The Psychology of Evaluation Affective Processes

in Cognition and Emotion 1st Edition Jochen

Musch download

https://ebookname.com/product/the-psychology-of-evaluation-
affective-processes-in-cognition-and-emotion-1st-edition-jochen-
musch/

Get the full ebook with Bonus Features for a Better Reading Experience on ebookname.com
Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Handbook of Cognition and Emotion 1st Edition Michael

D. Robinson (Ed.)

https://ebookname.com/product/handbook-of-cognition-and-
emotion-1st-edition-michael-d-robinson-ed/

Child Development at the Intersection of Emotion and

Cognition 1st Edition Susan D. Calkins

https://ebookname.com/product/child-development-at-the-
intersection-of-emotion-and-cognition-1st-edition-susan-d-
calkins/

Cognition and Chance The Psychology of Probabilistic

Reasoning 1st Edition Raymond S. Nickerson

https://ebookname.com/product/cognition-and-chance-the-
psychology-of-probabilistic-reasoning-1st-edition-raymond-s-
nickerson/

How Do Refrigerators Work Science in the Real World 1st

Edition Christina Wilsdon

https://ebookname.com/product/how-do-refrigerators-work-science-
in-the-real-world-1st-edition-christina-wilsdon/
Sustainable Protein Sources 1st Edition Sudarshan
Nadathur

https://ebookname.com/product/sustainable-protein-sources-1st-
edition-sudarshan-nadathur/

Papermaking Dard Hunter

https://ebookname.com/product/papermaking-dard-hunter/

Reporter Genes A Practical Guide 1st Edition Siobhan

Loughna

https://ebookname.com/product/reporter-genes-a-practical-
guide-1st-edition-siobhan-loughna/

Microcontroller Projects Using the Basic Stamp 2nd

Edition Al Williams (Author)

https://ebookname.com/product/microcontroller-projects-using-the-
basic-stamp-2nd-edition-al-williams-author/

Migrating to IPv6 A Practical Guide to Implementing

IPv6 in Mobile and Fixed Networks 1st Edition Marc
Blanchet

https://ebookname.com/product/migrating-to-ipv6-a-practical-
guide-to-implementing-ipv6-in-mobile-and-fixed-networks-1st-
edition-marc-blanchet/
Quantum Mechanics A Complete Introduction 1st Edition
Alexandre Zagoskin

https://ebookname.com/product/quantum-mechanics-a-complete-
introduction-1st-edition-alexandre-zagoskin/
 THE PSYCHOLOGY
OF EVALUATION
Affective Processes in Cognition and Emotion
This page intentionally left blank
 THE PSYCHOLOGY
OF EVALUATION
Affective Processes in Cognition and Emotion

Edited by

Jochen Musch
Karl Christoph Klauer
University of Bonn

LAWRENCE ERLBAUM ASSOCIATES, PUBLISHERS

2003 Mahwah, New Jersey London
Copyright © 2003 by Lawrence Erlbaum Associates, Inc.
All rights reserved. No part of this book may be reproduced in
any form, by photostat, microform, retrieval system, or any other
means, without the prior written permission of the publisher.

Lawrence Erlbaum Associates, Inc., Publishers

10 Industrial Avenue
Mahwah, New Jersey 07430

Cover design by Kathryn Houghtaling Lacey

Library of Congress Cataloging-in-Publication Data

The psychology of evaluation: affective processes in cognition and emotion,

edited by Jochen Musch and Karl Christoph Klauer.
Includes bibliographical references and indexes.
ISBN 0-8058-4047-8 (c)

Copyright information for this volume can be obtained by contacting the Library of Congress.

Books published by Lawrence Erlbaum Associates are printed on acid-free paper,

and their bindings are chosen for strength and durability.

Printed in the United States of America

10 9 8 7 6 5 4 3 2 1
Contents

List of Contributors ix

1 The Psychology of Evaluation: An Introduction 1

Jochen Musch and Karl Christoph Klauer

PART I: MECHANISMS, BOUNDARY CONDITIONS,

AND THEORIES OF AUTOMATIC EVALUATION

2 Affective Priming: Findings and Theories 7

Karl Christoph Klauer and Jochen Musch

3 The "Meddling-In" of Affective Information:

A General Model of Automatic Evaluation 51
Dirk Wentura and Klaus Rothermund

4 Reverse Priming: Implications for the (Un)conditionality

of Automatic Evaluation 87
Jack Glaser

5 The Hidden Vicissitudes of the Priming Paradigm

in Evaluative Judgment Research 109
Klaus Fielder
vi CONTENTS

PART II: EVALUATIVE JUDGMENTS

AND THE ACQUISITION OF EVALUATIONS

6 On the Acquisition and Activation of Evaluative

Information in Memory: The Study of Evaluative Learning
and Affective Priming Combined 139
Dirk Hermans, Frank Baeyens, and Paul Eelen

7 The Constructive Nature of Automatic Evaluation 169

Melissa Ferguson and John Bargh

8 The Hedonic Marking of Processing Fluency:

Implications for Evaluative Judgment 189
Piotr Winkielman, Norbert Schwarz, Tedra Fazendeiro,
and Rolf Reber

PART III: INDIVIDUAL DIFFERENCES

AND INDIRECT MEASURES OF EVALUATION

9 A Structural Analysis of Indirect Measures of Attitudes 219

Jan De Houwer

10 Beyond Verbal Self-Report: Priming Methods

in Relationship Research 245
Rainer Banse

11 Putting Process Into Personality, Appraisal, and Emotion:

Evaluative Processing as a Missing Link 275
Michael D. Robinson, Patrick T. Vargas, and Emily G. Crawford

PART IV: THE ROLE OF EVALUATION IN MOOD,

EMOTION AND BEHAVIOR

12 What Is Primed by Emotion

Concepts and Emotion Words? 307
Paula M. Niedenthal, Anette Rohmann, and Nathalie Dalle

13 The Parallel Worlds of Affective Concepts and Feelings 335

Gerald Clore and Stanley Colcombe
CONTENTS vii

14 Motor Compatibility: The Bidirectional Link

Between Behavior and Evaluation 371
Roland Neumann, Jens Forster, and Fritz Strack

Author Index 393

Subject Index 407

This page intentionally left blank
List of Contributors

Frank Baeyens, Department of Psychology, University of Leuven, Belgium

Rainer Banse, Department of Psychology, Humboldt University, Berlin, Germany
John Bargh, Department of Psychology, New York University, USA
Gerald Clore, Department of Psychology, University of Virginia, USA
Stanley Colcombe, Department of Psychology, University of Illinois, USA
Emily Crawford, Department of Psychology, University of Illinois, USA
Nathalie Dalle, Department of Psychology, University of Clermont-Ferrand, France
Jan De Houwer, Department of Psychology, University of Southampton, United
Kingdom
Paul Eelen, Department of Psychology, University of Leuven, Belgium
Tedra Fazendeiro, Department of Psychology, University of Denver, USA
Melissa Ferguson, Department of Psychology, New York University, USA
Klaus Fiedler, Department of Psychology, University of Heidelberg, Germany
Jens Forster, Department of Psychology, University of Wurzburg, Germany
Jack Glaser, Goldman School of Public Policy, University of California, Berkeley,
USA
Dirk Hermans, Department of Psychology, University of Leuven, Belgium
Karl Christoph Klauer, Department of Psychology, University of Bonn, Germany
Jochen Musch, Department of Psychology, University of Bonn, Germany
X LIST OF CONTRIBUTORS

Roland Neumann, Department of Psychology, University of Wurzburg, Germany

Paul Niedenthal, CNRS and the University of Clermont-Ferrand, France
Rolf Reber, Department of Psychology, University of Bern, Switzerland
Michael Robinson, Department of Psychology, North Dakota State University, USA
Anette Rohmann, Department of Psychology, University of Munster, Germany
Klaus Rothermund, Department of Psychology, University of Trier, Germany
Norbert Schwarz, Institute for Social Research, University of Michigan, USA
Fritz Strack, Department of Psychology, University of Wurzburg, Germany
Patrick Vargas, Department of Psychology, University of Illinois, USA
Dirk Wentura, Department of Psychology, University of Munster, Germany
Piotr Winkielman, Department of Psychology, University of Denver, USA
 Chapter 1
The Psychology of Evaluation:
An Introduction
Jochen Musch
Karl Christoph Klauer
Rheinische Friedrich-Wilhelms-Universitat Bonn

Evaluative and affective information processing in individuals has long

been a fundamental issue in social and cognitive psychology. The concepts
affect, valence, and attitude are all fundamentally linked to the most basic psy-
chological dimensions of good versus bad, positive versus negative, ap-
proach versus avoidance. The processing of stimulus valence, that is, the act
of determining the location of a stimulus on the affective dimension, is at
the heart of most current theories in cognition and emotion. Accordingly,
there has been a dramatic increase in interest in evaluative processes in the
late 20th century. Research on the nature of evaluative processes is now one
of the most rapidly growing endeavours of psychology and provides a unify-
ing focus for researchers working in a variety of disciplines such as social,
cognitive, and personality psychology.
Of particular interest has been the question whether evaluations are elic-
ited automatically, without intent, effort, and conscious awareness, and how
these evaluations influence subsequent information processing. Much of
this research has been conducted in the framework of the affective priming
paradigm and has sought to identify conditions under which evaluations
are processed automatically. Another major concern has been the conse-
quences of the activated evaluations on the perceiver's judgments and
behaviors. In addition, theoretical progress has revealed a number of sur-
prising parallels and connections between affective priming and other par-
adigms such as evaluative conditioning, Stroop-analogous tasks, the Simon
task, and the mere exposure paradigm, to name just a few. Finally, these in-
2 MUSCH AND KLAUER

sights have been used to develop unobtrusive measures of implicit attitudes

such as the Implicit Association Test (IAT) and other tasks based on the af-
fective priming paradigm. Out of this work, a common theoretical founda-
tion for evaluative information processing is beginning to emerge. The
present book seeks to provide an informative, scholarly, yet readable over-
view of what we know today about the nature of evaluation and affective
processes in cognition and emotion. It summarizes all recent advances in
the field, based on invited contributions from an eminent group of investi-
gators.
However, this is not simply an edited book in the usual sense. Rather, it is
the result of an ongoing discussion between a number of researchers
united by a joint and continued interest in the psychology of evaluation. Ac-
cordingly, the idea for this book evolved in a number of interconnected fo-
rums and is the result of extensive and fruitful theoretical exchanges that
took place on various occasions. In particular, from 1994 to 2000, several re-
search projects within the research program "Information processing in its
social context", which was initiated by Klaus Fiedler and Fritz Strack and im-
plemented by the German Research Foundation DFG, dealt with affective
and emotional processes. In June 1997, many of the contributors to the
present volume met in Konstanz at a symposium on affective priming
hosted by the social psychology division of the German Psychological Soci-
ety. In 1998, a special issue of the German Zeitschrift fur Experimentelle
Psychologie was devoted to affective priming. In December 1998, Jan De
Houwer and Dirk Hermans organized a workshop on affective processing
in Leuven, as a part of the Scientific Research Network "Acquisition and
representation of evaluative judgments and emotions". In June 1999, a sym-
posium on affective priming took place in Kassel at the occasion of the bi-
annual meeting of the German social psychologists. Finally, in March 2001,
a special issue of Cognition and Emotion dealt with the psychology of evalua-
tion, and in May 2001, another workshop supported by the Fund for Scien-
tific Research (Flanders, Belgium) was being held in Le Lignely.
The book is organised into four main sections. The first section deals
with the mechanisms, boundary conditions, and theories of automatic
evaluation processes. In a comprehensive review of findings obtained in
the affective priming paradigm, Klauer and Musch examine the evidence
for different mechanisms that have been proposed to underlie automatic
evaluation effects. Their review is structured around the impacts of the
major procedural variables: Prime variables, target variables, variables re-
lated to the prime-target pairs, list-context variables, and task-related vari-
ables. Major explanations of affective priming effects and their respective
empirical support are explored. The chapter concludes with a model of
the evaluative system that comprises a process of automatic evaluation ac-
tivation and two mechanisms, assumed to operate in parallel, that mediate
1. INTRODUCTION 3

