved 5 minutes and caught the plane, whereas Al is less easily able to do so.
For similar reasons, people are more likely to be bothered by a mishap if it
results from an atypical behavior than if it results from a typical one. For example, people imagine being more upset by having a traffic accident while
driving home on a route they take infrequently than by having an accident
while driving home on their normal route. Moreover, people may be more
upset by getting robbed by a hitchhiker if they only rarely pick up hitchhikers than if they regularly do so (Kahneman & Tversky, 1982). In each case,
the difference results from the fact that it is easier to imagine how the event
could have been avoided in the first instance than it is in the second.
The ease of constructing a narrative can also come into play in other
judgments and decision phenomena. Research by Miller (1977) provided an
example. That is, it is easier to imagine how ones donation of money would
help a particular family whose home has been destroyed by fire than to
imagine how it would help to alleviate the suffering of thousands of people
whose homes have been destroyed in Bosnia. Therefore, people are more
likely to contribute money in the first case than the latter. Millers (1977) research provided evidence that this is the case. Evidence that this contingency is recognized outside the laboratory is suggested by the appeals of
charitable organizations to adopt a child, a whale, or even a highway.
This strategy presumably stimulates the construction of a narrative in
which ones own behavior helps an individual person, animal, or highway
maintenance project. Because this narrative is easy to construct, it may increase the willingness to provide assistance.
When two or more alternative narratives are equally easy to construct,
the one that is generated may depend on a number of factors that affect the
relative accessibility of the knowledge that enters into their construction.
Several of these factors, which could be informational, situational, or motivational, are discussed in the sections that follow.
Informational Determinants of Narrative Construction
1. Framing Effects
The narrative that is constructed from information for use in making an
inference can depend to a large extent on features of the information itself
that make different sets of concepts accessible in memory. Examples of this
259
contingency abound. To give a well-known example from research on decision making, consider the following two choice situations:
1. One thousand people are in danger of being infected by a deadly virus.
You have the choice of administering two serums. One serum, A, is effective but in short supply. If you administer A, 350 persons will be
saved. If you choose B, there is a probability of .65 that everyone will be
saved but a probability of .35 that no one will be saved.
2. One thousand people are in danger of being infected . . . If you administer A, 650 persons will die. If you choose B, there is a probability of .35
that everyone will die but a probability of .65 that no one will die.
The choice alternatives are actually identical in each case. In the first
case, however, the description of the alternatives activates concepts associated with saving lives, and so people are likely to construct scenarios in
terms of these concepts. As a result, they are more likely to prefer A, which
guarantees that lives will be saved, than B. In the second case, on the other
hand, the descriptions emphasize the loss of life and activate concepts associated with the avoidance of this negative outcome. In this case, therefore, people typically prefer B, which describes a situation in which no one
will die, than A.
The implications of the narratives that people use to make inferences on
the basis of verbal information can also depend on the perspective from
which the narrative is formed. (For evidence of perspective differences in
the mental representations of verbal information, see Black, Turner, &
Bower, 1979, described in chap. 4.) A study by Read (1985) provides an example. Consider the following vignette:
Helen was driving to work along a three-lane road, where the middle lane is
used for passing by traffic from both directions. She changed lanes to pass a
slow-moving truck, and quickly realized that she was headed directly for another car coming in the opposite direction. For a moment, it looked as if a collision was inevitable. However, this did not occur. Please indicate in one sentence how you think the accident was avoided.
The situation is objectively similar for both cars. However, the story as written focuses the readers attention on Helen, leading a narrative to be constructed in which she is the central figure. Therefore, people who are called
upon to explain how the accident was avoided typically attribute the responsibility of avoiding the accident to Helen rather than the other driver.
If the story had been written from a different perspective, this tendency
would presumably have been less likely.
260
CHAPTER 9
People who overhear this conversation are likely to assume that the
questioner is seeking new information. That is, they may infer that Joans
passing the exam is somewhat unexpected and, therefore, requires an explanation. They may further assume that the explanation is informative;
that is, it describes behaviors or situational factors that do not normally exist. In interpreting the answer, therefore, they may construct a narrative of
Joan as someone who typically stays out late before exams and is usually
too tired to perform adequately, but who in the present instance went to
bed at a reasonable hour and was sufficiently rested the next morning to do
well. In contrast, consider the conversation:
Q.
A.
261
262
CHAPTER 9
structing these narratives. Norenzayan and Schwarz (1999) provided provocative evidence of these differences in an analysis of the expectations
that participants generate in psychology experiments. Participants read a
New York Times article about either (a) a disgruntled postal employee who
went on a mass murder rampage or (b) a man who was accused of murder
for bombing a federal office building. The letterhead of the questionnaire
being administered suggested that the study was being conducted by either
the Institute of Personality Research or the Institute of Social Research.
In each case, participants after reading the story listed five reasons why the
protagonist might have committed the crime.
It seems reasonable to suppose that people who believed that the study
was sponsored by a personality research institute would activate concepts
associated with characteristics of the actor and, therefore, would construct
a narrative account of the individual in terms of these characteristics. In
contrast, persons who were told that the study was being conducted by a
social research institute are more likely to activate concepts about social
(i.e., situational) factors and to construct a narrative in terms of these
features. These narratives should then influence the explanations that participants generate. This was in fact the case. Participants gave more dispositional explanations when they believed that the study was conducted
by a personality research institute (M = 4.4) than when they believed it was
conducted by a social research institute (M = 2.6), and gave fewer situational explanations in the former case than the latter (5.8 vs. 9.4).
It is important to mention these differences occurred not only when the
nature of the research institute was identified before participants read the
story but also when it was identified afterwards. As the authors point out,
this could indicate that the differences in attributions reflected a tendency
for participants to tailor their responses to the apparent interests of the individuals to whom they were communicating, and might not reflect a difference in the underlying representation they formed of the interaction. This
possibility cannot be discounted. However, it seems equally likely that participants mentally reviewed and reconstructed the crime after they had
read the story, and that they based their explanations on this reconstruction rather than the representation they had formed earlier.
Be that as it may, these considerations have implications for the interpretation of numerous psychology experiments in which experimenters implicitly or explicitly induce expectations for the criteria that should be
applied in providing information and, therefore, the narrative-based representations they are likely to construct for use as a basis for their responses.
As Schwarz, Strack, Hilton, and Naderer (1991) argued, studies that are ostensibly conducted by psychologists can dispose participants to construct
explanations in terms of personality characteristics of the protagonists
rather than employing other criteria that are less relevant to psychology. If
263
the studies were conducted in a different situational context, this might not
be the case. To demonstrate this, Schwarz et al. introduced some participants to a judgment task with instructions similar to those employed in an
earlier experiment by Kahneman and Tversky (1973):
A panel of psychologists have interviewed and administered personality tests
to 30 engineers and 70 lawyers . . . On the basis of these interviews, thumbnail
descriptions of the [100 individuals] have been written. In your form, you will
find 5 descriptions, chosen at random. . . . For each description, please indicate the probability that the person described is an engineer. . . . the same
task has been performed by a panel of experts who were highly accurate in
assigning probabilities . . .
A second condition was similar except that the first line of the instructions referred to researchers rather than to psychologists, and the last
line referred to statisticians rather than experts. Following these instructions, participants read a sketch of the following type;
Hans K. is 45 years old. He is married and has four children. He is generally
conservative, careful and ambitious. He shows no interest in social and political issues. He spends most of his time on his many hobbies, which include . . .
solving mathematical puzzles.
The authors speculated that participants who believed that the story was
associated with psychologists would construct an account of the individual
in terms of personality and life style characteristics and would use this account as a basis for predicting his vocation. In contrast, associating the task
with statistics was expected to lead participants to use a statistical criterion
(e.g., base rates) as a basis for prediction. This was in fact the case. That is,
.76 of the participants predicted that the individual was an engineer when the
study referred to psychologists, whereas only .55 of the participants made
this prediction when the instructions referred to statisticians.
These data obviously do not provide evidence that a narrative representation of the target person was constructed in one case but not the other.
The results nevertheless provide a general indication that contextual factors can play an important role in the criteria that underlie judgments and,
to this extent, the likelihood that narrative-based explanations are used.
Motivational Influences on Explanation-Based
Narrative Constructions
The scenarios that people construct in the course of explaining events may
not be entirely dispassionate. This is particularly true when the events are
ones in which people are personally involved in the events and the explana-
264
CHAPTER 9
tions they give could reflect positively or negatively on themselves. Individuals in Western cultures typically make defensive attributions. That is, they
take responsibility for their own success but attribute their failures to other
factors that have relatively few negative implications for their self-esteem
(Harvey & Weary, 1984; Zuckerman, 1975). These factors may determine the
narratives that people construct.
The self-serving motives that underlie the construction of scenarios
about oneself may be attributed to others as well. To give yet another example from Tversky and Kahneman (1982):
Tom and Jim were both eliminated from a tennis tournament. Both were eliminated on a tiebreaker. Tom lost when his opponent served an ace. Jim lost on
his own unforced error. Who will spend more time thinking about the match
that night?
265
Personally Failed
Others Succeeded
Others Failed
Others Succeeded
Others Failed
6.63
6.38
6.13
6.75
7.13
4.13
4.25
4.63
Note. High scores indicate greater attributions of responsibility to oneself. Adapted from
Arkin, Gleason, and Johnston (1976).
This study obviously does not provide direct evidence that participants
attributions were mediated by a narrative representation that they constructed of the situation. It is nevertheless plausible to assume that participants who did well spontaneously activated a scenario of themselves as
competent individuals who typically succeed in achievement activity and
considered this scenario to be a sufficient explanation for the outcome.
Therefore, because the scenario reflected positively on themselves, they
used it as a basis for judgments without searching for alternative explanations. In contrast, participants were reluctant to construct a scenario that
suggested they were responsible for failure if a plausible alternative could
be generated. Therefore, they only accepted responsibility for their failure
when they had personally chosen the therapy and had failed when others
had done well.
Arkin et al.s study provides only one example of a more pervasive tendency for individuals to construct representations of themselves and others that confirm a perception of themselves as admirable individuals and of
the world as a place in which they can be happy and successful. This tendency, which has been documented in detail by Kunda (1990), is discussed
more fully in the next chapter.
CONCLUDING REMARKS
The studies reviewed in this chapter confirm the assumption that people often attempt to construct a narrative representation of events on the basis
of information they receive about them, and that their judgments are based
on implications of this representation. The events may be ones they imagine themselves or another person experiencing, events they either read
about, or ones they directly observe. It would obviously be wrong to assume that all social inferences are based on these representations. At the
266
CHAPTER 9
C H A P T E R
10
The Impact of Implicational Molecules
and Implicit Theories on Inferences
About Oneself and Others
The construction of narrative representations from new information obviously requires prior knowledge about the type of persons, objects, and
events to which the information refers. This knowledge, which often has implications for features that are not specified in the information presented,
can sometimes include episode models of specific past experiences that have
features in common with the new one. However, it can also consist of generalized narrative representations that are composed of object and event concepts that the new information exemplifies. As noted in chapter 1, these latter representations often function as implicit theories that individuals apply in
comprehending new experiences and construing their implications.
This chapter reviews theory and research in a number of areas in which
implicit theories are likely to play a role, including reconstructions of the
past, the use of judgmental heuristics, perceptions of social support, the
correlates of marital satisfaction, and the effects of stereotypes on overt behavior. Before embarking on a discussion of specific inference phenomena,
however, it may be useful to provide a simplified conceptualization of the
processes we assume to underlie the use of previous acquired knowledge
representations in making inferences.
GENERAL CONSIDERATIONS
The generalized narrative representations that function as implicit theories
may be abstractions of a number of specific instances of the sequence of
events they characterize (although, as we indicated in chap. 6, these repre267
268
CHAPTER 10
sentations may be formed less often than one might intuitively expect). Others may be formed in the course of goal-directed cognitive activity that
leads the events that compose them to be associated. Many theory-related
narrative representations may be similar to story skeletons of the sort
proposed by Schank and Abelson (1995). Other, simpler theories may be
conceptualized as implicational molecules of the sort described earlier in
this volume. These molecules theoretically consist of two or more verbally
coded segments that combine to describe a temporally and causally related
sequence of events or states of affairs. Although the nature of implicational
molecules has been discussed in earlier chapters, their centrality to the issues of concern in the present chapter arranges a brief recapitulation of
their structure and use in making judgments and decisions.
In combination, the segments of a molecule often exemplify a generalization about the world in which we live. To reiterate an earlier example, the
generalization that smoking causes lung cancer might be represented in the
two-segment molecule:
[P smokes; P has (will get) lung cancer].
Similarly, the generalization that people get what they deserve might be
conveyed in a just-deserts molecule; i.e.:
[P does a bad thing; P encounters misfortune]
or, alternatively,
[P does a good thing; P has good fortune].
The completion principle theoretically holds regardless of which propositions are instantiated by the information available and which must be inferred. Thus, in the previous example, people who learn that a person has
committed a nefarious act might use a just-deserts molecule to infer that
the person will be punished or will otherwise experience misfortune. However, they might also use this molecule to infer that a person who has expe-
269
rienced misfortune has done a bad thing or is reprehensible for other reasons. This possibility is consistent with the just world hypothesis (Lerner et
al., 1976; Lerner & Simmons, 1966). In support of this hypothesis, Walster
(1966) found that people evaluate the innocent victim of an automobile accident more unfavorably if the individual was seriously injured than if he or
she was only slightly harmed. Research with similar implications was reported by Lerner and Simmons (1966) and also by Wyer, Bodenhausen, and
Gorman (1985) in a study described in chapter 3.
A study by Spiro (1977) also suggested an application of the completion
principle. Protagonists read a story about an engaged couple with instructions that the experiment concerned reactions to situations involving interpersonal relations. In some versions of the story, the man informed the
woman that he didnt want children, whereupon the woman expressed considerable upset and a bitter argument ensued. After reading the story, participants engaged in routine activities unrelated to the experimental task.
During this period, the experimenter, who ostensibly knew the couple incidentally noted that they eventually married and were still happily together.
Participants who are motivated to comprehend the couples relationship
are likely to construct a scenario of how this unexpected outcome might
have occurred. In doing so, they might invoke a previously formed kissand-make-up molecule of the form:
[People love one another; people fight; people resolve conflict; people make
up].
To this extent, they may infer an instantiation of the proposition people resolve conflict, which is not implied by the information available. For example, they might speculate that one of the partners had a change of heart,
that the woman found she couldnt have children, or that some other unforeseen event occurred that stimulated the couple to resolve their conflict.
The inferred event might then be added to the mental representation that
participants construct of the couples relationship and store in memory.
Participants after completing the activities they were asked to perform
were dismissed. However, they returned for a second experimental session
a few days or several weeks later, at which time they were asked to recall
the story they had read. They were explicitly told to report only things that
were mentioned in the story and not to include any personal reactions or
inferences they might have made. Nonetheless, participants made frequent
errors, the number of which increased over time. These errors were typically of the sort one might expect as a result of attempts to reconcile the experimenters incidental comment with the original information. For example, one person recalled that the problem was resolved when they found
that [the woman] could not have children anyway. Another reported that
270
CHAPTER 10
although one person thought the matter was important, the other did not
(see Spiro, 1977, for additional examples).
The completion principle has potentially broad applicability. However,
additional considerations arise in applying it to the inference phenomena of
concern in this chapter. For one thing, more than one implicit theory can
potentially be brought to bear on a particular observation. Or, in terms of
our present discussion, a proposition may be contained in more than one
molecule. This is particularly likely when an event has more than one possible cause. For example, peoples theories about poor academic performance might be reflected in the molecules:
[P doesnt study; P gets poor grades]
and
[P is stupid; P gets poor grades].
271
at the time an individual is asked to predict the likelihood of B, the individual is likely to identify and use A as a basis for this inference. Suppose, however, that the individual is uncertain that A is true. Then, the person is
likely to consider not only the likelihood that B would occur not only if A is
true but also if A is not true. Thus, in our example, people who are asked the
likelihood that American will go to war might base their belief on not only
the likelihood that this will occur if war is good for the economy, but might
also take into account the possibility that America would go to war even if
war were not good for the economy. The relative weight attached to these
two conditional beliefs may be a function of the relative likelihood that A is
or is not true (i.e., that war is or is not good for the economy, respectively).
The quantification of this process proposed by Wyer (1974; Wyer &
Goldberg, 1970) assumes that if beliefs are defined in units of probability,
then the belief in the event described by one proposition, B, is related to beliefs in a possible cause of this event, A, according to the equation:
P(B) = P(A)*P(B/A) + P(~A)*P(B/~A),
[10.1]
where P(B) is the belief that A is true, P(A) and P(~A) [= 1 - P(A)] are beliefs
that A is and is not true, respectively, and P(B/A) and P(B/~A) are beliefs
that B is true if A is and is not true, respectively.
This equation provides a remarkably accurate quantitative description
of the relations among these beliefs and, therefore, the influence of beliefs
in one proposition on beliefs in a second (for summaries of this evidence,
see Wyer & Carlston, 1979; Wyer & Hartwick, 1980). In one study, for example (Wyer, 1970), participants read scenarios composed of two parts. The
first part led participants to believe that a particular event (A) was unlikely
to occur and, in addition, implicitly conveyed the likelihood that a second
event, B, would occur if A did and did not occur. The second part of the
story provided additional information that increased beliefs that A would
occur without changing its implications for B. In neither part of the story,
however, was the target event itself explicitly mentioned. One story, for example, concerned the likelihood that a student riot will occur at a particular
university (A), and the likelihood that the university president would be
fired (B). The first part of the story described the president as coercive, establishing a low initial belief that a riot would occur. However, the second
part provided information that substantially increased the belief that a riot
would occur without affecting perceptions of whether the president would
remain in office.
After reading each part of the story, participants reported their belief
that B would occur, followed by their belief in A and the two conditional beliefs defined in Equation 10.1. These estimates, along a scale from 0 (not at
all likely) to 10 (extremely likely), were divided by 10 to convert them to
272
CHAPTER 10
units of probability. The predicted belief in B, P(B), was then computed for
each participant separately by combining his or her beliefs in the manner
described by the right side of Equation 10.1. Mean obtained values of P(B)
are plotted as a function of mean predicted values in Fig. 10.1. Changes in
beliefs in B after reading the second part of the story are also shown in this
figure. In each case, the standard error of the difference between mean predicted and mean obtained values was less than .05 (half a scale unit). This
accuracy was obtained without the use of any ad hoc curve-fitting parameters.
Further evidence bearing on the validity of Equation 10.1 as a description
of the relations among beliefs in causally related propositions is provided
by research on the Socratic effect (McGuire, 1960; Rosen & Wyer, 1972).
That is, peoples a priori beliefs in a set of propositions may not conform to
the relations specified in Equation 10.1 because the propositions have not
recently been thought about in relation to one another. However, if peoples beliefs in the propositions are called to their attention in close temporal contiguity, they may recognize their inconsistency and, consequently,
may modify one or more of the beliefs to eliminate this inconsistency. This
should be indicated by an increase in the quantitative accuracy of Equation
10.1. Rosen and Wyer (1972) found evidence that this is the case. That is,
participants reported their beliefs in causally related propositions in two
experimental sessions 1 week apart. The accuracy of Equation 10.1 in describing the relations among these beliefs was significantly greater in the
second session than it was at first, suggesting that the consistency of the
beliefs increased over time.
A later study (Henninger & Wyer, 1976) is potentially more relevant to
the concerns of this chapter. Participants in some conditions were asked to
report their belief in a proposition occupying the position of A in Equation
10.1 (i.e., an antecedent) before reporting their belief in B (the consequent).
These participants appeared to use A as a basis for their belief in B when
they encountered it later in the questionnaire (as inferred from the accuracy of Equation 10.1 in describing their relationship). Other participants,
however, reported their belief in B at the outset, when A was not salient to
them. These participants typically used a different criterion for evaluating
the validity of B, and so their belief in this proposition was inconsistent
with the belief in A that they reported later.
Henninger and Wyers findings suggest that Equation 10.1 might be used
to determine empirically the criteria that people bring to bear on their beliefs in a proposition they are asked to evaluate. Suppose two or more
implicational molecules could potentially provide the basis for an inference. A comparison of the quantitative accuracy of Equation 10.1 in describing the relations among the beliefs in the propositions composing each molecule could provide an indication of which molecule was actually activated
273
FIG. 10.1. Beliefs in a Proposition B, and changes in these beliefs, as a function of predicted values based on Equation 10.1. Based on data reported by Wyer (1970).
274
CHAPTER 10
and used. The utility of the equation is obviously restricted to theories that
concern only two causally related events or states. Many implicit theories
are more complex. However, if these theories can be broken down into subsets of causally related propositions, the conceptualization could be useful
in evaluating the assumptions underlying the application of these theories
as well.
275
276
CHAPTER 10
In the second case, the observer may activate an implicit theory that people do things because they have to; e.g.,
[A controls P; A desires X; P advocates X].
The first theory has implications for the actors belief, but the second theory does not. Consequently, the observer may infer that the first person is
more likely to believe in capital punishment than the second.
However, now suppose the two individuals in our example are asked to
report their own beliefs in capital punishment. If people use similar implicit
theories to infer their own attributes from their behavior, the first person
should report a stronger belief in capital punishment than the second. Although there are obviously alternative interpretations of this difference
(Festinger, 1957), Bems conception is certainly plausible.
Bem and McConnell (1970) constructed an intriguing empirical demonstration of self-perception processes that also has implications for the role
of implicational molecules in reconstructive memory. Some participants
wrote a counterattitudinal essay under conditions in which they were given
the right to refuse, whereas others were told to write the essay without being given a choice. Then, some participants in each condition were asked to
report their attitude toward the position they had advocated. These participants reported more favorable attitudes toward the position under freechoice conditions than under no-choice conditions, consistent with numerous other studies.
Other participants, however, were asked to recall the attitude they had
reported in an earlier experimental session. These participants recall of
their prebehavior attitudes was affected in the same way that postbehavior
attitudes of the first group of participants were influenced. That is, participants under free-choice conditions recalled their prebehavior attitudes as
similar to the position they advocated in the essay they had written subsequently. This was not the case under forced-choice conditions. Participants
in this study, like those in Rosss studies and the research by Goethals and
Reckman (1973), may have invoked an implicit theory about the attitudinal
antecedents of their behavior and used this theory to infer not only their
postbehavior attitudes but also what their attitudes must have been before
engaging in this behavior.
Further Considerations
The interpretation of research on the effects of behavior on attitudes is often equivocal. That is, it is often difficult to know whether these effects occur after the behavior has been performed or in the course of deciding to
277
engage in it. In the latter case, the results would not indicate an effect of the
behavior per se, but rather, would simply reflect the results of cognitive deliberations that were involved in deciding whether or not to engage in it.
A study by Albarracin and Wyer (2000) circumvented this ambiguity. Participants were told that the experiment was designed to test a new computerized procedure for identifying unconscious behavioral tendencies. The
procedure ostensibly consisted of presenting statements about social issues subliminally and having participants respond to them without being
consciously aware of their content. Participants were told that although the
stimulus statements would appear to be only flashes of light, they would
elicit unconscious feelings that would give rise to a more conscious intuition. They were informed that to provide a measure of their behavior,
they should follow their intuition and generate a yes or no response
to each statement, which would then be interpreted by the computer as a
vote either in favor of or against the issue to which it pertained.
With this preamble, participants were ostensibly exposed to 14 statements that ostensibly concerned seven different university policies. They
were told that the statements might express either support for or opposition to a policy and that the computer program would take this into account in determining the implications of their votes. After the 14 statements had been judged, participants received computerized feedback
about both the policies to which they were subliminally exposed and their
responses to these policies. Participants a priori positions on six of the policies (e.g., maintaining civil liberties on campus, receiving free tickets to
sports events, raising tuition, etc.) were self-evident. Participants were
told that they had voted in favor of or in opposition to each of these policies, depending on which vote was normative. (This established the credibility of the assessment procedure.) The seventh policy, instituting comprehensive examinations on campus, was one on which participants held
varying opinions. Independently of these opinions, however, some participants were told they had voted in favor of instituting the exams and others
were told they had voted against it.
Participants after receiving the feedback were told that to understand
their unconscious decisions, we needed to know their personal feelings
about the policies and, on this pretext, asked them to report their attitudes
toward instituting the exams. In addition, they estimated the likelihood and
desirability of a number of possible consequences of instituting the exams.
Finally, we indicated that because participants had an opportunity to think
about comprehensive examinations during the experiment, we wanted to
see how students might actually vote in a referendum that was likely to be
held later in the semester. Participants were then left alone with instructions to select a slip of paper that represented their choice and to place it in
a secured ballot box. (The box appeared to be partially full. In fact, how-
278
CHAPTER 10
ever, it contained only blank slips of paper, and so we could determine after
the experiment how the participants had voted.)
Thus, the procedure ensured that participants had not thought about
comprehensive examinations before they received feedback about their ostensible voting behavior. Nevertheless, participants reported more favorable attitudes toward the institution of the exams if they had ostensibly
voted in favor of instituting them than if they had voted against. Furthermore, this difference was evident regardless of the attitudes that participants had reported before to the experiment. Finally, their postbehavior attitudes, once formed, were used as bases for their actual voting decisions
that they made at the end of the experiment. (Path analyses indicated that
behavior feedback did not have an effect on participants judgments of specific behavioral consequences, nor did it have direct influence on participants voting behavior. Rather, this latter influence was mediated by the impact of the feedback on participants attitudes.)
This study provided no direct evidence of the cognitive mediators of the
effects we observed. It nevertheless seems reasonable to assume that participants invoked an implicit theory that people publicly advocate positions
they favor, as embodied in the implicational molecule:
[P favors X; P votes in favor of X].
Participants may have used this molecule not only to infer their attitude toward comprehensive examinations from their unconscious behavior, but
also, once their attitude was inferred, as a basis for the overt behavioral decision they made at the end of the experiment.
279
riages end in divorce, they nevertheless are convinced that they personally
will remain married to their first spouse for life (Kunda, 1987). This suggests
that individuals who are motivated to believe that they will be successful in
marriage, construct implicit self-theories that support this belief. To demonstrate this, Kunda (1987) gave participants information about a target person who was either happily married or divorced and whose demographic
and personality characteristics either matched or did not match those of
the participants themselves. After receiving this information, participants
indicated which of the targets attributes were most likely to contribute to
his or her marital status. Participants were more inclined to attribute the
success of happily married targets to characteristics that matched their
own than to characteristics that did not. Correspondingly, they were less
likely to attribute the failure of divorced targets to the former characteristics. Thus, they constructed theories about themselves that were consistent with the outcome they wished to attain.
People who are motivated to construct and maintain a self-serving theory may selectively search memory for information that supports it. In a
study by Santioso, Kunda, and Fong (1990), some participants were told
that extroversion was conducive to success after leaving college, whereas
others were told that introversion was more conducive to success. Then, in
an ostensibly unrelated study, they were asked to list all of the behaviors
they had performed in the past along a related trait dimension (shy vs. outgoing). Participants listed more behaviors that were congruent with the
trait they were told was conducive to success than behaviors that were incongruent with this trait. In a second study, participants after exposure to
the first task were shown a series of behaviors and were asked in each case
to press a button as soon as they thought of a personal experience that exemplified it. Participants responded more quickly to behaviors that exemplified the success-related trait than to behaviors that exemplified the opposite. Thus, people who are told that a particular trait is associated with
success selectively searched memory for personal experiences that confirmed their possession of this trait. Consequently, these experiences came
to mind more easily when they were called upon to report instances of
their behavior at a later point in time.
Research in other paradigms has similar implications. For example, people are more likely to report engaging in a particular activity (drinking coffee, brushing ones teeth, etc.) if they are told that the activity is healthy
than if they are told it is not (Ross, McFarland, & Fletcher, 1986; B. Sherman
& Kunda, 1989). These studies do not provide direct evidence that these responses were mediated by the construction of an implicit theory. However,
it seems reasonable to suppose that the construction of self-serving theories is a result of similar processes.
280
CHAPTER 10
Thus, people may use this molecule not only to infer the ease of recalling instances of an event from knowledge of how often it has occurred, but also
to infer the frequency of an events occurrence from the difficulty of recalling instances of it.
Two quite different sets of phenomena can be conceptualized as resulting from an application of this molecule. One concerns the effects of television on perceptions of social reality. The other concerns the effects of ease
of retrieving knowledge about oneself and others on judgments of the attributes to which this knowledge is relevant. Research on each set of phenomena is discussed in turn.
Perceptions of Social Reality
Americans watch an average of more than 4 hours of television each day
(Nielsen, 1995). As a result, television is a major source of the episode models that people form. On the other hand, people often fail to distinguish
clearly between the sources of the knowledge they acquire (Hasher, Goldstein, & Toppin, 1977; Jacoby, Kelly, Brown, & Jasechko, 1989; Johnson,
Hastroudi, & Lindsay, 1993). As a result, they are likely to use the episode
models they form while watching television as a basis for judgments without considering the context in which these models were constructed.
