1 RFT Learning

US $49.
95 PSYCHOLOGY
A Clear and Comprehensive Guide
LE ARNING RFT
to Relational Frame Theory
Learning
Relational frame theory, or RFT, is the little-understood behavioral theory be-
hind a recent development in modern psychology: the shift from the cognitive
paradigm underpinning cognitive behavioral therapy to a new understanding of
language and cognition. Learning RFT presents a basic yet comprehensive intro-
duction to this fascinating theory, which forms the basis of acceptance and com-
mitment therapy. The book also offers practical guidance for directly applying
RFT
RFT in clinical work.
In the book, author Niklas Törneke presents the building blocks of RFT: language
as a particular kind of relating, derived stimulus relations, and transformation of
stimulus functions. He then shows how these concepts are essential to understand-
ing acceptance and commitment therapy and other therapeutic models. Learning
RFT shows how to use experiential exercises and metaphors in psychological treat-
ment and explains how they can help your clients. This book belongs on the book-
shelves of psychologists, psychotherapists, students, and others seeking to deepen
their understanding of psychological treatment from a behavioral perspective.
“There is no better place to start learning about RFT than this excellent
book. Törneke teaches the principles of RFT simply and elegantly . . . I
wish a book like this had existed when I first learned about RFT.”
—Russ Harris, author of The Happiness Trap and ACT Made Simple An Introduction to
NIKLAS TÖRNEKE, MD, is a psychiatrist and licensed psychotherapist in private practice in
Relational Frame Theory
and Its Clinical Application
Kalmar, Sweden. Together with Jonas Ramnerö, Ph.D., he has previously authored The ABCs
of Human Behavior.
Foreword writer DERMOT BARNES-HOLMES, PH.D., is foundation professor of psychol-

TÖRNEKE
ogy at the National University of Ireland, Maynooth, and coauthor of Relational Frame Theory.
Afterword writer STEVEN C. HAYES, PH.D., is University of Nevada Foundation Professor

of Psychology at the University of Nevada, Reno, and has authored and coauthored numerous
books, including Acceptance and Commitment
Therapy, Relational Frame Theory, and Get Out of ISBN: 978-1-57224-906-6
Your Mind and Into Your Life.
NIKLAS TÖRNEKE, MD
54995
Context Press Foreword by Dermot Barnes-Holmes, Ph.D.

An Imprint of New Harbinger Publications, Inc.
www.newharbinger.com 9 781572 249066 Afterword by Steven C. Hayes, Ph.D.
Context
Press
“There is no better place to start learning about RFT than this excellent book.
Törneke teaches the principles of RFT simply and elegantly, using a wealth
of clinical examples to make it accessible and stimulating. I wish a book like
this had existed when I first learned about RFT; it would have saved me many
hours of hard work, frustration, and confusion.”
—Russ Harris, author of The Happiness Trap and ACT Made
Simple
“On rare occasions, the skills of writer, therapist, and theorist combine to
give the field a sophisticated yet highly practical book. This much-awaited
translation shows relational frame theory as an accessible, powerful tool for
all who use talk therapy. A must-read for those interested in contemporary
behaviorism.”
—Kelly Koerner, Ph.D., creative director at Evidence Based
Practice Institute in Seattle, WA
“Since RFT first appeared in the experimental literature, it has been hailed
as a breakthrough in our scientific understanding of language and cognition
with direct and important implications for clinical psychological practice. Yet,
descriptions of RFT, written largely for technical audiences, have been, at best,
curiously baffling, and at worst, maddeningly incomprehensible. In this book,
Törneke has solved the puzzle of RFT! He summarizes the history of RFT, its
key features, and its clinical implications with language that is user-friendly
and easily understandable. I believe this book will make a huge difference
for clinicians who wish to understand RFT and its implications for clinical
practice. It also may be a useful learning tool for researchers and RFT experts
themselves who wish to learn and see a beautiful example of how RFT can be
presented clearly and comprehensively.”
—Jonathan Kanter, associate professor at the University
of Wisconsin-Milwaukee and director of its Depression
Treatment Specialty Clinic
“For years, clinicians have asked me for recommendations about what they
should read to learn RFT. There was really no good advice I could give except
‘be persistent.’ Finally, I have a better answer. If you want to understand
relational frame theory, this is the place to start. Törneke’s RFT primer is both
masterful and accessible.”
—Kelly G. Wilson, Ph.D., associate professor of psychology
at the University of Mississippi, coauthor of Acceptance and
Commitment Therapy and author of Mindfulness for Two
“At times, while reading Törneke’s book, I have felt as though I were in
the middle of a thriller about the psychopathological behaviors of humans.
Clues to unraveling the mystery embedded in complex concepts like
‘arbitrarily applicable relational responding’ have alerted me, as the reader,
to what is coming up next. Our ability for relational framing and for
rule-governed behavior may at first glance seem fabulous—a gift from the
gods—but darkness lurks around the corner. Our ability to problem-solve is
the villain. This book helps me make sense of it all.”
—Maria Midbøe, M.Sc., candidate in psychology at
Stockholm University in Stockholm, Sweden
“Until now, explanations of relational frame theory have remained largely