the effects of activated evaluations on subsequent evaluative and noneval-

uative processing.
In chapter 3, Wentura and Rothermund discuss consequences of the au-
tomatic processing of valence that go beyond temporarily increasing the ac-
cessibility of associated concepts. Because of its global relevance, they
argue, the automatic processing of valence is strongly tied to response proc-
esses and is therefore likely to interrupt ongoing behavior by modifying the
probability of responses and redirecting behavior. Wentura and Rother-
mund make a strong case for this power of positive and negative stimuli to
meddle with ongoing processes of behavior formation, and present a theo-
retical framework in which this meddling-in of valent stimuli is seen as the
common ground to several automatic evaluation phenomena, including af-
fective priming, Simon, and Stroop effects.
In chapter 4, dealing with boundary conditions of automatic evaluation
processes, Glaser suggests that the automatic evaluative response is more
complex than a simple binary orientation. Specifically, he argues that auto-
matic evaluative responses can be automatically overridden when the prim-
ing stimulus is obtrusive and when accuracy motivation is high. Such find-
ings have implications for the important debate on the conditionality of
automatic evaluation. Glaser suggests that positions holding that automatic
evaluation will occur only for those stimuli toward which a reasonably
strong attitude is held, and positions holding that automatic evaluation is
unconditional and will occur with equal facility for strong and weak attitude
objects, may represent a false dichotomy. According to Glaser, all stimuli
can elicit an automatic evaluative response, but the strength of the attitude
will moderate the magnitude of the response.
In a thought-provoking chapter that concludes the more theoretically
oriented first section of the book (chapter 5), Fiedler points to hidden vicis-
situdes of the priming paradigm in evaluative judgment research in a re-
view that integrates findings from different priming paradigms ranging
from simple perception and word-recognition experiments to more com-
plex measures of decision making, manifest action, and goal orientation.
Fiedler outlines an enriched framework for studying priming effects on
evaluative judgments. Within this framework, he argues for the separation
of the evaluative judgment domain from the original paradigm of priming
in associative memory and highlights the distinction between afferent and
efferent process components.
The second section of the book investigates how evaluations are ac-
quired and how evaluative judgments are arrived at. Hermans, Baeyens,
and Eelen (chapter 6) highlight parallels between evaluative learning and
affective priming research and demonstrate the relevance of the study of
evaluative conditioning for a better understanding of the processes that are
involved in the cycle that encompasses the acquisition, the representation,
4 MUSCH AND KLAUER

and the activation of evaluative information in memory. In particular, they

show that associative acquisition procedures are capable of inducing eval-
uative changes that can be assessed by indirect measures of stimulus valence
such as the IAT and the affective priming procedure.
Ferguson and Bargh (chapter 7) argue against the assumption that an
evaluation consists of a single, affective representation associated in mem-
ory with the object. Studies showing that participants are able to automati-
cally evaluate novel, unfamiliar objects for which there are no previously
stored, corresponding representations, are consistent with the claim that
an evaluation represents a combination of numerous evaluations of various
features of the object. In addition, these studies suggest that automatic eval-
uations can be spontaneously and immediately constructed on the spot,
rather than being dependent on previous experience with, and conscious
appraisal of, the objects.
Winkielman, Schwarz, Fazendeiro, and Reber (chapter 8) propose that
one particular source of relevant information for the assessment of valence
is the fluency with which information about the target can be processed.
They propose that high fluency is associated with more favorable evalua-
tions and present empirical evidence consistent with their proposal. Subse-
quently, they explore possible reasons for the link between fluency and af-
fective reactions and discuss boundary conditions of the fluency-affect link.
The third section of the book considers indirect measures of individual
differences in the evaluation of social objects. In his structural analysis of
different indirect measures, De Houwer (chapter 9) focuses on four reac-
tion time tasks that provide potential ways to measure attitudes indirectly:
affective priming tasks, the emotional Stroop task, the Implicit Association
Test, and the affective Simon task. De Houwer presents a taxonomy of these
indirect measures of attitudes that reveals the essential similarities and dif-
ferences between them, as well as their relation to existing compatibility
tasks. He concludes by discussing the implications of this structural analysis
for the measurement of attitudes.
In chapter 10, Banse presents the results of his research on unobtrusive
measures of relationship quality. In an overview of experimental methods
that have been used in relationship research, it is shown how attachment
theories can be tested using indirect measures based on implicit associa-
tions and automatic evaluations that are not distorted by self-presentation
concerns. However, problems and limitations of the priming approach to
the investigation of the mental representation of relationships are also con-
sidered.
Robinson, Vargas, and Crawford (chapter 11) explore individual differ-
ences in evaluative processing. These differences in, for example, the speed
to recognize rewards or threats have the potential to influence emotional
behavior and experience. The authors therefore recommend to supple-
1. INTRODUCTION 5

ment self-report measures of personality traits by evaluative processing par-

adigms in order to develop a more complete understanding of how and
why people differ in their emotional reactions. An agenda is set for future
evaluative processing research in which individual differences play a central
role.
The relationship between evaluations on one hand and mood, emotion
and behavior on the other hand is explored in the fourth section of the
book. Niedenthal, Rohmann, and Dalle (chapter 12) review the research
on the automatic activation of evaluative responses and emotional states
and discuss the theoretical distinction between evaluations and emotional
responses. They report experimental evidence suggesting that the experi-
ence of emotional feelings and the activation of emotion concepts do not
have the same consequences for subsequent information processing. They
argue that a powerful theory of conceptual representation and processing
is required in order to understand the conditions under which the process-
ing of emotional words and concepts will result in the reexperience of some
affect.
Clore and Colcombe (chapter 13) discuss the mood-like effects that
sometimes occur when evaluative concepts are unconsciously primed. They
propose that moods and primed evaluative concepts have parallel effects,
because affective feelings and affective meaning obey the same rules. Both,
affective feelings engendered by mood states, as well as unconsciously
primed affective meaning can exert broad influence, because the implied
evaluation is not tied to a particular source. Moreover, they argue that the
information from affective mood and the information from affective prim-
ing share an important phenomenological quality that make them both es-
pecially compelling: in the absence of a salient, external source, they are ex-
perienced as internally generated.
In the concluding chapter, Neumann, Forster, and Strack (chapter 14)
discuss how emotions and attitudes serve adaptive functions in preparing
individuals to act in accordance with their needs and the requirements of
their environment. In their view, evaluative processes underlying emotions
and attitudes are directly linked to motor representations of either ap-
proach or avoidance responses. Approach or avoidance behavior is facili-
tated whenever compatible evaluative contents are processed. This link be-
tween evaluation and behavioral dispositions seems to be bidirectional in
nature, however, in the sense that the execution of approach or avoidance
behavior facilitates compatible evaluative processes. From this perspective,
approach and avoidance behavior is not only a consequence, but also a
cause of evaluative processes.
This page intentionally left blank
I: Mechanisms, Boundary Conditions,
and Theories of Automatic Evaluation
This page intentionally left blank
Affective Priming:
Findings and Theories
Karl Christoph Klauer
Jochen Musch
Rheinische Friedrich-Wilhelms-Universitat Bonn

Environmental events directly and automatically activate three interactive

but distinct psychological systems, responsible, respectively, for perceptual,
evaluative, and motivational analysis according to a model proposed by
Bargh (1997). These systems' automatic reactions to environmental events
influence perceptual interpretations of other people's behavior, they color
the evaluations of perceived objects and persons, and they inhibit or ener-
gize behavioral responses. Automaticity of a social phenomenon is a power-
ful finding because it implies that a person is not in conscious control of the
behavior or perception in question, cannot escape the automatic process-
ing once it is elicited by appropriate trigger stimuli, and ultimately cannot
be held fully responsible for the ensuing biases in perceptions, judgments,
and behavior (Bargh, 1999; Greenwald & Banaji, 1995).
Bargh (1997) distinguished preconscious from postconscious and goal-
dependent forms of automaticity. Preconscious effects require only the
presence of the triggering environmental event. They do not depend on a
prepared or receptively tuned cognitive state. In contrast, postconscious
and goal-dependent effects require special mental states in addition to the
mere presence of triggering objects or events. For example, goal-depend-
ent automaticity is conditioned on the individual intending to perform the
mental function, but given this intention, the processing occurs immedi-
ately and autonomously in the presence of the triggering stimulus (e.g.,
Pendry & Macrae, 1996).
The evaluative system comprises a process of automatic activation of eval-
uations that is triggered by the mere presence of an object in one's field of
8 KLAUER AND MUSCH

perception. A major tenet is therefore that the process of evaluation activa-

tion is preconscious. A second postulate is that the evaluative system is func-
tionally dissociable from the perceptual and the motivational system. The
evidence for both assumptions stems in large part from the affective prim-
ing paradigm.
Affective priming refers to the phenomenon that processing of an evalu-
atively polarized target word (e.g., love) is facilitated, that is, proceeds faster
and more accurately, when it is preceded by an evaluatively consistent prime
word (e.g., sunshine) rather than an evaluatively inconsistent prime word
(e.g., death). Since the seminal demonstrations by Fazio, Sanbonmatsu,
Powell, and Kardes (1986), more than 80 studies have been conducted in
that paradigm extending it in many ways and probing deeply into the dynam-
ics and mechanisms of evaluative processing. An overview of the studies re-
viewed in this chapter is given in the appendix. Affective priming effects con-
trast evaluatively consistent and inconsistent prime-target pairs: They are
defined by the interaction of prime and target valence. Both evaluatively con-
sistent and inconsistent prime-target pairs comprise positive and negative
words and thus, affective priming can be expected to provide a relatively
pure measure of evaluative processing uncontaminated by nonevaluative dif-
ferences between the sets of positive and negative stimuli in, for example, fa-
miliarity, informational diagnosticity, concreteness, and others. In recent
years, the paradigm has also received attention as providing an unobtrusive
measure for assessing evaluations (Banse, chap. 10, this volume; De Houwer,
chap. 9, this volume; Fazio, Jackson, Dunton, & Williams, 1995; Hermans,
Baeyens, & Eelen, chap. 6, this volume; Hermans, Vansteenwegen, Crombez,
Baeyens, & Eelen, in press; Otten & Wentura, 1999).
The present chapter is organized as follows. We begin with a review of
findings obtained in the affective priming paradigm. The review is struc-
tured around the impacts of the major procedural variables: Prime vari-
ables, target variables, variables related to the prime-target pairs, list-con-
text variables, and task-related variables. The next section introduces the
major explanations of affective priming effects and explores their respec-
tive empirical support vis a vis the just-reviewed findings. The chapter con-
cludes with a model of the evaluative system that comprises a process of au-
tomatic evaluation activation and two mechanisms, assumed to operate in
parallel, that mediate the effects of activated evaluations on subsequent
evaluative and nonevaluative processing.

AFFECTIVE PRIMING: FINDINGS

In Fazio's (1989) attitude theory, attitudes are seen as object-evaluation as-

sociations stored in memory. The strength of the association determines
the likelihood that the evaluation will be activated on encountering the atti-
2. AFFECTIVE PRIMING 9

tude object. The associative strength is termed attitude accessibility, and it is

measured by the speed with which attitude objects can be evaluated.
In Fazio et al.'s (1986) seminal demonstrations of affective priming,
there were two experimental phases. In the first phase, attitude accessibili-
ties of a number of attitude objects were assessed, and attitude objects high
and low in accessibility were selected for each participant to serve as primes
in the second phase of the experiment. In the second phase, evaluatively
consistent and inconsistent prime-target pairs as well as pairs with neutral
primes were presented. The participants' task was to classify the target word
as good or bad as fast as possible; a task that we will refer to as the evaluative
decision task. Targets were 10 clearly positive and 10 clearly negative adjec-
tives. Primes were nouns that referred to attitude objects. There was also a
baseline condition, in which primes were presumably neutral letter strings
such as BBB.
In Experiment 1, prime-target onset asynchrony (SOA) was 300 ms. A
priming effect emerged for primes high in attitude accessibility as assessed
and selected in the first phase, but not for primes low in attitude accessibil-
ity. This effect was replicated in Experiment 2. A new condition realized a
long SOA of 1000 ms and revealed an absence of affective priming effects
both for low-accessibility as well as for high-accessibility primes. In Experi-
ment 3, attitude accessibility was manipulated experimentally in the first
phase in which participants evaluated some attitude objects repeatedly,
thereby momentarily increasing these objects' attitude accessibility. Affec-
tive priming occurred at the short SOA (300 ms), but not at the long SOA,
and it was stronger for the primes with heightened attitude accessibility.