L. J. Shrum and his colleagues (OGuinn & Shrum, 1997; Shrum, OGuinn,
Semenik, & Faber, 1998; Shrum, Wyer, & OGuinn, 1998) have examined im-
281
plications of this possibility in the context of their research on the cultivation effect (Gerbner, Gross, Morgan, & Signorielli, 1994). That is, people
tend to overestimate the occurrence of events in the real world that are disproportionately represented on television. Moreover, the magnitude of
their overestimation increases with the amount of television they watch. Instances of events that occur frequently on television are easily accessible in
memory, and this is particularly true for heavy viewers. Therefore, if individuals are asked to infer the frequency of occurrence of these events, and
if they apply the ease-of-retrieval molecule, heavy users are likely to make
higher estimates than light viewers do. Thus, for example, heavy viewers
are relatively more likely than light viewers to overestimate the proportion
of Americans who belong to a country club or have a swimming pool in
their back yard (OGuinn & Shrum, 1997). On the other hand, they are also
more likely to overestimate the incidence of violent crime and the size of
the police force.
Several studies by Shrum and his colleagues support this interpretation.
In one study (OGuinn & Shrum, 1997), frequent and infrequent viewers of
soap operas were asked to estimate the incidence of characteristics that
are associated with an affluent life style (belonging to a country club, owning a luxury car, having a swimming pool, etc.). Frequent viewers not only
estimated the incidence of these characteristics to be higher than infrequent viewers did, but also made their estimates more quickly. These findings are consistent with the hypothesis that frequent television viewers find
instances of these characteristics easier to retrieve from memory and estimate their incidence to be greater as a result.
Other interpretations of these effects should of course be considered.
For example, a third variable, such as education level or socioeconomic
class, could have independent effects on both television watching and frequency estimates, producing an artifactual relation between these variables that would not otherwise exist. However, OGuinn and Shrum (1997;
see also Shrum, Wyer, & OGuinn, 1998) found that the relation between frequency estimates and television viewing was evident even when the potentially confounding effects of these variables were controlled. Moreover,
Shrum et al. (1998) found that when participants were asked to report their
television viewing habits before making frequency estimates (thus calling
their television watching to their attention), the impact of television viewing on frequency estimates was significantly reduced. The effects of watching television on frequency estimates can also be reduced or eliminated by
increasing participants motivation to make correct judgments (Shrum,
1999). Thus, people appear able to distinguish between events they see on
television and those they learn about from other sources, and thus can
avoid the bias produced by television when they are motivated to do so.
Normally, however, this motivation does not exist.
282
CHAPTER 10
283
assertive when they have been asked to recall 12 instances than when they
were asked to recall only 6. This was in fact the case.
The use of an ease-of-retrieval heuristic as a basis for judgment is quite
pervasive (for a review, see Schwarz, 1998). Further examples are described later in this chapter. Its implications can be quite ironic. For example, people may be less likely to believe that a proposition is true if they
have attempted to generate a large number of reasons for its validity than
if they have thought about only a few. Research by Wnke, Bless, and
Biller (1996) supported this speculation. Some participants were asked to
generate either 3 or 7 arguments that either favored or opposed a specific
issue, after which they were asked to report their own position on the issue. Other, yoked participants read the arguments that individuals in the
first group had written. The latter participants reported themselves to be
more in favor of the position advocated when they had read 7-argument
responses than when they had read 3-argument responses, confirming the
assumption that the substantive implications of the 7-argument sets were
relatively more persuasive. Nevertheless, the participants who had actually generated the arguments judged themselves to be less in favor of the
position when they had generated 7 arguments than when they had generated only 3. Thus, the effects of ease of retrieval overrode the effects of actual knowledge.
People do not always ignore the implications of their past knowledge, of
course. However, their computation of a belief on the basis of these implications is cognitively effortful. Consequently, they may only do so when ease
of retrieval is likely to be an unreliable criterion. In other conditions of
Schwarz et al.s (1991) research, for example, participants generated instances of assertiveness in the presence of distracting background music.
In this case, participants apparently attributed their difficulty of generating
instances to the distraction and to their lack of knowledge. In these conditions, therefore, they judged themselves more assertive when they had generated 12 instances rather than 6.
1. Effects of Information About Others Ease
of Retrieving Information
As suggested in our discussion of self-perception processes, people can
use implicit theories to construe the implications of others behavior as
well as their own. Support for this assumption in the present paradigm was
obtained by Menon and Raghubir (1998) in a study of consumer judgment.
Participants read an advertisement that described the attractive features of
a computer. Later, they were asked to recall either 2 or 8 of these features.
Either before or after the recall task, however, they were given information
about the difficulty that others had in recalling the products attributes. If
284
CHAPTER 10
Eight Attributes
Recalled
5.9
5.4
5.6
4.9
4.4
4.9
5.4
4.9
5.3
4.3
4.8
5.9
4.9
5.4
5.6
4.6
285
286
CHAPTER 10
Thus, suppose people perform a task and fail. This event (P fails)
instantiates a segment of each of these theories. Individuals who activate
and use the first theory to interpret the event may infer that they did not
work hard enough and that if they exert more effort, they might ultimately
succeed. As a result, they may be stimulated to try the task again. However,
suppose individuals activate the second theory instead. Then, they are
likely to infer that their failure reflects low ability and that repeating the
task will have the same outcome. Therefore, they may be disinclined to try
a second time. Individual differences in the chronic accessibility of these
theories, and their effects on responses to failure, have been elaborated by
Dweck and Leggett (1988).
Incremental Versus Entity Theories
of Personality and Ability
Dweck (1991; Dweck & Leggett, 1988) suggest that the aforementioned performance-related theories exemplify two more general theories of personality and behavior. Incremental theories imply that peoples attributes and
287
abilities are malleable and can be modified, whereas entity theories imply
that attributes are fixed and resistant to change. These different theories
have numerous implications for judgment and behavior, many of which
have been examined by Dweck, Chiu, and Hong (1995a, 1995b; Dweck, Hong,
& Chiu, 1993; Chiu, Dweck, Tong, & Fu, 1997; Chiu, Hong, & Dweck, 1997;
Hong, Chiu, Dweck, & Sacks, 1997).
Hong et al. (1997), for example, found evidence that people who were
classified as entity theorists on the basis of an independent measure were
more influenced by a target persons personality and ability test scores
than incremental theorists were, suggesting that they attached relatively
more significance to measures that ostensibly assess stable attributes and
abilities. People may also have incremental vs. entity theories of morality
(implying that individuals moral character can or cannot be easily modified) and of the world in which they live. A series of studies by Chiu et al.
(1997) provided evidence that individuals with entity theories of morality
and the world are more inclined than individuals with incremental theories to evaluate others behavior in terms of its fulfillment of obligations
rather than in terms of individual rights. Thus, for example, participants in
one study were asked to give open-ended descriptions of how they would
handle hypothetical classroom situations. Some scenarios described socially desirable behavior (e.g., The teacher asked Jerry to remove old papers and notices on the students notice boards, and Jerry does what she
requested as soon as possible.). Others described undesirable actions
(e.g., The teacher asked Peter to remove old papers and notices. . . . A
week has passed and Peter does not do it.). Entity theorists, who believe
that the fulfillment of responsibilities is normative, were less inclined than
incremental theorists to indicate that Jerry should be rewarded for fulfilling his obligations. Correspondingly, they were more inclined than incremental theorists to recommend that Peter be punished for not meeting his
responsibilities.
As Dweck et al. (1995a, 1995b) pointed out, the implicit theories that individuals apply may often be domain specific. Moreover, several alternative theories may coexist, the implications of which can conflict. The individual differences in implicit theories that Dweck and her colleagues
identified in the aforementioned studies could reflect differences in the
chronic accessibility of these theories. To this extent, situational factors
that make one or another theory more accessible in memory might sometimes override these chronic differences. Although this possibility was
not examined in the studies cited in this section, evidence that the effects
of transitory situational factors can often dominate the effects of chronic
accessibility (at least in the short run) is provided by other research described presently.
288
CHAPTER 10
289
These contingencies were identified by Mankowski and Wyer (1996). Participants were told that the study was concerned with person impression
formation. On this pretense, they were asked to read a story similar to the
one just described, in which one person was in need of support and the
other was in a position to provide it. In third-person perspective conditions,
the story was written from the perspective of a disinterested observer. In
provider-perspective conditions, however, the provider was referred to as
you and the recipient was identified by a name that was ambiguous with
respect to gender (Chris or Terry). In recipient-perspective conditions, the recipient was referred to as you and the potential provider was identified
by name. After reading the story, participants in each perspective condition
wrote down three adjectives that they would use to describe the provider
and then rated the provider with respect to several traits that were relevant to social support (supportive, warm, rejecting, accepting, sensitive,
etc.). Finally, participants were completed a general index of the support
they perceived to be personally available (Cutrona & Russell, 1987).
Participants perceptions of support availability could reflect two
things. First, it could indicate the frequency with which participants had
personally encountered support-relevant situations in the past and, therefore, the likelihood that semantic concepts associated with social support
were chronically accessible in memory. To this extent, participants perceptions of support availability should have a positive impact on their use
of support-related concepts to interpret the providers ambiguous behavior, and this should be true in all three perspective conditions. On the
other hand, participants perceptions of support availability could also indicate the likelihood that they have constructed an implicit theory of
themselves as recipients of support. If participants who read the story
from the recipients perspective activate and use this theory as a standard
of comparison in construing the storys implications, they should perceive
the situation o be less supportive when their self-referent theory implies
that they typically receive support than when it does not. This contrast effect of the storys implications on judgments could offset the positive effect of semantic concepts on its interpretation that occurs at an earlier
stage of processing.
Results were consistent with this analysis. Table 10.2 shows the supportiveness implied by both participants open-ended trait descriptions of the
support provider and their ratings of the providers supportiveness. When
participants took the perspective of either the provider or third person,
their perceptions of the providers supportiveness increased with their perception of their own support availability. When they took the role of the recipient, however, this was not the case. In these latter conditions, the contrast effect that resulted from using an implicit theory of their personal
experiences as a standard of comparison appeared to offset the positive ef-
290
CHAPTER 10
TABLE 10.2
Perceptions of the Providers Supportiveness as a Function
of Perspective and Social Support Availability
Perspective
Provider
Third Person
Recipient
0.39a
0.18
0.21
0.14
0.26
0.40
0.44
0.26
0.18
0.81b
0.35
0.46
0.26
0.44
0.70
1.07
1.16
0.09
fects of semantic concepts on the interpretation of the scenario that was evident in other perspective conditions.
291
sonal goals. As Briley and Wyer (2001) pointed out (see also Ho & Chiu,
1994; Rhee, Uleman, & Lee, 1996; Triandis & Gelfand, 1998), the individualismcollectivism construct is multidimensional, and the dimensions that
are often assumed to distinguish individualists and collectivists (i.e., individuality, cooperation, competitiveness, self-sacrifice, etc.) are not strongly
related (Briley & Wyer, 2001). Nevertheless, a general cultural difference in
the tendency to think of oneself as an independent individual vs. a member
of a group is fairly pervasive (Markus & Kitayama, 1994).
However, the implicit theories that reflect these different cultural orientations are not the only ones that people acquire. As Hong et al. (2000) suggested, representatives of a given culture are often exposed to norms and
values that pervade other cultures as well. Moreover, many individuals are
bicultural, having lived for many years in more than one country, or being
residing in a country (e.g., Hong Kong) in which both Western and Asian
norms and values are pervasive. Hong et al. (2000) conceptualize culture as
a dynamic construct, the effects of which can depend on situational factors
that influence the extent to which cultural norms and values are accessible
in memory. Several studies provide evidence of this dependence.
Attributions of Causality
Hong et al. (2000) showed that the implicit theories that Hong Kong participants bring to bear on the interpretation of information and the inferences
they draw from this information can be induced by exposing participants to
cultural symbols of Western and Asian cultures prior to the judgment task.
For example, Westerners (individualists) typically acquire an implicit theory of causality that places responsibility on the individual, whereas Asians
(collectivists) acquire theories that place responsibility on the society or
social context in which the behavior occurs. Evidence supporting this difference was reported by Morris and Peng (1994; see also, Choi, Nisbett, &
Norenzayan, 1999). Nevertheless, both implicit theories are likely to exist in
memory, and the theory that is activated and applied may be influenced by
situational factors that affect its accessibility.
To demonstrate this, Hong et al. constructed stimulus materials consisting of a school of fish in which one fish was swimming ahead of the others.
Thus, the situation could be interpreted as an indication that either the first
fish was leading the others (implying a dispositional cause of the fishs behavior) or, alternatively, that the first fish was being chased by the others
(a situational cause of its behavior). Hong Kong Chinese undergraduates
were exposed to these pictures with instructions to explain why the one
fish was swimming ahead of the others. Before doing so, however, they
were exposed to either a series of American cultural icons or a series of
Chinese icons as part of an ostensibly different experiment. Exposure to
292
CHAPTER 10
293
uation-based theory, whereas thinking about negative life events should activate a person-based theory. Consequently, Asians should be less inclined
to use their current life satisfaction to predict their future in the former condition than the latter.
Oishi et al. (2000) confirmed these predictions. Participants were European Americans and Asian Americans whose life satisfaction was assessed
in an ostensibly unrelated experiment. Participants first completed a lifeevent inventory in which they were asked to write about either a positive
life experience they had had or a negative one. Then, after doing so, they
were given a list of 10 positive and 10 negative events (getting the best
grade on an exam, finding that your friends were talking about you behind
your back, etc.), and estimated the likelihood that each event would occur
to them in the near future.
Multiple regression procedures were used to analyze the favorableness
of participants predictions of the future as a function of a number of orthogonal contrasts corresponding to the main and interactive effects of life
satisfaction (a continuous variable), cultural background and the valence of
the life experience that participants recalled. This analysis yielded a significant interaction of all three variables. The implications of this interaction
can be seen in the left half of Table 10.3, which shows estimated values of
the favorableness of participants future outlook as a function of culture,
the valence of the stories that participants wrote about themselves, and
current life satisfaction. European Americans life satisfaction had a greater
effect on their future outlook when they had written about a positive life experience than when they had written about a negative one. In contrast,
Asian Americans life satisfaction had more influence on their future outTABLE 10.3
Regression-Based Estimates of Future Outlook as a Function
of Current Life Satisfaction, Cultural Background, and the
Valence of Stories About Self and Other
Valence of Stories
About Self
European Americans
High current life satisfaction
Low current life satisfaction
Difference
Asian Americans
High current life satisfaction
Low current life satisfaction
Difference
Valence of Stories
About Others
Favorable
Unfavorable
Favorable
Unfavorable
0.36
2.04
2.40
1.26
0.26
1.52
1.44
1.18
0.26
1.84
0.38
2.22
0.80
0.32
1.12
1.54
1.10
2.64
1.74
0.56
1.18
0.98
0.24
0.74
294
CHAPTER 10
look when they had written about a negative experience than when they
had written about a positive one.
A second study was analogous to the first but focused on the effects of
activating concepts about others. If European Americans are motivated by
a desire to be superior to others, they may tend to attribute others successes to situational factors but others negative outcomes to dispositional
factors (Snyder, Stephan, & Rosenfield, 1976). In contrast, if Asians are typically self-effacing and tend to promote others interests at their own expense, they may be inclined to attribute others successes to dispositional
causes (Yamauchi, 1988). To this extent, a person-based theory of causality
should be activated in European Americans by stimulating them to think
about others failures, but should be activated in Asian Americans by inducing them to think about others successes.
To explore this possibility, participants participated in a study that was
very similar to the first. Rather than writing about a personal experience,
however, they read a newspaper story describing either positive or negative job prospects for the current years college graduates. (Participants
were freshmen, and so the story was not personally relevant.) Regression
analyses of these participants future outlook as a function of experimental
variables yielded a three-way interaction of cultural background, life satisfaction, and the valence of others experiences, the implications of which
are shown in the right half of Table 10.3. The pattern of these data is exactly
the opposite of the first set of data. That is, European Americans future outlook was most strongly influenced by their current life satisfaction when
they had read about others negative experiences, but Asians future outlook was most strongly influenced by life satisfaction when they had read
about others positive experiences.
Cultural Differences in Inference Strategies
The preceding discussion suggests cultural differences in the content of the
implicit theories that are applied by members of different cultures. However,
more general differences may exist in the conditions in which these theories
are used. Choi et al. (1998) suggested that general cultural differences between Western and Asian cultures exist in the tendency to use analytic reasoning in generating explanations for social events. To this extent, the inference processes that are implicitly assumed in much of our discussion of the
use of implicational molecules might not generalize over cultures.
Some support for this possibility was obtained in a study by Norenzayan
and Kim (2000) in research on the Socratic effect (McGuire, 1960). In an initial experimental session, European Americans and Asians completed a belief questionnaire containing causally related propositions of the sort to
which Equation 10.1 is applicable. They then returned 1 week later and com-
295
296
CHAPTER 10
faction might be related to these theories. To this end, we identified 32 married couples who had been in their relationship for an average of 9 years,
and whose marital satisfaction (based on the LockeWallace, 1959, scale of
marital adjustment) had been determined in an earlier study. In 16 of these
couples, both partners expressed satisfaction with their relationship, and in
the other 16, both partners reported being dissatisfied. Participants were
told that the studys purpose was to understand peoples perceptions of
how interpersonal relationships were likely to change over time. On this
pretense, participants plotted six graphs showing the changes that typically occurred in marriage partners feelings toward one another during the
first 10 years of their relationship. Three graphs described the typical
mans feelings of romantic love, commitment to the marriage, and affection
for his partner. Three others described the typical womans feelings along
the same dimensions. To aid them in constructing their graphs, participants
were given axes in which the x-coordinate pertained to the year of the relationship and the y-coordinate referred to the attribute being rated along a
scale from 0 (none) to 10 (very much). Partners were told to place an x
above the number denoting each year of the relationship to indicate their
perception of the partners feelings during that year.
Satisfied partners theories about the prognosis of marital relationships
were expected to be more similar to one another than dissatisfied partners
theories. Data pertaining to this hypothesis are conveyed in Figs. 10.2 and
10.3, which show the composite graphs drawn by satisfied and dissatisfied
partners along each dimension. Analyses of these data revealed an interaction of relationship satisfaction, the partner being rated, and attribute dimension, F(2, 60) = 4.29, p < .05 and an higher order interaction involving
these variables, participant sex, and time, F(18, 540) = 1.88, p < .01. These interactions are most clearly interpretable by considering data for satisfied
and dissatisfied couples separately.
As expected, satisfied husbands and wives had generally similar theories
about how a typical marriage partners feelings were likely to change. For
example, they agreed that the typical partners feelings of romantic love decreased to a much greater extent over time than his or her feelings of affection. Moreover, they agreed that the typical mans feelings of love would
tend to increase during the later years of the relationship without a corresponding increase in affection, whereas the typical womans feelings of affection were likely to increase in later years without a corresponding increase in feelings of romantic love.
In contrast, dissatisfied partners perceptions of the typical marriage differed much more dramatically. These differences were reflected in an interaction of respondent sex, the sex of the partner being rated, attribute dimension, and time, F(18, 270) = 2.80, p < .01, in an analysis of judgments by
dissatisfied partners alone. Three aspects of these data are noteworthy.
297
FIG. 10.2. Satisfied partners perceptions of the typical husbands and typical
wifes feelings over the first 10 years of a relationship.
298
CHAPTER 10
FIG. 10.3. Dissatisfied partners perceptions of the typical husbands and typical wifes feelings over the first 10 years of a relationship.
299
results demonstrate the utility of comparing the implicit theories that are
likely to characterize happy and unhappy couples. Differences in these theories may be worth more careful attention in future research.
300
CHAPTER 10
cations. For example, if people are asked to help another person, two types
of theories could influence their decision to help. One, person-focused theory might be captured by the molecule:
[P is a kind person; O asks P for assistance; P helps].
Note that if people who are asked to help activate the first theory and consider themselves to be kind, they are likely to provide help independently
of other considerations. If they activate the second theory, however, their
decision to help may depend on his or her perception of the requests
legitimacy.
An imaginative study by Langer and Abelson (1972) can be conceptualized in terms of these considerations. Shoppers were approached outside a
supermarket by a woman who had ostensibly hurt her leg. In some cases,
the woman asked the shopper to call her husband to come and pick her up.
This request was likely to be considered legitimate under the circumstances. In other cases, she asked the shopper to call her boss and tell him
she would be latesomething that was not sufficiently important to justify
asking a stranger. The second variable manipulated in this study was the
order in which the woman (a) described her adversity and (b) made her request. Specifically, under request-first conditions, the woman said: Would
you do something for me? Please do me a favor and call my husband to ask
him to pick me up. My knee is killing me. I think I sprained it. Under needfirst conditions, however, the womans words were: My knee is killing me. I
think I sprained it. Would you do something for me? Please call my husband
and ask him to pick me up.
Thus, the descriptive content of the womans statements was identical in
the two cases. Under request-first conditions, however, the victims initial
statement (Would you do something or me . . .) focuses attention on the
recipient and, therefore, is likely to activate a person-focused theory of the
sort described earlier. To this extent, shoppers who typically consider
themselves to be kind are likely to grant the request regardless of other
considerations. In contrast, the womans initial statement under need-first
conditions (My leg is killing me . . .) focuses attention on the victim and,
therefore, is likely to activate the second theory, in which the legitimacy of
the request comes into play. In this case, therefore, shoppers should comply with the request only when the request appears legitimate. Langer and
Abelsons findings are consistent with this analysis. That is, the likelihood
301
of helping the victim increased with the requests legitimacy under needfirst conditions but not under request-first conditions.
Effects of Anticipated Outcomes on Decisions
People who are confronted with a decision are likely to base it on an implicit theory about the consequences of the alternative courses of action
available. These theories are often domain specific. Peoples theories about
the consequences of betting on a horse race are likely to differ from their
theories about the consequences of studying for a final examination. However, more general theories can exist as well. As we discussed in chapter 3,
Higgins (1997) postulates two general motivational orientations that can underlie goal-directed information processing. One, promotion focus, is characterized by an emphasis on the rewards that can potentially result from a behavioral decision irrespective of the costs that might be incurred. The
other, prevention focus, is characterized by a concern with avoiding negative consequences of ones behavior independently of its potential benefits.
These different orientations could be embodied in two implicit theories,
each of which has implications for a different behavioral objective:
[P performs A; A has positive consequences; P attains benefits],
and
[P performs B; B avoids negative outcomes; P avoids misfortune].
302
CHAPTER 10
the relative accessibility of the theories and, therefore, the likelihood they
are applied. One factor, discussed in chapter 3, can be the salience of ones
group membership. That is, consciousness of belonging to a group appears
to activate feelings of responsibility to others and, therefore, to increase attention to negative consequences of behavior. Thus, it may increase the relative accessibility of a prevention-focus molecule that, once activated, influences decisions in both interpersonal and individual choice situations
(Briley & Wyer, 2002). Situational factors that emphasize the importance of
success as opposed to the importance of avoiding failure could have a similar influence. Finally, chronic individual differences in the accessibility of
the two theories can also exist (Aaker & Lee, 2001; Briley et al., 2000). (This
research is also discussed in chapter 3.)
However, the implicit theories on which decisions are based can pertain
to not only the objective consequences that are likely to occur but also the
emotional reactions that occur in response to these consequences. People
may anticipate feeling happy if their decision has positive consequences
but disappointed if it has negative consequences. In addition, they might
anticipate experiencing regret if they decide against an option that would
have benefited them and might expect to feel relief if they decide to avoid
an option that would have turned out badly if they had taken it. The decision they make might depend to a greater extent on their implicit theories
about the occurrence of these reactions than by the outcomes that elicit
them (cf. Bell, 1982; Loomes & Sugden, 1986; Mellers, Schwartz, & Ritov,
1999). However, both situational and individual differences undoubtedly occur in the conditions in which these theories are invoked.
Fong and Wyer (2003) examined these differences in simulated choice
situations similar to those that participants were likely to encounter in daily
life. Both North American and Hong Kong Chinese students participated. In
one study, participants imagined a financial situation in which they could
decide either (a) to make an investment that could potentially yield a considerable profit but also ran the risk of taking a substantial loss, or (b) to
make a bank deposit that would yield a low rate of interest but would avoid
the risk of loss. In a second study, they imagined an academic situation in
which they could decide whether (a) to study a particular topic that would
ensure a high grade on an exam if the topic actually appeared but a low
grade if it did not, or (b) to study other topics, guaranteeing an average
grade regardless of whether or not the special topic was on the exam. In
each experiment, the magnitude of the gain or loss that might occur as a result of taking the risk, and information about the decisions that others had
typically made, were varied over conditions.
Participants after reading the scenario reported the likelihood that they
would choose the risky option and estimated the risk of doing so. Then,
303
they estimated the emotional reactions they would experience if they (a)
took the risk and benefited, (b) took the risk and incurred misfortune, (c)
did not take the risk and avoided misfortune, or (d) did not take the risk
and missed out on an opportunity to benefit. These emotional reactions
were analogous to happiness, disappointment, relief and regret, respectively.
Participants decisions in each study were typically influenced by their
perception of the risk involved in doing so, the importance of the consequences, and the decisions that others ostensibly made. However, the effects of these variables were largely mediated by their influence on participants anticipated emotional reactions. Regression analyses indicated that
participants anticipated reactions accounted for a sizable proportion of
variance in decisions over and above the effects of situational variables,
whereas the latter variables contributed only a small and often nonsignificant proportion of variance in decisions over and above the effects of
anticipated emotional reactions. In other words, participants implicit theories about the emotional reactions they would experience in response to alternative decision outcomes were the primary determinant of their choices.
However, the relative impact of the four anticipated reactions varied
over both scenarios and cultural groups. For example, decisions in the financial situation were based primarily on the consequences of taking the
risk (i.e., happiness and disappointment), whereas decisions in the academic situation were influenced by anticipated relief and regret as well.
Moreover, Americans decisions in the latter situation were influenced to
the greatest extent by the happiness they anticipated if they benefited as a
result of taking the risk, and to the least extent by the relief they anticipated
if they avoided a negative outcome by not taking the risk. In contrast, Chinese participants decisions were influenced least by the happiness they anticipated as a result of receiving a positive outcome and most by the relief
they anticipated as a result of avoiding a negative one. These results suggest that chronic cultural differences exist in the tendency to invoke theories that focus on positive versus negative decision outcomes. To this extent, they confirm conclusions drawn by Briley et al. (2000; see also Aaker &
Lee, 2001).
Effects of Stereotypes on Social Behavior
One of the most heavily researched areas in social psychological research
surrounds the determinants and effects of social stereotypes (for reviews
of theory theoretical analyses, see Bodenhausen & Macrae, 1998; Fiske,
1998; Hamilton & Sherman, 1994). In most of this research, a stereotype has
been conceptualized as a cluster of traits and behavioral dispositions that
are associated with a social or ethnic category (women, African Americans,
304
CHAPTER 10
fraternity members, etc.). Generalized representations of this sort presumably influence persons attitudes toward individual members of the category and the attributes they assign to these members (Hamilton & Sherman, 1994). Moreover, a trait-based stereotype that is activated in one
situation can later influence the interpretation of information about other
individuals to whom the stereotype is objectively inapplicable (Devine,
1989; Lepore & Brown, 1997). Finally, active attempts to suppress the use of
the stereotype can have boomerang effects, actually increasing the use of
the stereotype as a basis for judgment once these conscious suppression
mechanisms are deactivated (Bodenhausen & Macrae, 1998).
An understanding of the impact of trait-based stereotypes on social judgments, and the conditions in which they are applied, is of obvious importance. As Reid and Wyer (1998) noted, however, there is surprisingly little
evidence that these stereotypes directly influence individuals actual behavior toward members of the stereotyped group. The classic study by
LaPiere (1934) provides an example of a situation in which individuals behavioral decisions in a social context do not correspond to the attitudes
and behavioral dispositions they express outside of this context. In fact, it
seems somewhat implausible that people who encounter a member of a stereotype group would expend the cognitive energy required to activate a
trait-based stereotype and construe its implications in anticipation of deciding how to behave toward members of the group. Instead, they are likely to
activate a situation stereotype of the events that are likely to occur in the
situation at hand, and to base their decision on the implications of this
event representation.
A situation stereotype is likely to be activated by a configuration of features that exist in the situation rather than by individual attributes. Thus,
its activation can depend on features of persons involved in a situation and
features of the situation in combination, rather than on either set of features
in isolation. For example, suppose a woman sees an African-American man
approaching her late at night on an empty Chicago street. She is unlikely to
activate a trait-based stereotype of African Americans as aggressive or hostile before deciding to cross the street. Rather, she is more likely to activate
a prototypic event representation of the events that occur in this particular
situation (i.e., People who see a woman walking alone late at night are
likely to mug or rape them, and so it is best to avoid contact with them if
possible.) This stereotypic event representation may function as an implicit theory that is activated and applied in the particular situation at hand.
If the woman encountered this same individual in a university library she
would be far less likely to engage in avoidance behavior. This is because a
different situation stereotype is activated that has quite different implications for behavior.
305
FINAL COMMENT
The research reviewed in this chapter has cut across a very wide variety of
concerns, ranging from counterfactual reasoning and reconstructive memory to perceptions of social reality, cultural differences in future outlook,
perceptions of social support, marital satisfaction, and stereotyping. The
role of implicit theories in these phenomena is admittedly somewhat conjectural. That is, few studies provide direct evidence of the cognitive basis
of the judgments and decisions that were made. To the extent these theories exist, however, they are a useful construct in integrating a diversity of
empirical findings within a common conceptual framework.
Much of our discussion was guided by the assumption that many different theories can often be brought to bear on a judgment or decision, and
that the theory that is activated and applied is determined by its relative accessibility. The determinants of knowledge accessibility summarized in
chapter 3 obviously come into play in predicting the conditions in which
these theories are used.