esoteric and even impenetrable to all but the most specialized scholars. For
the first time, this extraordinary book provides a highly accessible account
of relational frame theory, including its larger context within psychology,
the current research in the field, and its many potential applications.
Törneke strikes a fine balance between doing justice to relational frame
theory and making the theory, research, and its implications readily
comprehensible to the non-expert. This unique book is a must-read for
scholars of human cognition, as well as clinicians, educators, others seeking
to harness the power of basic psychological principles in their applied work,
and anyone interested in the renaissance of modern behavior analysis.”
—James D. Herbert, Ph.D., professor of psychology at
Drexel University in Philadelphia, PA, and director of
Anxiety Treatment and Research
Learning
RFT
An Introduction to
Relational Frame Theory
and Its Clinical Application
NIKLAS TÖRNEKE, MD
CONTEXT PRESS
An Imprint of New Harbinger Publications, Inc.
Publisher’s Note
This publication is designed to provide accurate and authoritative information in regard to
the subject matter covered. It is sold with the understanding that the publisher is not engaged
in rendering psychological, financial, legal, or other professional services. If expert assistance or
counseling is needed, the services of a competent professional should be sought.
Distributed in Canada by Raincoast Books
Copyright © 2010 by by Niklas Törneke
New Harbinger Publications, Inc.

5674 Shattuck Avenue
Oakland, CA 94609
www.newharbinger.com
All Rights Reserved
Acquired by Catharine Sutker; Cover design by Amy Shoup; Edited by Jasmine Star
PDF ISBN: 978-1-57224-908-0

The Library of Congress has cataloged the print edition as:

Törneke, Niklas.
[Relational frame theory. English]
Learning RFT : an introduction to relational frame theory and its clinical applications /
Niklas Törneke ; foreword by Dermot Barnes-Holmes ; afterword by Steven C. Hayes.
p. cm.
Includes bibliographical references.
ISBN 978-1-57224-906-6
1. Cognitive therapy. 2. Acceptance and commitment therapy. I. Title.
RC489.C6313T67 2010
616.89’1425--dc22
2010024048
In remembrance of my father, David
Contents

Foreword to the U.S. Edition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

Foreword to the Swedish Edition. . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii

A Personal Word of Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
PART 1
Background
CHAPTER 1
Radical Behaviorism and Fundamental Behavior Analytic Principles. . . . . 9
CHAPTER 2
Thinking and Human Language . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
CHAPTER 3
Is the Power of Thinking a Clinically Relevant Issue?. . . . . . . . . . . . . . 49
PART 2
Relational Learning
CHAPTER 4
Derived Relational Responding as the Fundamental Element in
Human Language. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Learning RFT
CHAPTER 5
Analogies, Metaphors, and Our Experience of Self. . . . . . . . . . . . . . . . 91
CHAPTER 6
Relational Framing and Rule-Governed Behavior. . . . . . . . . . . . . . . . 113
CHAPTER 7
The Dark Side of Human Languaging. . . . . . . . . . . . . . . . . . . . . . . .133
PART 3
Clinical Implications
CHAPTER 8
Learning Theory and Psychological Therapies. . . . . . . . . . . . . . . . . . 155
CHAPTER 9
General Guidelines for Clinical Behavior Analysis. . . . . . . . . . . . . . . 171
CHAPTER 10
Altering the Context with a Focus on Consequences. . . . . . . . . . . . . .193
CHAPTER 11
Altering the Context with a Focus on Antecedents. . . . . . . . . . . . . . .209

Afterword. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .239

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .261
vi

Clarity, simplicity, depth.