Generality of Affective Priming

Subsequent research confirmed and extended many of the original find-

ings. For example, whereas Fazio et al. (1986) asked participants to memo-
rize the prime word and to recite it aloud after they had evaluated the tar-
get, many subsequent studies showed that the affective priming effect is not
dependent on this requirement. In fact, in the majority of studies reviewed
below participants were simply asked to ignore the prime word. In Her-
mans, Van den Broeck, and Eelen's (1998) Experiment 4 as well as in
Bargh, Chaiken, Govender, and Pratto's (1992) Experiment 3, one group
of participants was required to memorize and recite the prime word,
whereas no such requirement was given for the participants of a second
group. In both experiments, affective priming effects were not affected by
the memory instruction. In a related experiment, Hermans, Crombez, and
Eelen (2000) found that affective priming effects were not moderated by
the requirement to hold a number of digits in memory for the duration of
each priming trial.
10 KLAUER AND MUSCH

Similarly, the effect can be obtained when there is no first phase of acces-
sibility assessment or manipulation and when primes are instead selected
from evaluation norms as having strong affective connotations. In fact, af-
fective priming effects have now been obtained with a variety of positively
and negatively evaluated stimuli as primes: With color slides of objects, per-
sons, and animals (which were also used as targets; Hermans, De Houwer,
& Eelen, 1994, Exp. 1); with photographs of self, significant others, and dis-
liked persons (Banse, 2000); with black-and-white line drawings (Giner-
Sorolla, Garcia, & Bargh, 1999); with non-words that were presented as
Turkish translations of positive or negative words before the priming trials
(De Houwer, Hermans, & Eelen, 1998); with positive and negative odors
(Hermans, Baeyens, & Eelen, 1998). Finally, the effect does not appear to
depend on the number of letters of prime and target words. Musch, Elze,
and Klauer (1998) compared priming by short and long primes crossed
with either short or long targets and found no moderating effect of prime
and target length on the affective priming effect.
In a related paradigm that is also often called an affective-priming para-
digm, Murphy and Zajonc (1993) presented smiling and scowling faces for
either 4 ms or 1000 ms followed immediately by Chinese ideographs. Lik-
ing and evaluative ratings of Chinese ideographs were found to be assimi-
lated to the affect displayed by the prime face under the brief, but not the
long presentation duration. Similarities and differences between affective
priming sensu Fazio et al. (1986) and related priming paradigms are dis-
cussed by Fiedler (chap. 5, this volume).
A recent innovation in procedure is the response window technique pro-
posed by Greenwald, Draine, and Abrams (1996). The response window
technique pushes participants toward responding within a narrow time
frame after the presentation of the target. As Greenwald et al. (1996)
pointed out, it has the major benefit of controlling for speed-accuracy
tradeoff problems by forcing all response latencies to be relatively similar,
thereby avoiding the dilution of the priming effect amongst both response
latency and accuracy. The dependent variable with this procedure is the
percentage of correct responses, and the technique typically leads to a large
increase in the size of accuracy priming. It has been used extensively by
Abrams and Greenwald (2000), Klauer and Musch (in press), Klinger, Bur-
ton, and Pitts (2000), Musch (2000), and Musch and Klauer (2001) to se-
cure affective priming effects under a wide range of conditions.

Prime Strength

An early debate concerned the question of whether affective priming is

found only for strongly accessible primes (Glaser, chap. 4, this volume).
Fazio et al. (1986) consistently found stronger effects for strongly accessible
2. AFFECTIVE PRIMING 11

than for weakly accessible primes given a short SOA of 300 ms. These ef-
fects were replicated by Bargh et al. (1992; Exp. 1). Interestingly, strong af-
fective priming effects also emerged for primes selected for consistency in
evaluations across participants rather than for strong accessibility (Bargh et
al., 1992; Exps. 1 & 2). Moreover, when the first phase of accessibility assess-
ment was separated from the priming trials by two days, the moderation by
accessibility was eliminated (Exp. 2), as further supported by a study by
Chaiken and Bargh (1993). Thus, it appears to make a difference whether
prime accessibility is assessed two days in advance of, or immediately be-
fore, the priming trials. A parsimonious explanation might be that the eval-
uation latencies used to assess accessibility are themselves only moderately
stable over a period of 2 days. If so, a regression effect would work to attenu-
ate the accessibility manipulation and any effects associated with it over that
period.
De Houwer et al. (1998) had participants learn translations of ostensibly
Turkish words that were unknown to the participants. The translations
were strongly positive or negative words. In three experiments with the
evaluative decision task and an SOA of 300 ms, the Turkish words engen-
dered affective priming effects when used as primes after the learning
phase. When their evaluation latencies, the measure of attitude accessibil-
ity, were assessed in a separate study (Exp. 5), the Turkish words' evalua-
tions were found to be less accessible than those associated with strongly
and even with only moderately positive and negative words. It was con-
cluded that affective priming can also be obtained even when attitude ac-
cessibility is low.
Just as attitude objects can differ in the strength of the object-evaluation
association, people may differ in the chronic accessibility of evaluations.
Building on this idea, Hermans, De Houwer, and Eelen (2001; Exp. 3) used
the Need to Evaluate Scale (NES) by jarvis and Petty (1996) to form two
groups that differed in their disposition to engage in evaluative respond-
ing. Under an SOA of 300 ms, affective priming emerged in evaluative deci-
sions of the group with high NES scores, but not in the group with low NES
scores.
Although the strength of the object-evaluation association is logically in-
dependent of the extremity of the evaluation itself, accessibility and evalua-
tion extremity tend to be correlated and therefore, studies that have manip-
ulated the extremity of prime evaluations are also relevant to this debate.
Giner-Sorolla et al. (1999) used line drawings that were either strongly or
weakly polarized in normative evaluations and found a tendency for affec-
tive priming to be stronger for strong rather than weak primes in their Ex-
periment 1. A number of studies have used the pronunciation task, in
which participants are asked to name the target as quickly as possible, and
manipulated prime accessibility and/ or extremity to obtain weak, moder-
12 KLAUER AND MUSCH

ate, and strong primes. Bargh, Charken, Raymond, and Hymes (1996) con-
sistently found equivalent affective priming effects for weak and strong
primes in naming latencies as did Giner-Sorolla et al. (1999) in their Exper-
iment 2. Glaser and Banaji (1999), on the other hand, obtained reversed
affective priming effects with extreme primes (i.e. faster responses for in-
consistent rather than consistent prime-target pairs), and normal or no af-
fective priming effects for trials based on moderately evaluated primes.
Using Bargh et al.' s (1996) procedures and stimuli, Glaser (chap. 4, this
volume) in contrast finds a weak normal effect for both weak and strong
primes. Finally, in an attempt to replicate the Bargh, et al. (1996) findings,
Klauer and Musch (2000; Exp. 3) did not obtain affective priming effects in
the pronunciation task irrespective of prime strength.
To summarize, there is some evidence that affective priming in the
evaluative decision task is moderated by attitude accessibility and related in-
dicators of prime strength. However, affective priming can also be obtained
with primes low in accessibility and strength. Based on their findings with
the pronunciation task, Bargh et al. (1996) have argued that the impact of
accessibility and extremity is modulated by the presence of a goal to evalu-
ate. Such a goal is clearly present in the evaluative decision task, where ac-
cessibility effects are typically found, and presumably absent in the pronun-
ciation task, where accessibility effects are often not found. The pattern of
findings obtained with the pronunciation task is however complex and will
be discussed further in subsequent sections.

Prime Presentation Parameters: Masked Priming

In this section, studies of two kinds are considered. The first kind is given by
studies that seek to establish the existence of affective priming effects when
primes are rendered invisble by suitable masks. The second class of studies
aim at demonstrating functional dissociations between visible and masked
priming effects by showing that they are differentially affected by experi-
mental manipulations.

Does Subliminal Affective Priming Occur? Studies demonstrating affective

priming when primes are rendered invisible by suitable masks provide
strong support for a role of automatic processes. Greenwald, Klinger, and
Liu (1989) report three experiments in which primes were masked di-
choptically by presentation of a random letter-fragment pattern to the
dominant eye, either rapidly following the prime (Exp. 1) or presented si-
multaneously with the prime (Exps. 2 and 3). The effectiveness of the mask-
ing procedure was demonstrated by participants' inability to discriminate
the left versus right position of a test series of words viewed under the same
masking conditions as the prime stimuli. In each experiment, there were
2. AFFECTIVE PRIMING 13

additional trials in which the prime words were presented without masks
and thus were clearly visible. In all three experiments, significant masked
priming effects were obtained with the chosen SOA of 500 ms, whereas af-
fective priming by visible primes was not obtained in Experiment 2.
Studies of unconscious cognition typically compare direct and indirect
measures of stimulus effects. Indirect effects are uninstructed influences of
the task stimuli on behavior. Direct effects measure instructed influences of
the relevant stimuli. In the Greenwald et al. (1989) studies, for example, in-
direct effects were the priming effects, and direct effects of prime words
were measured by the position discrimination task. Under the assumption
that the sensitivity of the direct measure for conscious stimulus effects is at
least as great as that of the indirect measure (Reingold & Merikle, 1988), an
indirect-without-direct effect pattern provides evidence for unconscious
stimulus processing.
Greenwald, Klinger, and Schuh (1995) proposed a regression method to
secure the desired indirect-without-direct effect pattern. In that technique,
an indirect measure is regressed on an appropriate direct measure. It is
tested whether the intercept of this regression function is significantly
larger than zero, implying indirect effects in the absence of a direct effect.
The new method has instigated considerable debate (Dosher, 1998; Green-
wald & Draine, 1998; Klauer, Greenwald, & Draine, 1998; Klauer & Green-
wald, 2000; Merikle & Reingold, 1998; Miller, 2000), and refinements of
the method were proposed by Klauer, Draine, and Greenwald (1998).
Using the new method, Greenwald et al. (1995) compared visible and
masked priming over an impressive series of experiments. With SOAs be-
tween 250 ms and 300 ms, there was no priming by visible primes and only
weak evidence for some amount of priming in the masked condition that
was present to the extent to which there was also evidence for direct effects
(i.e., zero intercept, but positive slope of the regression function relating
indirect and direct measure). These results thus provide little evidence for
subliminal affective priming.
Recent research has tended to use so-called sandwich masks rather than
dichoptical masking. In sandwich masking, the prime word is preceded by a
meaningless letter string (forward mask), and is followed by another mean-
ingless letter string (backward mask). Combining sandwich masking with
the response window technique, Draine and Greenwald (1998; cf. Green-
wald et al., 1996) studied affective priming for masked primes. Direct ef-
fects were assessed by participants' ability to discriminate between prime
words and letter strings of alternating Xs and Gs viewed under the same
masking conditions as were used in the priming trials (Exps. 1 to 3). In Ex-
periment 4, another direct measure was also realized, namely to make
evaluative decisions on the masked prime words. With SOAs between 34 ms
and 67 ms and masked primes presented for 17 ms, 33 ms, or 50 ms, inter-
14 KLAUER AND MUSCH

cepts significantly larger than zero were reliably obtained for the larger pre-
sentation durations of 33 ms and 50 ms in four experiments. An analysis of
the aggregated data revealed the desired indirect-without-direct pattern
also for the 17 ms prime duration (cf. also Klauer et al., 1998).
Thus, affective priming effects can be obtained with primes rendered in-
visible by suitable masks. Extending this work, Abrams and Greenwald
(2000) investigated which aspects of prime words were crucial for these
masked effects and found that letter strings composed of subword frag-
ments of earlier-viewed targets functioned as effective evaluative primes.
For example, after repeated evaluation of the targets angle and warm, the
nonword anrm acted as an evaluatively positive prime (Exp. 1), and there
were indications that when fragments were combined to yield words with
evaluations that were the opposite of the parent words, the evaluation of
the parent words prevailed in priming effects (Exp. 2). Thus, smile worked
as an evaluatively negative prime after repeated classification of the targets
smut and bile. In addition, priming was obtained only when primes did con-
tain parts of earlier targets (Exp. 3), a finding that is further discussed in
the section on open questions.