Although the phenomena discussed in this and the preceding chapter do
not pretend to be exhaustive, they are representative of those in which narrative representations are likely to be constructed and used as a basis for
judgments and decisions in daily life. However, an obvious and important
factor that influences judgments and behavior in social situations has been
totally ignored, both in these chapters and elsewhere in this volume. We
now turn to a consideration of this factor.
C H A P T E R
11
The Role of Affect
in Information Processing
Our focus throughout this volume has been on the way we comprehend information about people, objects, and events of the sort we encounter in
daily life. In our discussion, we have emphasized the role of narrative representations of knowledge in conceptualizing new information and in making
judgments of the people and events to which the representations refer. We
have also considered the role of pictures and visual images in the comprehension and use of information. In doing so, however, we have virtually ignored an obvious but important factor that pervades information processing outside the laboratory.
Specifically, many situations we encounter in daily life elicit affective reactions. These reactions can be stimulated by direct experiences with a
person, object, or event with which they are associated. However, imagining these stimuli, recalling past experiences in which they occurred, or writing stories about them can also elicit these reactions. When these affective
reactions are experienced, they can influence judgments and overt behavior toward the stimuli that gave rise to them. Thus, we behave antagonistically to a person who makes us angry, and attempt to increase our contact
with someone who makes us feel happy.
The cognitive determinants and consequences of affect and emotion have
been a major concern of social psychological research and theorizing for
more than 25 years. The issues investigated have ranged from the influence
of affect on the encoding and organization of new information in memory
(Bower, 1981, 1991) to its impact on the retrieval of previously acquired concepts and knowledge (Blaney, 1986; Niedenthal & Setterlund, 1994), its use as
a basis for judgments and decisions (Clore, 1992; Schwarz & Clore, 1983,
1988), and its influence on the strategies that individuals use when perform306
307
ing cognitive tasks (Isen, 1987; Mackie & Worth, 1989; Nisbett, Schwarz, &
Bless, 1991). These issues have been investigated in research on not only
judgment and decision making but also creativity (Isen, Daubman, & Nowicki,
1987; Isen, Johnson, Mertz, & Robinson, 1985), communication and persuasion (Albarracin & Wyer, 2001; Bless, Bohner, Schwarz, & Strack, 1990; Mackie
& Worth, 1989l; Wegener, Petty, & Smith, 1995); stereotyping (Bodenhausen,
1993; Bodenhausen, Sheppard, & Kramer, 1994), self-evaluations (Levine,
Wyer, & Schwarz, 1994); political judgment (Abelson, Kinder, Peters, & Fiske,
1982; Ottati & Isbell, 1996; Ottati & Wyer, 1993), and consumer behavior
(Adaval, 2001; Pham, 1998; Pham, Cohen, Pracejus, & Hughes, 2001).
Despite the diverse conditions in which affect has been shown to influence information processing, the cognitive mechanisms through which it
exerts this influence may be more limited than is often assumed. These
mechanisms have been difficult to isolate. However, the research reported
to date converges on two general conclusions:
1. Positive and negative affect can be preconditions for cognitive operations for cognitive productions of the sort that compose procedural knowledge (J. Anderson, 1983; E. R. Smith, 1990) and that govern behavior in specific types of situations to which they are relevant. Once activated, these
productions can influence responses to new information without conscious
awareness.
2. The affective reactions that one experiences at a given moment can be
used as information about ones attitude toward oneself, other people, situations with which one is confronted, the outcomes of behavior, or the appropriateness of certain strategies for attaining behavioral outcomes. It can also
influence perceptions of the validity of other affect-eliciting information and,
therefore, can influence the weight that is attached to this information in
making judgments.
The validity of these conclusions is discussed with reference to research
and theory in several different areas, with particular emphasis on the impact of affect on behavioral decisions. In this context, a performance feedback model is proposed that accounts for the influence of affect on behavior and judgment in a number of different situations that occur in both the
laboratory and daily life.
GENERAL CONSIDERATIONS
Definition
Wyer, Clore, and Isbell (1999) conceptualized affect as a configuration of
positively or negatively valenced subjective reactions that a person experiences at a given point in time and perceives as either pleasant or unpleas-
308
CHAPTER 11
ant feelings. These feelings can be elicited by proprioceptive cues (e.g., facial expressions; see Strack, Martin, & Stepper, 1988; Zajonc, Murray, &
Inglehart, 1989), physical stimulation, or drugs. More commonly, however,
they are internal responses to a set of new or previously formed cognitions.
Several general classes of affective reactions can be distinguished (for a detailed discussion of these distinctions, see Clore, Schwarz, & Conway, 1994;
Ortony, Clore, & Collins, 1988; Wyer et al., 1999).
1. Moods and Emotions
Wyer et al. (1999) assumed that emotions are affective states that occur
in response to cognitions about persons, objects, and events, and that are
consciously attributed to a specified source. In contrast, moods are typically experienced in the absence of specific cognitions about their source
and are not attributed to any particular cause unless one is called upon to
do so. Thus, the clusters of subjective reactions that exemplify emotions
and moods are similar except that one is consciously attributed to a given
source and the other is not. As Wyer et al. (1999) suggested, feelings that
are elicited by a stimulus event take time to dissipate. Therefore, feelings
that are initially experienced as an emotion can become a mood once their
source is no longer considered.
The preceding remarks suggest an important distinction between emotions and affect per se. That is, emotions, unlike moods or affective reactions more generally, have a strong cognitive component. Forgas (2002), for
example, assumed that emotions are defined in terms of a cognitive appraisal of a situation, and that physiological and autonomic reactions characteristic of that emotion are responses to this appraisal (p. 104). Thus,
emotions, unlike other affective states, are necessarily accompanied by
cognitions about the situational factors that give rise to them. To this extent, there may be concepts in memory whose features include a representation of ones affective reactions as well as alternative names of both the
emotion and its behavioral manifestations.
The concept angry, for example, might be metaphorically represented as
shown in Fig. 11.1. The concept itself is denoted by a central node to which
the features that define and characterize it (a cognitive appraisal of the conditions that elicit the emotion, a mental representation of the affective reactions that are experienced, names, behaviors, etc.) are associatively linked.
The concept of a mood might be similar to that shown in the figure except
that a characterization of its eliciting conditions would not be included
among its features. In each case, thinking about one or more of the features
associated with it could presumably activate the mood or emotion concept.
Thus, consciousness of the cluster of physiological reactions that define an
emotion concept might activate it, in much the same way as the features of
309
any concept can activate this concept. Note, however, that these physiological reactions per se do not activate the concept. This activation occurs
only if the physiological reactions are thought about or otherwise cognitively represented.
2. Affect Versus Evaluation
It is important to distinguish between affective reactions to a stimulus
and evaluations of the stimulus along a goodbad dimension (e.g., attitudes;
see Eagly & Chaiken, 1993, 1998). Although ones feelings about a person or
object can often be the basis of such an evaluation, not all evaluative judgments are based on this criterion (cf. Adaval, 2001; Wyer et al., 1999; Zanna
& Rempel, 1988). The distinction between affect and evaluation is not always made in theories of affect-based information processing (cf. Abelson
et al., 1982; Bargh, Chaiken, Govender, & Pratto, 1992; Fiske & Pavelchak,
1986). In the present context, however, this distinction is important to keep
in mind. For one thing, evaluations are cognitions and, as such, can be features of the mental representations of the entities to which they refer. Affective reactions, however, may not be.
3. Affect Versus Cognition
As implied by our discussion of moods and emotions, people have concepts of different types of affect, and cognitions about different configurations of subjective reactions can be features of these concepts (see Fig.
11.1). These latter cognitions, however, may not always have verbal labels.
Moreover, it is important to distinguish between cognitions about ones affective reactions and the affective reactions per se. Although affective reactions may sometimes be learned responses to cognitions, they are not part
310
CHAPTER 11
of the cognitive system. This distinction is made by LeDoux (1994) in conceptualizing the relation of emotion and memory:
It is important to distinguish emotional memory [the learned or conditioned
emotional significance of an object that automatically triggers an emotion]
and memory for emotion. The latter is a declarative, conscious memory of an
emotional experience. It is stored as a fact about an emotional episode. Emotional memory (mediated by the amygdala) and memory for emotion (mediated by the hippocampus) can be reactivated in parallel on later occasions,
thereby giving new declarative memories an emotional flavor. . . . Because
emotional memory and memory for emotion are stored through different
brain systems, it is possible that the information stored will not always coincide. (p. 312)
311
312
CHAPTER 11
313
314
CHAPTER 11
315
perform at the time they encounter stimuli that is not affect eliciting in and
of itself. That is, people may spontaneously categorize stimuli as either
good or bad. (This automatic categorization process could reflect an unlearned, evolution-based disposition to identify stimuli as potentially harmful or not harmful in order to survive.) However, affective reactions might
not provide the basis for this categorization. In fact, these reactions might
only be elicited by a cognitive appraisal of the stimulus and its implications
that occurs after the stimulus has been categorized. If this is so, the effects
observed by Bargh et al. (1992; see also Fazio et al., 1986) would not conflict
with the implications of Niedenthal et al.s (1997) findings.
This conclusion can be summarized in the following postulate:
Postulate 11.1. Affective reactions may activate a specific affect-relevant concept that can be used to interpret them. However, affective reactions do not
spontaneously influence the accessibility in memory of other similarly
valenced semantic concepts and declarative knowledge.
316
CHAPTER 11
317
318
CHAPTER 11
319
320
CHAPTER 11
321
The cat still had not come back from his night on the prowl, and M wondered
where he was. Then the phone rang. . . . It was an old friend . . . who had recently left town. He was calling to say that things were going fine and that he
liked his new job. . . . He would not be back in town over Thanksgiving but
would probably be back over Christmas break. he said he had a great gift for
her in mind.
322
CHAPTER 11
323
An Initial Test
An initial demonstration of the use of affect as information was performed
by Schwarz and Clore (1983, Experiment 1). This study not only established
a methodology for identifying the impact of affective reactions on judgments, but also identified conditions in which the reactions were likely to
be employed.
Schwarz and Clore (1983) recognized that the affect people experience at
any given time can come from different sources. However, people often cannot distinguish easily between the affect elicited by one source and the affect elicited by another. Consequently, if they base their judgments of a
stimulus on their affective reactions to it, they are likely to confuse these reactions with affect they are experiencing for other, irrelevant reasons, and
so the latter affect can have an influence on these judgments. On the other
hand, suppose people do not base their judgment on affect. Then, extraneous affect should have no impact in any condition.
To examine these possibilities, Schwarz and Clore asked participants to
write about a happy or sad past experience while they were seated in a
strange-looking soundproof booth. Some participants were told that people
often found that the booth made them feel mildly elated after they had been
in it for a while. Others were told that the booth made people feel mildly depressed, and others were given no information about the booths effects.
Participants after writing their essays were then asked to judge their satisfaction with their life as a whole. These judgments are summarized in Table
11.1 as a function of the type of experience participants wrote about and
the ostensible effects of the booth.
Participants who were given no information about the booths effects reported greater life satisfaction when they had written about a happy experience than when they had written about an unhappy experience. Considered
in isolation, this difference might be attributed to the content of the story
that participants wrote rather than the affect they were experiencing. If this
were the case, however, the alleged effects of the booth on their feelings
should have had no impact on their judgments. In fact, with participants who
TABLE 11.1
Judgments of Life Satisfaction as a Function of the Affective Quality of a
Recalled Personal Experience and Ostensible Characteristics of the Booth
Ostensible Effect of Booth
Depressing
No Effect
Elating
9.6
8.6
8.6
5.7
9.7
4.4
324
CHAPTER 11
had written about a sad experience, telling them that the booth would make
them feel sad eliminated the negative impact of the story on their judgments
of life satisfaction, whereas telling them that the booth would make them
happy increased these judgments relative to no-information conditions. In
other words, participants who were told that the booth made them unhappy
attributed their feelings to the booth and did not use it as a basis for judgment. In contrast, participants who experienced unhappiness despite being
told that the booth would make them happy believed that they must be particularly unhappy about their life as a whole and, therefore, judged themselves to be even less satisfied than they would otherwise.
It is interesting to note that the effects of the booth were not apparent
when participants had written about a positive life experience and, therefore,
were feeling happy. These participants appeared to base their judgments of
life satisfaction on their feelings regardless of the information they were
given about the booth. The asymmetric effects observed in the two affectinduction condition could be due to factors similar to those underlying Arkin
et al.s (1976) research described in chapter 9. That is, participants whose affect had favorable implications for themselves were inclined to accept these
implications and use them as a basis for judgment without seeking alternative explanations for their feelings. In contrast, people whose affect had undesirable implications resisted using it as a basis for judging themselves unless no alternative explanation for their feelings was available.
The Role of Narratives in Affect Elicitation
Affective reactions can be elicited in many ways, as noted earlier. Affect
that is induced by proprioceptive feedback (cf. Strack, Martin, & Stepper,
1988; Zajonc, Murray, & Inglehart, 1989) can also influence judgments to
which it is relevant (e.g., the amusement elicited by cartoons). The role of
narratives in affect elicitation was confirmed in a series of studies by Strack,
Schwarz, and Gschneidinger (1985). They found that when people listed
past experiences in abstract terms that were unlikely to elicit affect, they
used these experiences as standards of comparison in judging their current
life satisfaction, reporting themselves to be less satisfied when the past experiences were favorable than when they were unfavorable. When participants wrote detailed narratives of their past experiences, however, this activity elicited affect, and this affect had a positive influence on satisfaction
judgments.
In a particularly provocative study (Strack et al., 1985, Experiment 3),
participants were either asked to explain why a particular experience occurred or to describe how it occurred. Writing explanations for happy and
sad events did not elicit affect, and the favorableness of the experiences
that participants explained had contrast effects on the life satisfaction they
325
reported later. However, writing descriptions of how the experience occurred did elicit affect and consequently had effects on judgments similar
to those reported by Schwarz and Clore (1983).
326
CHAPTER 11
Judgment of
Achievement Competence
Judgment of
Social Competence
4.77
4.37
4.38
4.38
4.87
4.67
4.46
3.46
327
Consumer Judgment
Judgments of competence can often be based on either descriptive or affective criteria, and so participants are likely to use whichever criterion comes
to mind most easily at the time judgments are made. This may also be true
in the consumer domain. Evaluations of commercial products, for example,
are often based on utilitarian considerations as well as hedonic or affective
criteria. A refrigerator, for example, is likely to be evaluated on the basis of
its ability to perform the functions for which it is designed. On the other
hand, a cashmere sweater, or strawberry shortcake, could be judged on the
basis of either affective criteria (how it feels or tastes) or utilitarian considerations (durability, healthfulness, etc.). In the latter case, people are most
likely to base their evaluations on affect if (a) the stimulus is one that is typically judged on the basis of affective (hedonic) criteria, (b) the information
about the stimulus is conveyed in a way that spontaneously elicits positive
or negative feelings, or (c) peoples objectives at the time of judgment require the use of affective criteria.
The latter contingency was demonstrated by Pham (1998). Participants
who had been induced to feel happy or unhappy by recalling a past experience were asked to read a description of a new movie and indicate if they
would like to attend a preview of it. Before reading the description, some
participants were given instructions that emphasized hedonic criteria for a
decision (e.g., having a good time, enjoying life, etc.), whereas others were
given a utilitarian reason for seeing it (its relevance to participating in a
subsequent study for pay). Participants in the first condition expressed
stronger intentions to see the movie if they were feeling happy than if they
were not, suggesting that they used the affect they were experiencing as a
basis for their judgment. In contrast, participants intentions in the second
condition were unaffected by the affect they were experiencing.
The use of hedonic or utilitarian criteria can also depend on their relative accessibility at the time of judgment. However, unless situational factors make descriptive criteria particularly relevant (as in Levine et al.s 1994
study), these criteria may be relatively inaccessible. In contrast, the affective reactions can be activated very quickly, before a detailed cognitive
analysis of the utilitarian features of the description is performed. (For a
more general argument that affective responses to a stimulus as a whole occur more quickly than responses to its individual features, see Zajonc,
1980.) One implication of this is that if a stimulus can potentially be evaluated on the basis of both hedonic and utilitarian criteria, hedonic criteria
may often take priority. That is, utilitarian considerations may come into
play only when people have both the motivation to evaluate these criteria
and the ability to do so.
A study by Shiv and Fedorikhin (1999) supported this contention. Participants were given a choice between a chocolate cake and a fruit salad as a
328
CHAPTER 11
reward for participating in the study. The authors assumed that the sight of
the chocolate cake would elicit positive affect, and that the choice of this
option would reflect the use of these reactions as a basis for their decision.
In contrast, the choice of a fruit salad was more likely to reflect the use of
utilitarian (health-related) criteria. The authors further speculated that the
latter criteria would come into play only if participants had the motivation
and ability to think carefully about their decision.
To evaluate this possibility, some participants were placed under high
cognitive load at the time they made their choice by being asked to remember a multiple-digit number for use in a later part of the study. Others were
put under low load. Participants in the latter condition tended to choose the
fruit salad, suggesting that they based their decision on utilitarian (i.e.,
health-related) criteria. Participants who were distracted from thinking about
their decision, however, were more inclined to choose the chocolate cake.
The implications of these findings may extend beyond the domain of
consumer judgment. For example, a faculty member who wishes to hire a
research assistant may consider utilitarian criteria (background, experience, etc.) to be particularly relevant. In a personal interview, however, a
candidates personality and physical attractiveness might elicit affective reactions that are used as a basis for judgment in addition to other considerations. Moreover, if the interviewer meets the candidate before receiving
descriptive information about his or her qualifications, these affective criteria may predominate, and descriptive criteria may be taken into account
only if the interviewer has the motivation and ability to think more carefully
about the candidates credentials.
329
evaluations of the product but also can bias the search for additional information about it.
1. Experiment 1
To examine these possibilities, participants in one experiment took part
in three conditions. In no-picture conditions, participants who had been induced to feel either happy or unhappy as a result of writing about a life experience were given verbal descriptions of a pair of sports shoes consisting
of three favorable and three unfavorable features. Some participants were
told to base their judgment on hedonic criteria (how it would feel to wear
the shoes), and others were told to use utilitarian criteria (durability, workmanship, etc.). Participants evaluations of the products in these conditions, shown in the top left panel of Table 11.3, indicate that the extraneous
affect that participants were experiencing influenced their evaluations
when they were told to use hedonic criteria but had little effect when they
were told to use utilitarian criteria. These data, which are consistent with
those reported by Pham (1998), confirmed the assumption that when participants are told to use hedonic criteria as a basis for evaluations, they imagine the feelings they would experience when using the product, and the affect they were experiencing for other reasons influence their assessment of
these feelings. When participants were told to use utilitarian criteria, however, they did not consult their feelings and so the extraneous affect they
were experiencing had little effect.
In two other conditions of this study, however, participants were
shown an attractive picture of the product before they received descripTABLE 11.3
Product Evaluations and Recall of Attribution Information
as a Function of Induced Affect, Judgment Criterion,
and Information Presentation Conditions
Affect Induced
Before Picture
No Picture
Positive
Affect
Negative
Affect
Positive
Affect
Negative
Affect
Positive
Affect
Negative
Affect
5.25
5.25
3.75
3.78
3.46
3.89
3.50
4.33
4.67
6.38
3.78
4.88
(Experiment 2)
1.32
2.13
1.80
1.62
4.38
3.63
5.78
4.11
5.30
4.48
1.32
1.29
1.68
1.32
1.59
1.26
4.10
4.12
330
CHAPTER 11
331
332
CHAPTER 11
Yeung (2003) obtained evidence of this contingency in a study that paralleled the experiments described in the previous paragraphs. Rather than
giving participants a picture of the product, however, she told them that
the product they would judge was being produced by a company that was
well known for either hedonic products (Pepsi) or utilitarian ones (Bonaqua, a brand of distilled water). Either before or after conveying this information, participants were induced to feel happy or unhappy by writing
about a past experience. Finally, they were given information about a pair
of sports shoes that were ostensibly about to be produced by the company
whose name had been provided earlier.
When the brand was associated with a product that was judged on the
basis of affective criteria, results paralleled those obtained under comparable conditions of the earlier studies. That is, extraneous affect had a positive influence on participants evaluations of the brand extension when it
was induced before participants learned of the brand, but not when it was
induced afterward. When the brand was associated with a product that was
judged on the basis of utilitarian criteria, however, the extraneous affect
that participants were experiencing had no influence regardless of when it
was induced.
These results further strengthen the conclusion that judgments of a stimulus are often induced by an initial appraisal that is made at the time the
stimulus is first encountered independently of more specific information
about the stimulus that is conveyed afterwards. In these conditions, however, affective reactions only enter into this process when the appraisal
elicits affect. Moreover, extraneous affect that participants happen to be experiencing only has an impact when it is experienced at the time the appraisal is made and, therefore, becomes confused with the affect that is automatically elicited by the appraisal-inducing stimulus.
AFFECT-CONFIRMATION PROCESSES
In the research by Yeung (2003; Yeung & Wyer, in press) affect was presumably elicited by a spontaneous global appraisal of the product as a whole
that was independent of any specific product features. However, affect can
obviously be elicited by specific features of a stimulus as well. A politicians
position on a particular issue, for example, could elicit positive or negative
affective reactions in someone who favors or opposes this position. In the
product domain, a sweaters material can convey how nice it would feel to
wear it, and could elicit affect for this reason. Individuals who wish to evaluate such stimuli on the basis of affective considerations are likely to pay attention to features that elicit affect and use these features rather than others as a basis for judgment. If the individuals are also experiencing affect
333
for other reasons, however, this affect could have two effects. First, they
might confuse their affective reactions to these features with the affect they
are experiencing for other reasons, and consequently could interpret these
features as more or less favorable. On the other hand, their extraneous feelings could influence the importance that they attach to the affect-eliciting
features and, therefore, the weight they attach to these features when computing a judgment.
This latter possibility is formalized in the affect-confirmation theory proposed by Adaval (2001). According to this theory, participants who base
their judgments of a stimulus on hedonic considerations assess not only
their affective reactions to the stimuluss individual features but also the extent to which their perceptions of these reactions are valid. If extraneous affect that they happen to be experiencing is similar to the affect that is elicited by a stimulus feature, it may appear to confirm their reactions to this
feature and, therefore, may give them more confidence in the implications
of these reactions. On the other hand, if extraneous affect they are experiencing differs from that elicited by a stimulus feature, they might interpret
this difference as ambivalence and thus might decrease their confidence in
the validity of their assessment of the features affective implications. Consequently, if individuals base their judgments of the stimulus on their affective reactions to it, they should attach greater weight to features that elicit
affect similar to the feelings they are experiencing for other, unrelated reasons than to features that elicit different reactions.
Empirical Evidence
Two studies by Adaval (2001) evaluated these hypotheses. In one, participants who had been induced to feel either happy or unhappy received information about a pair of jeans consisting of either a favorable or an unfavorable brand name (Levis vs. Rustler) and either a favorable or an
unfavorable attribute (preshrunk vs. not preshrunk). In some cases, participants were told to base their evaluations of the jeans on hedonic criteria
(how it would feel to wear them) and the attribute was described in a way
that emphasized its implications for comfort. In other cases, they were told
to base their judgments on utilitarian criteria, and the attribute description
emphasized its implications for the quality of construction. Participants after reading the product description evaluated the product and then reported their confidence that this evaluation was correct.
Note that differences in the impact of the attribute information on product evaluations under different affect-induction conditions could indicate either that the attribute information was weighted more heavily in making
their judgments or, alternatively, that the attribute information was interpreted differently (i.e., as more or less extreme) under these conditions. To
334
CHAPTER 11
avoid this ambiguity, Adaval inferred the relative weight attached to the attribute information from the impact of the brand information that accompanied it. If affect information receives relatively greater weight when its
evaluative implications are consistent with the extraneous affect that participants were experiencing, the relative influence of the brand information
that accompanies it should be relatively less.22
Data summarized in the top half of Table 11.4 confirm this hypothesis.
That is, when participants were told to evaluate the product on the basis of
hedonic criteria, brand information had less impact on their judgments
when the attribute description was evaluatively consistent with the extraneous affect that participants were experiencing than when it was inconsistent, indicating that the attribute information received greater weight in the
former condition. When participants were told to use utilitarian criteria,
however, brand had greater influence in the former condition than the latter. This reversal suggests that participants in this condition may have intentionally avoided basing their judgments on affective criteria and, in fact,
overcompensated for its influence, attaching less weight to the attribute information than they otherwise would. Interestingly, participants reported
greater confidence in their judgments when the evaluative implications of
the attribute information and those of the extraneous affect they were experiencing were consistent, and this was true regardless of the objective they
were pursuing. Thus, participants were not only more confident about using their affective reactions to the attribute information as a basis for judgments when their evaluations were based on hedonic criteria but were also
more confident about avoiding its use when the basis for their judgment
was utilitarian.
A second experiment confirmed these predictions under conditions in
which the criterion that participants used was activated spontaneously by
the type of product that participants were asked to consider. Normative
data suggested that in the absence of explicit instructions, participants
were likely to evaluate jeans on the basis of hedonic criteria and backpacks
on the basis of utilitarian criteria. A second experiment was conducted using these two product categories but without instructing participants about
the criteria they should employ. The bottom half of Table 11.4 shows that as
expected, brand had less impact on evaluations of jeans when the
evaluative implications of the attribute information were consistent with
the extraneous affect that participants were experiencing, whereas it had
greater impact on evaluations of backpacks in this condition.
22
The use of this criterion assumes that judgments are a weighted average of the implications of the two pieces of information considered separately (cf. N. Anderson, 1971, 1981). Although this assumption is often questionable (cf. Wyer & Carlston, 1979), the validity of the assumption in product evaluation situations is viable (cf. Bettman, Capon, & Lutz, 1975; Troutman
& Shanteau, 1976).
22
335
Attribute Inconsistent
With Induced Affect
2.56
1.85
4.2
1.3
Intrinsic motivation was inferred from responses to a self-report index of the strength of
participants partisanship (i.e., their party identification). To induce extrinsic motivation, participants were told that the individual they would evaluate was a current candidate for office in
their area, and that we were consequently interested in whether they would vote for him. In contrast, low motivation was induced by telling participants we were primarily interested in
whether the article was clearly written, independently of its specific content.
23
336
CHAPTER 11
pants after reading the news article reported their impressions of the candidate and then indicated their personal agreement with each of the issue
stands he had espoused, along with an equal number of positions on which
the candidates stands were unknown.
When participants were neither intrinsically nor extrinsically motivated
to evaluate the candidate, they judged him more favorably when they were
feeling happy than when they were not. Moreover, they reported stronger
agreement with the candidates issue stands in the former case than the latter. (In contrast, their agreement with issue positions that were not associated with the candidate were not influenced by the affect they were experiencing.) Thus, participants mood influenced their perceptions of the
favorableness of the candidates issue stands and these perceptions, in
turn, determined their liking for the candidate. This conclusion was confirmed by covariance analyses. Specifically, the effects of experimental variables on candidate evaluations were eliminated when their effects on issue
agreement were covaried, whereas the effects of these variables on issue
agreement remained significant when evaluations of the candidate were
covaried.
Isbell and Wyer (1999) concluded that participants interpreted their reactions to the candidates issue stands as generally more favorable when
they were feeling happy than when they were not. However, affectconfirmation processes could also underlie these findings. For example,
happy participants may have had more confidence in the validity of their
reactions to issue stands that elicited positive affect (i.e., stands with which
they agreed) than in the validity of their reactions to issue stands that elicited negative affect (stands with which they disagreed). Unhappy participants, on the other hand, may have had less confidence in their reactions
to the former statements than their reactions to the latter. The differential
weighting of the issue stands that resulted from these differences in confidence could have influenced candidate evaluations for much the same reasons suggested by Adaval (2001). Isbell and Wyers study does not distinguish between these interpretations.
337
extraneous sources of affect, they might try to adjust for this biasing influence. As noted in chapter 10, however, people are often unclear how much
adjustment is necessary. If people do not adjust enough, the extraneous affect they are experiencing would continue to have a positive influence on
their judgments. If they adjust too much, on the other hand, it would appear
to have a contrast effect.
The latter appeared to be the case in Isbell and Wyers study. That is, although the extraneous affect that participants were experiencing had a positive impact on their judgments when they were unmotivated to make accurate evaluations, it had precisely the opposite effect on judgments by
individuals who were either intrinsically or extrinsically motivated to be
correct (see footnote 24).
Effects of Ability
Even if people are motivated to adjust for the biasing influence of extraneous affect, they may be unable to do so. Ottati and Isbell (1996), for example, found that extraneous affect had a positive influence on evaluations of
a candidate by participants with relatively little knowledge about politics,
but had a contrast effect on evaluations by participants with high political
expertise. Perhaps both groups of participants adjusted for the biasing influence of the extraneous affect they were experiencing. However, participants with little expertise had difficulty evaluating the implications of the
candidates issue positions and the cognitive resources they had to devote
to this activity prevented them from adjusting sufficiently. On the other
hand, experts found it easy to evaluate the implications of the candidates
issue positions and so they could devote more effort to distinguishing their
reactions to these stands from other sources of affect. However, these participants adjusted too much, resulting in a contrast effect of extraneous affect on their evaluations.