This is what I seek to achieve and unite.
Excluding none of the three.
Hence the difficulty.
—Pär Lagerkvist (translated by Elizabeth Ask de Lambert)

Foreword to the U.S. Edition
A PRAGMATIC THEORY OF HUMAN

LANGUAGE AND COGNITION
Behavior analysis is an extremely unusual approach to psychological science.
In stark contrast to mainstream psychology, the behavioral tradition refuses
to appeal to mediating mental representations and processes as the basis for
explaining human behavior. Instead, it adheres to a thoroughgoing functional
analytic approach in which the systematic analysis of the interactions between
an organism and its past and current environmental contexts provides the
framework for explaining all psychological events.
Somewhat surprisingly, perhaps, this unusual approach yielded consid-
erable success in the early years, particularly in improving the lives of indi-
viduals diagnosed with various learning disabilities. However, the same level
of success was not observed when behavior analysis turned its attention to
human language and cognition. Noam Chomsky’s damning review of B. F.
Skinner’s Verbal Behavior is well known, and is sometimes offered as “proof ”
that the nonmediational approach of behavior analysis could not stretch to
the more advanced or sophisticated aspects of human psychology (such as
language and thought). In fact, Skinner’s work did go on to provide the basis
for a number of language training programs, but once again success was
limited largely to learning-disabled populations.
The key problem with Verbal Behavior, and one that Chomsky highlighted,
is that it fails to address the highly generative nature of human language.
Although the book does not leave this issue completely untouched, it fails
Learning RFT
to provide a well-developed technical account of the almost infinite novelty

that language can generate. Furthermore, its treatment of advanced language
phenomena such as metaphor and analogy is unconvincing. For example,
in providing functional analytic interpretations of these verbal behaviors, it
interweaves lay terms with technical terms, and the resulting analyses thus
lack the required precision.
But of course Skinner wrote the book almost two decades before Murray
Sidman’s first study on equivalence class formation was conducted, with all
of the implications for the behavioral study of human language that emerged
from that seminal research. Skinner was clearly at a serious disadvantage in
not having access to this equivalence data set and the conceptual work that
followed. We now know that derived relational responding appears quite
early in the behavioral repertoires of young children, and modern behavioral
treatments of human language and cognition have emerged from this work.
Unfortunately for Skinner’s Verbal Behavior, this research was lacking in the
book. For the most part, therefore, Verbal Behavior was a direct contingency
account of human language that made only passing reference to the most
important defining feature of verbal behavior: derived relational responding.
The first book on relational frame theory, Relational Frame Theory: A
Post-Skinnerian Account of Human Language and Cognition (Hayes, Barnes-
Holmes, & Roche, 2001), published over thirty years after Skinner’s work,
aimed to present a modern behavior analytic account of human language and
cognition. The theory embraced derived relations and indeed put them at the
very heart of the account. Nevertheless, RFT remained a natural extension of
earlier conceptual and empirical research within behavior analysis. The core
concept in the book, arbitrarily applicable relational responding, was based
solidly on Skinner’s concept of the operant and drew heavily on Sidman’s
seminal work on equivalence classes. Specifically, equivalence class formation
was seen as the result of a history of operant conditioning (a learned response
class), and based on this argument, the possibility of multiple forms of such
response classes (relational frames) was predicted.
The 2001 RFT book explains how the basic analytic units of human
verbal behavior—relational frames—may be combined into more complex
units, giving rise to the relating of relational frames, increasingly complex
relational networks, and the relating of entire relational networks to other
relational networks. These types of highly abstract concepts are used in the
book to provide a nonmediational and purely functional analytic account
of the full range of human verbal abilities, including naming, storytelling,
humor, abstract logic, the verbal construction of self, and spirituality.
The primary purpose behind the 2001 RFT book involved provid-
ing more than a modern behavioral interpretation of human language and
x
cognition. Its purpose was intensely pragmatic. Among other objectives, the
book aimed to stimulate both basic and applied research on human language
and cognition, and to provide a set of functional analytic terms that would
facilitate communication among researchers and practitioners. It appears to
have been relatively successful in achieving the first goal, but the latter objec-
tive, I believe, requires another book: the one you are currently reading.
Relational Frame Theory: A Post-Skinnerian Account of Human Language
and Cognition is intensely academic, full of jargon, and littered with highly
abstract concepts. Learning RTF contains many of those concepts and some
of the jargon, but it presents the material in a very accessible manner and,
critically, does complete justice to the subject matter.
The first section of Learning RFT starts with a succinct but well-worked
introduction to the philosophical and conceptual underpinnings of behav-
ior analysis, an understanding of which is essential in grappling with what is
to follow. The topics of thinking and human language are then introduced,
and the traditional Skinnerian perspective on these topics is explained and
contrasted with that of traditional cognitive therapy. In this examination it is
proposed that neither approach has fully dealt with the role of thinking and
language, at least in the clinical domain. The first section of the book will
serve as a strong motivator for the reader, and particularly the clinician, to
delve into the next and perhaps most challenging section.
In the second part of Learning RFT, the theory itself is presented, but in
a highly accessible way. The chapters in this section strike a perfect balance
between providing an appropriate level of technical detail and keeping the
writing lively, light, and a pleasure to read. Furthermore, although the earlier
chapters in this section focus necessarily on the more abstract features of
RFT, the writing progresses rapidly and with relative ease to issues that will
be of more interest to the practicing clinician, dealing with topics such as
self and perspective taking. The final chapter in this section, “The Dark Side
of Human Languaging,” will be particularly relevant to clinicians in that it
explains how RFT means that human language and cognition may be the
source of much human suffering.
The third and final section of the book focuses on the clinical implica-
tions of RFT. The section begins with traditional behavior therapy and how
it relates to other therapeutic approaches, then it explains how RFT makes a
unique contribution to our understanding of psychotherapy itself. The reader
is now prepared for the final chapters of the book, which work systematically
through the application of modern behavior analysis to clinical psychology.
This material provides a powerful review of clinical behavior analysis and
in particular explains how RFT supplements and extends the traditional
behavior therapeutic approach. It is only in this final part of the book that
xi
Learning RFT
the intensely pragmatic nature of RFT is fully revealed. This highly abstract
and arcane theory allows the practitioner to conceptualize human language
and thought as composed of behavioral units that may be subjected to func-
tional analyses and behavioral intervention strategies. In short, Learning RFT
clearly illustrates in a very powerful way how RFT can contribute toward the
conceptualization and treatment of human suffering. In truth, this is a book
I would love to have written.
—Dermot Barnes-Holmes
National University of Ireland, Maynooth
Dermot Barnes-Holmes is foundation Professor of Psychology at the National