Functional Dissociations Between Masked and Visible Affective Priming. A

number of studies have compared visible and masked priming and found
that both kinds of priming effects were affected differently by experimental
manipulations. Musch (2000) used the response window technique and
sandwich masking to compare the so-called consistency proportion effect
for masked and visible primes. The consistency proportion effect refers to
the observation that affective priming effects tend to increase as the pro-
portion of evaluatively consistent prime-target pairs increases relative to the
proportion of inconsistent pairs (Klauer, RoBnagel, & Musch, 1997). With
an SOA of 71 ms, Musch (2000; Exp. 5) obtained the consistency propor-
tion effect for visible primes, thereby replicating findings of his previous ex-
periments in this series. However, when primes were masked, the effect did
not show although a substantial priming effect was found, and this differ-
ence in the consistency proportion effects for masked and visible primes
was significant.
Greenwald et al. (1996) contrasted visible and masked priming as a func-
tion of SOA using the response window technique and sandwich masks.
They found a rapid decline in masked priming for SOAs exceeding 100 ms,
whereas priming by visible primes was undiminished over SOAs varying
from 100 ms to 400 ms. A dissociation between visible and masked priming
was also demonstrated with respect to a so-called sequential effect in the
studies summarized by Greenwald et al. (1996). When the trial preceding
the current trial had presented an evaluatively inconsistent rather than con-
sistent prime-target pair, affective priming was diminished in the current
2. AFFECTIVE PRIMING 15

trial. This effect occurred with visible primes and an SOA of 150 ms, but was
absent when primes were masked by sandwich masks and SOA was 67 ms in
experiments that employed the response window procedure.
Another interesting dissociation has recently been reported by Banse
(2000; Exp. 2). Comparing priming engendered by pictures and names of a
liked person (Charlie Chaplin), a disliked person (Saddam Hussein), and a
neutral stranger, normal priming was found when primes were visible
whereas reversed effects emerged when primes were masked (SOA 42 ms;
evaluative decision task). Similarly, Hermans (1996), obtained indications
of reversed effects in masked priming in two of three studies under condi-
tions that led to normal priming effects when primes were visible (Hermans
et al., 1994, Exp. 1).
To summarize, indirect without direct effects of prime words can be reli-
ably produced when short SOAs and the response window technique are
used. In addition, local sequential effects and consistency proportion ef-
fects occur for visible, but not for masked primes.

Target Presentation Parameters

Only a few studies have looked at effects of target variables. Musch and
Klauer (2001) presented primes and targets simultaneously at different lo-
cations on the screen. For half of the participants, the location of the up-
coming target was signalled by an advance cue appearing 600 ms and 150
ms before target onset in Experiments 1 and 2, respectively; for the other
half, the advance cue was uninformative with respect to target location. In
both experiments, affective priming was observed only in the latter group
(i.e. under locational uncertainty), but not when participants could pre-
pare for the target's location.
While prime and target were always shown in different locations in these
experiments, Hermans et al. (2001; Exp. 3) compared centered versus
uncentered presentation of prime and target. In centered trials, primes
and targets were presented at the same central location on the screen with
an SOA of 300 ms. In uncentered trials, primes and targets were written on
different adjacent lines, one above the other. Affective priming was ob-
tained for the centered, but not for the uncentered trials. It appears from
these studies that affective priming effects are weakened when a perceptual
and/or attentional separation of primes and targets is supported by the
manner of presentation.
In an experiment by De Houwer, Hermans, and Spruyt (2001), half of
the participants were presented targets in a degraded fashion (e.g.,
%U%G%L%Y%) and the other half saw undegraded targets (e.g., UGLY).
Using the pronunciation task and an SOA of 250 ms, affective priming ef-
16 KLAUER AND MUSCH

fects were found in the degraded condition, but not in the standard unde-
graded condition.

SOA: The Time Course of Affective Pruning

Stimulus-onset asynchrony is an important moderator of affective priming

effects as already indicated by the original Fazio et al. (1986) studies.
Klauer, RoBnagel, and Musch (1997; Exp. 1) varied SOA in six steps (-100
ms, 0 ms, 100 ms, 200 ms, 600 ms, and 1200 ms) in a between-participants
design. SOA had a significant effect, and individually significant priming ef-
fects were found at SOAs 0 ms and 100 ms, but not at the other SOAs. These
findings were corroborated in a study by Hermans et al. (2001; Exp. 1) us-
ing a within-participants design and five levels of SOA (-150 ms, 0 ms, 150
ms, 300 ms, 450 ms). Individually significant priming effects emerged at the
short SOAs of 0 ms and 150 ms and were absent at the other SOA levels.
Using the pronunciation task and SOAs of 150 ms, 300 ms, and 1000 ms in a
second experiment, an individually significant priming effect was obtained
at the SOA 150 ms, but not at the longer SOAs. Taken together with the
above-reviewed findings by Greenwald et al. (1996; section on masked
priming), these different findings indicate that affective priming effects are
obtained most robustly for short SOAs well below 300 ms. Thus, the activa-
tion of prime evaluations appears to be quite short-lived.

List Context Effects

Is affective priming a local phenomenon that depends only on the charac-

teristics of the current trial, or is it modulated by the wider context in which
the prime-target pair is placed? As already explained, sequential effects of
the trial preceding the current one were in fact found with visible primes in
the experiments by Greenwald et al. (1996). Wentura (1999) also demon-
strated that responses to the target of a given trial are influenced by charac-
teristics of the preceding trial in the evaluative decision task. Specifically,
when the prime in the preceding trial matched the current target in va-
lence, responses were inhibited in two experiments with SOA 300 ms; a pat-
tern of results that is known as negative priming (e.g., Fox, 1995).
Klauer et al. (1997) manipulated the proportion of evaluatively consis-
tent prime-target pairs presented in the evaluative decision task. In a be-
tween-participants design, three proportions were realized in their Experi-
ment 2: 25%, 50%, and 75%. A second between-participants factor was SOA
(0 ms, 200 ms, and 1000 ms). As already explained, a consistency propor-
tion (CP) effect is given if affective priming increases as a function of con-
sistency proportion. A CP effect was found in the latency data for the 0 ms
2. AFFECTIVE PRIMING 17

SOA and in the error data for the 200 ms SOA, but not at the long SOA of
1000 ms. Musch and Klauer (1997) replicated the CP effect at SOA 0 ms.
In a series of experiments employing the response window procedure
and the evaluative decision task with SOA 71 ms, Musch (2000) further ex-
plored the CP effect. In a first experiment, the CP effect was again repli-
cated. The second experiment addressed the question whether local se-
quential contingencies were responsible for the effect. It has been argued
that participants can shift the weight given to the prime information on a
trial-by-trial basis. Specifically, they might rely more strongly on the prime
information when it has just been seen to provide valid information regard-
ing the target valence in the last trial, that is, when that trial presented an
evaluatively consistent prime-target pair, as is indeed suggested by the se-
quential effect observed by Greenwald et al. (1996; cf. Greenwald &
Rosenberg, 1978). In lists with a high CP, trials are frequently preceded by
trials with evaluatively consistent prime-target pairs, simply because such
pairs are generally more frequent when CP is high. Therefore, the CP effect
could reflect a trial-by-trial adjustment of the weight given to the prime in-
formation. In Experiment 2, CP was manipulated in three steps (25%, 50%,
and 75%) between participants. A subset of trials was, however, balanced
with respect to evaluative consistency versus inconsistency of the preceding
prime-target pair in each CP condition. Nevertheless, the CP effect
emerged in full strength even when only the balanced subset of trials was
considered. The notion that the CP effect reflects context-dependencies
that extend over a wider range than just the preceding trial was further sup-
ported by Experiment 3, in which a final block of trials realized a 50% CP
regardless of the CP that was given in the previous five blocks (either 25%,
50%, and 75%). The CP effect was found to be as strong in this last block as
in the ones preceding it. Because the actual CP was 50% in this block, a lo-
cal explanation of the effect is ruled out. In a fourth experiment, only in-
consistent prime-target pairs and pairs with neutral letter string primes
were employed to explore the extent to which the CP effect might elimi-
nate affective priming. Although the CP was 0% in this study, a residual af-
fective priming effect was nevertheless found. In a final experiment (Exp.
5), masked and visible primes were compared in a between-participants de-
sign that also varied CP. The usual CP effect emerged for visible primes, but
was absent in the masked priming groups, and this difference was reliable.
To summarize, the CP effect moderates affective priming, but cannot over-
ride it. In addition, it appears to rely on learning processes that extend over
a wider range of trials and require prime visibility to occur.
It can be concluded that visible affective priming in the evaluative deci-
sion task is modulated both by the local trial-by-trial context in which a
given trial is placed and by the global list context in which the trials are em-
bedded.
18 KLAUER AND MUSCH

Task

Another kind of context-dependency is given by task-dependence. The pat-

tern of findings differs pronouncedly between tasks. It is helpful to catego-
rize tasks in three groups: Tasks requiring the identification of certain tar-
get attributes, tasks that require affirmative and negative responses, and the
pronunciation task.

Identifying Target Attributes. Most of the studies have relied on the origi-
nal evaluative decision task in which the prime valence has a direct relation-
ship to the required response. When targets had to be classified with re-
spect to nonevaluative features, however, affective priming usually did not
occur. For example, Hermans, Van den Broeck, and Eelen (1998; Exp. 3)
contrasted evaluative decisions with a color-naming task. Targets appeared
in one of four colors, and in the color naming trials, participants were to
name the color in which the target was written. Affective priming effects
were found for the evaluative decision task, but not in the color naming
task, and this difference between the two kinds of trials was significant. Pre-
vious experiments in this series had used only the color-naming task and
did not obtain affective priming as was the case in two experiments con-
ducted by Rothermund and Wentura (1998) with the color-naming task.
In a related study, De Houwer, Hermans, Rothermund, and Wentura
(2000; Exp. 2) contrasted two groups. In the first group, participants cate-
gorized targets as persons versus animals (semantic classification); in the
second group, participants were required to make evaluative decisions on
these same targets. An affective priming effect was found for the evaluative
decision task, but not for the semantic-classification task. This difference
between tasks in the pattern of priming effects was significant. Similarly,
Klinger et al. (2000; Exp. 2) had members of different groups either classify
targets as denoting living versus nonliving things (animacy decisions) or
make evaluative decisions on targets. Affective priming effects were found
in evaluative decisions, but not in the animacy decisions. Conversely, the
animacy category of the prime elicited congruency effects in the animacy
decision task, but not in the evaluative decision task. That is, deciding
whether a target denoted a living versus nonliving thing was facilitated
when prime and target were congruent (both living versus both nonliving)
rather than incongruent (one living, the other nonliving) in this respect;
this kind of congruency had no effect on evaluative decisions, however.
Double dissociations of this kind constitute strong evidence for the task-
dependence of affective priming. In four experiments by Klauer and Musch
(in press), primes and targets could vary orthogonally with respect to their
valence and with respect to their values on a second dimension. The second
dimension was spanned, respectively, by two presentation locations of
2. AFFECT IVE PRIMING 19

prime and target stimuli (an upper vs. a lower line; Exp. 1), by two colors
in which prime and target were presented (blue vs. brown; Exp. 2), by let-
ter case (small vs. capital letters; Exp. 3), and by grammatical category (ad-
jective vs. noun; Exp. 4). In each experiment, one group was asked to
make evaluative decisions on targets, whereas members of the second
group were required to decide which of the two values of the second di-
mension was realized by a given target. In each experiment, affective
priming effects were found in evaluative decisions, but not for the group
making decisions on nonevaluative target features. Conversely, equivalent
nonevaluative congruency effects engendered by the prime value on the
second dimension were found in the group making decisions with respect
to that dimension, but not in the group making evaluative decisions across
all four experiments.