Adjustment Versus Discounting
Adjustments for bias are most likely to occur when people consider their affective reactions to the primary basis for their judgments of an object but
cannot distinguish clearly between these reactions and affect from other
sources. In many instances, however, nonaffective criteria for judgment
may be as or more appropriate bases for judgment. In these instances, people who believe that the affective reactions are biased by extraneous factors might ignore their feelings entirely and default to other, nonaffective
criteria. In this case, the affect they are experiencing would not have a contrast effect on judgments. Rather, it would have no influence at all.
338
CHAPTER 11
339
340
CHAPTER 11
Furthermore, these alternative productions might be activated spontaneously when the affect-based preconditions for their occurrence are met, and
might guide behavior with little if any conscious cognitive mediation.
The different productions that are activated by positive and negative affect require different types of knowledge. The procedures activated by positive affect involve the use of general concepts and knowledge structures. In
contrast, the procedures activated by negative affect may require more
specific concepts that can be used to interpret informational details. This
difference should be reflected in the type of concepts and knowledge that
are used to interpret and organize new information. At least two studies
support this possibility.
Bless et al. (1996) reasoned that if people who experience positive affect
spontaneously activate general knowledge structures to use in processing information, they should be able to process this information more efficiently.
Moreover, their use of these knowledge structures should free cognitive resources for use in other, unrelated cognitive activity. To evaluate this possibility, participants who were induced experimentally to feel happy or unhappy listened to two tape-recorded stories about daily life events (e.g.,
making a telephone call from a public phone booth) while they performed a
clerical task. Both recognition memory for the events mentioned in the audiotapes and accuracy in performing the clerical task were recorded.
As expected, happy participants were more likely than unhappy ones to
use a prototypic script (Schank & Abelson, 1977) to comprehend the taped
information. That is, happy participants were more likely than unhappy
ones to report prototype-consistent events as having been mentioned, but
were less likely than unhappy participants to report atypical events. To the
extent that participants apply a general prototype in comprehending the
tape-recorded information, however, they should be able to devote more
cognitive resources to the clerical task. This was also the case. Happy participants performed better on the secondary, clerical task than unhappy
participants did.
A quite different study by Dienes (1996) has similar implications. In this
study, happy and unhappy participants were asked to learn a list of words
that exemplified either easily identifiable concepts (animals, countries, etc.)
or obscure ones (e.g., things made of wood). Learning the list could be facilitated by (a) attending to the details of the individual words, (b) mentally
organizing the words in terms of the concepts they exemplify, or (c) both.
(For evidence that the organization of information items into categories improves learning, see Hunt & Einstein, 1981.) Dienes reasoned that when the
concepts exemplified by the stimulus words were obvious, both happy and
unhappy participants would spontaneously identify and use these concepts
to organize the stimuli. When the concepts were more obscure, however,
only happy participants should identify them spontaneously, so only these
participants learning of the stimulus words should be facilitated. Consis-
341
tent with this reasoning, happy participants had better memory than unhappy participants for the hard-to-categorize words but not for the easy-tocategorize words.
The conclusions that persons who experience positive affect bring categorical and general knowledge structures to bear on tasks they perform
does not imply that affective reactions per se cue the retrieval of these
knowledge structures. Rather, this retrieval is mediated by cognitive productions that require different types of declarative knowledge. The preconditions for activating a production include situational features as well as
affective reactions. In some cases, the cognitive productions that are activated by situation-specific features might not require the types of knowledge of concern in the studies by Dienes and Bless et al. In other cases, affective reactions might not be a feature of the preconditions that elicit goalrelevant productions. Consequently, there are many instances in which affect is likely to have little influence on the strategies that people use in responding to information.
In this regard, Murray, Sujan, Hirt, and Sujan (1990) found that although
happy participants used fewer categories to classify stimuli than control participants when they were told to focus on similarities, they used relatively
more categories to classify the stimuli when they were told to focus on differences. These latter findings qualify the conclusion that happy persons typically use broader categories to classify stimuli than unhappy persons. Further experiments by Murray et al. (1990) indicated that differences in
categorization observed in both task conditions were likely to result from a
more general tendency for positive affect to increase the number of attribute
dimensions that participants brought to bear on the particular task they
were asked to perform. Thus, it increased their ability to make distinctions
among the stimuli they were considering as well as to identify similarities. As
Murray et al. suggest, these results could reflect mood-induced differences in
the motivation to perform the task at hand rather than differences in the cognitive productions that participants spontaneously apply under the conditions being investigated. We consider this possibility in more detail shortly.
Implications for Judgment
If positive affect increases the disposition to rely on categorical criteria for
comprehending and responding to stimuli, this disposition should be reflected in judgments to which the criteria are relevant. However, this influence could be localized at two stages of processing. First, if the information
that is available for judging a stimulus consists of both categorical descriptions and more detailed features, positive affect might increase the relative
weight that is attached to the categorical information at the time the implications of the two types of information are combined to form a judgment. An al-
342
CHAPTER 11
343
Several factors could have been hypothesized to account for this difference. Mackie and Worth (1989), for example, proposed that positive affect is
likely to activate a large number of concepts and knowledge representations
in memory, taxing participants capacity to provide new information and
making them less able to attend to details of the information available. This
interpretation, however, is called into question by Niedenthal and Setterlunds (1994) finding that affective reactions activate only those concepts that
the specific type of reactions being experienced exemplify, and do not increase the accessibility of valenced concepts and knowledge in general.
A somewhat more viable conceptualization is consistent with the use of
affect as information. Specifically, Schwarz (1990; Schwarz, Bless, & Bohner,
1991) argued that because positive affect is typically associated with pleasant states of affairs, people who experience these feelings in a new situation
may appraise the situation as benign and, therefore, may not pay careful attention to details. In contrast, negative affect is associated with aversive
states of affairs and so the experience of it stimulates an interpretation of a
new situation as potentially problematic. Although these associations could
give rise to cognitive productions that are applied automatically, as noted
in a previous section, they could potentially stimulate deliberative cognitive activity as well. Consequently, people may be less motivated to pay attention to the details of a message they receive when they feel happy than
when they feel unhappy, and so the quality of the arguments contained in
the message has less impact on their judgments.
Although this interpretation seems plausible, it is contradicted by evidence that people engage in as much cognitive activity in responding to a
persuasive message when they are in a good mood as when they are in a
bad mood. In Bless et al.s (1990) study, for example, thought listing data indicated that the overall amount of this activity did not depend on the mood
that participants were in. That is, participants in a good mood tended to
counterargue both weak and strong arguments (and thus were equally influenced by both),whereas participants in a bad mood elaborated strong arguments and counterargued weak ones, thus being differently persuaded
by them. The total number of thoughts generated in the two mood conditions was the same in both cases, however. Research on impression formation by Isbell (2000) also suggested that happy and sad individuals do not
differ in the amount of cognitive activity they perform in response to information, but differ only in the nature of this activity.
A third conceptualization is based on the notion that people who are feeling happy are motivated to maintain their mood and, consequently, avoid activities that are likely to bring them down. Unhappy individuals, on the other
hand, are motivated to divest themselves of the negative feelings they are experiencing, and engage in activities that distract themselves from thinking
about these feelings and the events that gave rise to them. Because persua-
344
CHAPTER 11
sive messages are typically counterattitudinal, thinking about them is particularly likely to destroy ones good mood. Consequently, people who feel
happy tend to avoid this cognitive activity and, as a result, are insensitive to
differences in the quality of arguments contained in the messages.
If this interpretation is correct, processing differences should be eliminated or even reversed by leading happy persons to believe that thinking
about the information they receive will be enjoyable and, therefore, will
maintain or enhance their positive mood state. Wegener, Petty, and Smith
(1995) showed this to be the case. Participants listened to one of two communications. One message advocated a position with which participants
were inclined to disagree and was introduced with instructions that it was
likely to make them unhappy. The quality of arguments contained in this
message had less impact on participants attitudes when they were happy
than when they were not. The other message advocated a position that participants typically favored and was introduced with a comment that reading
it was likely to be enjoyable. In this case, the quality of the arguments had
greater effect on participants attitudes when they were happy.
Wegener et al.s (1995) results therefore qualified the conclusion that people who experience positive affect are generally unmotivated to engage in extensive cognitive processing. There are also qualifications on the assumption
that people who experience negative affect are motivated to process new information extensively. Cognitive activity may only help to eliminate negative
affect if the affect is mild and nothing can be done about the events that produced it. Extremely unhappy individuals are likely to ruminate about the
events that created their unhappiness instead of distracting themselves by
engaging in irrelevant activity (Ellis & Ashbrook, 1988; Klinger, 1975; Martin &
Tesser, 1996). To this extent, these individuals may be just as insensitive to
details of affect-irrelevant information as happy people are.
The information processing implied by this interpretation is deliberative.
To this extent, it would not contradict the possibility that affective reactions can stimulate cognitive procedures of the sort noted in the previous
section, which are performed automatically in situations in which they apply. However, the deliberative processes that are suggested by moodmaintenance and mood-repair motives have an alternative interpretation.
That is, these processes could be the byproduct of a more general tendency to use affect as information. We now turn to this possibility.
A PERFORMANCE-FEEDBACK MODEL
OF AFFECT AS INFORMATION
In most research that has been performed to evaluate the use of affect as information, people are assumed to make judgments of a particular person,
object, or event, and the affect they experience at the time of judgment is
345
used as information about their reaction to the stimulus they are judging.
However, peoples affective reactions can provide information about other
things as well. In situations outside the laboratory, people may use their affect as a basis for behavioral decisions. For example, they may decide to engage in a behavior that gives them pleasure, while avoiding activities they
consider to be aversive. In achievement situations, on the other hand, peoples decisions to continue on a task can also be influenced by their feelings
about the quality of their performance. A woman who is writing a term paper, for example, might have little objective basis for deciding whether her
output is sufficient to attain the grade she is seeking. Therefore, her decision to continue working on the paper may be based on her subjective feelings about the work she has done. The extraneous affect she happens to be
experiencing at this time could influence her perceptions of these feelings
and, as a consequence, her decision to stop working or continue.
These possibilities were formalized by Wyer et al. (1999) in a performance-feedback model of affect. This conceptualization, which was stimulated in large part by Martin, Achee, Ward, and Wyers (1993) previous
work, assumes that people spontaneously monitor their behavior and the
outputs it generates to determine if they have satisfactorily attained the
goal they are pursuing. If they conclude that this is the case, they terminate
their goal-directed activity. Otherwise, they continue processing, using the
same or a different strategy, until they generate an output they consider to
be satisfactory.
When objective criteria exist for deciding whether a goal has been attained, the decision to engage in goal-directed activity is obviously based
on these criteria. More often than not, however, the criteria for goal attainment are ambiguous. When this is the case, people may base their decision
on the affect they happen to be experiencing and attribute to their feelings
about their attainment of the goal they are pursuing.
The most interesting implications of this analysis derive from the fact
that peoples perceptions of their affective reactions depend on the particular goal they happen to be pursuing. In the preceding example, the womans
goal was to write a paper of high quality. In this case, she may ask herself if
she feels the paper meets this standard and is more likely to stop work if
her feelings are positive than if they are negative. On the other hand, suppose the woman focuses her attention on whether she is having fun rather
than the quality of the paper she is writing. Then, her feelings are likely to
have precisely the opposite effect. That is, she may be more inclined to
stop writing if her feelings are negative than if they are positive.
Two studies by Martin et al. (1993) provided direct support for these possibilities. In one, happy or sad participants were asked to form an impression of someone on the basis of a series of behaviors. Each behavior was
printed on a separate card of a deck that participants were given. Partici-
346
CHAPTER 11
347
by the experimenter. Under these conditions, participants may spontaneously employ a performance criterion. That is, they may base their decision
to persevere on whether they feel that their responses fulfill these externally imposed task requirements. This hypothesis was supported in a third
condition of Martin et al.s second experiment in which participants were
simply told to stop generating birds when they wanted to. The results in
this condition were very similar to those obtained when participants were
given an explicit performance criterion; that is, happy participants stopped
sooner than unhappy ones. These results confirm the speculation that participants in many experimental situations spontaneously focus their attention on whether they have fulfilled the implicit task objectives they are
given. Nevertheless, when an experimental task is particularly interesting,
or when achievement goals are de-emphasized, an enjoyment criterion
might be applied as well.
Some evidence in support of the latter possibility was reported by
Murray et al. (1990) in a study mentioned earlier in this chapter. Some participants were asked to list similarities between two popular television programs, whereas others were asked to list differences between the programs. The task was intrinsically interesting to college-age participants and,
therefore, was likely to stimulate them to use an enjoyment criterion in deciding how much effort they should expend. To this extent, participants
should use the affect they were experiencing as a basis for inferring their
enjoyment. Consistent with this assumption, happy participants reported
enjoying the task more than control participants did. Correspondingly,
happy participants listed more similarities between the television programs when they were asked to focus on similarities, and listed more differences between the programs when they were focused on differences, than
control participants did. These results would be hard to explain on the basis of either the assumption that positive affect generally increases the accessibility of global concepts in memory. Nor is it consistent with the assumption that positive affect reduces cognitive resources. It seems more
reasonable to suppose that happy participants inferred from the affect they
were experiencing that they were enjoying the task, and therefore devoted
more effort to an identification of both similarities and differences in the
stimuli they were judging than other participants did.
Murray et al.s finding that happy participants reported greater enjoyment of the task than other participants raises an additional consideration.
That is, when several goals are available in a situation, the affect that persons happen to be experiencing could influence the goal they choose to
pursue as well as their inference that they have or have not attained it (for
a discussion of this possibility, see Schwarz & Clore, 1996). For example,
suppose happy individuals are generally more inclined to adopt an enjoyment objective than unhappy individuals are (Isen, 1987). If this is so, happy
348
CHAPTER 11
persons may persist on a task because they infer they are enjoying it,
whereas unhappy persons, who adopt a performance criterion, may persist
because they feel they have not yet attained an adequate level of achievement. To this extent, happy and unhappy persons might persist to an equal
extent, albeit for different reasons.
Despite this ambiguity, the influence of affect on behavior in a large number of situations both in and outside the laboratory can be conceptualized
in terms of its mediating influence on the answers to questions that persons
implicitly ask themselves about how they feel about their behavior and its
outcomes. The broader implications of this conceptualization can be summarized in two postulates:
Postulate 11.3. When people engage in behavior for the purpose of attaining
an external goal and objective criteria for inferring the attainment of this goal
do not exist, they use the affect they are experiencing to infer whether they feel
the strategy they have used in pursuit of this goal, and the outputs they have
generated by employing it, are appropriate. Therefore, they are more likely to
infer they have met the criterion for satisfactory performance if they are experiencing positive affect than if they are not.
Postulate 11.4. When peoples primary reason for engaging in behavior is enjoyment, they use the affect they are experiencing as a basis for inferring this
enjoyment. Therefore, they will continue the activity longer if they are experiencing positive affect than if they are not.
An application of these postulates calls attention to several contingencies in the impact of affect that have not been investigated in the research
performed to date. Several areas of research are worth reviewing from this
perspective.
Creativity
A creativity task often employed in the laboratory requires participants to
generate instances of a concept. This task is similar to that constructed by
Martin et al. (1993, Experiment 2) except that the concept involved has unclear boundaries, thus allowing participants to generate novel responses.
Thus, for example, participants might be asked to generate uses of a brick
or things they could use to sit on. In this situation, like that constructed by
Martin et al., participants might assume they should stop when they have
generated enough good responses to fulfill the demands of the task. However, the task itself might often be intrinsically enjoyable, as it requires a
degree of imaginativeness. Thus, many participants might spontaneously
apply an enjoyment criterion as well. Consistent with this speculation, Isen
and Daubman (1984) found that happy participants generated more re-
349
sponses when performing such a task than control participants did. This
suggests that an enjoyment criterion predominated in their study.
A somewhat different interpretation of Isen and Daubmans results is
suggested by the fact that happy participants also tend to generate more
novel responses than other participants do. When the criterion for an appropriate response is unclear, participants may inhibit reporting responses
they feel are poor exemplars of the concept they are considering. Specifically, they may subjectively generate a response (e.g., a potential use for
a brick), ask themselves if they feel their response is a good one, and only
report it if they answer this question affirmatively. Thus, they are more
likely to record their response if they are happy than if they are not. Evidence that happy persons perform better then others on the Remote Associates Test (Isen, Daubman, & Nowicki, 1987) is consistent with this interpretation. Furthermore, the interpretation could explain the different
results obtained by Isen and Daubmans (1984) study and those reported by
Martin et al. (1993). That is, the criterion for membership in the category
that participants considered in Martin et al.s experiment (birds) was
clearly defined, and so the use of affect to evaluate the appropriateness of
individual exemplars was less likely.
Impression Formation
In impression-formation research (for a review, see Wyer & Carlston, 1994),
participants typically receive information about several behaviors a person
has performed with instructions to form an impression of the person. A large
amount of information is typically presented (e.g., 2530 items; see Wyer &
Martin, 1986), and so it is difficult to form an impression that takes all of it
into account. Consequently, participants in these experiments may often ask
themselves at various points if the information they have received is a sufficient basis for a concept of what the individual is like. If they answer this selfgenerated question affirmatively, they may base their concept of the person
on this information without taking into account the implications of the information they receive later (see Postulate 3.1). If this is so, and if happy participants are more likely than unhappy participants to respond affirmatively to
such a question, they should use relatively less information to form their impressions. Moreover, the information they use should typically occur at the
beginning of the sequence rather than at the end.
Martin et al.s (1993, Experiment 1) results confirmed the assumption
that happy participants with an impression-formation objective use less information to form an impression than sad participants do. The prediction
that happy persons are more likely than unhappy ones to base their impression on the first information they receive was confirmed in research by
Isbell (2000; Isbell, Clore, and Wyer (1998). Participants who were feeling ei-
350
CHAPTER 11
ther happy or sad were asked to form an impression of someone who was
described at the outset as either extraverted or introverted. This general
description was followed by a series of behaviors, an equal number of
which exemplified each of these traits. After receiving the information, participants judged the targets extraversion and then recalled the behaviors
they had received.
Isbell expected that the general description of the target that participants received at the outset would activate a concept of either an extravert
or an introvert, and that participants would implicitly ask themselves if
they felt this concept was an appropriate basis for their impression. Happy
participants, who should answer this question affirmatively, should be
more inclined than unhappy participants to apply the concept to the target
and to base the judgments they report later on this concept. This was in
fact the case, as indicated earlier. That is, the initial target description had a
substantial effect on happy participants judgments of the targets extraversion but had very little impact on unhappy participants judgments.
The fact that happy participants based their judgments on the initial description of the target does not mean that they failed to think about the behavioral information they received later. To the contrary, persons who
apply this concept to the target should think about the behavioral information they receive with reference to it (Srull & Wyer, 1989). Moreover, they
should think more extensively about behaviors that are inconsistent with
the concept in order to reconcile their occurrence (Hastie, 1980; Srull &
Wyer, 1989; Wyer & Srull, 1989). As implied by this reasoning, these participants recalled more inconsistent behaviors than consistent ones. Unhappy participants, who did not apply the concept activated by the initial
trait description to the target, did not engage in this inconsistency resolution. Consequently, they recalled behaviors equally well regardless of their
consistency with the initial trait description.
In many cases, peoples initial impressions of someone are based largely
on their feelings toward this person. This is particularly likely if their first
encounter with the person elicits a spontaneous appraisal on the basis of
his or her physical appearance (cf. Yeung & Wyer, in press). In such cases,
the affect they experience at the time they judge the person could potentially have a dual function. First, it may convey judgment-relevant information about the individual being judged (Schwarz & Clore, 1983). Second, it
could be used as an indication of whether this affect is a sufficient basis for
making the judgment. Thus, people who experience positive affect at the
time they are asked to judge someone might not only infer they feel favorably toward the target, but also that these feelings are a sufficient basis for
the judgment they are asked to make. In contrast, people who experience
negative affect may infer they feel negatively toward the target but also
may conclude that these feelings are not a sufficient basis for the judgment.
351
This implies that the effects of positive affect on impression judgments may
often be greater than the effects of negative effect.
There is a contingency in these predictions. That is, the contention that
happy people are inclined to decide that the initial information they receive
is a sufficient basis for judgment assumes that the implications of this initial
information, considered in isolation, are unambiguous. If these implications
are mixed or unclear, they may not be considered sufficient bases for judgments by either happy or unhappy individuals. A recent study by Isbell
(2003) bears on this contingency. She found that happy participants were
more likely than unhappy participants to base their evaluations of the target on a general trait characterization of him even when this characterization was presented after the targets behaviors were described. In this
study, however, the behavior descriptions were both favorable and unfavorable, and so a clear initial impression of the target was difficult to form
on the basis of these descriptions alone. Therefore, participants were unlikely to consider the behaviors to be a sufficient basis for judgment regardless of their affective state. The greater use of categorical information by
happy participants under these conditions could reflect their generally
greater tendency to apply broad categories in interpreting information, as
suggested by Adaval (2003) and Bless et al. (1996) and noted earlier.
Stereotyping
Our interpretation of Isbell et al.s (1998) findings also has implications for
the effects of affect on the use of stereotypes. In a study by Bodenhausen
(1993; see also Bodenhausen et al., 1994), for example, happy and sad participants were asked to judge the culpability of a defendant on the basis of information about the defendants background and the conditions surrounding the crime. The description of the crime was preceded by the targets
name and hometown, which either identified him as Hispanic or provided
no clue as to his ethnicity. It seems reasonable to suppose that the defendants ethnicity spontaneously activated a stereotype, and that participants
implicitly asked themselves if they felt this was an appropriate basis for
their judgment, answering affirmatively if they were happy and negatively if
they were sad. Consistent with this reasoning, the stereotype had more impact on happy participants judgments than on unhappy participants. Note
that if the stereotype had not been activated until after the detailed crimerelated information was presented, participants would not be in a position
to ask themselves at the outset whether the stereotype was a sufficient basis for judgment. If anything, they should ask themselves if they felt the
crime-related information was sufficient. To this latter extent, happy participants in this condition might be more influenced by the crime-related infor-
352
CHAPTER 11
mation, and less influenced by the stereotype, than sad participants are. To
my knowledge, this possibility has not been examined.
Self-Evaluation Processes
The determinants and effects of self-judgments have been a major focus of
social-psychological research for many years (cf. Greenwald & Pratkanis,
1984; Klein & Loftus, 1993; McGuire & McGuire, 1988). Because of the general
interest in these matters, the implications of the performance-feedback
model for self-judgment processes are worth considering in some detail.
The influence of affect on four different types of self-judgmentsjudgments
of global life satisfaction, self-worth, specific trait judgments, and estimates
of personal riskare therefore considered from the perspective of this
model.
Several alternative cognitive processes could underlie these judgments,
each of which can involve the use of a different type of self-knowledge. The
effects of these procedures can be conceptualized in terms of the general
processing model outlined at the beginning of this volume. Specifically,
when a self-judgment is required, the goal specification may resonate with a
number of different cognitive procedures for generating this judgment.
These processes could include (a) the search for a previously formed judgment-relevant trait description, (b) the identification of a judgment-related
past experience and a construal of its implications for the judgment to be
made, and (c) an assessment of the implications of the affective reactions
one is experiencing. Situational or individual difference factors that make
one process easier to use than another will increase the likelihood that this
process is applied. An additional consideration, however, is implied by the
performance-feedback conceptualization. That is, participants who have
generated a response on the basis of the winning procedure may spontaneously ask themselves if they feel that this response is a sufficient basis
for judgment. participants who are experiencing positive affect may answer
this question affirmatively, whereas those who are experiencing negative
affect may not. Therefore, happy persons are more likely than unhappy persons to accept and apply the first criterion they are able to generate. The
following discussion is based on this assumption.
1. Judgments of Life Satisfaction
A request to judge ones life satisfaction could activate three different
processes. First, respondents might search for a previously formed concept
of their life as a whole that is directly applicable to the judgment. Second,
they might recall one or more previous life events that are relevant to the
judgment and base their response on the affect elicited by these memories.
Finally, they might assess the affect they happen to be experiencing at the
353
FIG. 11.2. Processing stages involved in making (A) global judgments of life
satisfaction and self-esteem, and (b) specific trait judgments. Based on Wyer et
al. (1999).
time and use this as an indication of their feelings about life in general without conducting a search of memory at all. These processes, which could occur simultaneously, are summarized in Fig. 11.2a.
On a priori grounds, it is unclear which of these processes will predominate. For example, if participants in an experiment have a previously
354
CHAPTER 11
formed concept of their life as a whole, a request to estimate their life satisfaction might spontaneously activate this concept and so a judgment can
be quickly computed on the basis of the affect it elicits. On the other hand,
people are infrequently called upon to make this type of judgment in daily
life. Consequently, they usually do not have a previously formed concept of
their life as a whole stored in memory, and must compute a judgment at the
time they are asked to report it. In this case, it would normally take less
time to compute a judgment on the basis of the transitory affect one is experiencing at the time of judgment than to do so on the basis of other, less
easily accessible criteria, and so judgments are more likely to be based on
the former computation. Evidence that people use transitory affect as a basis for life satisfaction judgments independently of any specific self-knowledge they have available (Schwarz & Clore, 1983; Strack, Schwarz, &
Gschneidinger, 1985) is consistent with this hypothesis.
However, the performance-feedback model we have proposed raises a
further consideration. People who experience positive affect for reasons
that are objectively unrelated to their life situation might not only use these
feelings as a basis for judgment but also consider this criterion to be sufficient. In contrast, people who experience negative affect might be disposed
to question the sufficiency of this criterion. Therefore, they might either
seek other judgment-relevant information or, if no alternative criteria are
readily available, might adjust for the bias that results from their use of feelings as information.
Schwarz and Clores (1983, Experiment 1) findings support this conjecture. To reiterate, happy persons appeared to use their feelings as a basis
for judging their life satisfaction regardless of the situational factors that
had implications for their validity as a basis for judgment. In contrast, unhappy participants adjusted their judgments to compensate for the effects
of these factors. Perhaps both happy and unhappy participants spontaneously considered the affect they were experiencing as a potential basis for
judgment. However, whereas happy participants concluded that these feelings were a sufficient basis for judging without thinking about situational
factors that might have influenced them, unhappy participants considered
that an assessment of their feelings per se was insufficient and took situational factors into account as well.
2. Judgments of Self-Worth
Estimates of self-worth, like estimates of life satisfaction, may typically be based on feelings. These feelings are likely to be estimated by
either (a) retrieving a previously formed concept of oneself with which affective reactions have become associated through learning, (b) recalling
descriptive self-knowledge to which affective reactions have become condi-
355
The retrieval and use of this molecule, however, is likely to take longer than
computations based on descriptive criteria. This is particularly true if a
trait-relevant experience is easily accessible in memory at the time of judg-
356
CHAPTER 11
ment. Levine et al.s (1994) findings, described earlier, are consistent with
this hypothesis. That is, the descriptive implications of participants recalled life experiences were used as a basis for judgments of attributes to
which they were relevant. The affective implications of recalled experiences only influenced judgments of attributes in a domain to which the experiences were unrelated.
The implications of the performance-feedback model become of interest
in this context. If trait judgments are computed more quickly on the basis of
descriptive criteria than on the basis of affective criteria, people who experience positive affect should be likely to infer that these criteria are sufficient, whereas people who experience negative effect are likely to consider
additional (e.g., affective) criteria. Thus, in contrast to judgments of life satisfaction and self-worth, trait inferences are more likely to be influenced by
negative affect than by positive affect.
Communication and Persuasion
Suppose people read a persuasive communication for the purpose of deciding if the position it advocates is valid. They are likely to have an initial impression of the positions validity position before they evaluate the communications contents. This impression could be based on the source of the
message, on surface features of the communication itself (e.g., writing
style), or on a previously formed attitude toward the position in question. If
people who have made formed this impression ask themselves if it is an appropriate basis for accepting the position advocated, they should be more
inclined to answer this question affirmatively if they are happy than if they
are sad. This has two implications. First, happy persons are more likely
than unhappy persons to rely on the source of a message as a basis for accepting the position advocated. Second, happy persons are less inclined
than sad persons to make a careful evaluation of the arguments contained
in a message. Consequently, their acceptance of the position advocated is
less likely to be influenced by the quality of these arguments. Support for
the first hypothesis has been obtained by Roselli (1995), and the second hypothesis is consistent with results of several studies described previously
(Bless et al., 1990; Mackie &Worth, 1989).
A performance-feedback conceptualization of these results suggests two
contingencies, however. First, the different cognitive activities performed
by happy and unhappy persons should only occur if these persons perceive their feelings to have implications for the validity of their initial impression. If participants attribute their feelings to other, impression-irrelevant factors, the difference in processing should theoretically not occur.
Results reported by Sinclair, Mark, and Clore (1994) confirmed this prediction. They found that the effect of argument strength on participants accep-
357
358
CHAPTER 11
text. In three studies, participants were asked to indicate how they would
phrase a verbal request in a hypothetical social interaction. It seems reasonable to suppose that people who want another person to do something
(e.g., to repay a loan) will spontaneously generate a verbal phrasing of the
request that corresponds directly to the desired action (e.g., Give me the
money you owe me) and may implicitly ask themselves if it is appropriate
to express the request in these terms. If they answer this question positively, they are likely to alter the phrasing in a way that corresponds to normative standards of politeness. Otherwise, they might not. If this is so, the
performance-feedback conceptualization we propose suggests that happy
persons are more likely to express their requests impolitely than unhappy
persons are. This, in fact, is what Forgas (1998) found. Although his studies
were performed in the laboratory, they seem likely to generalize to other
situations as well.