University of Ireland, Maynooth. He has published approximately two
hundred scientific articles, book chapters, and books, the vast majority of
which have focused on the study of human language and cognition from a
behavior analytic perspective.
xii

Foreword to the Swedish
Edition
Is there a need for a book devoted to relational frame theory? Can we learn
anything about language by endorsing the principles of learning? Honestly, is
it not the case that cognitive psychologists do a much better job at explain-
ing how we think? My answers to the first two questions are affirmative, and
Learning RFT is an important contribution, giving a thorough description of
how a behaviorist framework can help us understand cognition and language.
While some readers might not find this topic attractive at first sight, perhaps
because they have a clinical focus, my conviction is that this book can be
helpful for clinicians as well as researchers and that the principles described
are put forward in a reader-friendly manner, facilitating comprehension of the
sometimes difficult-to-grasp concepts in relational frame theory.
I received my training as a clinical psychologist in the mid-1980s, and
most of the textbooks I read stated that behaviorism was dead and that the
cognitive revolution had taken over after the “dark years” under the rule of
B. F. Skinner. However, not all my teachers shared that opinion, and at the
department of psychology in Uppsala, Sweden, I got the basics of applied
behavior analysis and developed an interest in behaviorism. I guess I can say
I acquired what might best be called a “behaviorist framework,” which influ-
enced my clinical work and research, by that time focused on hearing loss in
the elderly. In my work I found operant psychology very useful and ended up
with a thesis entitled Hearing as Behavior. However, I was also painfully aware
of the fairly low status of operant psychology in mainstream psychology, and
while I found the work by Skinner useful, I could not fully appreciate his
book Verbal Behavior. On the other hand, the work by Steven Hayes and his
Learning RFT
colleagues on rule-governed behavior was in my awareness, but I never saw it