Tasks Requiring Affirmative and Negative Responses. Thus, affective prim-

ing effects were reliably found only when the task itself was focused on eval-
uations. An exception to this rule may, however, be given by tasks that re-
quire affirmative and negative responses as in yes/no decisions. For
example, in the lexical decision task, targets can be words or meaningless
letter strings. Participants are asked to decide whether or not the target is a
word and to respond "yes, word" in the first case and "no, not a word" in the
second case. In trials in which primes and targets are words, evaluative con-
sistency of prime and target can be manipulated to test for affective prim-
ing effects. Klinger et al. (2000, Exp. 2) used masked primes and found no
evidence for affective priming in the lexical decision task. With visible
primes, however, Wentura (1998, 2000) reported effects of affective prim-
ing in that task. In addition, priming was moderated by the assignment of
"yes" and "no" responses to words and nonwords, respectively. In the usual
"word=yes" condition, Wentura (1998, 2000) obtained affective priming ef-
fects in three experiments. These effects were however reversed in ten-
dency (Wentura, 1998) and significantly (Wentura, 2000; Exp. 2) when par-
ticipants were required to respond "no" if the target was a word and "yes" if
it was a nonword, A similar reversal of affective priming effects from trials
requiring yes-responses to those requiring no-responses was reported by
Klauer and Stern (1992) in an early study based on grammatical classifica-
tions of prime-target pairs.
In Experiments 5 to 8 by Klauer and Musch (in press), primes and tar-
gets varied with respect to evaluations and orthogonally with respect to the
same nonevaluative second dimensions used in Experiment 1 to 4, respec-
tively, as just detailed previously. Participants were required to compare
primes and targets with respect to valence (first group) or with respect to
their values on the second dimension realized in each experiment (second
group) and to respond "yes, same" in the case of a match and "no, differ-
20 KLAUER AND MUSCH

ent" in the case of a mismatch. Pervasive affective priming effects (i.e.

facilitatory effects of a match of prime and target valence), were found in
the group making nonevaluative comparisons for trials requiring "yes,
same" responses. The affective priming effects were eliminated or reversed
in tendency for the "no, different" response. Equivalent congruency effects
of matches versus mismatches on the second, nonevaluative dimension
were not found in the group making evaluative comparisons. In a ninth
study, participants made decisions on the grammatical category of the tar-
get and were required to respond "yes" for one category (e.g., in the case of
a noun) and "no" for the other category of targets (e.g., in the case of an ad-
jective). Again, affective priming was found for trials requiring "yes" re-
sponses, and the effect was reversed in tendency for trials requiring "no" re-
sponses. This is a remarkable finding as there were no effects of the
irrelevant evaluations in a previous study in this series (Exp. 4) that used ex-
actly the same task with the only difference that the responses were directly
labelled adjective and noun, respectively. Taken together, there is some evi-
dence that the affective match (mismatch) of prime and target can bias
nonevaluative "yes" ("no") responses.

The Naming Task. Finally, a number of studies have employed the pro-
nunciation task to study affective priming. Hermans et al. (1994; Exp. 2)
found an affective priming effect using that task and an SOA of 300 ms. As
already mentioned, Hermans et al. (2001; Exp. 2) manipulated SOA in
three levels (150 ms, 300 ms, 1000 ms) in a study that found no overall af-
fective priming effect and no main effect of SOA. An individually signifi-
cant effect emerged at the short SOA, however. Bargh et al. (1996; Exps. 1
to 3) and Giner-Sorolla et al. (1999; Exp. 2) consistently observed affective
priming effects in the pronunciation task for weak and strong primes using
SOAs between 250 ms and 300 ms. Klauer and Musch (2001) ran five statis-
tically powerful experiments with the pronunciation task that varied prime-
set size (10 primes vs. infinite set size) and target-set size (2, 10, infinite;
SOA 200 ms; Exp. la and 1b), SOA (0 ms, 50 ms, 100 ms; Exp. 2), prime
strength (weak vs. strong; SOA 300 ms; Exp. 3) and language of primes and
targets (English vs. German; SOA 300 ms; Exp. 4) and did not observe affec-
tive priming in any of these experiments although traditional semantic
priming (Neely, 1991) was obtained. De Houwer, Hermans, and Spruyt
(2001) found affective priming when targets were presented in a degraded
manner, but not under the standard undegraded presentation mode with
SOAs of 250 ms. Similarly, De Houwer et al. (1998; Exp. 2) did not obtain
an affective priming effects for stimuli that were associated with positively
and negatively valenced words in the experimental context although these
stimuli engendered priming effects in evaluative decisions (Exps. 1, 3, and
4). Finally, in a series of five experiments with SOA 150 ms and the pronun-
2. AFFECTIVE PRIMING 21

ciation task, Glaser and Banaji (1999) found reversed affective priming ef-
fects by strong primes (Exps. 1 to 5), and normal (Exps. 2 and 3) or no
(Exps. 4 and 5) affective priming effects for moderately strong primes.
Using stimuli and procedures more similar to the Bargh et al. (1996) exper-
iments, Glaser (chap. 4, this volume) in contrast finds a weak normal prim-
ing effect for both weak and strong primes. There are also many unpub-
lished studies that failed to obtain an affective priming effect in the
pronunciation task (e.g., Hermans, 1996). We are not aware of published
studies that have compared the pronunciation task to other tasks within
one experiment.
This review of the effects of task settings supports the following conclu-
sions:

• When targets have to be classified into a small number of categories,

there are priming effects engendered by task-relevant prime catego-
ries, but prime categories that are not task-relevant do not give rise to
priming effects. There are no differences between evaluative and non-
evaluative classifications in this respect.
• When the task requires affirmative or negative responses, affective
priming effects occur for trials requiring affirmative responses, and
they tend to be eliminated or reversed for trials requiring negative re-
sponses.
• The pattern of findings obtained with the naming task is mixed and
complex, and there appear to be as yet unidentified factors determin-
ing whether normal priming effects, reversed effects, or no effects are
obtained.

AFFECTIVE PRIMING: EXPLANATIONS

Several mechanisms have been considered as underlying affective priming

effects (cf. De Houwer, chap. 9, this volume; Fiedler, chap. 5, this volume;
Wentura & Rothermund, chap. 3, this volume). The most prominent are
(1) a mechanism based on an analogy with the semantic priming paradigm
(Neely, 1991) and the notion of spreading activation, (2) a mechanism
based on an analogy with the Stroop paradigm (MacLeod, 1991) and the
notions of selective attention and response competition (Klauer, 1998;
Klauer, RoBnagel, & Musch, 1997; Musch, 2000; Rothermund & Wentura,
1998), and (3) an affective-matching hypothesis proposed by Klauer (1991;
Klauer, 1998; Klauer & Stern, 1992). In the account by spreading activation,
primes exert an influence by preactivating related target nodes in a lexical
or semantic network. In the Stroop mechanism, irrelevant evaluations exert
an effect by virtue of the observed response having an evaluative compo-
22 KLAUER AND MUSCH

nent. In the affective-matching mechanism, irrelevant evaluations exert an

effect by virtue of biasing yes/no-responses.

The Account by Spreading Activation

An early explanation of affective priming drew an analogy between affective

priming and semantic priming (Neely, 1991) using the concept of spread-
ing activation. Roughly, perceiving the prime is assumed to activate its rep-
resenting node in a lexical or semantic network (Bower, 1991; Fazio et al.,
1986), and the activation then spreads to nodes of evaluatively consistent
targets, but not of inconsistent targets, thereby facilitating processing of the
target whenever prime and target are evaluatively consistent. Spreading ac-
tivation can thereby account for affective priming effects.
Yet, there are many findings that are difficult to reconcile with the no-
tion of spreading activation. According to the account by spreading activa-
tion, pervasive and context-independent facilitation of target processing
should be a consequence of evaluative consistency of prime and target at
least when lexical processing of the target is required. This expectation is
not borne out by the findings of strong task-dependence (e.g., De Houwer
et al., 2000; Klauer & Musch, in press; Klinger et al., 2000; Rothermund &
Wentura, 1998) reviewed earlier. Nor does it agree well with the finding
that in tasks with affirmative and negative responses, the effects of evalu-
ative consistency are less pronounced and in tendency reversed when nega-
tive responses are required (Klauer & Musch, in press; Klauer & Stern,
1992; Wentura, 1998, 2000).
Similarly, list-context effects are also difficult to explain from the per-
spective of a spread of activation. Thus, the sequential effects observed by
Wentura (1999) and Greenwald et al. (1996) as well as the CP effects found
by Klauer et al. (1997; Musch, 2000; Musch & Klauer, 1997) at short SOAs
cannot be accounted for by this mechanism. For these and other, more the-
oretical reasons (e.g., Bargh, 1997; Klauer & Musch, 2001), the account by
spreading activation has been abandoned by most researchers in the field.
Hermans et al. (1998) proposed an affective-motivational account of af-
fective priming according to which evaluative inconsistency, but not evalu-
ative consistency, of two incoming stimuli delays any kind of cognitive proc-
essing, irrespective of the participants' current goals. This mechanism also
leads one to expect affective priming effects on target processing regardless
of the nature of the task that participants are required to perform on the
target word, and it encounters the same difficulties as the account by
spreading activation.
Spreading activation is, however, still relevant as a methodological ca-
veat. Spreading activation is often argued to underlie the reliable and
largely task-independent priming effects observed for strongly associated
2. AFFECTIVE PRIMING 23

primes and targets (e.g., bread and butter; Neely, 1991). Many researchers
have therefore taken explicit measures to control for associative relatedness
in prime-target pairs used in affective priming studies. In our own research,
we prepare large and diverse pools of positive and negative words, from
which each participant's list is randomly sampled. This is to ensure that
evaluative consistent and inconsistent word pairs do not differ systemati-
cally and substantially within and across participants in associative related-
ness. Another possibility is to look at each list of prime-target pairs and to
screen out associatively related pairs before presentation (e.g., Rother-
mund & Wentura, 1998). In some circumstances, a third possibility is to an-
alyze the data by items rather than by participants. If an affective priming
effect goes back to a few targets that by accident are paired with highly asso-
ciated primes for all participants, then the effect should not generalize over
targets although it might generalize over participants. That is, instead of
computing an affective priming effect for each participant by averaging
over items, the analysis by items computes the effect for each target by aver-
aging over participants (Clark, 1973; for examples in the context of affec-
tive priming, see Wentura, 2000). The target-wise priming effects are then
subjected to an analysis of variance with targets taking the role of partici-
pants to see whether the effects generalize over items or are concentrated
on a few unusual, or unusually paired, targets.

The Stroop Mechanism

In the classical Stroop task, words are presented in different colors. Naming
the color is delayed when the word itself denotes a color that differs from
the one that the word is written in (MacLeod, 1991). There are many vari-
ants of the task, some of which are structurally similar to the evaluative deci-
sion task in affective priming research. The so-called flanker task in particu-
lar works with two sets of stimuli (e.g., the letters H and K versus S and C),
which are assigned different responses (e.g., pressing the left key for H and
K, and pressing the right key for S and C). Irrelevant letters from the wrong
response set interfere with the response to the target (Eriksen & Eriksen,
1974) when they flank the target letter. Flanker effects are also found with
words from different categories (Shaffer & LaBerge, 1979) and the task to
categorize the target word (e.g., as a piece of furniture vs. a metal). When
primes are identified with flankers, affective priming in the evaluative deci-
sion task can thus be seen as an instance of flanker effects (cf. De Houwer,
chap. 9, this volume; Klauer et al., 1997; Rothermund & Wentura, 1998;
Wentura & Rothermund, chap. 3, this volume).
Adapting a prominent model of Stroop effects (Logan & Zbrodoff,
1979) to the present case, Musch (2000) assumed that both prime and tar-
get evaluations are activated and integrated in a random-walk process on a
24 KLAUER AND MUSCH

decision dimension related to the responses. Two response thresholds are

located on the decision dimension, and a response is made as soon as the
accumulated evidence falls outside the interval spanned by the two thresh-
olds. At each point in time, the available evidence is given by a weighted
sum of the accrued prime information and the accrued target information.
The weights themselves are sums of automatic and strategic components.
An automatic component of the prime weight is positive and reflects the au-
tomatic influence of irrelevant prime evaluations as found in masked prim-
ing. The prime weight also has a strategic component that reflects strategies
of attention allocation, requiring some amount of strategic, conscious proc-
essing, as occurs in contexts where the prime information is generally valid
and helpful (Cheesman & Merikle, 1986). The integration of prime and
target evaluations proceeds in a random-walk process in which the impacts
of prime and target are accrued in proportion to their weights.
The Stroop mechanism thereby explains visible and masked affective
priming in evaluative decisions, and its time course agrees well with analo-
gous findings from other Stroop-like tasks (MacLeod, 1991). In addition,
consistency proportion effects and sequential effects of the Greenwald et al.
(1996) variety can be explained through learning-induced, strategic shifts
in prime weights that require visibility of the irrelevant prime stimulus
(Cheesman & Merikle, 1986; cf. Musch, 2000). Similarly, separating prime
and target location, and allowing participants to prepare for the target loca-
tion supports a strategy of attention allocation in which to-be-ignored
prime information is given a small weight because it can be more effectively
ignored when perceivers can prepare for the location of the target (Musch
& Klauer, 2001). On the other hand, presenting prime and target at the
same rather than different locations is likely to hinder the strategic screen-
ing out of prime information, thereby explaining larger priming in the for-
mer condition (Hermans et al., 2001). Furthermore, negative priming
effects as demonstrated by Wentura (1999) for affective priming are gener-
ally found in Stroop-like tasks and thus fall under the scope of effects that
can be explained by the set of mechanisms that underlie findings in Stroop
tasks. Effects of prime strength follow naturally from Musch's (2000) model
in which prime and target evaluations are integrated in the form of a
weighted sum. Effects of prime accessibility follow from the temporal dy-
namics of this random-walk model, in which evaluations that are available
quickly are likely to exert a greater influence.
Finally, the mechanism is easily reconciled with the absence of affective
priming in tasks that require nonevaluative classification of target stimuli
(cf. section on task-dependence), because only response-relevant prime in-
formation is integrated. But for the same reason, it cannot explain affective
priming in nonevaluative tasks that require affirmative or negative re-
2. AFFECTIVE PRIMING 25

sponses such as the lexical decision task. Nor can it account for affective
priming in the pronunciation task for which task the pattern of findings is,
however, complex and mixed. To summarize, the Stroop mechanism inte-
grates the results obtained with the evaluative decision task, but fares less
well in accounting for affective priming in nonevaluative tasks.