It would nevertheless be hazardous to conclude that happy persons are
generally less likely to engage in extensive information processing than unhappy persons are. This difference should theoretically exist only when individuals have an implicit performance objective. In nonlaboratory situations, information processing is often performed for enjoyment. In these
circumstances, the performance-feedback implies that happy persons will
persevere longer, and will process information more analytically, than sad
persons will. Thus, for example, happy individuals might work harder than
sad individuals on crossword puzzles, and may read news articles in relatively greater depth. Moreover, they might make less hasty decisions about
other persons on the basis of their initial impressions of these persons, in
contrast to results observed in the laboratory (Isbell et al., 1998). The possibilities, which are not implied by other conceptualizations of the influence
of affect on performance, should be kept in mind when drawing conclusions
on the basis of experimental research in which performance objectives are
implicitly emphasized.
359
[11.1]
Where bi is the belief that the behavior will have a consequence i, and ei is
the evaluation of this consequence. (Not all possible consequences of the
behavior are taken into account, of course, but only a subset that is most
360
CHAPTER 11
salient at the time the attitude is formed.) To this extent, a communication should be effective in changing an attitude if it either increases beliefs
that the behavior has consequences that the recipient considers to be desirable or, alternatively, increases the evaluation of consequences that the
recipient already believes are likely to occur.
Affective reactions are likely to play a quite different role in the processing of attitude-focused communications than in the processing of belieffocused messages. Moreover, this role may depend in part on whether the
attitude in question is based on hedonic or utilitarian criteria (Zanna &
Rempel, 1988). When the criterion is hedonic, the role of affect could depend on whether (a) recipients think sufficiently about the consequences
described in a message for the affect associated with these consequences
to be elicited, (b) whether they can effectively combine the implications of
these consequences into a single attitudinal response in the manner specified in Equation 11.1, and (c) whether they can distinguish this affect from
the feelings they are experiencing for other, unrelated reasons. Once an attitude is formed in this manner it may be reactivated at the time the course
of action is actually contemplated (see Fazio, 1986, for a discussion of the
factors that influence this activation).
As noted earlier, however, the specific consequences of the position advocated in a communication may not be the only determinants of the attitude that recipients report. Affect could also be stimulated by the intrinsic
pleasantness or unpleasantness of the behavior being advocated, independently of any specific information about it. (The thought of taking comprehensive examinations, for example, might spontaneously elicit negative
affect independently of any specific consequences the activity might have.)
Finally, extraneous affect that recipients happen to be experiencing may be
misattributed to their feelings about this activity. Affect from all of these
sources could contribute to peoples perceptions of their attitude toward
the behavior being advocated.
Experimental Evidence
The role of affect in responses to attitude-focused persuasive messages was
evaluated in a series of studies by Albarracin and Wyer (2001). We told participants we were concerned with how people process information in natural settings, and that to do this, we would like them both to transmit and receive information in a simulated coffee shop situation. On this pretense, we
induced them to feel happy or unhappy by asking them to write a letter to a
friend describing a positive or negative life experience. Then, we asked participants to read a communication, ostensibly taken from a local newspaper, which summarized either strong or weak arguments in support of instituting comprehensive examinations and urging recipients to vote in favor
361
362
CHAPTER 11
TABLE 11.5
Effects of Argument Strength and Extraneous Affect on Behavioral
Intentions, Attitudes, and Beliefs and Evaluations of Specific
Behavioral Outcomes as a Function of Distraction
Outcome
Beliefs (bi )
Outcome
Evaluations (ei )
Attitudes
(Ab)
Behavior
Intentions
2.10a
2.08
<1
3.22
4.37
2.23
0.62
2.61
5.82*
1.77
3.91
3.69*
0.75b
0.92
4.87*
1.48
0.52
6.80*
1.79
0.73
9.39*
2.09
0.60
5.79*
traneous affect that participants were experiencing should have greater effect. This was also the case, as shown in the bottom half of Table 11.5. When
distraction was high, extraneous affect had an positive influence on not
only participants attitudes and intentions but also their estimates of the
likelihood and desirability of individual consequences. When distraction
was low, however, extraneous affect had slight contrast effects on these
measures.
3. Reciprocal Effects of Attitudes
on Outcome Beliefs and Evaluations
When participants were not distracted, they presumably computed beliefs and evaluations associated with each consequence described in the
persuasive message and then integrated these estimates to form an attitude
in the manner implied by Equation 11.1. To this extent, their outcome beliefs and evaluations should have multiplicative effects on their attitudes.
Under high-distraction conditions, however, participants presumably based
their attitudes on the extraneous affect they were experiencing rather than
the content of the persuasive message. Having done so, they may have attempted to justify the affect-based attitude they have formed using processes suggested by McGuire and McGuire (1991; see also Janis & King,
1954). For example, they may engage in rationalization (evaluating the desirability of the consequences in a manner that is evaluatively consistent
with their attitude) and wishful thinking (increasing their beliefs that these
363
FINAL COMMENTS
As noted at the outset, peoples affective reactions can come into play in
judgments and behavioral decisions in virtually every domain of social experience, both in and outside the laboratory. The inordinate length of this
chapter testifies to the diversity of its possible effects and the conditions in
which they occur. We have argued, however, that affective reactions per se
exert their influence primarily through the information they provide about
persons, objects, and events to which these reactions are directed, and the
likelihood that ones information-processing objectives have been satisfactorily attained.
The performance-feedback conceptualization suggested by Martin et
al.s (1993) findings can account for many effects of affect on judgments and
behavior. It therefore seems worthy of continued consideration in future research. Its ultimately utility is nonetheless something of a promissory note.
That is, a rigorous application of the conceptualization requires a task analysis of the cognitive processing steps that underlie a judgment or behav-
364
CHAPTER 11
ioral decision and the mental operations that are performed at each step.
An example of this analysis was provided in our analysis of the impact of affect on self-judgments. Similar analyses of other information processing situations may be fruitful.
Our analysis of the influence of affect on judgments and behavior has
largely been independent of our discussion in previous chapters. As noted,
however, the stories that people write about themselves, as well as stories
they encounter in conversations or in the media, are very likely to elicit affective reactions. To this extent, many of the inference process that are
based on narrative representations of knowledge (see chaps. 9 and 10)
could be based in large part on affect rather than the substantive implications of the information per se. The impact of pictures and visual images on
narrative-based processing could likewise be mediated by their influences
on peoples affective reactions and the use of these reactions as bases for
judgment. Therefore, although we have not explicated these possibilities in
this chapter, the relevance of the phenomena we have described to issues
raised elsewhere in this volume is hopefully apparent.
C H A P T E R
12
Epilogue: The Book, the Author,
and Philosophical Ruminations
The Book
The information we acquire in daily life can come from many sources: direct observation, conversations with friends and acquaintances, movies
and television, newspapers, and books. Some of this information is about
unknown or fictitious people and events. In many cases, however, we already know a lot about the things that are described, and we spontaneously recognize the information as either true or false. When information is
conveyed in pictures or is acquired through observation, we often form visual images and retain them in memory as part of our knowledge about
their referents. However, verbal descriptions of events can stimulate the
construction of visual images as well. Finally, the information we receive
can often elicit affective reactions both at the time it is acquired and when
we are later reminded of the events to which it refers, and these reactions
can be used as a basis for evaluating the events and the people or objects
involved in them. The information from all of these sources, and in all of
these modalities, combines to form our accumulated knowledge about the
world in which we live, and serves as a basis for our judgments and behavioral decisions.
This volume has provided a conceptualization of the way in which this
information is comprehended and stored in memory and how it is later retrieved and used. Much of our discussion has focused on the construction
and use of narratives (that is, temporally and thematically organized sequences of events that are stored in memory as a unit). These mental representations include episode models of specific events that we happen to en365
366
CHAPTER 12
EPILOGUE
367
more circumscribed sets of phenomena can often be viewed as components of this formulation. (That is, they specify the operations that are performed by one or another special purpose processing unit and represented
either in a goal schema or in a processing units library.) In this regard,
many recent conceptualizations of social cognition have recognized that
different cognitive processes can underlie responses to information (e.g.,
see Chaiken & Trope, 1999). However, the assumption that several alternative processes can be brought to bear on a judgment or decision, depending on social, informational, and individual difference variables that determine their applicability and relative accessibility, is also a central feature of
both the Wyer and Srull (1989) formulation and the present one.
The proposed conceptualization is not without flaws. For one thing, it is
metaphorical, and does not pretend to be an exact depiction of the physiology of the brain. As noted in chapter 1, however, most formulations of information processing are inherently metaphorical, and should be judged on
the basis of their utility and not their validity. The present formulation appears to be a useful one, and establishes directions for future research in
social information processing. If this book stimulates even a few researchers to pursue these directions, its objectives will be accomplished.
The Author
In the concluding remarks of his recent book, Constructing social psychology,
Bill McGuire (1999) commented: It may be best if each of us . . . makes his
or her own way . . . but ones own way should be taken deliberatively and
egosyntonically, not stumbled into by chance (p. 432). I wish that I could
say that my own career exemplifies this sage advice. I have, undoubtedly,
made my own way. But this way was hardly deliberative. Even my decision
to become a psychologist was determined by largely fortuitous events. Although my reflections about these events is perhaps less out of place in a
volume that emphasizes the role of narratives than it might otherwise be,
they are likely to be of little interest to anyone but myself. At the risk of
abusing a privilege, however, I would like to take advantage of this opportunity to indulge myself in these reflections.
When I was in high school, I wanted to be either a jazz musician or a
sportswriter. I was discouraged from both pursuits by my father. On one
hand, he pointed out that my skills as a clarinet and saxophone player
would probably qualify me for no better than third chair in a mediocre
dance band. He further argued that to be a sportswriter, I would inevitably
have to start off as a copy boy at a city newspaper and would have to work
my way up gradually over a period of many years. Moreover, he claimed
that this would be true regardless of whether I went to college or not.
Therefore, he argued that if I decided to go to college at all, I should major
368
CHAPTER 12
in something more useful. So, having had my primary options dashed, and
having received a scholarship to Rensselaer Polytechnic Institute, I became
an electrical engineer.
Four years in engineering school did little to stimulate my interest in the
field. Nevertheless, I took a job at Bell Laboratories, which offered me full
salary for going to school 3 days a week and working only 2. Although winding up 2 years later with a Masters degree, I spent most of my time improving my table tennis skills and imbibing in the intellectual atmosphere that
pervaded Greenwich Village in the late 1950s. These activities convinced
me that I should get a liberal arts education.
I took a battery of ability and interest tests, which indicated that the only
things for which I was suited were architecture and psychology. I hardly
knew what psychology was, having barely escaped a D in my single undergraduate course in the area. I nevertheless decided to apply to graduate
schools in psychology in physical locations I might enjoy, without knowing
anything at all about the schools themselves, and not really thinking that I
might actually go. Inevitably, one school that met my enjoyment criterion
was the University of Colorado. The application form, however, listed three
areas that were completely foreign to me: experimental (which sounded
like engineering), clinical (which sounded like medicine) and socialpersonality. Because I thought I understood what social and personality
meant, I checked the third area as my preference. By chance, this was the
one area at Colorado in which Public Health Service fellowships were available, and I was offered one. My bluff being called, I accepted it, knowing
nothing at all about what I was getting into and not really caring, but assuming that I would have the opportunity to expose myself to liberal arts in general.
This was in fact the case. The graduate program at Colorado in 1960 was
hardly rigorous, and I was able to get through the program in 3 years with a
total of only 27 hours of psychology (most of which was for thesis and independent study). I actually audited more courses in humanities, sociology,
history, and philosophy than I took in psychology. What Colorado lacked in
substance, however, it made up for in stimulation, due largely to OJ Harvey
(one of the most inspirational individuals I had ever met) and Bill Scott (my
dissertation advisor, who gave me an appreciation for the need to think rigorously as well as innovatively). In general, Colorados reactions to my
graduate school career were probably exemplified by one faculty members
prediction that unless I was always right, I would never succeed as a psychologist because I never listened to anything that anyone else was saying.
I obviously havent always been right. Whether that reflects a change in my
chronic disposition, however, is debatable.
I spent an additional year at Colorado in 1962 with Glenn Terrell, a developmental psychologist and Associate Dean, on research on nonintellective
EPILOGUE
369
370
CHAPTER 12
EPILOGUE
371
372
CHAPTER 12
EPILOGUE
373
continue to be profoundly influenced by their work. Thom Srull, first as student and then as colleague, was a major collaborator for many years. Other
students whose contributions to my career are particularly noteworthy include Galen Bodenhausen, Lee Budesheim, Bob Fuhrman, Alan Lambert,
Victor Ottati, and Dave Trafimow. Dolores Albarracin, Deborah Gruenfeld,
Linda Isbell, and L. J. Shrum were also major collaborators. Their research
on communication and persuasion, the pragmatic implications of information, affect, and perceptions of social reality, respectively, has had an obvious impact on my own. Rashmi Adaval, Stan Colcombe, and Eric Mankowski, in quite different ways, stimulated my interest in the role of
narratives in information processing. Most of these individuals have been
close friends as well as stimulating colleagues, making our collaboration
doubly rewarding.
It has been a long trip, and the journey is not quite ended. However,
these individuals and others too numerous to mention have made the trip
exciting and enjoyable beyond belief, and I will always be in their debt.
A P P E N D I X
Summary of Postulates
Res(k) =
n( pI k)
,
n( k)
SUMMARY OF POSTULATES
375
376
APPENDIX
Postulate 4.4. If a memory search activates a previously formed knowledge representation whose features include both the subject probe cues
and the predicate probe cues, the new information is spontaneously recognized as redundant with prior knowledge and a new representation is not
formed.
Postulate 4.5. If no previously formed model (i.e., knowledge unit) refers to both the subject and predicate of a target statement, the features
that are activated by the targets subject are compared to a weighted composite of the features associated with the targets predicate. This comparison yields a global estimate of similarity, S. A new situation model is formed
if either (a) S is greater than a minimum threshold value, TC, or (b) the targets subject contains a subset of features that are common to the subject
of all models activated by the probe. In the latter case, the new model is
constructed by combining the composite of features activated by the subject and the composite of features activated by the predicate. In all cases,
however, only the features that are weighted heavily in the composite are included in the new model that is constructed.
Postulate 4.6. A statement is comprehended spontaneously if it is redundant with prior knowledge or if a new model is constructed of it. If neither is the case, comprehension of the statement is not spontaneous, but
requires (Executor-controlled) goal-directed processing.
SUMMARY OF POSTULATES
377
378
APPENDIX
Postulate 8.6. The amount of humor that is elicited as a result of reinterpreting a stimulus event is a monotonic function of the amount of cognitive elaboration of the event and its implications that occurs subsequent to
its reinterpretation.
1. If recipients processing objective at the time is simply to comprehend
and enjoy experiencing the event, cognitive elaboration of the event will typically increase the humor it elicits.
2. If recipients goal is more restricted, the event is elaborated in terms of
its implications for this more specific objective. In this case, cognitive elaboration could either increase, decrease, or have no effect on the amount of humor elicited, depending on whether the humor-eliciting reinterpretation of
the event is relevant to the attainment of this objective.
Inference Postulate (Completion Principle; chap. 10)
Postulate 10.1. If the information available about a specific situation
instantiates all but one of the propositions that compose a molecule, an
instantiation of the remaining proposition will be spontaneously inferred.
Affect Postulates (chap. 11)
Postulate 11.1. Affective reactions may activate a specific affectrelevant concept that can be used to interpret them. However, affective reactions do not spontaneously influence the accessibility in memory of other
similarly valenced semantic concepts and declarative knowledge.
Postulate 11.2. The affective reactions that one experiences, along with
concepts and cognitions that are stimulated by an appraisal of a stimulus
situation, can compose a precondition for a previously learned sequence of
responses (i.e., a cognitive production). This response sequence may be activated spontaneously, with little conscious cognitive mediation, under conditions in which the precondition is met.
Postulate 11.3. When people engage in behavior for the purpose of attaining an external goal and objective criteria for inferring the attainment of
this goal do not exist, they use the affect they are experiencing to infer
whether they feel that the strategy they have used in pursuit of this goal,
and the outputs they have generated by employing it, are appropriate.
Therefore, they are more likely to infer that they have met the criterion for
satisfactory performance if they are experiencing positive affect than if they
are not.
SUMMARY OF POSTULATES
379
References
Aaker, J. L., & Lee, A. (2001). I seek pleasures, we avoid pains: The role of self-regulatory goals in
information processing and persuasion. Journal of Consumer Research, 27, 3349.
Abelson, R. P. (1959). Modes of resolution of belief dilemmas. Conflict Resolution, 3, 343352.
Abelson, R. P. (1976). Script processing in attitude formation and decision making. In J. S. Carroll
& J. W. Payne (Eds.), Cognition and social behavior (pp. 3346). Hillsdale, NJ: Lawrence
Erlbaum Associates.
Abelson, R. P., Kinder, D. R., Peters, M. D., & Fiske, S. T. (1982). Affective and semantic components in political person perception. Journal of Personality and Social Psychology, 42, 619630.
Abelson, R. P., & Reich, C. M. (1969). Implicational molecules: A method for extracting meaning
from input sentences. In D. E. Walker & L. M. Norton (Eds.), Proceedings of the International
Joint Conference on Artificial Intelligence.
Adaval, R. (2001). Sometimes it just feels right: The differential weighting of affect-consistent and
affect-inconsistent product information. Journal of Consumer Research, 7, 207245.
Adaval, R. (2003). How good gets better and bad gets worse: Understanding the impact of affect
on evaluations of known brands. Journal of Consumer Research.
Adaval, R., Isbell, L. M., & Wyer, R. S. (2003). Political impression formation: The role of pictures and
information presentation format on evaluations of politicians. Unpublished manuscript, Hong
Kong University of Science and Technology.
Adaval, R., & Monroe, K. B. (2002). Automatic construction and use of contextual information for
product and price evaluations. Journal of Consumer Research, 28, 572588.
Adaval, R., & Wyer, R. S. (1998). The role of narratives in consumer information processing. Journal of Consumer Psychology, 7, 207245.
Adaval, R., & Wyer, R. S. (2003). Memory for social interactions: Effects of post-information processing
objectives on the recognition of protagonists statements and nonverbal behaviors. Unpublished
manuscript, Hong Kong University of Science and Technology.
Albarracin, D., & Wyer, R. S. (2000). The cognitive impact of past behavior: Influences on beliefs,
attitudes, and future behavioral decisions. Journal of Personality and Social Psychology, 79,
522.
Albarracin, D., & Wyer, R. S. (2001). Elaborative and nonelaborative processing of a behaviorrelated persuasive communication. Personality and Social Psychology Bulletin, 27, 691705.
380
REFERENCES
381
Albrecht, J. E., & Myers, J. L. (1995). Updating a mental model: Maintaining both local and global
coherence. Journal of Experimental Psychology: Learning, Memory, and Cognition, 19,
10611070.
Anderson, A., Garrod, S. C., & Sanford, A. J. (1983). The accessibility of pronominal antecedents
as a function of episode shifts in narrative text. Quarterly Journal of Experimental Psychology,
35A, 427440.
Anderson, J. R. (1978). Arguments concerning representations for mental imagery. Psychological
Review, 85, 249277.
Anderson, J. R. (1982). Acquisition of cognitive skill. Psychological Review, 89, 369406.
Anderson, J. R. (1983). The architecture of cognition. Cambridge, MA: Harvard University Press.
Anderson, J. R., & Bower, G. H. (1973). Human associative memory. Washington, DC: V. H. Winston.
Anderson, N. H. (1971). Integration theory and attitude change. Psychological Review, 78, 171206.
Anderson, N. H. (1981). Foundations of information integration theory. New York: Academic Press.
Anderson, R. C., Reynolds, R. E., Schallert, D. L., & Goetz, F. T. (1977). Frameworks for comprehending discourse (Tech. Rep. No. 12). Urbana, IL: Laboratory for Cognitive Studies in Education,
University of Illinois at Urbana-Champaign.
Apter, M. J. (1982). The experience of motivation: The theory of psychological reversals. San Diego,
CA: Academic Press.
Arkin, R. M., Gleason, J. M., & Johnston, S. (1976). Effect of perceived choice, expected outcome
and observed outcome of an action on the causal attributions of actors. Journal of Experimental Social Psychology, 12, 151158.
Banks, W. P. (1977). Encoding and processing of symbolic information in comparative judgments. In G. H. Bower (Ed.), The psychology of learning and motivation (Vol. 11). New York: Academic Press.
Barclay, J. R. (1973). The role of comprehension in remembering sentences. Cognitive Psychology,
4, 229254.
Bargh, J. A. (1994). The four horsemen of automaticity: Awareness, intention, efficiency, and control in social cognition. In R. S. Wyer & T. K. Srull (Eds.), Handbook of social cognition (2nd ed.,
Vol. 1, pp. 140). Hillsdale, NJ: Lawrence Erlbaum Associates.
Bargh, J. A. (1997). The automaticity of everyday life. In R. S. Wyer (Ed.), Advances in social cognition (Vol. 10, pp. 162). Mahwah, NJ: Lawrence Erlbaum Associates.
Bargh, J. A., Bond, R. N., Lombardi, W., & Tota, M. E. (1986). The additive nature of chronic and
temporary sources of construct accessibility. Journal of Personality and Social Psychology, 50,
869878.
Bargh, J. A., Chaiken, S., Govender, R., & Pratto, F. (1992). The generality of the automatic attitude activation effect. Journal of Personality and Social Psychology, 62, 893912.
Bargh, J. A., Chaiken, S., Raymond, P., & Hymes, C. (1996). The automatic evaluation effect: Unconditionally automatic attitude activation in a pronunciation task. Journal of Experimental
Social Psychology, 32, 104120.
Bargh, J. A., Chen, M., & Burrows, L. (1996). Automaticity of social behavior: Direct effects of trait
construct and stereotype activation on action. Journal of Personality and Social Psychology,
71, 230244.
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, 437449.
Baron, R. A., & Byrne, D. (2000). Social psychology (9th ed.). Boston, MA: Allyn & Bacon.
Barsalou, L. W. (1990). On the indistinguishability of exemplar memory and abstraction in category representations. In T. K. Srull & R. S. Wyer (Eds.), Content and process specificity in the effects of prior experiences: Advances in social cognition (Vol. 3, pp. 6188). Hillsdale, NJ: Lawrence Erlbaum Associates.
Barsalou, L. W. (1993). Flexibility, structure, and linguistic vagary in concepts: Manifestations of
a compositional system of perceptual symbols. In A. F. Collins, S. E. Gathercole, M. A.
382
REFERENCES
Conway, & P. E. Morris (Eds.), Theories of memory (pp. 29102). Hillsdale, NJ: Lawrence
Erlbaum Associates.
Baumeister, R. F., & Newman, L. S. (1994). How stories make sense of personal experience: Motives that shape autobiographical memories. Personality and Social Psychology Bulletin, 20,
676690.
Bell, D. E. (1982). Regret in decision making under uncertainty. Operations Research, 30, 961981.
Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in experimental social
psychology (Vol. 6, pp. 162). New York: Academic Press.
Bem, D. J., & McConnell, H. K. (1970). Testing the self-perception explanation of dissonance phenomena: On the salience of premanipulation attitudes. Journal of Personality and Social Psychology, 14, 2331.
Bergson, H. (1911). Laughter: An essay on the meaning of the comic. New York: Macmillan.
Berlyne, D. E. (1969). Laughter, humor and play. In G. Lindzey & E. Aronson (Eds.), Handbook of
social psychology (Vol. 3, pp. 795852). Reading, MA: Addison-Wesley.
Berlyne, D. E. (1971). Aesthetics and psychology. New York: Appleton-Century.
Bettman, J. R., Capon, N., & Lutz, R. J. (1975). Multiattribute measurement models and
multiattribute attitude theory: A test of construct validity. Journal of Consumer Research, 1,
115.
Birnbaum, M. H. (1974). The nonadditivity of personality impressions. Journal of Experimental
Psychology, 102, 543561.
Black, J. B., Galambos, J. A., & Read, S. (1984). Comprehending stories and social situations. In
R. S. Wyer & T. K. Srull (Eds.), Handbook of social cognition (Vol. 3, pp. 4586). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Black, J. B., Turner, T., & Bower, G. H. (1979). Point of view in narrative comprehension, memory,
and production. Journal of Verbal Learning and Verbal Behavior, 11, 717726.
Blaney, P. H. (1986). Affect and memory: A review. Psychological Bulletin, 99, 229246.
Bless, H. (2001). Mood and the use of general knowledge structures. In L. L. Martin & G. L. Clore
(Eds.), Theories of mood and cognition: A users guidebook (pp. 926). Mahwah, NJ: Lawrence
Erlbaum Associates.
Bless, H., Bohner, G., Schwarz, H., & Strack, F. (1990). Mood and persuasion: A cognitive response analysis. Personality and Social Psychology Bulletin, 16, 331345.
Bless, H., Clore, G. L., Schwarz, N., Golisano, V., Rabe, C., & Woelke, M. (1996). Mood and the use
of scripts: Does being in a happy mood really lead to mindlessness? Journal of Personality and
Social Psychology, 71, 665679.
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: Interactive processes in group perception (pp. 1337). San Diego, CA: Academic Press.
Bodenhausen, G. V., & Macrae, C. N. (1998). Stereotype activation and inhibition. In R. S. Wyer
(Ed.), Advances in social cognition (Vol. 11, pp. 152). Mahwah, NJ: Lawrence Erlbaum Associates.
Bodenhausen, G. V., Sheppard, L. A., & Kramer, G. P. (1994). Negative affect and social judgment:
The differential impact of anger and sadness. European Journal of Social Psychology, 24, 4562.
Bodenhausen, G. V., & Wyer, R. S. (1985). Effects of stereotypes on decision-making and information-processing strategies. Journal of Personality and Social Psychology, 52, 871880.
Boorstin, D. (1973). The Americans: The national experience. New York: Random House.
Bower, G. H. (1981). Emotional mood and memory. American Psychologist, 36, 129148.
Bower, G. H. (1991). Mood congruity of social judgments. In J. P. Forgas (Ed.), Emotion and social
judgments (pp. 3153). Oxford, England: Pergamon.
Bower, G. H., Black, J. B., & Turner, T. J. (1979). Scripts in memory for texts. Cognitive Psychology,
11, 177220.
Bower, G. H., Gilligan, S. G., & Monteiro, K. P. (1981). Selectivity of learning caused by affective
states. Journal of Experimental Psychology: General, 110, 451473.
REFERENCES
383
Bower, G. H., & Mayer, J. D. (1985). Failure to replicate mood-dependent retrieval. Bulletin of the
Psychonomic Society, 23, 3942.
Bransford, J. D., Barclay, J. R., & Franks, J. J. (1972). Sentence memory: A constructive versus interpretative approach. Cognitive Psychology, 3, 193209.
Bransford, J. D., & Johnson, M. K. (1972). Contextual prerequisites for understanding: Some investigations of comprehension and recall. Journal of Verbal Learning and Verbal Behavior, 11,
717726.
Bransford, J. D., & Stein, B. S. (1984). The ideal problem solver: A guide to improving thinking, learning, and creativity. New York: Freeman.
Brewer, M. (1988). A dual process model of impression formation. In T. K. Srull & R. S. Wyer
(Eds.), Advances in social cognition: A dual process model of impression formation (Vol. 1, pp.
136). Hillsdale, NJ: Lawrence Erlbaum Associates.
Brewer, W. F. (1995). To assert that essentially all human knowledge and memory is represented
in terms of stories is certainly wrong. In R. S. Wyer (Ed.), Advances in social cognition (Vol. 8,
pp. 109119). Hillsdale, NJ: Lawrence Erlbaum Associates.
Brewer, W. H., & Lichtenstein, E. H. (1981). Event schemas, story schemas, and story grammars.
In J. Long & A. Baddeley (Eds.), Attention and Performance, IX (pp. 363379). Hillsdale, NJ: Lawrence Erlbaum Associates.
Briley, D. A., Morris, M., & Simonson, I. (2000). Reasons as carriers of culture: Dynamic versus
dispositional models of cultural influence on decision making. Journal of Consumer Research,
27, 157178.
Briley, D. A., & Wyer, R. S. (2001). Transitory determinants of values and decisions: The utility (or
nonutility) of individualism and collectivism in understanding cultural differences. Social
Cognition, 19, 198229.
Briley, D. A., & Wyer, R. S. (2002). The effect of group membership salience on the avoidance of
negative outcomes: Implications for social and consumer decisions. Journal of Consumer Research, 29, 400415.
Brown, N. R., Rips, L. J., & Shevell, S. K. (1985). The subjective dates of natural events in verylong-term memory. Cognitive Psychology, 17, 139177.
Byrne, D. (1971). The attraction paradigm. New York: Academic Press.
Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality and Social Psychology, 42, 116131.
Carlston, D. E. (1980). Events, inferences and impression formation. In R. Hastie, T. Ostrom, E.
Ebbesen, R. Wyer, D. Hamilton, & D. Carlston (Eds.), Person memory: The cognitive basis of social perception (pp. 89119). Hillsdale, NJ: Lawrence Erlbaum Associates.
Carlston, D. E. (1994). Associated Systems Theory: A systematic approach to cognitive representations of persons. In R. S. Wyer (Ed.), Advances in social cognition (Vol. 7, pp. 178). Hillsdale,
NJ: Lawrence Erlbaum Associates.