mentioned in the cognitive psychology literature. While I could understand
the objections to Skinner’s analysis of language, I thought Chomsky’s review
from 1959 was overly negativistic. All of that said, the psychology of learning
and behavior still makes a difference, and the development of relational frame
theory (RFT) clearly indicates that it was premature to dismiss behaviorism
as a framework for the understanding of language and cognition. When many
researchers and students of psychology believed behaviorism was long gone,
interesting activities continued to occur in this field. Learning RFT provides
an excellent summary of what happened after Skinner, and also does justice
to what Skinner probably meant with his analysis of verbal behavior.
What does RFT add, then? I am not too convinced that RFT is necessary
for how we provide effective psychological treatments, but I do believe we can
benefit from a good theory, and in addition to wanting to be good clinicians,
we also want to be able to understand and explain human behavior. Since
language and thinking are integral parts of how we understand each other
in lay terms, it makes sense that behavior therapists were attracted to cogni-
tive therapy, in which thoughts and beliefs were prioritized. However, not all
behavior therapists were convinced. In clinical psychology, and in particu-
lar in the field of psychotherapy, it could be argued that basic psychological
science, such as cognitive psychology, is used to inform clinical practice while
not really being part of clinical science. For example, clinicians sometimes
regard Aaron Beck’s cognitive therapy as part of cognitive psychology, but in
reality, basic cognitive psychology on constructs such as working memory has
only more recently been applied in clinical research within the psychotherapy
field. Most of what has been written about cognitive therapy has very little in
common with basic cognitive psychology. Within the behavior therapy com-
munity, this has sometimes led to a negative attitude toward what cognitive
psychology can contribute, as we sometimes falsely assume that cognitive
therapy equals cognitive psychology. This might be a mistake, as overviews
of the scientific status of different subfields in psychology clearly show that
cognitive psychology is the leading branch of psychology when it comes to
research funding, publications, and citations. But perhaps the emergence
of RFT will change the situation for behavior therapists with an interest in
language and cognition. It is probably fair to say that behaviorists had not
finished their work on language and cognition in the lab, and I hope the work
on RFT will inspire therapists to look more closely at basic research in lan-
guage and cognition. Even if knowledge of RFT isn’t necessary to do a good
job as a clinician, good explanations of human behavior definitely are. Also,
because of our background in psychology, we need data to be convinced that
xiv
something is likely to be true. With RFT, we now have more tools to explain
the verbal behavior we confront in our clinical work.
You may wonder about the third question I raised: Is it not true that cog-
nitive psychologists are doing a better job at explaining cognition? Learning
RFT is excellent in this respect, as the author does not ignore the substantial
literature on the cognitive psychology of language; rather, he relates it to RFT.
For me, and most likely for many other psychologists, this makes it easier to
understand RFT and take it seriously as a major contribution to psychology.
So here’s the answer to my last question: To date, cognitive psychology is, if
not the best, at least the most productive when it comes to language and cog-
nition. But RFT need not be seen as in opposition to the rest of psychology,
and it can provide us with important clues to further our understanding of
language and cognition. It is possible that a behaviorist renaissance is on the
horizon. Niklas Törneke’s book is one of the building blocks in that venture.
—Gerhard Andersson, Ph.D., professor of clinical psychology in the
Department of Behavioral Sciences and Learning at Linköping
University, in Linköping, Sweden
xv

Acknowledgments
In 1998 I went to an international conference in Ireland and for the first time
heard two people speak who, more than any others, are behind the ideas and
the research this book is based on: Steven Hayes and Dermot Barnes-Holmes.
From that time, both of them have generously answered my questions and
helped me become familiar with an outlook and research tradition that, up
until then, had been essentially unknown to me. Many thanks to both!
The person who first told me I should write a book like this is Kelly
Wilson. He is also the person I have mainly learned ACT from in practice. I
owe him warm thanks, as well.
A fourth person who has meant a lot to me in the process leading to
this book is Carmen Luciano. She too is a leading figure in the international
network of researchers and clinicians bound together by a common interest in
RFT and ACT. In recent years she has been a never-ending source of knowl-
edge and inspiration to me.
Several individuals in Sweden have also been of particular help to me.
First and foremost, Jonas Ramnerö. Ever since we met at a conference in
Dresden in 1999, we have been engaged in an ongoing dialogue on the role
of behavioral psychology in psychotherapy. This dialogue has made a decisive
contribution to my writing this book. Jonas has also helped me by reading
and commenting on the Swedish manuscript, as have Jonas Bjärehed, Martin
Cernvall, and Billy Larsson.
As for the English version, Kelly Koerner, Rainer Sonntag, and Ian
Stewart have all read parts of an earlier version of the manuscript and made
many valuable suggestions. All deficiencies remain, of course, my own respon-
sibility. Elizabeth Ask de Lambert did most of the work translating the book
Learning RFT
into English, and Jasmine Star made the editing process extremely helpful
and smooth. Görel Gunnarsson and her colleagues in the medical library
at the county hospital in Kalmar, Sweden, have been of invaluable help in
obtaining articles and other literature. My heartfelt and sincerest thanks to
one and all!
xviii

A Personal Word of
Introduction
As a psychotherapist, I am a child of my time. I grew up in a psychodynamic

world dominated by a strong emphasis on understanding, but not—as far
as I could see—well anchored in scientific research. It also lacked concrete
guidance regarding workable therapeutic interventions. My encounter with
cognitive therapy at the end of the 1980s was therefore a liberating experi-
ence. Scientific foundation was paramount, and the therapeutic strategies
were applicable in my everyday work within psychiatry. Cognitive theory as
a basis of psychotherapy has since been victorious—not only in my world,
but in the world of psychotherapy at large. Cognitive therapy has gradually
been integrated with behavior therapy under the designation of “cognitive
behavioral therapies” (CBT), but the different hybrids are dominated by the
theories underlying the cognitive model.
During the 1990s, I progressively found what I saw as shortcomings in
cognitive theory. It was difficult to get a clear idea of what the basic termi-
nology was and what scientific support it had. Different cognitive theorists
liberally used their different terms to describe what was taking place in the
“psyche,” and this obscurity and lack of consensus became all the more trou-
blesome due to the assumption that the core of psychological problems was
located somewhere inside this unknown realm. In the mid-1990s, I was still
largely unaware of any vigorous alternative. I did, however, run into some
works that aroused my interest, notably Marsha Linehan’s 1993 book on
Learning RFT
dialectical behavior therapy (DBT), which revealed a strong influence from