The Affective-Matching Mechanism

A mechanism that predicts a broader range of effects of irrelevant evalua-

tions is the affective-matching mechanism. It was originally proposed to ac-
count for tendencies toward evaluative consistency in social judgments
(e.g., Abelson & Rosenberg, 1958; Cooper, 1981; cf. Klauer, 1991; Klauer &
Stern, 1992; Nisbett & Wilson, 1977) and is adapted from so-called post-
lexical mechanisms as discussed in the context of semantic priming (e.g.,
de Groot, 1984; Neely, Keefe, & Ross, 1989).
The affective-matching model makes three assumptions:

1. It is assumed that the evaluations of both prime and target are acti-
vated automatically and are spontaneously compared for evaluative consis-
tency regardless of the perceiver's current goals or tasks.
2. Evaluative consistency of two words (e.g., sunshine, friendly) gives
rise to a feeling of plausibility, evaluative inconsistency (e.g., sunshine, sick)
engenders a feeling of implausibility.
3. A spontaneous feeling of plausibility facilitates making affirmative re-
sponses, whereas a spontaneous feeling of implausibility inhibits such re-
sponses. Conversely, a spontaneous feeling of implausibility facilitates mak-
ing negative responses, whereas a spontaneous feeling of plausibility inhibits
such responses. That is, a feeling of plausibility biases affirmative responses
and a feeling of implausibility biases negative responses.

An effect of evaluative consistency is predicted by this mechanism when-

ever affirmative or negative responses are required with respect to both
evaluatively consistent as well as inconsistent word pairs. Consider, for ex-
ample, the case of lexical decisions. In the lexical decision task, participants
decide whether target letter strings constitute words or not. When prime
and target are evaluatively consistent words, the "yes, word" response is fa-
cilitated, via a feeling of plausibility, according to the affective-matching
model. The "yes" response is inhibited, via a feeling of implausibility, when
prime and target are evaluatively inconsistent words. As a result, an affective
priming effect is predicted for word targets although the evaluations of
prime and target are irrelevant in the lexical decision task.
26 KLAUER AND MUSCH

Wentura (1998, 2000) has used the lexical decision task to perform tests
of this model. As just explained, affective priming effects for word targets
are expected and were in fact obtained. The crucial test of the model con-
sisted of a condition in which the assignment of "yes" responses and "no"
responses to words and nonwords, respectively, was reversed. Because of As-
sumption 3, the affective-matching model predicts a reversal of priming ef-
fects as a consequence of this manipulation. For the original "word=yes"
condition, the affective priming effect emerged, whereas the data pattern
was in fact reversed for the "word=no" condition. That is, "no" responses to
word targets preceded by evaluatively inconsistent rather than consistent
primes were now made faster. The model predicts this reversal because
evaluative inconsistency is expected to facilitate negative responses via a
feeling of implausibility, whereas evaluative consistency should inhibit neg-
ative responses. Analogous reversals of the effects of irrelevant evaluations
were reported by Klauer and Stern (1992) and using a different task by
Klauer and Musch (in press).
Applying the mechanism to the evaluative decision task is somewhat
complicated. The response good or positive for positive words can be classi-
fied as affirmative, the response bad or negative for negative words is nega-
tive. A priming effect is therefore expected for positive targets, and a re-
versed priming effect is expected for negative targets. The priming effect is
expected to be reversed for negative targets, because the required response
negative is negative, and negative responses are facilitated by evaluative in-
consistency and inhibited by evaluative consistency under the affective-
matching model. Overall, the net priming effect, averaged over positive
and negative targets, should be zero. In addition, it is not easy to test the
prediction of reversed priming effects for positive versus negative targets in
the evaluative decision task. Any difference between priming for positive
targets and priming for negative targets is perfectly confounded with the
main effect of prime valence as is not difficult to see. It is therefore impossi-
ble to disentangle possible effects of prime valence from differences caused
by affective matching. Even the absence of a difference between priming
for positive targets and priming for negative targets cannot be interpreted
unambiguously because of this confounding.
To summarize, the affective-matching model predicts affective priming
in tasks requiring responses that can be classified as affirmative or negative,
the paradigmatic case being binary "yes/no" decisions, and it explains the
reversal of affective priming from trials requiring "yes" responses to those
requiring "no" responses. The affective-matching mechanism is thereby less
task-dependent than the Stroop mechanism in the sense that it does not
presuppose an intention to evaluate the stimuli themselves as good or bad.
It cannot, however, account for the pattern of findings obtained with the
evaluative decision task.
2. AFFECTIVE PRIMING 27

A MODEL OF THE EVALUATIVE SYSTEM

A large portion of the research is integrated by a model of the evaluative sys-

tem that has at its core a process of preconscious evaluation activation. The
evaluations, once activated, exert their influence on subsequent processes
through mediating mechanisms. We assume that there are at least two such
mechanisms that were termed the Stroop mechanism and the affective-
matching mechanism. Both are assumed to operate simultaneously and in
parallel. By their properties they determine the scope and generality of ob-
servable priming effects.
As already discussed, the Stroop mechanism explains most of the find-
ings obtained with the evaluative decision task, but because of its strong
task-dependence it fails to account for affective priming in tasks without a
strong evaluative component. Neumann (1984) reviewed findings on goal
dependence obtained in classical Stroop paradigms and concluded that "to
a large degree, a distractor causes interference not because of its intrinsic
properties but because it is related to the intended action" (p. 269). At this
point, the affective-matching model comes into play to explain affective
priming in wider contexts. As detailed above, the affective-matching mech-
anism is not expected to contribute to affective priming in the evaluative
decision task itself. However, it complements the Stroop mechanism to ac-
count for affective priming in tasks that require nonevaluative affirmative
and negative responses, the paradigmatic case being binary yes/no deci-
sions. Considerable evidence for its operation has by now accrued in the
form of the telltale pattern of a reversal of priming effects from trials re-
quiring an affirmative response to those requiring a negative response.
This model, comprising a preconscious process of evaluation activation
and two mediating mechanisms, offers differentiated answers to two issues
of considerable debate: (a) The extent to which the effects of irrelevant
evaluations are goal-dependent, and (b) whether the processing of evalu-
ative information must be assumed to differ from cognitive processing.