Carlston, D. E., & Smith, E. R. (1996). Principles of mental representation. In E. T. Higgins & A. W.
Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 184210). New York:
Guilford Press.
Ceci, S. J., & Bruck, M. (1993). Suggestibility of the child witness: A historical review and synthesis. Psychological Bulletin, 113, 403439.
Chaiken, S. (1980). Heuristic versus systematic information processing in the use of source versus message cues in persuasion. Journal of Personality and Social Psychology, 39, 752766.
Chaiken, S. (1987). The heuristic model of persuasion. In M. P. Zanna, J. M. Olson, & C. P. Herman
(Eds.), Social influence: The Ontario Symposium (Vol. 5, pp. 339). Hillsdale, NJ: Lawrence
Erlbaum Associates.
Chaiken, S., & Bargh, J. A. (1993). Occurrence versus moderation of the automatic attitude activation effect: Reply to Fazio. Journal of Personality and Social Psychology, 64, 759765.
Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic information processing
within and beyond the persuasion context. In J. S. Uleman & J. A. Bargh (Eds.), Unintended
thought (pp. 212252). New York: Guilford Press.
384
REFERENCES
Chaiken, S., & Trope, Y. (1999). Dual-process theories in social psychology. New York: Guilford
Press.
Chaiken, S., Wood, W., & Eagly, A. H. (1996). Principles of persuasion. In E. T. Higgins & A. W.
Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 702742). New York:
Guilford Press.
Chapman, A. J., & Foot, H. C. (Eds.). (1976). Humor and laughter: Theory, research and application.
New York: Wiley.
Chapman, A. J., & Foot, H. C. (Eds.). (1977). Its a funny thing, humor. Elmsford, NY: Pergamon
Press.
Chartrand, T. L., & Bargh, J. A. (1996). Automatic activation of impression formation and memorization goals: Nonconscious goal priming reproduces effects of explicit task instructions.
Journal of Personality and Social Psychology, 71, 464478.
Chartrand, T. L., & Bargh, J. A. (2002). Nonconscious motivations: Their activation, operation
and consequences. In A. Tesser & D. Stapel (Eds.), Self and motivation: Emerging psychological
perspectives (pp. 1341). Washington DC: American Psychological Association.
Chernev, A. (2001). The impact of common features on consumer preferences: A case of confirmatory reasoning. Journal of Consumer Research, 27, 475488.
Chiu, C., Dweck, C. S., Tong, J. Y., & Fu, J. H. (1997). Implicit theories and conceptions of morality.
Journal of Personality and Social Psychology, 73, 923940.
Chiu, C., Hong, Y., & Dweck, C. S. (1997). Lay dispositionism and implicit theories of personality.
Journal of Personality and Social Psychology, 73, 1930.
Chodorow, N. (1979). Feminism and difference: Gender, relation and difference in psychoanalytic perspective. Socialist Review, 46, 4264.
Choi, I., Nisbett, R. E., & Norenzayan, A. (1999). Causal attribution across cultures: Variation and
universality. Psychological Bulletin, 125, 4763.
Clore, G. L. (1992). Cognitive phenomenology: Feelings and the construction of judgment. In L. L.
Martin & A. Tesser (Eds.), The construction of social judgments (pp. 133163). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Clore, G. L., & Byrne, D. (1974). A reinforcement affect model of attraction. In T. L. Huston (Ed.),
Foundations of interpersonal attraction (pp. 143170). New York: Academic 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. 323417). Hillsdale, NJ: Lawrence Erlbaum Associates.
Colcombe, S. J., & Wyer, R. S. (2001). The effects of image-based priming on lexical access, conceptual activation, and behavior. Unpublished manuscript, University of Illinois at UrbanaChampaign.
Colcombe, S. J., & Wyer, R. S. (2002). The role of prototypes in the mental representation of temporally-related events. Cognitive Psychology, 44, 67103.
Collins, A. M., & Loftus, E. F. (1975). A spreading-activation theory of semantic processing. Psychological Review, 82, 407428.
Conway, M., & Ross, M. (1984). Getting what you want by revising what you had. Journal of Personality and Social Psychology, 47, 738748.
Costley, C. L., & Brucks, M. (1992). Selective recall and information use in consumer preferences.
Journal of Consumer Research, 11, 464474.
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research.
Journal of Verbal Learning and Verbal Behavior, 11, 671684.
Cutrona, C., & Russell, D. (1987). The provisions of social relationships and adaptation to stress.
In W. H. Jones & D. Perlman (Eds.), Advances in personal relationships (Vol. 1, pp. 3767).
Greenwich, CT: JAI.
DePaulo, B. M., & Friedman, H. S. (1998). Nonverbal communication. In D. T. Gilbert, S. T. Fiske, &
G. Lindzey (Eds.), Handbook of social psychology (4th ed., Vol. 2, pp. 340). New York: McGrawHill.
REFERENCES
385
DePaulo, B. M., & Kirkendol, S. E. (1989). The motivational impairment effect in the communication of deception. In J. Yuille (Ed.), Credibility assessment (pp. 5170). Belgium: Kluwer.
Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components.
Journal of Personality and Social Psychology, 56, 518.
Dienes, B. P. A. (1996). Mood as information: Affective cues for cognitive processing styles. Unpublished doctoral dissertation, University of Illinois at Urbana-Champaign.
Dijksterhuis, A., & van Knippenberg, A. (1998). The relation between perception and behavior, or
how to win a game of Trivial Pursuit. Journal of Personality and Social Psychology, 74, 865877.
Duckworth, K. L., Bargh, J. A., Garcia, M., & Chaiken, S. (2002). The automatic evaluation of novel
stimuli. Psychological Science, 13, 514520.
Dweck, C. S. (1975). The role of expectations and attributions in the alleviation of learned helplessness. Journal of Personality and Social Psychology, 31, 674685.
Dweck, C. S. (1991). Self-theories and goals: Their role in motivation, personality, and development. In R. Dienstbier (Ed.), Nebraska Symposium on Motivation: Vol. 38. Perspectives on motivation (pp. 199235). Lincoln: University of Nebraska Press.
Dweck, C. S., Chiu, C., & Hong, Y. (1995a). Implicit theories and their role in judgments and reactions: A world from two perspectives. Psychological Inquiry, 6, 267285.
Dweck, C. S., Chiu, C., & Hong, Y. (1995b). Implicit theories: Elaboration and extension of the
model. Psychological Inquiry, 6, 322333.
Dweck, C. S., Hong, Y., & Chiu, C. (1993). Implicit theories: Individual differences in the likelihood
and meaning of dispositional inference. Personality and Social Psychology Bulletin, 19, 644656.
Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to motivation and personality.
Psychological Review, 25, 109116.
Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort Worth, TX: Harcourt Brace
Jovanovich.
Eagly, A. H., & Chaiken, S. (1998). Attitude structure and function. In D. T. Gilbert, S. T. Fiske, & G.
Lindzey (Eds.), Handbook of social psychology (4th ed., Vol. 1, pp. 269322). Boston, MA:
McGraw-Hill.
Ebbesen, E. B. (1980). Cognitive processes in understanding ongoing behavior. In R. Hastie, T.
Ostrom, E. Ebbesen, R. Wyer, D. Hamilton, & D. Carlston (Eds.), Person memory: Cognitive basis of social perception (pp. 179226). Hillsdale, NJ: Lawrence Erlbaum Associates.
Edell, J. A., & Staelin, R. (1983). The information processing of pictures in print advertisements.
Journal of Consumer Research, 10, 4561.
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 (pp.
2543). Toronto: C. J. Hofgrefe.
Ellsworth, P. C., & Carlsmith, M. (1968). Effects of eye contact and verbal content on affective response to a dyadic interaction. Journal of Personality and Social Psychology, 10, 1520.
Emler, N., Renwick, S., & Malone, B. (1983). The relationship between moral reasoning and political orientation. Journal of Personality and Social Psychology, 45, 10731080.
Englis, B. G. (1994). The role of affect in political advertising: Voter emotional responses to the
nonverbal behavior of politicians. In E. M. Clark et al. (Eds.), Attention, attitude, and affect in response to advertising (pp. 223247). Hillsdale, NJ: Lawrence Erlbaum Associates.
Erdley, C. A., & Dweck, C. S. (1993). Childrens implicit personality theories as predictors of their
social judgments. Child Development, 64, 863878.
Fazio, R. H. (1986). How do attitudes guide behavior? In R. M. Sorrentino & E. T. Higgins (Eds.),
Handbook of motivation and cognition (pp. 204243). New York: Guilford.
Fazio, R. H. (1990). Multiple processes by which attitudes guide behavior: The MODE model as
an integrative framework. In M. P. Zanna (Ed.), Advances in experimental social psychology
(Vol. 23, pp. 75109). San Diego, CA: Academic Press.
Fazio, R. H., Sanbonmatsu, D. M., Powell, M. C., & Kardes, F. R. (1986). On the automatic activation
of attitudes. Journal of Personality and Social Psychology, 50, 229238.
386
REFERENCES
Festinger, L. (1957). A theory of cognitive dissonance. Palo Alto, CA: Stanford University Press.
Fiedler, K. (1988). Emotional mood, cognitive style, and behavior regulation. In K. Fiedler & J.
Forgas (Eds.), Affect, cognition and social behavior (pp. 100119). Toronto: Hofgrefe International.
Fischoff, B. (1975). Hindsight foresight: The effect of outcome knowledge on judgment under
uncertainty. Journal of Experimental Psychology: Human Perception and Performance, 1,
288299.
Fischoff, B. (1982). For those condemned to study the past: Heuristics and biases in hindsight. In
D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases
(pp. 332351). New York: Cambridge University Press.
Fishbein, M. (1963). An investigation of the relationships between beliefs about an object and attitude toward that object. Human Relations, 16, 233239.
Fishbein, M. (1967). A behavior theory approach to the relations between beliefs about an object
and the attitude toward the object. In M. Fishbein (Ed.), Readings in attitude theory and measurement (pp. 389400). New York: Wiley.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory
and research. Reading, MA: Addison-Wesley.
Fishbein, M., & Hunter, R. (1964). Summation versus balance in attitude organization and
change. Journal of Abnormal and Social Psychology, 69, 505510.
Fishkin, J., Keniston, K., & MacKinnon, C. (1973). Moral reasoning and political ideology. Journal
of Personality and Social Psychology, 27, 109119.
Fiske, S. T. (1998). Stereotyping, prejudice, and discrimination. In D. T. Gilbert, S. T. Fiske, & G.
Lindzey (Eds.), Handbook of social psychology (4th ed., Vol. 2, pp. 357414). New York: McGraw
Hill.
Fiske, S. T., & Pavelchak, M. A. (1986). Category-based versus piecemeal-based affective responses: Developments in schema-triggered affect. In R. M. Sorrentino & E. T. Higgins (Eds.),
Handbook of motivation and cognition (pp. 167203). New York: Guilford.
Fletcher, G. J. O., Danilovics, P., Fernandez, G., Peterson, D., & Reeder, G. D. (1986). Attributional
complexity: An individual differences measure. Journal of Personality and Social Psychology,
51, 875884.
Fong, C. P. S., & Wyer, R. S. (2003). Cultural, social and emotional determinants of decisions under uncertainty. Organizational Behavior and Human Decision Processes,
Forgas, J. P. (1991). Affect and cognition in close relationships. In G. J. O. Fletcher & F. D. Fincham
(Eds.), Cognition in close relationships (pp. 735). Hillsdale, NJ: Lawrence Erlbaum Associates.
Forgas, J. P. (1992). Affect in social judgments and decisions: a multiprocess model. In M. P.
Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 25, pp. 227275). San Diego, CA:
Academic Press.
Forgas, J. P. (1995). Mood and judgment: The affect infusion model (AIM). Psychological Bulletin,
117, 3966.
Forgas, J. P. (1998). Asking nicely? The effects of mood on responding to more or less polite requests. Personality and Social Psychology Bulletin, 2, 173185.
Forgas, J. P. (2001). Feeling and doing: Affective influences on interpersonal behavior. Psychological Inquiry, 13, 128.
Forgas, J. P. (2002). Feeling and doing: Affective influences on interpersonal behavior. Psychological Inquiry, 13, 128.
Forgas, J. P., & Bower, G. H. (1987). Mood effects on person perception judgments. Journal of Personality and Social Psychology, 53, 5360.
Forgas, J. P., Bower, G. H., & Krantz, S. (1984). The influence of mood on perceptions of social interactions. Journal of Experimental Social Psychology, 20, 497513.
Freud, S. (1928). Humor. International Journal of Psychoanalysis, 9, 16.
Freud, S. (1960/1905). Jokes and their relation to the unconscious. New York: Norton.
REFERENCES
387
Fry, P. S., & Ghosh, R. (1980). Attributions of success and failure: Comparison of cultural differences between Asian and Caucasian children. Journal of Cross-Cultural Psychology, 11,
343363.
Gaelick, L., Bodenhausen, G. V., & Wyer, R. S. (1985). Emotional communication in close relationships. Journal of Personality and Social Psychology, 49, 12461265.
Garnham, A. (1981). Mental models as representations of text. Memory & Cognition, 9, 560565.
Gasper, K., & Clore, G. L. (1998). The persistent use of negative affect by anxious individuals to
estimate risk. Journal of Personality and Social Psychology, 74, 13501363.
Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1994). Growing up with television: the cultivation perspective. In J. Bryant & D. Zillmann (Eds.), Media effects: Advances in theory and research (pp. 1741). Hillsdale, NJ: Lawrence Erlbaum Associates.
Gick, M., L., & Holyoak, K. J. (1983). Schema induction and analogical transfer. Cognitive Psychology, 15, 138.
Gilbert, D. T. (1991). How mental systems believe. American Psychologist, 46, 107119.
Gilligan, C. (1982). In a different voice: Psychological theory and womens development. Cambridge,
MA: Harvard University Press.
Giner-Sorolla, R., & Chaiken, S. (1997). Selective use of heuristic and systematic processing under defense motivation. Personality and Social Psychology Bulletin, 23, 8497.
Glenberg, A. M. (1997). What memory is for. Behavioral & Brain Sciences, 20, 155.
Glenberg, A. M., Kruley, P., & Langston, W. E. (1994). Analogical processes in comprehension:
Simulation of a mental model. In M. A. Gernsbacher (Ed.), Handbook of psycholinguistics (pp.
609640). San Diego, CA: Academic Press.
Glenberg, A. M., Meyer, M., & Lindem, K. (1987). Mental models contribute to foregrounding during text comprehension. Journal of Memory and Language, 26, 6983.
Goethals, G. R., & Reckman, R. F. (1973). The perception of consistency in attitudes. Journal of Experimental Social Psychology, 9, 491501.
Gohm, C. L., & Wyer, R. S. (1998). Narrative representations of long-term relationships: An evaluation
of assessment procedures and implications for marital satisfaction. Unpublished manuscript,
University of Illinois at Urbana-Champaign.
Goldstein, J. H., & McGhee, P. (Eds.). (1972). The psychology of humor. New York: Springer-Verlag.
Goldstein, J. H., Suls, J. M., & Anthony, S. (1972). Enjoyment of specific types of humor content:
Motivation or salience. In J. Goldstein & P. McGhee (Eds.), Its a funny thing, humor (pp.
159171). New York: Springer-Verlag.
Gordon, S. E., & Wyer, R. S. (1987). Person memory: Category-set-size effects on the recall of a
persons behaviors. Journal of Personality and Social Psychology, 53, 648662.
Graesser, A. C. (1981). Prose comprehension beyond the word. New York: Springer-Verlag.
Graesser, A. C., Gordon, S. E., & Sawyer, J. D. (1979). Memory for typical and atypical actions in
scripted activities: Test of a script pointer + tag hypothesis. Journal of Verbal Learning and
Verbal Behavior, 18, 319322.
Graesser, A. C., & Nakamura, G. V. (1982). The impact of a schema on comprehension and memory. In G. H. Bower (Ed.), The psychology of learning and motivation: Advances in research and
theory (Vol. 16, pp. 66109). New York: Academic Press.
Graesser, A. C., Singer, M., & Trabasso, T. (1994). Constructing inferences during narrative text
comprehension. Psychological Review, 101, 371395.
Green, G. M. (1989). Pragmatics and natural language understanding. Hillsdale, NJ: Erlbaum.
Green, M. C., & Brock, T. C. (2000). The role of transportation in the persuasiveness of public narratives. Journal of Personality and Social Psychology, 79, 701721.
Greenwald, A. G. (1968). Cognitive learning, cognitive responses to persuasion and attitude
change. In A. G. Greenwald, T. C. Brock, & T. M. Ostrom (Eds.), Psychological foundations of attitudes (pp. 147170). New York: Academic Press.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review, 102, 427.
388
REFERENCES
Greenwald, A. G., & Pratkanis, A. R. (1984). The self. In R. S. Wyer & T. K. Srull (Eds.), Handbook of
social cognition (Vol. 3, pp. 129178). Hillsdale, NJ: Lawrence Erlbaum Associates.
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics:
Speech acts (pp. 4158). New York: Academic Press.
Gruenfeld, D. H., & Wyer, R. S. (1992). Semantics and pragmatics of social influence: How affirmations and denials affect beliefs in referent propositions. Journal of Personality and Social Psychology, 62, 3849.
Gruenfeld, D., & Wyer, R. S. (1993). A two-set flow model of social communication: The influence of
information source and communication objectives on memory, message content and delivery,
and recipients beliefs. Unpublished manuscript, University of Illinois at Urbana-Champaign.
Hamilton, D. L., & Gifford, R. K. (1976). Illusory correlation in interpersonal perception: A cognitive basis of stereotypic judgment. Journal of Experimental Social Psychology, 13, 392407.
Hamilton, D. H., Katz, L. B., & Leirer, V. O. (1980). Organizational processes in impression formation. In R. Hastie, T. Ostrom, E. Ebbesen, R. Wyer, D. Hamilton, & D. Carlston (Eds.), Person
memory: The cognitive basis of social perception (pp. 121153). Hillsdale, NJ: Lawrence
Erlbaum Associates.
Hamilton, D. L., & Sherman, J. W. (1994). Stereotypes. In R. S. Wyer & T. K. Srull (Eds.), Handbook
of social cognition (2nd ed., Vol. 2, pp. 168). Hillsdale, NJ: Lawrence Erlbaum Associates.
Harkins, S. G., Latane, B., & Williams, K. (1980). Social loafing: Allocating effort or taking it easy?
Journal of Experimental Social Psychology, 16, 457465.
Harvey, J. H., & Weary, G. (1984). Current issues in attribution theory and research. Annual Review of Psychology, 35, 427459.
Harvey, J. H., Weber, A. L., & Orbuch, T. L. (1990). Interpersonal accounts. Cambridge, MA: Basil
Blackwell.
Hasher, L., Goldstein, D., & Toppin, T. (1977). Frequency and the conference of referential validity. Journal of Verbal Learning and Verbal Behavior, 16, 107112.
Hastie, R. (1980). Memory for behavioral information that confirms or contradicts a personality
impression. In R. Hastie, T. Ostrom, E. Ebbesen, R. Wyer, D. Hamilton, & D. Carlston (Eds.),
Person memory: The cognitive basis of social perception (pp. 141172). Hillsdale, NJ: Lawrence
Erlbaum Associates.
Hastie, R., & Kumar, P. A. (1979). Person memory: Personality traits as organizing principles in
memory for behaviors. Journal of Personality and Social Psychology, 37, 2538.
Hastie, R., & Pennington, N. (1995). The big picture: Is it a story? In R. S. Wyer (Ed.), Advances in
social cognition (Vol. 8, pp. 133138). Mahwah, NJ: Lawrence Erlbaum Associates.
Hawkins, S. A., & Hastie, R. (1990). Hindsight: Biased judgment of past events after the outcomes
are known. Psychological Bulletin, 107, 311327.
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Heine, S. J., & Lehman, D. R. (1997). Culture, dissonance, and self-affirmation. Personality and Social Psychology Bulletin, 23, 389400.
Heine, S., Lehman, D., Markus, H., & Kitayama, S. (1999). Is there a universal need for positive
self-regard? Psychological Review, 106, 766794.
Henninger, M., & Wyer, R. S. (1976). The recognition and elimination of inconsistencies among
syllogistically-related beliefs: Some new light on the Socratic effect. Journal of Personality
and Social Psychology, 34, 680693.
Herr, P. M. (1986). Consequences of priming: Judgment and behavior. Journal of Personality and
Social Psychology, 51, 11061115.
Higgins, E. T. (1981). The communication game: Implications for social cognition and persuasion. In E. T. Higgins, C. P. Herman, & M. P. Zanna (Eds.), Social cognition: The Ontario Symposium (Vol. 1, pp. 342392). Hillsdale, NJ: Lawrence Erlbaum Associates.
Higgins, E. T. (1996). Knowledge activation: Accessibility, applicability, and salience. In E. T. Higgins & A. Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 133168). New
York: Guilford.
REFERENCES
389
Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 55, 12171233.
Higgins, E. T. (1998). Promotion and prevention: Regulatory focus as a motivational principle. In
M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 30, pp. 146). San Diego,
CA: Academic Press.
Higgins, E. T., Bargh, J. A., & Lombardi, W. (1985). The nature of priming effects on categorization. Journal of Experimental Psychology: Learning, Memory, and Cognition, 11, 5969.
Higgins, E. T., & Lurie, L. (1983). Context, categorization and recall: The change-of-standard effect. Cognitive Psychology, 15, 525547.
Higgins, E. T., & Rholes, W. S. (1978). Saying is believing: Effects of message modification on
memory and liking for the person described. Journal of Experimental Social Psychology, 14,
363378.
Higgins, E. T., Rholes, W. S., & Jones, C. R. (1977). Category accessibility and impression formation. Journal of Experimental Social Psychology, 13, 141154.
Hilgard, E. R. (1951). Methods and procedures in the study of learning. In S. S. Stevens (Ed.),
Handbook of experimental psychology. New York: Wiley.
Hintzman, D. L. (1986). Schema abstraction in a multiple-trace model. Psychological Review, 93,
411428.
Ho, D. Y. F., & Chiu, C. Y. (1994). Component ideas of individualism, collectivism, and social organization: An application in the study of Chinese culture. In U. Kim, C. Kagitcibasi, S. C. Choi, &
G. Yoon (Eds.), Individualism and collectivism: theory, method and applications (pp. 137156).
Thousand Oaks, CA: Sage.
Hofstede, G. H. (1991). Empirical models of cultural differences. In N. Bleichrodt & P. J. D. Drenth
(Eds.), Contemporary issues in cross-cultural psychology (pp. 420). Amsterdam, Netherlands:
Swets & Zeitlinger.
Holmberg, D., & Holmes, J. G. (1994). Reconstruction of relationship memories: A mental models
approach. In N. Schwarz & S. Sudman (Eds.), Autobiographical memory and the validity of retrospective reports (pp. 267288). New York: Springer-Verlag.
Holmes, J. G., & Murray, S. L. (1995). Memory for events in close relationships: Applying Schank
and Abelsons story skeleton model. In R. S. Wyer (Ed.), Advances in social cognition (Vol. 8,
pp. 193210). Hillsdale, NJ: Lawrence Erlbaum Associates.
Hong, Y., Chiu, C., Dweck, C. S., & Sacks, R. (1997). Implicit theories and evaluative processes in
person cognition. Journal of Experimental Social Psychology, 33, 296303.
Hong, Y., Morris, M. W., Chiu, C., & Benet-Martinez, V. (2000). Multicultural minds: A dynamic
constructivist approach to culture and cognition. American Psychologist, 55, 709720.
Humphreys, M. S., & Kashima, Y. (2002). Connectionism and self: Distributed representational
systems and their implications for self and identity. In Y. Kashima, M. Foddy, & M. Platow
(Eds.), Self and identity: Personal, social and symbolic (pp. 2754). Mahwah, NJ: Lawrence
Erlbaum Associates.
Hunt, R. R., & Einstein, G. O. (1981). Relational and item-specific information in memory. Journal
of Verbal Learning and Verbal Behavior, 20, 497514.
Hyman, I. E., & Loftus, E. F. (2002). False childhood memories and eyewitness memory errors. In
M. Eisen & J. Quas (Eds.), Memory and suggestibility in the forensic interview (pp. 6384).
Mahwah, NJ: Lawrence Erlbaum Associates.
Innes-Ker, A., & Niedenthal, P. M. (2002). Emotion concepts and emotional states in social judgment and categorization. Journal of Personality and Social Psychology, 83, 804816.
Isbell, L. M. (2000). Beyond heuristic information processing: Systematic processing in happy and sad
moods. Unpublished doctoral dissertation, University of Illinois at Urbana-Champaign.
Isbell, L. M., Clore, G. L., & Wyer, R. S. (1998). Mood-mediated uses of stereotyping in impression formation. Unpublished manuscript, University of Illinois at Urbana-Champaign.
Isbell, L. M., & Wyer, R. S. (1999). Correcting for mood-induced bias in the evaluation of political
candidates. Personality and Social Psychology Bulletin, 25, 237249.
390
REFERENCES
Isbell, L. M., Wyer, R. S., & Collins, J. E. (2002). The cognitive dynamics of humor elicitation: The role
of diminishment, disparagement and cognitive elaboration in responses to jokes and witticisms.
Unpublished manuscript, Hong Kong University of Science and Technology, Hong Kong.
Isen, A. M. (1984). Toward understanding the role of affect in cognition. In R. S. Wyer & T. K. Srull
(Eds.), Handbook of social cognition (Vol. 3, pp. 179236). Hillsdale, NJ: Lawrence Erlbaum Associates.
Isen, A. M. (1987). Positive affect, cognitive processes, and social behavior. In L. Berkowitz (Ed.),
Advances in experimental social psychology (Vol. 20, pp. 203253). New York: Academic Press.
Isen, A. M., & Daubman, L. A. (1984). The influence of affect on categorization. Journal of Personality and Social Psychology, 47, 12061217.
Isen, A. M., Daubman, L. A., & Nowicki, G. P. (1987). Positive affect facilitates creative problem
solving. Journal of Personality and Social Psychology, 52, 11221131.
Isen, A. M., Johnson, M., Mertz, E., & Robinson, G. (1985). Positive affect and the uniqueness of
word association. Journal of Personality and Social Psychology, 48, 14131426.
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, 112.
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, 326338.
Janis, I. L., & King, B. T. (1954). The influence of role playing on opinion change. Journal of Abnormal and Social Psychology, 49, 211218.
Johnson, M., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring. Psychological Bulletin,
114, 328.
Johnson-Laird, P. N. (1983). Mental models: Towards a cognitive science of language, inference and
consciousness. Cambridge, MA: Harvard University Press.
Johnson-Laird, P. N. (1989). Mental models. In M. I. Posner (Ed.), Foundations of cognitive science
(pp. 469500). Cambridge, MA: MIT Press.
Jones, E. E., & Davis, K. J. (1965). From acts to dispositions: The attributional process in person
perception. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp.
220266). New York: Academic Press.
Kahneman, D., & Miller, D. T. (1986). Norm theory: Comparing reality to its alternatives. Psychological Review, 93, 136153.
Kahneman, D., Slovic, P., & Tversky, A. (Eds.). (1982). Judgment under uncertainty: Heuristics and
biases. New York: Cambridge University Press.
Kahneman, D., & Tversky, A. (1982). The simulation heuristic. In D. Kahneman, P. Slovic, & A.
Tversky (Eds.), Judgment under uncertainty: Heuristics and biases (pp. 201208). New York:
Cambridge University Press.
Kane, T. R., Suls, J., & Tedeschi, J. T. (1977). Humor as a tool of social interaction. In A. J. Chapman & H. C. foot (Eds.), Its a funny thing, humor. Elmsford, NY: Pergamon Press.
Kelley, C. M., & Lindsay, D. S. (1993). Remembering mistaken for knowing: Ease of retrieval as a
basis for confidence in answers to general knowledge questions. Journal of Memory and Language, 32, 124.
Kelley, H. H. (1967). Attribution theory in social psychology. In D. Levine (Ed.), Nebraska symposium on motivation, 15, 192238.
Kelley, H. H. (1987). Causal schemata and the attribution process. In E. Jones, D. Kanouse, H.
Kelley, R. Nisbett, S. Valins, & B. Winer (Eds.), Attribution: Perceiving the causes of behavior.
Hillsdale, NJ: Lawrence Erlbaum Associates. (Originally published 1971).
Kihlstrom, J. F. (1980). Posthypnotic amnesia for recently learned material: Interactions with
episodic and semantic memory. Cognitive Psychology, 12, 227251.
Kintsch, W. (1998). Comprehension: A paradigm for cognition. Cambridge, England: Cambridge
University Press.
REFERENCES
391
Klein, S. B., & Loftus, J. (1993). Behavioral experience and trait judgments about the self. Personality and Social Psychology Bulletin, 19, 740745.
Klein, S. B., Loftus, J., Trafton, J. G., & Fuhrman, R. W. (1992). Use of exemplars and abstractions
in trait judgments: A model of trait knowledge about the self and others. Journal of Personality
and Social Psychology, 63, 739753.
Klinger, E. (1975). Consequences of commitment to and disengagement from incentives. Psychological Bulletin, 82, 223231.
Koestler, A. (1964). The act of creation. New York: Macmillan.
Kohlberg, L. (1976). Moral stages and moralization: The cognitive-developmental approach. In T.
Lickona (Ed.), Moral development and behavior: Theory, research and social issues. New York:
Holt, Rinehart, & Winston.