classical behavior therapy.
Through my interest in affect theory and the use of metaphors, I came
into contact with acceptance and commitment therapy (ACT), a therapy
model related to DBT in many respects, but with a much more elaborate
theoretical and experimental foundation. Once again I was provided with
new, useful tools. In addition, I found myself in more extensive contact with
the new theoretical approach ACT had been built on, as well as with classical
learning theory as a basis for psychological therapy. I realized that one cannot
really understand ACT, much less its theoretical foundation, relational frame
theory (RFT), without an understanding of basic behavioral principles—that
is, operant and respondent conditioning. Once I gradually became familiar
with these principles, I discovered what I have come to regard as the most
promising psychological model available when it comes to understanding
human behavior in a way that also contributes directly to the therapeutic
work for change. This led to an inspiring dialogue with my good friend Jonas
Ramnerö, who’d had a somewhat longer history with behavioral psychology
than I had. This dialogue resulted in our writing a book together: The ABCs
of Human Behavior: Behavioral Principles for the Practicing Clinician (Ramnerö
& Törneke, 2008).
Recent years have seen a growing interest in behavioral principles, not
least among psychotherapists with cognitive training. This new interest has
given me a reason to work more thoroughly on the issue that has engaged
me for quite a few years by now: the issue of the power of thinking in rela-
tion to other human behavior and the role this plays in the problems that
make people seek help in psychotherapy. My purpose in this book is to elu-
cidate the behavioral perspective on this question. You might say that this
book attempts to respond to the questions raised by cognitive theory and
therapy, although the answers are given from a different point of departure
than those common to cognitive approaches. Instead, the starting point here
is the foundation laid down years ago in the shape of operant and respondent
conditioning. Behavioral psychology has long had problems in applying its
basic agenda of prediction and influence in this area. I think these problems
are coming to an end. The increasing basic research performed to describe
relational conditioning and the theoretical structure that has evolved around
this phenomenon (RFT) provides new answers and opens the door to new
interventions in respect to cognition and human language. All this is done
from the agenda of classical behavioral therapy, with new interventions being
founded upon data from experimental research.
For me to begin this book by contrasting its content with psychodynamic
and cognitive models may seem provocative and deprecating. This is not my
2
A Personal Word of Introduction
intention. I am aware that the psychodynamic and cognitive theories that

I earlier found unsatisfactory have not stayed unchanged since my point of
departure; they have been developing in their own way. Moreover, I believe
that a behavioral perspective is integrative in nature. Behavioral psychology
is not about one specific model of therapy; it describes fundamental, universal
principles of behavior. It is legitimate to approach anything and everything
a person does from this perspective. Behavioral psychology is not limited in
scope to phenomena that are easy to observe and define, like when someone
avoids air travel or washes her hands obsessively. It also applies to behaviors
in the close interactions between people that are harder to capture, like estab-
lishing and maintaining close relationships or behaving toward the therapist
much like one earlier behaved toward a parent. Even behavior that in many
ways seems concealed to anyone other than the person who is doing it, like
dwelling on past grudges or struggling with feelings of dejection, can be
approached from this perspective. The same goes for behaviors that are rarely
the focus in psychotherapy, like playing the flute or writing poetry.
Thus, nothing human is alien to behavioral psychology. All human phe-
nomena that take place in time and space and that can be the subject of atten-
tion in psychodynamic or cognitive therapy can also be approached from this
perspective, hence its integrative nature. But the starting point of this analysis
is a given theoretical position: the one which, since the days of Skinner, has
been designated “radical behaviorism.” Some of this is novel, but the founda-
tion was laid quite a few years back.
A WORD ON TERMINOLOGY
Skinner called the science of behavior that he developed behavior analysis.
However, this term is used in slightly different ways. Within behavior therapy
in Europe, it is sometimes used synonymously with the word “conceptualiza-
tion.” In this usage, a behavior analysis is understood as an initial phase of
behavior therapy. I will be using the term in the way Skinner did, which is
how it is still used in the United States. Used in this way, “behavior analysis”
is a designation of the science, as a whole, that aims at predicting and influ-
encing behavior, along with the practical work involved in doing this. There
is usually a distinction made between two branches within behavior analysis:
experimental behavior analysis and applied behavior analysis. Experimental
analysis of behavior is the type of experimental activity usually connected
with Skinner: Under carefully monitored conditions, different factors are
varied to determine whether an organism’s behavior can be predicted and
influenced. In applied behavior analysis, the basic principles that can be
3
Learning RFT
described following the experimental work are applied to different types of