Goal Dependence of Affective Priming

Turning first to the issue of goal dependence, Holender (1992) has

pointed out that a necessary condition for Stroop-like congruity effects
such as flanker effects is an overlap between the ensemble of task-relevant
attributes of the target stimuli or the required responses on the one hand
and the attributes of the irrelevant primes on the other hand. The overlap
endows irrelevant primes with the power to prime a response from the set
of responses, either the same response as that required by the target or a
different one, thereby facilitating or inhibiting, respectively, the task-appro-
priate response. The Stroop mechanism (i.e., the mechanisms that under-
Exploring the Variety of Random
Documents with Different Content
 XIII.] ASSASSINATION OF KA'B, SON OF AL-ASHRAF 247
throwing the victim off his guard by fair words and pretence of
unfriendliness to the Prophet's rule. Abu Na'ila, fosterbrother of Ka'b,
being deputed to pave the way, complained to him of the calamities
and poverty which the advent of Mohammad had brought upon
them, and begged that he would advance corn and dates for the
sustenance of himself and a party like-minded with him. Ka'b, taken
in the snare, demanded security ; Abu Na'ila agreed that they should
pledge their arms, and appointed a late hour of meeting at the
house of Ka'b, when the bargain would be completed. Towards
evening the conspirators assembled at the house of Mohammad. It
was a bright moonlight night, and the Prophet accompanied them to
the outskirts of the town. As they emerged from the low shrubs of
the Muslim burying-ground, he bade them god-speed : ' Go ! ' said
he ; ' the blessing of God be with you, and assistance from on High!'
The house of Ka'b was near one of the Jewish suburbs, two or three
miles from the city. When they reached it he had retired to rest. Abu
Na'ila called aloud for him to come down, and Ka'b started from his
couch. His bride (for he had been lately married, and the
biographers omit nothing that adds to the heartlessness of the
affair) caught him by the skirt, and warned him not to go. ' It is but
my brother Abu Na'ila,' he said ; and, as he pulled the garment from
her, gaily added the verse : ' Shall a warrior be challenged and not
respond ? ' Descending, he was not alarmed to find the party armed,
as the weapons were to be left with him in pledge. They wandered
along, conversing on the misfortunes of Medina, till they reached a
waterfall, and upon its bank they proposed to pass some part of the
moonlight night. Meanwhile, his foster-brother, having thrown his
arm around Ka'b, was familiarly drawing his hand through his long
locks, and praising their sweet scent, which Ka'b said was that of his
bride. Suddenly the traitor seized his hair, and dragging him to the
ground, shouted! ''Slay him I Slay the enemy of God!' Drawing their
swords, they fell upon the victim. The wretched man clung so close
to his foster-brother that he was with difficulty put to death. As he
received the fatal wound he uttered a fearful scream, which
resounded far and near amongst the strongholds of the Jews, and
lights were seen at the windows
 assassination 248 THE YEAR AFTER BEDR [chap. of the
affrighted inhabitants. The assassins, fearful of pursuit, retired in
haste, carrying in their arms one of their number who had received
two sword-cuts aimed at Ka'b. As they regained the burying-ground,
they shouted the well-known tekbtr, ' Great is the Lord ' ; which
Mohammad hearing knew that their work had been successfully
accomplished. At the gate of the Mosque he met them, saying : '
Welcome ; for your countenances beam of victory.' ' And thine also,
O Prophet,' they exclaimed, as they cast the ghastly head of their
victim at his feet. Then Mohammad praised God for what had been
done, and comforted the wounded man. Reflections I have been
thus minute in the details of the murder of on Ka'b's Ka'b, as it
faithfully illustrates the ruthless fanaticism into which the teaching of
the Prophet was fast drifting. It was a spirit too congenial with the
passions of the Arabs not to be immediately caught up by his
followers. The strong religious impulse under which they acted
hurried them into excesses of barbarous treachery, and justified that
treachery by the interests of Islam and approval of the Deity. I am
far from asserting that every detail in the foregoing narrative, either
of instigation by Mohammad, or of deception by the assassin, is
beyond question. The actors, indeed, in such scenes were not slow
to magnify and embellish their own services at the expense of their
imagination. There may also have been the desire to justify an act of
perfidy that startled even the loose morality of the day, by casting
the burden of it on the infallible Prophet. But, after due weight given
to both considerations, enough remains to prove some of the worst
features of assassination, and the presumption that these were
countenanced, if not in some instances directly prompted, by
Mohammad himself.^ 1 There can be little doubt that some Muslims
were at times scandalised by crimes like this ; though it is not in the
nature of tradition to preserve the record of what they said. The
present is one of the few occasions on which such murmurs have
come to light. When Merwan was Governor of Medina, he one day
asked Benjamin, a convert from Ka'b's tribe, in what manner Ka'b
met his death. 'By guile and perfidy,' said Benjamin. Now
Mohammad, son of Maslama (the assassin), by this time a very aged
man, was sitting by. He exclaimed: 'What, O Merwan ! could the
Prophet of the Lord, thinkest thou, be guilty of perfidy? By the Lord !
we did not kill him but by command of the
 XIII.] MURDER OF IBN SUNEINA >49 On the morning after
the murder of Ka'b, Mohammad, exasperated at the opposition (or,
as tradition puts it, the treachery) of the Jews, accorded a general
permission to his followers to slay them wherever met. Accordingly,
Muheisa, having encountered Ibn Suneina, a Jewish merchant, slew
him, though a confederate of his tribe.' The occurrence is alluded to
by the biographers rather for the purpose of explaining the sudden
conversion of the assassin's brother Huweisa, than to record the
murder of a petty Jewish trader. When Huweisa upbraided his
brother for killing the confederate Jew, and appropriating his wealth
; — ' By the Lord ! ' replied Muheisa, ' if he that commanded me to
kill him had commanded to kill thee also, I would have done it.' '
What ! ' Huweisa cried ; ' wouldst thou have slain thine own brother
at Mohammad's bidding ? ' ' Even so,' answered the fanatic ' Strange
indeed ! ' Huweisa responded ; ' hath the new religion reached to
this? Verily, it is a wonderful faith.' And Huweisa was converted from
that very hour. The progress of Islam begins to stand out in
unenviable contrast with that of early Christianity. Converts were
gained to the faith of Jesus by witnessing the constancy with which
its confessors suffered death ; they were gained to Islam by the
spectacle of the readiness with which its adherents inflicted death.
In the one case conversion imperilled the believer's life; in the other,
it was the only means of saving it. The Jews were now in extreme
alarm. None ventured abroad. Every family lived in fear of a night
attack ; every individual dreaded the fate of Ka'b and Ibn Suneina. A
deputation of their principal men waited upon Mohammad and
complained that he had treacherously cut off one of their chiefs
without fault or apparent cause. ' Had Ka'b Prophet. I swear that no
roof, save that of the Mosque, shall hereafter cover thee and me.'
Then, turning to Benjamin, he swore that if he had had a sword in
his hand, he would have cut off his head. The unfortunate Benjamin
could not thenceforward quit his house without first sending a
messenger to see that Mohammad was out of the way ; but one day
he was caught at a funeral by Mohammad, who seized a bundle of
date branches from a woman passing by, and broke them every one
over the face and back of Benjamin. Thus were murmurers against
such acts silenced in the early days of Islam. ^ Ibn Hisham, p. 553 f.
; At-Tabari, i. 1372 ; Al-Wakidi, p. 97 f. Murder of Ibn Suneina, a
Jew, Causes conversion of the murderer's brother New treaty with
the Jews
 250 THE YEAR AFTER BEDR [chap. Mohammad marries
Hafsa, A.H.'lII. November, A.D. 624 Marriage of Fatima with 'All, A.D.
624 ; birth of Al-Hasan and Alii osein conducted himself,' replied
Mohammad, *as ye have done, he would not have been cut off. But
he offended me by his seditious speeches and his evil poetry. And if
any one amongst you,' he added, ' doth the same, verily the sword
shall be again unsheathed.' At the same time he invited them to
enter into a fresh compact with him, such as he might deem
sufficient for the interests of Islam. So a new treaty was written out
and deposited with 'All. Nevertheless, adds Al-Wakidi, the Jews
thenceforward lived (as well they might) in a state of depression and
disquietude. Towards the close of the year 624 the Prophet took to
himself a third wife, Hafsa, the daughter of 'Omar, then about
twenty years of age. She was the widow of Khoneis, an early
convert, who had died six or seven months previously. By this
marriage Mohammad not only gratified the passion for fresh
espousals, a leading feature of his advancing years, but bound
himself closer in friendship to her father 'Omar. Abu Bekr and 'Omar
were now connected equally with the Prophet, and through their
daughters had_ access to his ear. There was much rivalry between
'A'isha and Hafsa ; but youth, vivacity, and beauty maintained the
supremacy of 'A'isha, The marriages contracted by Mohammad at
Medina were all unfruitful. But meanwhile his house was built up in
the female line of Khadlja's progeny. We hear of no issue, certainly
of none that survived, by his daughters Zeinab and Um Kulthum,
though the name of the latter would imply maternity. Rokeiya bore
'Othman a son, two or three years before the Flight, but his eyes
were pecked out at Medina by a fowl, and he died still a child. It was
through Fatima alone that the Prophet's race, the famous Seiyids or
nobility of Islam, was to be perpetuated. 'All was now five-andtwenty
years of age. Though not above middle stature, he was broad and
powerful in make, with a ruddy complexion, and a thick and comely
beard. He had already given proof of daring gallantry and prowess
on the field of Bedr. Endowed with a clear intellect, warm in
affection, and confiding in friendship, he was from boyhood devoted
heart and soul to the Prophet. Simple, quiet, and unambitious, when
in after days he obtained the rule of half the Muslim world, it was
rather thrust upon him than sought. Shortly
 '^"^•1 'ALT MARRIES FATIMA 251 after the field of Bedr
(some authorities say before it) Mohammad gave him the hand of
Fatima his youngest daughter, now seventeen or eighteen years of
age in marriage. Within the next twelve months she gave'' birth to
Al-Hasan, the first grandson born to Mohammad that survived, and
the year after to Al-Hosein ;-names famous m Islam.
 Koreish resolve to avenge defeat at Bedr Mohammad
receives intimation from Al'Abbas Koreish march from Mecca. A.H.
III. January, A.D. 621; CHAPTER XIV THE BATTLE OF OHOD ^
Shauwdl, A.H. III. — January, A.D. 625 /ETAT. 56 The third year of
Mohammad's life at Medina had nearly closed, and the winter had
again set in, when a storm clouded the horizon. Twelve months had
elapsed since the battle of Bedr. The cry of revenge had ever since
resounded in the valley of Mecca ; and the long-suspended threat
was now put into execution. Rumours of a threatened attack had for
some time been reaching Mohammad ; but the first authentic notice
of impending invasion with a sealed letter placed in his hands, while
at the Mosque in Koba, by a messenger from Mecca. It was from his
uncle Al-'Abbas, who, as usual holding with both sides, had engaged
the courier, by a high reward, to deliver it in three days. The letter,
read aloud on the spot, contained the startling intelligence that
Koreish, 3,000 strong, were on the point of marching. Mohammad
enjoined secrecy ; but the tidings could not be suppressed. The
Prophet communicated the news privately to Sa'd, the Khazraji '
Leader,' and his wife overheard it. Whether thus, or otherwise, the
coming attack was soon noised abroad, and caused great
excitement, especially among the Jews and those who sympathised
with them. The movement did, indeed, justify alarm. Koreish had
unanimously agreed to devote the profits of the caravan, whose
precious freight was still retained in the Council- Hall as it were in
bond, and for which so much blood had been 1 Ibn Hisham, p. 555
ff. ; At-Tabari, i. 1383 ff. ; Al-Wakidi, p. loi ff. ; Ibn Sa'd, p. 25 IT. 252
 CHAP. XIV] KOREISH MARCH AGAINST MEDINA 253 shed
at Bedr, towards avenging their defeat. These profits amply sufficed
for the equipment and provisioning of a great army. Emissaries were
despatched throughout the Bedawl tribes, connected with Koreish by
alliance or descent, inviting them to join the enterprise. At length, in
the month of January, they commenced their march, 3,000 strong ;
700 were mailed warriors, and 200 well-mounted cavalry ; the
remainder rode on camels. The Beni Zuhra (who had, on the
previous occasion, retired before reaching Bedr) alone remained
behind ; but the army was reinforced by 100 men from At-Ta'if. The
chiefs of Koreish all joined the force. After a sharp discussion,
women were allowed to accompany them ; and fifteen, including
two wives of Abu Sufyan, availed themselves of the permission.
Taking timbrels in their hands, they sang to their wild cadence songs
of vengeance for kinsmen slain at Bedr. Foremost, Hind, the wife of
Abu Sufyan, thirsting for the blood of Hamza who slew her father in
that field, had engaged an Ethiopian^ with his deadly javelin, to
make sure of her victim. There was also with the army a band of
Medina citizens under Abu 'Amir, 'the Monk,' who, it will be
remembered, went over to Mecca in disgust at the enthusiastic
reception of Mohammad, and now boasted that his simple presence
with the army would produce an immediate reaction amongst his
former fellow-citizens. The army took the ordinary route by the
seashore, and. And en after ten days, reached Dhu'l-Huleifa, in the
valley of Al- ScSnT' «Aklk, about five miles west of Medina. It was
Thursday Thursday morning ; and the same day, fetching a circuit to
the left, and then marching northward for a few miles, they
encamped in the fertile plain beneath the hill of Ohod. The corn was
cut down as forage for the horses ; and the camels, set loose to
graze, trampled the rich fields around. Friday was passed Friday
inactively. Between the city and the plain were several rocky ridges,
which rendered it secure from direct attack on that side ; but the
Syrian highway, sweeping eastward under Ohod, and then south,
reached the northern suburb by an easy circuit. Koreish feared to
advance by this route, as the houses upon it afforded their
adversaries a position of dangerous offence. They hoped rather to
draw them to the outskirts, and overpower them there by superior
numbers upon
 254 BATTLE OF OHOD [chap. Proceedings in Medina.
Thursday Resolution to remain within the city. Friday equal ground.
Perhaps, also, they expected by delay to create some dangerous
diversion in the city. Meanwhile Mohammad, by his spies, was kept
apprised of the enemy's movements. Al-Hobab reconnoitred their
camp and brought back an alarming estimate of its strength, which
the Prophet desired him to keep secret. The farmers, with their
cattle and their stuff, had affected a timely retreat; but the
destruction of their fields was a trial sore to bear. Still, there was no
ebullition of feeling against Mohammad as the cause of their
misfortune. Indeed, so great was the hold he had already gained,
that, the elements of disaffection notwithstanding, he was at once
recognised throughout the city as the leader and director in its
defence. Several chief men, with an armed band of Citizens, posted
themselves at the great Mosque, and kept watch throughout the
night by his door. The sleep of Mohammad was troubled. He
dreamed that, securely clad in mail, he rode upon a ram, when
suddenly his sword was broken at its point, and a steer was
slaughtered in his sight. The next day, Friday, the people came