Kosslyn, S. M. (1980). Image and mind. Cambridge, MA: Harvard University Press.
Kunda, Z. (1987). Motivated inference: Self-serving generation and evaluation of causal theories.
Journal of Personality and Social Psychology, 53, 636647.
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480498.
LaFave, L., Haddad, J., & Maesen, W. A. (1976). Superiority, enhanced self-esteem and perceived
incongruity humor theory. In A. J. Chapman & H. C. Foot (Eds.), Humor and laughter: Theory,
research and applications (pp. 6391). New York: Wiley.
Lampel, A. K., & Anderson, N. H. (1968). Combining visual and verbal information in an impression-formation task. Journal of Personality and Social Psychology, 9, 16.
Langer, E. J., & Abelson, R. P. (1972). The semantics of asking a favor: How to succeed in getting
help without really dying. Journal of Personality and Social Psychology, 24, 2632.
LaPiere, R. T. (1934). Attitudes vs. actions. Social Forces, 13, 230237.
Lazarus, R. S. (1982). Thoughts on the relations between emotion and cognition. American Psychologist, 37, 10191024.
Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford University Press.
LeDoux, J. E. (1994). Memory vs. emotional memory in the brain. in P. Ekman & R. J. Davidson
(Eds.), The nature of emotion: Fundamental questions (pp. 311312). New York: Oxford University Press.
Lefcourt, H. M., & Martin, R. A. (1986). Humor and life stress: Anecdote to adversity. New York:
Springer-Verlag.
Lepore, L., & Brown, R. (1997). Category and stereotype activation: Is prejudice inevitable? Journal of Personality and Social Psychology, 72, 275287.
Lerner, M. J. (1970). The desire for justice and reactions to victims. In J. Macaulay & L. Berkowitz
(Eds.), Altruism and helping behavior (pp. 205229). New York: Academic Press.
Lerner, M. J., & Miller, D. T. (1978). Just world research and the attribution process: Looking
back and ahead. Psychological Bulletin, 85, 10301051.
Lerner, M. J., Miller, D. T., & Holmes, J. G. (1976). Deserving and the emergence of forms of justice. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 9). New York: Academic Press.
Lerner, M. J., & Simmons, C. H. (1966). Observers reaction to the innocent victim: Compassion
or rejection? Journal of Personality and Social Psychology, 4, 203210.
Levine, S. R., Wyer, R. S., & Schwarz, N. (1994). Are you what you feel? The affective and cognitive
determinants of self-judgments. Experimental Journal of Social Psychology, 24, 6377.
Locke, H. L., & Wallace, K. (1959). Short marital adjustment and prediction tests: Their reliability
and validity. Marriage and Family Living, 21, 251255.
Loftus, E. F. (1975). Leading questions and the eyewitness report. Cognitive Psychology, 7, 560
572.
Loftus, E. F. (2000). Remembering what never happened. In E. Tulving (Ed.), Memory, consciousness and the brain: The Tallinn conference (pp. 106118). Philadelphia, PA: Taylor & Francis.
392
REFERENCES
Loftus, E. F., & Marburger, W. (1983). Since the eruption of Mt. St. Helens, has anyone beaten you
up? Improving the accuracy of retrospective reports with landmark events. Memory & Cognition, 11, 114120.
Loftus, E. F., & Palmer, S. J. (1974). Reconstruction of automobile destruction. Journal of Verbal
Learning and Verbal Behavior, 2, 467471.
Logan, G. L. (1988). Toward an instance theory of automatization. Psychological Review, 95,
492527.
Loken, B. A. (1984). Attitude processing strategies. Journal of Experimental Social Psychology, 20,
272296.
Loken, B. A., & Wyer, R. S. (1983). Effects of reporting beliefs in syllogistically related propositions on the recognition of unmentioned propositions. Journal of Personality and Social Psychology, 45, 306322.
Lombardi, W. J., Higgins, E. T., & Bargh, J. A. (1987). The role of consciousness in priming effects
on categorization. Personality and Social Psychology Bulletin, 13, 411429.
Long, D. L., & Graesser, A. C. (1988). Wit and humor in discourse processing. Discourse Processes,
11, 3560.
Loomes, G., & Sugden, R. (1986). Disappointment and dynamic consistency in choice under uncertainty. Review of Economic Studies, 53, 271282.
Lynch, J. G., Chakravarti, D., & Mitra, A. (1991). Contrast effects in consumer judgments: Changes
in mental representations or in the anchoring of rating scales? Journal of Consumer Research,
18, 284297.
Mackie, D. M., & Worth, L. T. (1989). Cognitive deficits and the mediation of positive affect in persuasion. Journal of Personality and Social Psychology, 57, 2740.
Macrae, C. N., Bodenhausen, G. V., Milne, A. B., & Jetten, J. (1994). Out of mind but back in sight:
Stereotypes on the rebound. Journal of Personality and Social Psychology, 67, 808817.
Magliano, J. P., Dijkstra, K., & Zwaan, R. A. (1996). Generating predictive inferences while viewing
a movie. Discourse Processes, 22, 199224.
Maheswaran, D., Mackie, D. M., & Chaiken, S. (1992). Brand name as a heuristic cue: The effects of
task importance and expectancy confirmation on consumer judgments. Journal of Consumer
Psychology, 1, 317336.
Mandler, J. (1979). Categorical and schematic organization in memory. In C. R. Puff (Ed.), Memory, organization and structure. New York: Academic Press.
Mandler, J., & Johnson, N. S. (1977). Remembrance of things parsed: Story structure and recall.
Cognitive Psychology, 9, 111151.
Mankowski, E., & Rappaport, J. (1995). Stories, identity, and the psychological sense of community. In R. S. Wyer (Ed.), Advances in social cognition (Vol. 8, pp. 211226). Hillsdale, NJ: Lawrence Erlbaum Associates.
Mankowski, E. S., & Wyer, R. S. (1996). Cognitive processes in perceptions of social support. Personality and Social Psychology Bulletin, 22, 894905.
Markus, H., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion and
motivation. Psychological Review, 98, 224253.
Martin, L. L., Achee, J. W., Ward, D. 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, 317326.
Martin, L. L., Seta, J. J., & Crelia, R. A. (1990). Assimilation and contrast as a function of peoples
willingness and ability to expend effort in forming an impression. Journal of Personality and
Social Psychology, 59, 2737.
Martin, L. L., & Tesser, A. (1996). Some ruminative thoughts. In R. S. Wyer (Ed.), Ruminative
thoughts: Advances in social cognition (Vol. 9, pp. 148). Mahwah, NJ: Lawrence Erlbaum Associates.
McAdams, D. (1988). Power, intimacy and the life story: Personalogical inquiries into identity. New
York: Guilford Press.
REFERENCES
393
McClelland, D. C., Atkinson, J. W., Clark, R. A., & Lowell, E. L. (1953). The achievement motive. New
York: Appleton-Century-Crofts.
McGhee, P., & Goldstein, J. (Eds.). (1983). Handbook of humor research: Basic issues. New York:
Springer-Verlag.
McGuire, W. J. (1960). A syllogistic analysis of cognitive relationships. In M. J. Rosenberg & C. I.
Hovland (Eds.), Attitude organization and change (pp. 140162). New Haven, CT: Yale University Press.
McGuire, W. J. (1972). Attitude change: An information processing paradigm. In C. G. McClintock
(Ed.), Experimental social psychology (pp. 108141). New York: Holt, Rinehart, and Winston.
McGuire, W. J. (1981). The probabilogical model of cognitive structure and attitude change. In
R. E. Petty, T. M. Ostrom, & T. C. Brock (Eds.), Cognitive responses in persuasion (pp. 291307).
Hillsdale, NJ: Lawrence Erlbaum Associates.
McGuire, W. J. (1999). Constructing social psychology: Creative and critical processes. New York:
Cambridge University Press.
McGuire, W. J., & McGuire, C. V. (1988). Content and process in the experience of self. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 21, pp. 97144). San Diego,
CA: Academic Press.
McGuire, W. J., & McGuire, C. V. (1991). The content, structure and operation of thought systems.
In R. S. Wyer & T. K. Srull (Eds.), Advances in social cognition (Vol. 4, pp. 178). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Mellers, B., Schwartz, A., & Ritov, I. (1999). Emotion-based choice. Journal of Experimental Psychology: General, 128, 332345.
Menon, G., & Raghubir, P. (1998). When automatic accessibility meets conscious content: Implications for judgment formation. Unpublished manuscript, New York University.
Miller, D. T. (1977). Altruism and threat to a belief in a just world. Journal of Experimental Social
Psychology, 13, 113124.
Miller, L. C., & Read, S. R. (1991). On the coherence of mental models of persons and relationships: A knowledge structure approach. In G. J. O. Fletcher & F. D. Fincham (Eds.), Cognition
in close relationships (pp. 6999). Hillsdale, NJ: Lawrence Erlbaum Associates.
Miller, P. J. (1994). Narrative practices: Their role in socialization and self-construction. In U.
Neisser & R. Fivush (Eds.), The remembering self: Construction and accuracy in the self-narrative. Emory symposia in cognition (Vol. 6, pp. 158179). New York: Cambridge University Press.
Miller, P. J. (1995). Personal storytelling in everyday life: Social and cultural perspectives. In R. S.
Wyer (Ed.), Advances in social cognition: Knowledge and memory (Vol. 8, pp. 177184). Mahwah, NJ: Lawrence Erlbaum Associates.
Miller, P. J., Fung, H., & Mintz, J. (1996). Self-construction through narrative practices: A Chinese
and American comparison of early socialization. Ethos, 24, 237280.
Miller, P. J., Sandel, T. L., Liang, C.-H., & Fung, H. (2001). Narrating transgressions in Longwood:
The discourses, meanings and paradoxes of an American socializing practice. Ethos, 29,
159186.
Miller, P. J., Wiley, A. R., Fung, H., & Liang, C. H. (1997). Personal storytelling as a medium of socialization in Chinese and American families. Child Development, 68, 557568.
Mindess, H., Miller, C., Turek, J., Bender, A., & Corbin, S. (1985). The Antioch Humor test: Making
sense of humor. New York: Avon Books.
Miniard, P. W., Bhatla, S., Lord, K. R., Dickson, P. R., & Unnava, H. R. (1991). Picture-based persuasion processes and the moderating role of involvement. Journal of Consumer Research, 18,
82107.
Miniard, P. W., & Cohen, J. B. (1981). An examination of the Fishbein-Ajzen behavioral intentions
models concepts and measures. Journal of Experimental Social Psychology, 17, 309339.
Morris, M. W., & Peng, K. (1994). Culture and cause: American and Chinese attributions for social
and physical events. Journal of Personality and Social Psychology, 67, 949971.
394
REFERENCES
Moskowitz, G. B., & Skurnik, I. W. (1999). Contrast effects as determined by the type of prime:
Trait versus exemplar primes initiate processing strategies that differ in how accessible constructs are used. Journal of Personality and Social Psychology, 76, 911927.
Mueller, J. H., Grove, T. R., & Thompson, W. B. (1991). Mood-dependent retrieval and mood
awareness. Cognition and Emotion, 5, 331349.
Murray, N., Sujan, H., Hirt, E. R., & Sujan, M. (1990). The influence of mood on categorization: A
cognitive flexibility interpretation. Journal of Personality and Social Psychology, 59, 411425.
Murray, S. L., & Holmes, J. G. (1996). The construction of relationship realities. In G. J. O. Fletcher
& J. Fitness (Eds.), Knowledge structures in close relationships: A social psychological approach
(pp. 91120). Mahwah, NJ: Lawrence Erlbaum Associates.
Mussweiler, T. (2003). Comparison processes in social judgment: Mechanisms and consequences. Psychological Review,
Mussweiler, T., & Strack, F. (1999a). Hypothesis-consistent testing and semantic priming in the
anchoring paradigm: A selective accessibility model. Journal of Experimental Social Psychology, 35, 136164.
Mussweiler, T., & Strack, F. (1999b). Comparing is believing: A selective accessibility model of
judgmental anchoring. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 10, pp. 135167). Chichester, England: Wiley.
Mussweiler, T., & Strack, F. (2000a). The use of category and exemplar knowledge in the solution
of anchoring tasks. Journal of Personality and Social Psychology, 78, 10381052.
Mussweiler, T., & Strack, F. (2000b). The relative self: Informational and judgmental consequences of comparative self-evaluation. Journal of Personality and Social Psychology, 79,
2338.
Nassi, A. J., Abramowitz, S. I., & Youmans, J. E. (1983). Moral development and politics a decade
later: A replication and extension. Journal of Personality and Social Psychology, 45, 11271135.
Nehrhardt, G. (1976). Incongruity and funniness: Toward a new descriptive model. In A. J. Chapman & H. C. Foot (Eds.), Humor and laughter: Theory, research and application (pp. 5562). New
York: Wiley.
Nelson, K. (Ed.). (1989). Stories from the crib. Cambridge, MA: Harvard University Press.
Nelson, K. (1993). The psychological and social origins of autobiographical memory. Psychological Science, 4, 714.
Nelson, K., & Greundel, J. (1981). Generalized event representations: Building blocks of cognitive
development. In M. Lamb & A. Brown (Eds.), Advances in developmental psychology (Vol. 1,
pp. 131158). Hillsdale, NJ: Lawrence Erlbaum Associates.
Newtson, D. A. (1973). Attribution and the unit of perception of ongoing behavior. Journal of Personality and Social Psychology, 28, 2838.
Newtson, D. A. (1976). Foundations of attribution: The perception of ongoing behavior. In J.
Harvey, W. Ickes, & R. Kidd (Eds.), New directions in attribution research (Vol. 1, pp. 223247).
Hillsdale, NJ: Lawrence Erlbaum Associates.
Niedenthal, P. M., Halberstadt, J. B., & Setterlund, M. B. (1997). Being happy and seeing happy:
Emotional state mediates visual word recognition. Cognition and Emotion, 11, 403432.
Niedenthal, P. M., & Setterlund, M. B. (1994). Emotion congruence in perception. Personality and
Social Psychology Bulletin, 20, 401411.
Nielsen, A. C. (1995). Nielsen report on television. Northbrook, IL: A.C. Nielsen.
Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies and shortcomings of social judgments.
Englewood Cliffs, NJ: Prentice Hall.
Norenzayan, A., & Kim, B. (2000). Unpublished data.
Norenzayan, A., & Schwarz, N. (1999). Telling what they want to know: Participants tailor causal
attributions to researchers interests. European Journal of Social Psychology, 29, 10111020.
Norman, D. A., & Bobrow, D. G. (1979). Descriptions: An intermediate stage in memory retrieval.
Cognitive Psychology, 11, 107123.
REFERENCES
395
Nottenburg, G., & Shoben, E. J. (1980). Scripts as linear orders. Journal of Experimental Social Psychology, 16, 329347.
OGuinn, T. C., & Shrum, L. J. (1997). The role of television in the construction of social reality.
Journal of Consumer Research, 23, 278294.
Oishi, S., Wyer, R. S., & Colcombe, S. (2000). Cultural variation in the use of current life satisfaction to predict the future. Journal of Personality and Social Psychology, 78, 434445.
Ortony, A., Clore, G. L., & Collins, A. (1988). The cognitive structure of emotions. New York: Cambridge University Press.
Ostrom, T. M., & Upshaw, H. S. (1968). Psychological perspective and attitude change. In A. G.
Greenwald, T. M. Ostrom, & T. C. Brock (Eds.), Psychological foundations of attitudes. New
York: Academic Press.
Ottati, V. C., & Isbell, L. M. (1996). Effects of mood during exposure to target information on subsequently reported judgments: An on-line model of misattribution and correction. Journal of
Personality and Social Psychology, 71, 3953.
Ottati, V. C., & Wyer, R., S. (1993). Affect and political judgment. In S. Iyengar & W. J. McGuire
(Eds.), Current approaches to political psychology (pp. 296320). Durham, NC: Duke University
Press.
Parducci, A. (1965). Category judgment: A range-frequency model. Psychological Review, 72,
407418.
Park, J. W., & Wyer, R. S. (1993). The cognitive organization of product information: Effects of attribute category set size on information recall. Journal of Consumer Psychology, 2, 329357.
Park, J. W., Yoon, S. O., Kim, K. H., & Wyer, R. S. (2001). Effects of priming a bipolar attribute concept on dimension versus concept-specific accessibility of semantic memory. Journal of Personality and Social Psychology, 81, 405420.
Parrott, G., & Sabini, J. (1990). Mood and memory under natural conditions: Evidence for mood
and incongruent recall. Journal of Personality and Social Psychology, 59, 321336.
Pennington, N., & Hastie, R. (1986). Evidence evaluation in complex decision making. Journal of
Personality and Social Psychology, 51, 242258.
Pennington, N., & Hastie, R. (1988). Explanation-based decision making: Effects of memory structure on judgment. Journal of Experimental Psychology: Learning, Memory and Cognition, 14,
521533.
Pennington, N., & Hastie, R. (1992). Explaining the evidence: Tests of the Story Model for juror
decision making. Journal of Personality and Social Psychology, 62, 189206.
Pennington, N., & Hastie, R. (1993). A theory of explanation-based decision making. In G. Klein, J.
Orasanu, R. Calderwood, & C. E. Zsambok (Eds.), Decision making in action: Models and methods (pp. 188204). Norwood, NJ: Ablex.
Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes
to attitude change. New York: Springer-Verlag.
Pham, M. T. (1998). Representativeness, relevance and the use of feelings in decision making.
Journal of Consumer Research, 25, 144159.
Pham, M. T., Cohen, J. B., Pracejus, J., & Hughes, G. D. (2001). Affect monitoring and the primacy
of feelings in judgment. Journal of Consumer Research, 28, 167188.
Picek, J. S., Sherman, S. J., & Shiffrin, R. M. (1975). Cognitive organization and coding of social
structures. Journal of Personality and Social Psychology, 31, 758768.
Prinz, W. (1990). On dynamic pertinence models. In H. G. Geissler & M. H. Miller (Eds.), Psychophysical explorations of mental structures (pp. 411421). Kirkland, WA: Hofgrefe & Huber.
Pylyshyn, Z. W. (1973). What the minds eye tells the minds brain: A critique of mental imagery.
Psychological Bulletin, 80, 124.
Pyszczynski, T. A., & Greenberg, J. (1987). Toward an integration of cognitive and motivational
perspectives on social inference: A biased hypothesis-testing model. In L. Berkowitz (Ed.),
Advances in experimental social psychology (Vol. 20, pp. 297340). San Diego, CA: Academic
Press.
396
REFERENCES
Radvansky, G. A., Gerard, L. D., Zacks, R. T., & Hasher, L. (1990). Younger and older adults use of
mental models as representations for text material. Psychology and Aging, 5, 209214.
Radvansky, G. A., Wyer, R. S., Curiel, J. M., & Lutz, M. F. (1997). Mental models and abstract relations. Journal of Experimental Psychology: Learning, Memory, and Cognition, 23, 12331246.
Radvansky, G. A., & Zacks, R. T. (1991). Mental models and the fan effect. Journal of Experimental
Psychology: Learning, Memory and Cognition, 17, 940953.
Ratcliff, R. (1978). A theory of memory retrieval. Psychological Review, 85, 59108.
Read, D. (1985). Determinants of relative mutability. Unpublished research, University of British
Columbia, Vancouver, Canada.
Regan, D., & Totten, J. (1975). Empathy and attribution: Turning observers into actors. Journal of
Personality and Social Psychology, 32, 850856.
Reid, L. D., & Wyer, R. S. (1998). Stereotype suppression in a social context. In R. S. Wyer (Ed.),
Advances in social cognition (Vol. 11, pp. 177192). Mahwah, NJ: Lawrence Erlbaum Associates.
Reyes, R. M., Thompson, W. C., & Bower, G. H. (1980). Judgmental biases resulting from differing
availabilities of arguments. Journal of Personality and Social Psychology, 39, 212.
Rhee, E., Uleman, J., & Lee, H. K. (1996). Variations in collectivism and individualism by ingroup
and culture: Confirmatory factor analysis. Journal of Personality and Social Psychology, 71,
10371054.
Roselli, F. (1995, August). Positive affect and heuristic cues in persuasion. Poster presented at the
American Psychological Society Convention, New York.
Rosen, N. A., & Wyer, R. S. (1972). Some further evidence for the Socratic effect using a subjective probability model of cognitive organization. Journal of Personality and Social Psychology,
24, 420424.
Ross, L. D., Greene, D., & House, P. (1977). The false consensus effect: An egocentric bias in social perception and attribution processes. Journal of Experimental Social Psychology, 13,
279301.
Ross, L., Lepper, M. R., & Hubbard, M. (1975). Perseverance in self-perception and social perception: Biased attributional processes in the debriefing paradigm. Journal of Personality and Social Psychology, 32, 880892.
Ross, L., Lepper, M. R., Strack, F., & Steinmetz, J. (1977). Social explanation and social expectation: Effects of real and hypothetical explanations on subjective likelihood. Journal of Personality and Social Psychology, 35, 817829.
Ross, M. (1989). Relation of implicit theories to the construction of personal histories. Psychological Review, 96, 341357.
Ross, M., McFarland, C., & Fletcher, G. J. O. (1981). The effect of attitude on recall of past histories. Journal of Personality and Social Psychology, 10, 627634.
Rubin, D. C. (1995). Stories about stories. In R. S. Wyer (Ed.), Advances in social cognition (Vol. 8,
pp. 153164). Hillsdale, NJ: Lawrence Erlbaum Associates.
Rusting, C. L., & DeHart, T. (2000). Retrieving positive memories to regulate negative mood:
Consequences for mood-congruent memory. Journal of Personality and Social Psychology, 78,
737752.
Sanitioso, R., Kunda, Z., & Fong, G. T. (1990). Motivated recruitment of autobiographical memories. Journal of Personality and Social Psychology, 59, 229241.
Sanna, L. J., & Schwarz, N. (2003). Debiasing the hindsight bias: The role of accessibility experiences and (mis)attributions. Journal of Experimental Social Psychology.
Sanna, L. J., Schwarz, N., & Small, E. M. (2003). Accessibility experiences and the hindsight bias: I
knew it all along versus it could never have happened. Memory & Cognition.
Sanna, L. J., Schwarz, N., & Stocker, S. L. (2002). What debiasing backfires: Accessible content
and accessibility experiences in debiasing hindsight. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 28, 497502.
REFERENCES
397
Schachter, S., & Singer, J. (1962). Cognitive, social and physiological determinants of emotional
state. Psychological Review, 69, 379399.
Schaeken, W., Johnson-Laird, P. N., & dYdewalle, G. (1996). Mental models and temporal reasoning. Cognition, 60, 205234.
Schank, R. C., & Abelson, R. P. (1977). Scripts, plans, goals and understanding. Hillsdale, NJ: Lawrence Erlbaum Associates.
Schank, R. C., & Abelson, R. P. (1995). Knowledge and memory: The real story. In R. S. Wyer (Ed.),
Advances in social cognition (Vol. 8, pp. 185). Hillsdale, NJ: Lawrence Erlbaum Associates.
Schick, C., McGlynn, R. P., & Woolam, D. (1972). Perception of cartoon humor as a function of familiarity and anxiety level. Journal of Personality and Social Psychology, 24, 2225.
Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing:
I. Detection, research and attention. Psychological Review, 84, 166.
Schooler, J. W., & Engstler-Schooler, T. Y. (1990). Verbal overshadowing of visual memories:
Some things are better left unsaid. Cognitive Psychology, 22, 3671.
Schwarz, N. (1990). Feelings as information: Informational and motivational functions of affective states. In E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of motivation and cognition:
Foundations of social behavior (Vol. 2, pp. 527561). New York: Guilford Press.
Schwarz, N. (1998). Accessible content and accessibility experiences: The interplay of declarative and experiential information in judgment. Personality and Social Psychology Review, 2,
8799.
Schwarz, N. (1994). Judgment in a social context: Biases, shortcomings, and the logic of conversation. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 24, pp. 123162).
San Diego, CA: Academic Press.
Schwarz, N., & Bless, H. (1991). Emotion and social judgments. In J. Forgas (Ed.), Emotion and social judgments (pp. 5571). Elmsford, NJ: Pergamon Press.
Schwarz, N., Bless, H., & Bohner, G. (1991). Mood and persuasion: Affective states influence the
processing of persuasive communications. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 24, pp. 161199). 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, 195202.
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,
513523.
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. 4462). Toronto:
Hofgrefe International.
Schwarz, N., & Clore, G. L. (1996). Feelings and phenomenal experiences. In E. T. Higgins & A.
Kruglanski (Eds.), Social psychology: A handbook of basic principles (pp. 433465). New York:
Guilford Press.
Schwarz, N., Strack, F., Hilton, D., & Naderer, G. (1991). Base rates, representativeness, and the
logic of conversation: The contextual relevance of irrelevant information. Social Cognition,
9, 6784.
Schwarz, N., & Wyer, R. S. (1985). Effects of rank ordering stimuli on magnitude ratings of these
and other stimuli. Journal of Experimental Social Psychology, 21, 3046.
Scott, C. K., Fuhrman, R. W., & Wyer, R. S. (1991). Information processing in close relationships. In
G. J. O. Fletcher & F. D. Fincham (Eds.), Cognition in close relationships (pp. 3768). Hillsdale,
NJ: Lawrence Erlbaum Associates.
Sentis, K. P., & Burnstein, E. (1979). Remembering schema-consistent information: Effects of a
balance schema on recognition memory. Journal of Personality and Social Psychology, 37,
22002212.
398
REFERENCES
REFERENCES
399
Snyder, M. L., Stephan, W. G., & Rosenfield, D. (1976). Egotism and attribution. Journal of Personality and Social Psychology, 33, 435441.
Sperber, D., & Wilson, D. (1986). Relevance: Communication and cognition. Oxford, England: Basil
Blackwell.
Sperling, G. (1960). The information available in brief visual presentations. Psychological Monographs, 74, 129.
Spiro, R. J. (1977). Remembering information from text: The state of schema approach. In R. C.
Anderson, R. J. Spiro, & W. E. Montague (Eds.), Schooling and the acquisition of knowledge.
Hillsdale, NJ: Lawrence Erlbaum Associates.
Srull, T. K. (1981). Person memory: Some tests of associative storage and retrieval models. Journal of Experimental Psychology: Human Learning and Memory, 7, 440463.
Srull, T. K., & Brand, J. F. (1983). Memory for information about persons: The effect of encoding
operations on subsequent retrieval. Journal of Verbal Learning and Verbal Behavior, 22,
219230.
Srull, T. K., Lichtenstein, M., & Rothbart, M. (1985). Associative storage and retrieval processes
in person memory. Journal of Experimental Psychology: Learning, Memory and Cognition, 11,
316345.
Srull, T. K., & Wyer, R. S. (1979). The role of category accessibility in the interpretation of information about persons: Some determinants and implications. Journal of Personality and Social
Psychology, 37, 16601672.
Srull, T. K., & Wyer, R. S. (1980). Category accessibility and social perception: Some implications
for the study of person memory and interpersonal judgments. Journal of Personality and Social Psychology, 38, 841856.
Srull, T. K., & Wyer, R. S. (1983). The role of control processes and structural constraints in models of memory and social judgment. Journal of Experimental Social Psychology, 19, 497521.
Srull, T. K., & Wyer, R. S. (1989). Person memory and judgment. Psychological Review, 96, 5863.
Stapel, D. A., & Koomen, W. (2000). How far do we go beyond the information given? The impact
of knowledge activation on interpretation and inference. Journal of Personality and Social Psychology, 78, 1937.
Stapel, D. A., Koomen, W., & van der Plight, J. (1997). Categories of category accessibility: The impact of trait concept versus exemplar priming on person judgments. Journal of Experimental
Social Psychology, 33, 4776.
Stein, N. L. (1982). The definition of a story. Journal of Pragmatics, 6, 487507.
Stein, N. L., & Glenn, C. G. (1979). An analysis of story comprehension in elementary school children. In R. O. Freedle (Ed.), New directions in discourse processing. Norwood, NJ: Ablex.
Storms, M. (1973). Videotape and the attribution process: Reversing actors and observers point
of view. Journal of Personality and Social Psychology, 27, 165175.
Strack, F. (1994). Response processes in social judgment. In R. S. Wyer & T. K. Srull (Eds.), Handbook of social cognition (2nd ed., Vol. 1, pp. 287322). Hillsdale, NJ: Lawrence Erlbaum Associates.
Strack, F., & Deutsch, R. (2002). Reflective and impulsive determinants of social behavior. Unpublished manuscript, University of Wrzburg, Germany.
Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating conditions of the human
smile: A nonobtrusive test of the facial feedback hypothesis. Journal of Personality and Social
Psychology, 54, 768777.
Strack, F., Schwarz, N., & Gschneidinger, E. E. (1985). Happiness and reminiscing: The role of time
perspective, affect and mode of thinking. Journal of Personality and Social Psychology, 49,
14601469.
Strack, F., Schwarz, N., Bless, H., & Kuebler, A. 1993). Awareness of the influence as a determinant of assimilation versus contrast. European Journal of Social Psychology, 23, 5362.
Suls, J. M. (1972). Two-stage model for the appreciation of jokes and cartoons: Informationprocessing analysis. In J. H. Goldstein & P. E. McGhee (Eds.), The psychology of humor (pp.
81100). San Diego, CA: Academic Press.
400
REFERENCES
REFERENCES
401
Wegner, D. M. (1994). Ironic processes of mental control. Psychological Review, 101, 3452.