problems “out there, in real life.” A branch of applied behavior analysis is clini-
cal behavior analysis. This is behavior analysis in the field commonly referred
to as psychotherapy. Consequently, in our book The ABCs of Human Behavior:
Behavioral Principles for the Practicing Clinician (2008), Jonas Ramnerö and
I called this field of application “behavioral psychotherapy.” In this book,
however, I will mainly be using the term “clinical behavior analysis.”
OUTLINE
This book is divided into three parts. Part 1 provides some important back-
ground. Chapter 1 offers a short account of basic and well-known principles of
learning from the viewpoint of radical behaviorism, with a particular empha-
sis on concepts that must be understood in order to become familiar with
RFT. Chapter 2 provides a survey of how behavior analysis had tried to tackle
“the power of thinking” before the experimental data on which RFT is based
were available. The bulk of this chapter consists of an overview of Skinner’s
analysis of verbal behavior. Although his analysis has limitations (described
here as well), it remains important as a backdrop to RFT. In chapter 3, argu-
ments for renewed inquiry into human cognition and language conclude part
1 of the book.
Part 2 of the book is its core; this is where RFT is described. Chapter 4
presents and defines RFT’s basic terminology and describes the type of experi-
ments the theory is based on. In essence, chapter 4 describes the fundamental
elements in human language. In chapters 5 and 6, I have attempted to show
how these building blocks are combined with an increasing degree of com-
plexity, and how they cast new light on complex human behavior. In chapter
7, part 2 concludes with an account of the problems that verbal (cognitive)
behavior creates for human beings, or the side effects of human language.
Part 3 of the book describes clinical applications. Chapter 8 takes a
look at psychological therapies in general from a behavioral perspective. The
remaining three chapters focus on clinical behavior analysis, with particular
emphasis on strategies and techniques based in RFT.
THE CHARACTER OF THE BOOK

The number of scientific texts presenting RFT and its experimental base is
growing rapidly. The same is true for books presenting ACT. This book has
the character of being in between these two categories of texts. Although the
4
A Personal Word of Introduction
purpose is to give an overall introduction to RFT, the book has its limita-
tions. The main limitation lies in maintaining both a theoretical and a clini-
cal perspective. Though RFT is based on experimental research, this book
does not present the experimental work in detail; it simply gives an outline of
the experimental work and devotes more attention to the conclusions drawn
from that work. This book is more focused on concepts than data and details,
partly to give a general introduction, and partly to give an understanding that
facilitates clinical work. The book does not include more detailed presenta-
tions of the experiments in their entirety, such as how they are arranged and
performed. I have tried, however, to frequently refer to literature that contains
such presentations so that the interested reader can find more in-depth mate-
rial. There is also a paradox involved in this limitation. There is a degree of
learning RFT that can be achieved only by engaging in experimental work.
Yet this is a book by someone who has never done that, written primarily for
others in the same situation. Experimental psychologists will probably find
lack of precision and technical detail. The same might be true of others who
are very well acquainted with the existing scientific literature. At the same
time, some readers will probably find parts of the book too technical and
abstract. Still, this kind of book, “in between,” is what I wished to read when
I first encountered RFT. Hopefully it will be helpful to others who are now
in the situation I was in then.
ACT has a central position in part 3 of the book, on clinical applications.
This is only natural, as this therapeutic model has evolved together with RFT.
Alongside ACT, other forms of clinical behavior analysis, especially functional
analytical psychotherapy and behavioral activation, have their place. It has
not been my goal, however, to present any of these individual models in their
entirety, or to carry out an in-depth comparison. I want to pursue the agenda
outlined in The ABCs of Human Behavior (Ramnerö & Törneke, 2008): to
describe psychological therapy from the broad perspective of radical behav-
iorism, and to describe the therapeutic tradition that can be called behavior
therapy, behavioral psychotherapy, or clinical behavior analysis. I want to do
this with a special emphasis on how an understanding of RFT adds some new
elements to this tradition.
5
PART 1
Background
CHAPTER 1
Radical Behaviorism and
Fundamental Behavior
Analytic Principles
Radical behaviorism is the philosophical basis on which B. F. Skinner built