together, and Mohammad discussed with them the course to be
pursued. He told them of his dream. ' The fracture in my sword
portendeth an injury to myself,' he said ; ' the slaughter of the steer,
some damage to the people ; riding upon the ram signifieth carnage
amongst the enemy ; and the impenetrable coat of mail is Medina
fortified and safe. Within the city we are secure: without it there is
risk and danger.' In this opinion the men of years and wisdom, both
Citizens and Refugees, agreed. 'Abdallah ibn Obei, who,
notwithstanding his jealousy of Mohammad, was equally concerned
in the defence, strongly supported the views of Mohammad : ' O
Prophet ! Our city,' he said, ' is a virgin inviolate. Quitting it, we have
ever suffered loss : remaining, we have beaten back attack. Leave
Koreish alone. If they remain, it will be in evil case. At length,
frustrated in their designs, they will retire.' It was resolved
accordingly to bring all outlying inhabitants within the walls, and, if
Koreish should venture near, to drive them back by a galling
discharge of arrows and stones from the walls and house-tops. The
decision was displeasing to the younger and more impetuous
Citizens. ' Shall we sit quietly here,' they asked
 XIV.] DECISION TO GIVE BATTLE TO KOREISH 255
indignantly, 'a laughing-stock to all Arabia, and look on in Set aside
by patience while our possessions are ravaged all around ? Dis-
youngs" grace will cleave to us ever after, and the enemy,
emboldened, converts will repeat the insult. Nay, we will go forth
and smite our foes, even as we did at Bedr.'^ There were not
wanting men even among the Refugees who sided with this party,
and their ardour was so great that Mohammad against his better
judgment at last gave way, and announced his readiness to offer
battle. Ascending the pulpit for the weekly service (the day was
Friday) he stirred up the people, in his discourse, to fight
courageously: 'If ye be steadfast,' he said, ' the Lord will grant you
victory.' Then he commanded to make ready for the battle. The most
part rejoiced greatly, but some were grieved that the first decision
had been set aside. By the time the afternoon prayer was ended, the
people Mohammad had assembled in the court of the Mosque,
armed for battle, l"^^^^ .^'^ Mohammad then retired with Abu
Bekr and 'Omar, to make ready. In a little while he issued from his
chamber clad in ^ As usual, we are overwhelmed with anecdotes of
believers bent on martyrdom, and dreams and pious anticipations of
rewards to be enjoyed in Paradise. These are the growth of after
years ; the halo pictured by tradition around the martyr's head.
There were nevertheless worldly motives enough to justify this party
in their desire to go forth. The Citizens were grieved at the
occupation of their fields ; the barley crops were being destroyed,
and the season for sowing was passing away. Even Hamza joined
them on political considerations. ' We fear,' he said, 'lest Koreish
should attribute our backwardness to cowardice, and that it will
embolden them ever after. We were but few at Bedr, and we are
many now. Verily, this is the day we have longed and prayed to the
Lord for ; and now He hath driven the enemy as a prey into our very
midst.' Some specimens of the martyr spirit may interest the reader.
One said to Mohammad : 'The slaughtered steer thou sawest was an
emblem of the dead amongst thy followers, and verily I shall be of
the number ; wherefore, hinder me not from Paradise. Let us go
forth ; surely, by the one God ! I shall quickly enter therein.' Again,
Khaithama told Mohammad that his son, whom he had lost at Bedr,
appeared to him in his sleep ; — 'A goodly appearance truly he had ;
he described to me the blessedness of Paradise ; all is true that our
Lord hath promised ; and he besought me to come quickly, and be
his companion there. And now, verily, I am old, and long for the
meeting with my Lord. Pray, therefore, that God would grant me
martyrdom, and reunite me with my son.' So Mohammad prayed ;
and Khaithama was slain at Ohod. Such are the tales which tradition
delights to embellish or haply to create.
 256 BATTLE OF OHOD [chap. mail and helmet, his sword
hanging from a leathern girdle,'^ and shield slung over his shoulder.
The Citizens, seeing him thus accoutred, repented of their rash
remonstrance, and prayed that he would even now do as seemed
good to him. But it was too late. ' I invited you to this,' he said, ' and
ye would not. It becometh not a prophet, when once he hath girded
himself to the battle, to lay his armour down again until the Lord
hath decided betwixt him and his enemies. Wait, therefore, on the
Lord. Only be steadfast, and He will send you victory.' Marches So
saying, he called for three lances, and fixed banners from Medina,
ypQj^ them. One for the Refugees he gave to Mus'ab, the and halts.
^ , , , . , , i i r i a i tVi • Friday night second and third to the leaders
ot the Aus and Khazraj. 'Abdallah ibn Um Mektum (the blind man of
whom we read at Mecca) was appointed to command the city, and
lead the public prayers. Just then the bier of a Citizen was brought
into the Mosque. Mohammad pronounced over it the usual service ;
then mounting his horse, and surrounded by his followers, he took
the road to Ohod. There was but one other horse with the Muslim
army. Arrived at an eminence, the Prophet turned round and saw
following, amid the palm plantations on the right, a rude and
disorderly band of men, and being told that they were the Jewish
confederates of 'Abdallah ibn Obei, he commanded that they should
go back ; ' for,' said he, ' ye shall not seek the aid of Unbelievers to
fight against the unbelieving.' He then passed onwards to Esh-
Sheikhain, half-way to Ohod,^ and having reviewed the force, and
sent back some striplings unequal to the fight, there halted for the
night. 'Abdallah ibn Obei, with his followers, encamped near at hand
; but, displeased at the rejection of his advice, and also at the
unfriendly treatment of his Jewish friends, kept sullenly aloof.
Mohammad passed the night with the Beni an-Najjar, and a guard of
faithful followers was stationed over him. Mohammad, son of
Maslama, patrolled the camp with fifty ^ This girdle was preserved
and handed down in the family of Abu Rafi', Mohammad's servant. 2
Burckhardt notices it as 'a ruined edifice of stones or bricks,' a mile
from the town, 'where Mohammad put on his coat of mail' ; i.e. on
the following morning. ' Farther on,' he tells us, there is a stone
where the Prophet 'leaned for a few minutes on his way to Ohod.'
 XIV.] MOHAMMAD HALTS AT OHOD 257 men. A similar duty
was performed for Koreish by 'Ikrima with a troop of horse ; these
approached close enough to alarm the Muslims by their neighing,
but did not venture over the ridge which still separated the two
armies. At early dawn the army of Medina, 1,000 strong, was in
Mohoammad motion. In the dim morning light they marched, by the
ohoj'^an? nearest path, through the intervening fields and
gardens,^ draws up and emerged upon the sandy plain beneath the
peaks of ^^^^°^ Ohod. The vicinity owes its verdure to a
watercourse, which Saturday, carries off the drainage of the country
lying to the south and January east The hill of Ohod, three miles
distant from Medina, is A.n. 624 a rugged and almost insulated
offshoot of the mountain range, projecting eastward for three or four
miles into the plain. The torrent, sometimes swollen so as quite to
inundate the adjacent tract, sweeps along its southern and western
face, and discharges its flood into the GJidba, or low basin lying
beyond. Now dry, its course was marked only by deep sand and
scattered stones. On the farther bank, upon a slightly sloping plain,
bare and stony, over which, as Burton tells us, 'the seared and
jagged flanks of Ohod rise like masses of iron,' Mohammad halted
his army. By this time it was daylight, and, although the columns of
the enemy were in sight, the cry for morning prayers was raised by
Bilal, and the whole army, led by the Prophet, prostrated itself in
worship. 'Abdallah ibn Obei at this moment wheeled suddenly round,
and, deserting the army with his 300 followers, took the road back
to the city. Mohammad was thus left with but 700 followers, of
whom only a hundred were clad in mail ; but they were all true men,
and, fighting in what they believed to be the cause of God, they
boldly faced a well-appointed enemy four times their number.
Advancing, they occupied the rising ground in front ; their 1 As he
passed through one of these gardens, its owner, a blind man,
murmured at the injury to his property, and cast dust at Mohammad.
One of the Aus sprang up and beat him. A chief of the Khazraj
resented the affront, and a fierce contention arose. It was ended by
a savage threat from Oseid, the Ausite ' Leader,' who said that had
he not known that it would be displeasing to Mohammad, he would
have cut the blind man's head off. There must, no doubt, have been
difficulty in keeping down these intestine quarrels and jealousies,
though, in the hands of a skilful administrator like Mohammad, they
were really elements of power. R
 2s8 BATTLE OF OHOD [chap. Army of Mecca advances
Battle opens with single combats rear was thus protected by the
frowning heights of Ohod, excepting on the left, where the rocks,
receding, afforded the enemy a dangerous opening, suited to the
movements of the Koreishite horse, Mohammad, therefore, posted
on an adjoining eminence the flower of his archery, and gave their
leader stringent orders on no possible contingency to quit the spot,
but steadily to check any attempts which Koreish might make to turn
his flank : ' Guard our rear,' he said, 'and stir not from this spot: if ye
see us pursuing and plundering the enemy, join not with us : if we
be pursued and even worsted, do not venture to our aid.' Then he
drew out his line, facing towards Medina ; — Mus'ab, with the
Refugee standard, being in the centre, and the Aus and Khazraj
clans forming either wing. He forbade his followers to engage the
enemy till he gave command ; for he knew that the strength of his
position would be sacrificed by a premature advance. Having thus
disposed his force, he put on a second coat of mail, and calmly
awaited the enemy's approach. Meanwhile Abu Sufyan, as hereditary
leader, brought up the Meccan army ; and, facing Ohod, marshalled
it in front of the Muslim line. The banner, which had been duly
mounted on its standard in the Council-Hall at Mecca, was borne by
Talha, grandson of *Abd al-'Ozza.^ The right wing was commanded
by Khalid ; the left by 'Ikrima, son of Abu Jahl. 'Amr ibn al-*As (the
famous 'Amr) was over the Koreishite horse. The women at first kept
to the front, and beat their timbrels to shrill martial song ; but as the
line advanced, they fell to the rear. The battle opened by the
inglorious advance of the exile Abu 'Amir, who vainly expected his
fellow-citizens of Medina ' This Talha is to be distinguished from
Talha, son of 'Obeidallah, who stood by Mohammad in the battle.
The Talha in the text was of the family of 'Abd ed-Dar, which
retained the right of carrying the Koreishite standard {vide p. xcvi.).
Abu Sufyan desired not only to lead the army, but to carry the
standard, or at least to raise a second banner ; but the descendants
of 'Abd ed-Dar would hear of no encroachment on their ancestral
privilege. There is a tradition that, as the enemy drew near,
Mohammad inquired who bore their standard. On being told that it
was one of the house of 'Abd ed-Dar, he exclaimed: 'Our side is
more worthy of the honour ; ' and, calling for Mus'ab (who was of
the same lineage), he placed the standard in his hands.
 XIV.] SINGLE COMBATS 259 to fraternise with him. He was
received with a shower of stones, and forced with his band of
followers to retire ; Talha crying out indignantly : — ' Get to the rear,
ye slaves ! Guard the camp, — a fitting employment for you!' Then,
flourishing the Koreishite banner, Talha advanced alone, and
challenged the enemy to single combat, shouting these words : —
The standard-bearer hath the right To dye its shaft in blood. Till it be
broken in his hand. 'All stepped forth, and, rushing on him, with one
blow of his sword brought him to the ground.^ Mohammad, who
had intently watched the rapid combat, exclaimed with loud voice,
Great is the Lord ! and the cry, taken up all round, arose in an
overwhelming shout from the whole Muslim army. Talha's brother,
'Othman, who was in charge of the women, then ran forward and
seized the banner which lay by the lifeless body. The women beat
their timbrels loudly, as they sang : — Daughters of the brave are
we, On carpets step we delicately ; Boldly advance, and we embrace
you ! Turn your backs and we will shun you, — Shun you with
disdain. Hamza responded to 'Othman's challenge, and, after a brief
encounter, brought him also lifeless to the ground. Then striding
proudly back to the Muslim ranks, he shouted : * I am the son of
him that gave the pilgrims drink,' — meaning of 'Abd al-Muttalib,
who had held that office. One after another, the family of Talha, two
brothers and three sons, seized the standard ; one after another,
they fell in single combat.^ ^ Mohammad declared that thus was
fulfilled that part of his vision in which he appeared to ride upon a
ram. Talha was the ram. 2 One of the sons was wounded by an
arrow, shot by 'Asim, The wounded lad was carried to his mother
Sulafa, at the rear. She asked him, as he was breathing his last, who
killed him. He said that as his foe shot the arrow, he heard him cry :
' Take that from me, the son [that is, grandson] of Al-Aklah ! ' 'By
the Lord ! ' Sulafa said, ' it was 'Asim, one of our own kin ' ; and she
vowed she would yet drink wine out of 'Asim's skull. The savage vow
was nearly being fulfilled, as we shall see hereafter.
 26o BATTLE OF OHOD [chap. Koreish are pressed, and
waver Day changed by charge of Khalid, and army of Medina routed
This Arab custom of single combat put the two armies on an
equaUty for the time. So long as it went on, the Koreish derived no
advantage from their superior numbers ; and the rapid destruction
of their standard-bearers carried dismay into their ranks. A general
engagement ensued ; and, pressed by the fierce ardour of the
Muslims, the Meccan army began to waver. Their horse sought
repeatedly to turn the left flank of Mohammad ; but they were each
time forced back by the galling archery of the little band which
Mohammad had posted there. The same daring contempt of danger
was displayed as at Bedr. The Meccan ranks might be seen to quiver
as Abu Dujana, distinguished by a red kerchief round his helmet,
swept along the enemy's ranks, and, with a sword given him by
Mohammad, dealt death on every hand.^ Hamza, conspicuous from
his waving ostrich feather ; *AlI, known by his long white plume,
and Az-Zubeir, by his bright yellow turban, like heroes of the Iliad, —
carried confusion wherever they appeared. Such were the scenes in
which were reared the great leaders of the Muslim conquests.^ But
now the Muslims pressed too hotly their success. Their line lost form
and order ; and a portion, piercing the enemy's ranks, fell to
plundering his camp. The archers, who had hitherto held the Meccan
horse in check, saw from their height the tempting opportunity, and,
casting the Prophet's strict injunction to the winds, as well as the
earnest expostulation of their leader, hurried to the spoil. The ready
eye of Khalid saw the chance, and he hastened to retrieve the day.
Wheeling his cavalry round the enemy's left, and ^ There is a mass
of tradition about Abu Dujana's prodigies of valour. At the
commencement of the action Mohammad held up his sword, and
said: 'Who will take this sword, and give to it its due?' 'Omar, Az-
Zubeir, &c., one after another, came forward and were rejected ; last
of all Abu Dujana offered, and Mohammad gave it to him; 'And he
clave therewith the heads of the Unbelievers.' After the battle, 'All,
giving his bloody sword to Fatima to wash, said : 'Take this sword,
for it is not to be despised,' alluding to his own acts of prowess that
day. Mohammad added: 'If thou hast done well, O 'All ! verily Al-
Harith and Abu Dujana have done well also.' Ibn Hisham, p. 588. 2
For example, in this battle we have Sa'd and Abu 'Obeida on the side
of Mohammad, and Khalid and 'Amr on the side of Koreish ; all
famous in after days.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookname.com