White, R. W. (1959). Motivation reconsidered: the concept of competence. Psychological Review,
66, 297333.
Wills, T. A. (1981). Downward comparison principles in social psychology. Psychological Bulletin,
90, 245271.
Wolfson, M. R., & Salancik, G. R. (1977). Observer orientation and actor-observer differences in
attributions for failure. Journal of Experimental Social Psychology, 13, 441451.
Wyer, R. S. (1970). The quantitative prediction of belief and opinion change: A further test of a
subjective probability model. Journal of Personality and Social Psychology, 16, 559571.
Wyer, R. S. (1973). Category ratings as subjective expected values: Implications for attitude formation and change. Psychological Review, 80, 446467.
Wyer, R. S. (1974). Cognitive organization and change: An information-processing approach. Hillsdale, NJ: Lawrence Erlbaum Associates.
Wyer, R. S. (1977). The role of logical and nonlogical factors in making inferences about category
membership. Journal of Experimental Social Psychology, 13, 577595.
Wyer, R. S., & Adaval, R. (2003). Message reception skills in social communication. In J. O. Greene
& B. Burleson (Eds.), Handbook of communication and social interaction skills (pp. 291354).
Mahwah, NJ: Lawrence Erlbaum Associates.
Wyer, R. S., Adaval, R., & Colcombe, S. J. (2002). Narrative-based representations of social knowledge: Their construction and use in comprehension, memory and judgment. In M. P. Zanna
(Ed.), Advances in experimental social psychology (Vol. 34, pp. 131197). San Diego, CA: Academic Press.
Wyer, R. S., & Bodenhausen, G. V. (1985). Event memory: The effects of processing objectives
and time delay on memory for action sequences. Journal of Personality and Social Psychology,
49, 304316.
Wyer, R. S., Bodenhausen, G. V., & Gorman, T. F. (1985). Cognitive mediators of reactions to rape.
Journal of Personality and Social Psychology, 48, 324338.
Wyer, R. S., Bodenhausen, G. V., & Srull, T. K. (1984). The cognitive representation of persons and
groups and its effect on recall and recognition memory. Journal of Experimental Social Psychology, 20, 445469.
Wyer, R. S., & Budesheim, T. L. (1987). Person memory and judgment: The impact of information
that one is told to disregard. Journal of Personality and Social Psychology, 53, 1429.
Wyer, R. S., Budesheim, T. L., & Lambert, A. J. (1990). Person memory and judgment: The cognitive representation of informal conversations. Journal of Personality and Social Psychology, 58,
218238.
Wyer, R. S., Budesheim, T. L., Lambert, A. J., & Martin, L. L. (1989). Person memory: The priorities
that govern the cognitive activities involved in person impression formation. Unpublished manuscript, University of Illinois at Urbana-Champaign.
Wyer, R. S., Budesheim, T. L., Lambert, A. J., & Swan, S. (1994). Person memory and judgment:
Pragmatic influences on impressions formed in a social context. Journal of Personality and Social Psychology, 66, 254267.
Wyer, R. S., Budesheim, T. L., Shavitt, S., Riggle, E. J., Melton, R. J., & Kuklinski, J. H. (1991). Image,
issues and ideology: The processing of information about political candidates. Journal of Personality and Social Psychology, 61, 533545.
Wyer, R. S., & Carlston, D. E. (1979). Social cognition, inference and attribution. Hillsdale, NJ: Lawrence Erlbaum Associates.
Wyer, R. S., & Carlston, D. E. (1994). The cognitive representation of persons and events. In R. S.
Wyer & T. K. Srull (Eds.), Handbook of social cognition (2nd ed., Vol. 1, pp. 4198). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Wyer, R. S., Clore, G. L., & Isbell, L. M. (1999). Affect and information processing. In M. P. Zanna
(Ed.), Advances in experimental social psychology (vol. 31, pp. 177). San Diego: Academic
Press.
402
REFERENCES
Wyer, R. S., & Collins, J. E. (1992). A theory of humor elicitation. Psychological Review, 99, 663688.
Wyer, R. S., & Goldberg, L. (1970). A probabilistic analysis of the relationships among beliefs and
attitudes. Psychological Review, 77, 100120.
Wyer, R. S., & Gordon, S. E. (1982). The recall of information about persons and groups. Journal of
Experimental Social Psychology, 18, 128164.
Wyer, R. S., & Gruenfeld, D. H. (1995). Information processing in social contexts: Implications for
social memory and judgment. In M. P. Zanna (Ed.), Advances in experimental social psychology
(Vol. 27, pp. 4991). San Diego, CA: Academic Press.
Wyer, R. S., & Hartwick, J. (1980). The role of information retrieval and conditional inference
processes in belief formation and change. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 13, pp. 241284). New York: Academic Press.
Wyer, R. S., & Hartwick, J. (1984). The recall and use of belief statements as bases for judgments:
Some determinants and implications. Journal of Experimental Social Psychology, 20, 6585.
Wyer, R. S., Lambert, A. J., Budesheim, T. L., & Gruenfeld, D. (1992). Theory and research on person impression formation: A look to the future. In L. Martin & A. Tesser (Eds.), The construction of social judgment (pp. 336). Hillsdale, NJ: Lawrence Erlbaum Associates.
Wyer, R. S., & Martin, L. L. (1986). Person memory: The role of traits, group stereotypes, and specific behaviors in the cognitive representation of persons. Journal of Personality and Social
Psychology, 50, 661675.
Wyer, R. S., & Radvansky, G. A. (1999). The comprehension and validation of social information.
Psychological Review, 106, 89118.
Wyer, R. S., Shoben, E. J., Fuhrman, R. F., & Bodenhausen, G. V. (1985). Event memory: The cognitive representation of social action sequences. Journal of Personality and Social Psychology,
49, 857877.
Wyer, R. S., & Srull, T. K. (1989). Memory and cognition in its social context. Hillsdale, NJ: Lawrence
Erlbaum Associates.
Wyer, R. S., Swan, S., & Gruenfeld, D. H. (1995). Impression formation in informal conversations.
Social Cognition, 13, 243272.
Wyer, R. S., & Unverzagt, W. H. (1985). The effects of instructions to disregard information on its
subsequent recall and use in making judgments. Journal of Personality and Social Psychology,
48, 533549.
Yamauchi, H. (1988). Effects of actors and observers roles on causal attributions by Japanese
subjects for success and failure in competitive situations. Psychological Reports, 63, 619626.
Yeung, C. W. M. (2003). Unpublished data.
Yeung, C. W. M., & Wyer, R. S. (in press). Affect, appraisal, and consumer judgment. Journal of
Consumer Research.
Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology, 9, 127.
Zajonc, R. B. (1980). Feeling and thinking. Preferences need no inferences. American Psychologist,
35, 151175.
Zajonc, R. B., Murray, S. T., & Inglehart, M. (1989). Feeling and facial efference: Implication of the
vascular theory of emotion. Psychological Review, 96, 395416.
Zanna, M. P., & Cooper, J. (1976). Dissonance and the attribution process. In J. Harvey, W. Ickes,
& R. Kidd (Eds.), New directions in attribution research (Vol. 1, pp. 199217). Hillsdale, NJ: Lawrence Erlbaum Associates.
Zanna, M., & Rempel, J. K. (1988). Attitudes: a new look at an old concept. In D. Bar-Tal & A.
Kruglanski (Eds.), The social psychology of knowledge (pp. 315334). Cambridge, England:
Cambridge University Press.
Zigler, E., Levine, J., & Gould, L. (1967). Cognitive challenge as a factor in childrens humor appreciation. Journal of Personality and Social Psychology, 6, 332336.
Zillman, D. (1983). Connections between sex and aggression. Hillsdale, NJ: Lawrence Erlbaum Associates.
REFERENCES
403
Zillmann, D., & Bryant, J. (1982). Pornography, sexual callousness, and the trivialization of rape.
Journal of Communication, 32, 1021.
Zillman, D., & Cantor, J. R. (1976). A disposition theory of humor and mirth. In A. J. Chapman &
H. C. Foot (Eds.), Humor and laughter: Theory, research and application (pp. 93116). New York:
Wiley.
Zinn, H. (1980). A peoples history of the United States. New York: Harper-Collins.
Zuckerman, M. (1975). Belief in a just world and altruistic behavior. Journal of Personality and Social Psychology, 31, 972976.
Zwaan, R. A., Magliano, J. P., & Graesser, A. C. (1995). Dimensions of situation model construction
in narrative comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 386397.
Zwaan, R. A., & Radvansky, G. A. (1998). Situation models in language comprehension and memory. Psychological Bulletin, 123, 162185.
Author Index
A
Aaker, J. L., 70, 148, 302, 303, 380
Abramowitz, S. I., 164, 394
Achee, J. W., 312, 345-349, 363, 392
Abelson, R. P., 3, 16, 18, 21, 28, 116,
122, 126, 146, 149, 168, 293, 300,
307, 309, 340, 380, 391, 397
Adaval, R., 29, 34, 47, 63, 86-88, 115,
139, 142-144, 151, 236, 239-241,
243-250, 274, 307, 309, 333, 335,
342, 351, 386, 401
Ajzen, I., 358, 359, 363, 386
Albarracin, D., 62, 121, 277, 307,
338, 360, 362, 363, 380
Albrecht, J. E., 39. 381
Anderson, A., 28, 381
Anderson, J. R., 6, 9, 39, 49, 73, 128,
133
Anderson, N. H., 35, 165, 235, 236,
334, 342, 381, 391.
Anderson, R. C., 51, 381
Anthony, S., 208, 387
Apter, M. J., 191, 193, 195, 196, 198,
200, 381
Arkin, R. M., 66, 264, 265, 278, 292,
324, 381
Ashbrook, P. W., 344, 385
Atkinson, J. W., 204, 210, 393
B
Banaji, M., 13, 387
Banks, W. P., 114, 381
Barclay, J. R.10, 29, 84, 151, 381
Bargh, J. A., 7, 13, 23, 41, 47, 50-53,
55, 57, 62, 69, 73-75, 101, 136, 199,
309, 314, 315, 381, 383-385, 389,
392
406
AUTHOR INDEX
Collins, J. E., 161, 198, 201-203, 205207, 212, 215, 216, 218, 221, 223,
224, 390, 401
Conway, M., 274, 308, 328, 384
Cooper, J., 66, 402
Corbin, S., 191, 194, 393
Corty, E., 280, 398
Costley, C. L., 236, 384
Craik, F. I. M., 162, 282, 384
Crelia, R. A., 55, 392
Curiel, J. M., 17, 109-111, 396
Cutrona, C., 289, 384
D
Danilovics, P., 163, 386
Daubman, L. A., 307, 348, 349,
390
Davis, K. E., 35, 187, 390
DeCoster, J., 39, 398
DeHart, T., 317, 397
DePaulo, B. M., 179, 384
Deutsch, R., 8, 399
Devine, P. G., 304, 385
Dickson, P. R., 236, 393
Dienes, B. P. A., 340, 385
Dijksterhuis, A., 47, 385
Dijkstra, K., 81, 392
Duckworth, K. L., 314, 385
Dweck, C. S., 4, 286, 287, 299, 384,
385, 389
d'Ydewalle, G., 81, 397
E
Eagly, A. H., 48, 188, 309, 383-385
Ebbesen, E. B., 103, 385
Edell, J. A., 235, 385
Einstein, G. O., 340, 389
Ellis, H. C., 344, 385
Ellsworth, P. C., 179, 385
Emler, N., 164, 385
Englis, B. G., 241, 385
Engstler-Schooler, T. Y., 138, 397
Erdley, C. A., 286, 385
407
AUTHOR INDEX
F
Faber, R. J., 280, 398
Fazio, R. H., 83, 314, 315, 358, 385
Fedorikhin, A., 66, 327, 338, 398
Fernandez, G., 163, 386
Festinger, L., 66, 276, 386
Fiedler, K., 386
Fischoff, B., 285, 386
Fishbein, M., 235, 322, 358, 363, 386
Fishkin, J., 164, 386
Fiske, S. T., 36, 47, 217, 303, 307,
309, 380, 386, 400
Fletcher, G. J. O., 163, 279, 386, 396
Fong, C. P. S., 302, 386
Fong, G. T., 279, 396
Foot, H. C., 191, 384
Forgas, J. P., 295, 308, 312, 314-316,
318-320, 322, 357, 358, 386
Franks, J. J., 29, 84, 151, 383
Freud, S., 191, 193, 204, 218, 386
Friedman, H. S., 179, 384
Fry, P. S., 292, 387
Fu, J. H., 287, 384
Fuhrman, R. W., 28, 113, 114, 132,
391, 387, 402
Fung, H., 145-147, 385
G
Gaelick, L., 28, 186, 187, 387
Galambos, J. A., 19, 382
Garcia, M., 314, 385
Garnham, A., 29, 84, 387
Garrod, S. C., 28, 381
Gasper, K., 338, 387
Gelfand, M. J., 291, 400
Gerard, L. D., 29, 396
Gerbner, G., 281, 387
Gerrod, S. C., 28, 381
Ghosh, R., 292, 387
Gick, M. L., 130, 387
Gifford, R. K., 282, 388
Gilbert, D. T., 102, 387
Gilligan, C., 183, 387
Gilligan, S. G., 316, 318, 320, 382
408
AUTHOR INDEX
J
Jackson, D. D., 160, 186, 400
Jacoby, L. L., 54, 280, 390
Janis, I. L., 362, 390
Jasechko, J., 54, 280, 390
Jetten, J., 56, 392
Johnson, M., 307, 390
Johnson, M. K., 84, 280, 383, 390
Johnson, N. S., 4, 19, 392
Johnson-Laird, P. N., 16, 80, 81, 390,
397
Johnston, S., 66, 264, 265, 278, 292,
324, 381
Jones, C. R., 47, 49, 52, 321, 389
Jones, E. E., 35, 187, 390
K
Kahneman, D., 35, 64, 257, 258, 260,
263, 264, 280, 390, 400
Kane, T. R., 191, 300
Kardes, F. R., 314, 315, 385
Karp, L., 311, 390
Kashima, Y.,39, 389
Katz, L. B., 10, 388
Kelley, C. M.,54, 280, 390
Kelley, H. H., 187, 390
Keniston, K., 164, 386
Kihlstrom, J. F., 318, 390
Kim, B., 294, 295, 394
Kim, K. H., 24, 58, 59, 395
Kinder, D. R., 307, 309, 380
King, B. T., 362, 390
Kintsch, W., 16, 39, 81, 390
Kirkendol, S. E., 385
Kitayama, S., 290-292, 388, 392
Klein, S. B., 38, 132, 352, 391, 398
Klinger, E., 344, 391
Klumpp, G., 282, 397
Koestler, A., 193, 196, 391
Kohlberg, L., 164, 391
Koomen, W., 52, 63, 321, 399
Kosslyn, S. M., 9, 81, 391
Kramer, G. P., 307, 351, 382
Krantz, S., 320, 386
409
AUTHOR INDEX
Kruley, P., 81, 387
Kuebler, A., 55, 399
Kuklinski, J. H., 251, 253, 401
Kumar, P.A., 27, 163, 166, 168, 388
Kunda, Z., 123, 265, 278, 279, 391,
396, 398
L
LaFave, L., 191, 193, 391
Lambert, A. J., 28, 163, 168, 169,
172, 176, 178, 401, 402
Lampel, A. K., 236, 391
Langer, E. J., 300, 391
Langston, W. E., 81, 387
LaPiere, R. T., 304, 391
Latane, B., 55, 388
Lazarus, R. S., 82, 154, 328, 391
LeDoux, J. E., 310, 391
Lee, A. Y., 70, 148, 302, 303, 380
Lee, H. K., 291, 393
Lefcourt, H. M., 191, 391
Leggett, E. L., 4, 286, 299, 385
Lehman, D. R., 290, 292, 388
Leirer, V. O., 10, 388
Lepore, L., 304, 391
Lepper, M. R., 102, 256, 257, 285,
396
Lerner, M. J., 66, 269, 391
Levine, J., 205, 402
Levine, S. R., 307, 325-327, 355, 391
Liang, C. H., 146, 147, 393
Liberman, A., 48, 383
Lichtenstein, E. H., 19, 383
Lichtenstein, M., 168, 399
Lindem, K., 85, 387
Lindsay, D. S., 54, 280, 390
Locke,H. L., 296, 391
Lockhart, R. S., 162, 282, 384
Loftus, E. F., 10, 24, 39, 113, 138,
146, 384, 389, 391, 392
Loftus, J., 38, 132, 352, 391
Logan, G. L., 36, 40, 392
Loken, B. A., 11, 122, 392
Lombardi, W., 41, 50-52, 55, 69, 199,
381, 389, 392
410
AUTHOR INDEX
O
OGuinn, T. C., 54, 101, 280, 281,
395, 398
Oishi, S., 4, 292, 293, 385
Ortony, A., 308, 395
Ostrom, T. M., 35, 65, 395
Ottati, V. C., 307, 337, 395
P
Palmer, S. J., 138, 392
Parducci, A., 35, 65, 395
Park, J. W., 24, 58, 59, 167, 395
Parrott, G., 316, 317, 395
Pavelchak, M.A., 36, 309, 386
Peng, K., 291, 393
Pennington, N., 4, 232-234, 255,
388, 395
Peters, M. D., 307, 309, 380
Peterson, D., 163, 386
Petty, R. E., 55, 307, 344, 346, 357,
383, 395, 400
Pham, M. T., 307, 327, 395
Picek, J. S., 122, 395
Pietromonaco, P., 53, 57, 62, 381
Powell, M. C., 314, 315, 385
Pracejus, J., 307, 395
Pratkanis, A. R., 352, 308
Pratto, F., 309, 314, 315, 381
Prinz, W., 75, 395
Pylyshyn, Z. W., 9, 81, 395
Pyszczynski, T. A., 66, 395
R
Rabe, C., 339, 340, 342, 351, 382
Radvansky, G. A., 17, 29, 81, 95, 98,
100, 108-111, 136, 396, 402, 403
Raghubir, P., 283, 284, 393
Rappoport, J., 4, 144, 148, 392
Ratcliff, R., 39, 396
Raymond, P., 314, 381
Read, D., 396
Read, S., 19, 259, 295, 382, 393
Reckman, R. F., 275, 276, 387
AUTHOR INDEX
Reeder, G. D., 163, 386
Regan, D., 261, 396
Reich, C. M., 21, 122, 380
Reid, L. D., 304, 396
Rempel, J. K., 309, 325, 360, 402
Renwick, S., 164, 385
Reyes, R. M., 29, 255, 396
Reynolds, R. E., 51, 381
Rhee, E., 291, 396
Rholes, W. S., 47, 49, 52, 144, 321,
389
Riggle, E. J., 251, 253, 401
Rips, L. J., 113, 383
Ritov, I., 302, 393
Rittenauer-Schatka, H., 282, 397
Robinson, G., 307, 390
Roselli, F., 356, 396
Rosen, N. A., 272, 295, 396
Rosenfield, D., 294, 399
Ross, L. 102, 133, 255-257, 280, 285,
394, 396
Ross, M., 4, 274, 275, 279, 384, 396
Rothbart, M., 168, 399
Rubin, D. C., 3, 396
Russell, D., 289, 384
Rusting, C. L., 317, 397
S
Sabini, J., 316, 317, 395
Sacks, R., 287, 389
Salancik, G. R., 261, 401
Sanbonmatsu, D. M., 314, 315, 387
Sandel, T. L., 146, 393
Sanford, A. J., 28, 381
Sanna, L. J., 285, 396
Santioso, R., 279, 396
Sawyer, J. D., 126, 127, 387
Schachter, S., 310, 397
Schaeken, W., 81, 397
Schallert, D. L., 51, 381
Schank, R. C., 3, 18, 28, 116, 126,
146, 149, 293, 340, 397
Schick, C., 267, 397
Schneider, W., 7, 397
Schooler, J. W., 138, 397
411
Scott, C. K., 28, 397
Schwartz, A., 302, 393
Schwarz, N., 8, 35, 55, 64, 262, 263,
280, 282, 283, 285, 304, 307, 308,
311, 312, 322-328, 338-340, 342,
343, 347, 350, 351, 354-356, 382,
384, 391, 394, 396, 397,399
Semenik, R. J., 280, 398
Sentis, K. P., 122, 397
Seta, J. J., 55, 392
Setterlund, M. B., 306, 312, 313, 317,
318, 335, 343, 394
Shalker, T. E., 311, 390
Shanteau, J. A., 334, 400
Shavitt, S., 251, 253, 401
Sheehan, P. W., 250, 398
Shepard, R. N., 81, 398
Sheppard, L. A., 307, 351, 382
Sherman, B. R., 279, 398
Sherman, J. W., 132, 303, 304, 388,
398
Sherman, S. J., 122, 256, 280, 395,
398
Shevell, S. K., 113, 393
Shiffrin, R. M., 7, 122, 395, 397
Shiv, B., 66, 327, 338, 398
Shoben, E. J., 18, 113, 114, 395, 402
Shurcliff, A., 201, 398
Shrum, L. J., 54, 101, 280, 281, 395,
398
Signorielli, N., 281, 387
Simmons, C. H., 66, 269, 391
Simons, A., 282, 397
Simonson, I., 69, 70, 72, 148, 302,
303, 383
Sinclair, R. S., 356, 357, 398
Singer, J., 310, 397
Singer, M., 28, 81, 231, 387
Skowronski, J. J., 240, 398
Skov, R. B., 256, 398
Skurnik, I. W., 63, 394
Slovic, P., 280, 390
Small, E. M., 284, 396
Smith, E. R., 6, 22, 26, 39, 166, 235,
307, 339, 383, 398
Smith, S. M., 307, 344, 357, 400
412
AUTHOR INDEX
413
AUTHOR INDEX
Wood, W., 48, 384
Woolam, D., 207, 396
Worth, L. T., 307, 342, 343, 356, 392
Wyer, R. S., 4, 5, 8-10, 12, 13, 16-19,
21, 22, 24, 26-29, 31, 32, 34-37, 3943, 47, 50, 52, 54, 57-62, 64, 65, 6772, 74, 75, 86-89, 91, 95, 98, 100,
101, 103, 104, 109, 111, 113-122,
124, 126-130, 132-136, 139, 142144, 151, 153, 157-159, 161-163,
165-167, 172, 173, 176, 178-181,
183, 185-187, 198, 199, 201-203,
205-207, 212, 215, 218, 221, 223,
224, 236, 239-241, 243-251, 253,
269-274, 277, 280, 281, 288-293,
295, 302, 304, 307-309, 311, 312,
314, 317, 320, 322, 325-329, 331,
332, 335-338, 345-351, 353, 355,
359, 360, 362, 363, 366, 367, 380,
382-384, 386-392, 395-401
Y
Yamauchi, H., 292, 294, 402
Yeung, C. W. M., 328, 329, 331, 332,
350, 402
Yoon, S. O., 24, 58, 59, 395
Youmans, J. E., 164, 394
Z
Zacks, R. T., 29, 108-110, 396,
Zajonc, R. B., 210, 308, 324, 327, 402
Zanna, M. P., 66, 309, 325, 360, 402
Zigler, E., 205, 402
Zillman, D., 191, 193, 201, 203, 218,
282, 402, 403
Zinn, H., 148, 403
Zuckerman, M., 264, 292, 403
Zwaan, R. A., 28, 81, 392, 403
Subject Index
A
Accessibility of knowledge in
memory
determinants of, 41, 48-49
frequency of use, 49-52
recency of use, 49-52
effects on,
of group membership
salience, 69-72
of prior judgments, 63-64
of thought suppression, 56-57
effects of,
on activation of bipolar
concepts, 57-60
on assimilation and contrast,
57-60, 62-63, 65
on avoidance of negative
outcomes, 69-72
on belief formation, 60-62
on beliefs in a just world, 6669
on comprehension time, 97-98
on goals and motives, 65-72
on interpretation of information, 57-60
on own behavior, 72-75
on perceptions of social reality,
280-282
on reactions to rape, 66-69
on selective attention, 57-60
on use of stereotypes, 70-72
on verification, 98-100
generalizeability of effects, 52-53
nonconscious vs. conscious
effects of, 53-56
Adjustment for bias, 55-56, 336-338
Affect
adjustments for influences of,
336-338
416
SUBJECT INDEX
417
SUBJECT INDEX
politeness, 175-178
truthfulness, 160-165
Conversational style, effects of
on information seeking, 184-186
on liking for the speaker, 180-186
on perception of message implications, 179-180
Cultural identity, effects of
on avoidance of negative outcomes, 72
Cultural norms and values
activation of, 66-67, 72
effect of,
on avoidance of negative outcomes, 72
on implicit theories, 290-295
on representation of self in
memory, 144-147
on prediction of future outcomes, 292-294
on regulatory focus, 72
on self-identity, 144-149
on socialization practices, 144149
on theories of personal causality, 292-294
D
Dual-processing models, 35-36
E
Ease of retrieval, role of
in hindsight bias, 285
in perception of self and others,
282-284
in perception of social reality,
280-282
Emotions; see also Affect
communication of, in close relationships, 186-188
anticipation of, in decisionmaking, 301-303
Episode models; see also Situation
models
418
SUBJECT INDEX
H
Humor elicitation
comprehension processes in, 198204
comprehension difficulty,
effects of, 204-207, 218-220
situational and individual
differences in, 207-208
diminishment, role in, 195-198
effects of disparagement on, 218227
effects of distraction on, 210-211
effects of repetition on, 210-211
elaboration processes in,
and responses to sexist humor,
215-217
situational influences on, 212215, 220-227
theoretical effects of, 209-211,
220-222
role of stereotypes in, 218-227
by sexist jokes, 222-227
source effects on, 218-227
attributions of motivation and,
223-224
theories of, 192-209
Apters reversal theory, 195-199
arousal reduction theories,
192-193
comprehension-elaboration
theory, 198-209
incongruity resolution
theories, 193-194
superiority and disparagement theories, 193
I
Imagery; see Visual imagery
Images of politicians
influence of,
on information processing,
240-250
on interpretation of issue
stands, 251-250
Implicational molecules
conceptualization of, 20-21
construction of, 121-122
examples of, 267-270
Implicit theories
cultural differences in, 290-295
effects of, on causal attributions, 291-292
effects of, on predictions of the
future, 292-294
of ease of retrieval, 280-285
effects of perspective on, 288-290
incremental vs. entity, 286-288
motivational influences on, 278279
of personal achievement, 286-287
of personal causality, 292-294
role of,
in anticipated emotional
reactions, 301-303
in behavioral decision making,
301-303
in helping decisions, 299-301
in hindsight, 285
in marital satisfaction, 295-299
in perception of social reality,
280-282
in perception of social support,
288-290
in personal achievement, 286288
in prevention and promotion
focus, 301-303
in reconstructing the past, 274275
in self perception, 275-276
in stereotype-based behavior,
303-305
Impression formation
of communicators, 168-175
effect of social context on, 165174
in get-acquainted conversations,
180-186
in nonsocial situations, 166-167
of politicians, 248-254
419
SUBJECT INDEX
role of affect in, 349-351
role of conversational norms in,
165-186
role of visual imagery in, 140-254
Information processing models
conceptualization of
comprehension, 89-92
conscious vs. nonconscious
processes, 36-37
dual processing, 35-36
free flow of thought, 43-45
knowledge accessibility, 41
reminding, 41-45
storage and retrieval processes, 37-41
Wyer-Srull (1989) theory, 26-37
limitations of, 26-29
modifications of, 22, 89-92
role of consciousness in, 36-37
Information processing strategies
piecemeal vs. holistic, 234-253
effect of presentation format on,
234-235
effect of visual images on, 236253
Informativeness principle
in emotional communication,
186-188
in responses to news, 157-159
in social communication, 155160, 180-196
Innuendo effects, 157-159
J
Just world, belief in
as an implicational molecule,
122, 268
role, in reactions to rape, 66-69
M
Marital relationships
communication of emotions in,
186-188
role of implicit theories in, 295-299
420
SUBJECT INDEX
S
Self and self-knowledge
cultural differences in, 144-149
representation in memory of,
133-136
and social identity, 144-149
socialization influences on, 144149
and story-telling, 144-145
Self-perception,
theory of, 275-276
role of affect in, 352-356
Self-referent statements,
effects on person impressions,
175-178
Self-serving bias,
in implicit theories, 178-179
Situation models; see also Episode
models
conceptualization of, 16-17
content and structure of, 80-83
of events, 81-83
role of visual imagery in, 81, 83-88
Social knowledge
approach to investigating, 14-16
generalized entity representations of, 22-24
generalized event representations of, 17-21, 116-123
histories, 20
stories, 19-20
implicational molecules, 20-21,
121-123
implicit theories, 20
independent storage of, in
memory, 37-39
modalities of, 9
representation in memory of, 1014
referents of, 7-8
situation models of, 16-17
types of, 6-7
validation of, 98-103
Social support, perceptions of, 288290
421
SUBJECT INDEX
Socialization, influence of
on content of self-narratives, 144149
on social identity, 144-145, 148149
Stereotypes,
nonconscious activation of, 7375
role in behavioral decision
making, 303-305
role in humor elicitation, 218-227
suppression of, 56-57
Sufficiency principle,
Role in comprehension and
judgment, 48
T
Thought suppression, effects of, 5657
Truthfulness, perception of, 98-104,
153-154
Truthfulness principle, effects of
violating
on humor elicitation, 161-162
on information seeking, 163-165