his psychology (1953). This term has caused a lot of debate over the years.
It has also frequently been misunderstood, sometimes to such a degree that
one might wonder if the term is useful at all or if it has actually become an
obstacle to introducing the psychology itself. It is common to see Skinner’s
views described as superficial and coarsely mechanistic, even in psychology
textbooks (Power & Dalgleish, 1997, pp. 35-36; Solso, MacLin, & MacLin,
2005, p. 329). This is in stark contrast to my own impression from reading
Skinner, and I have often wondered if writers who portray his positions in
this way have actually read his works. Be that as it may, his positions, and
the terms he uses to describe them, are controversial, and the term “radical
behaviorism” is a clear example of that. An alternative and more modern term
is functional contextualism (Gifford & Hayes, 1999). This alternative term
may better convey in what way this particular philosophy of science relates
to other modern approaches. It puts Skinner’s position in relation to alterna-
tive types of contextualism, such as social constructivism or certain types
of feminism (Roche & Barnes-Holmes, 2003; Gifford & Hayes, 1999). The
term “functional contextualism” emphasizes two essential elements in radical
behaviorism: The first is that behavior must always be understood in relation
to the setting, or context, in which it takes place. The second is that in order
Learning RFT
to understand and influence behavior, we need to study its function—that is,

what it is aimed at.
My reason for using the term “radical behaviorism” to start out with,
despite the above discussion, is that this is the term that has survived and is
widely accepted among those who follow in Skinner’s footsteps in their work.
It is linguistically correct, and it also brings out some essential elements of the
position on which this book is based.
BEHAVIORISM AND ITS PREMISES

Let’s begin with a discussion of the more general term “behaviorism.” This is
a broad term that encompasses many partly differing approaches (O’Donohue
& Kitchener, 1998). Yet these approaches have certain premises in common,
making the concept of behaviorism appropriate. The most fundamental premise
was made clear by Watson, who came up with the term “behaviorism.” In this
approach, what is in focus is behavior (Watson, 1929); that is, something a
person—or any other organism—is doing. The actions, or responses, of the
whole organism are in focus. Another premise that the different approaches
have in common is the method of seeking knowledge: The science is to be
built from the bottom up. The quest is for fundamental, universally valid prin-
ciples for understanding behavior. This means that laboratory experiments
have a strong position. When carrying out experiments, a key concept is to
try to minimize noncontrollable variables as much as possible before going on
to identify and systematically manipulate those variables that are essential. In
many ways this resembles how psychoanalysts are strict about the setting for
their sessions, so as to eliminate irrelevant disturbances and observe relevant,
governing phenomena in the interaction taking place. The most well-known,
classical examples of this method within behaviorism are perhaps Skinner’s
experiments with pigeons and rats. The environment is stripped (there is only
a box), and the relevant variables are few (the box contains a bar that enables
the animal to act in order to receive food, and a light that is turned on and
off). What is important here is not the laboratory experiments as such, or even
the actions of pigeons or rats; rather, the aim in using this method is to be able
to identify principles underlying the behavior of organisms—principles that
can then be used to understand more complex processes that perhaps cannot
be studied in laboratories.
This reveals another premise of behaviorism: that of an assumed continu-
ity across different organisms. Research on pigeons, for example, is used to
draw conclusions about human beings, at least in certain respects. This has
sometimes been a point of controversy, especially among psychotherapists.
10
Radical Behaviorism and Fundamental Behavior Analytic Principles
Is it possible to understand humans based on an understanding of animals?

This was the topic of heated discussion during the 1960s and 1970s, but a lot
of water has flowed under the bridge since then. Today, it is uncontroversial
to say that evolutionary psychology and neuropsychology, as well as ethology,
hold firm positions in shaping various theories within psychotherapy, regard-
less of the specific camp. Based on an evolutionary perspective, researchers
who have greatly influenced different psychotherapy traditions, such as John
Bowlby (attachment theory) and Joseph LeDoux (affect theory), take the
same view as what Skinner once maintained: simply stated, that evolution
continuously builds on what is already at hand. Functions that work are not
removed; they become building blocks in future developments. This is why
we can learn a lot about human beings by studying how gorillas relate to
their offspring (attachment theory) or by studying basic cerebral functions in
animals (affect theory).
Now, if behavior is what is to be studied, the question is how we define
this term. What should be counted as behavior? The answer to this question
can be expressed in somewhat different ways within behaviorism, and in
order to account for what is meant by behavior in this book, I will now turn
to Skinner’s view on this, as encapsulated in the term “radical behaviorism.”
WHAT IS RADICAL ABOUT RADICAL

BEHAVIORISM?
Being radical can be taken as being extreme. This is not what Skinner had
in mind, though, when he chose the term “radical.” In this context, “radical”
implies not “extreme” but “consistent.” Radical behaviorism entails not a
departure from fundamental behavioristic principles but the application of
them in an all-inclusive way. This has some consequences. For example, take
the principle that Skinner used to describe operant conditioning (more on
this below). This principle implies that our actions are influenced by the con-
sequences we have previously encountered following a particular action. The
probability that a pigeon will peck at a certain point increases if it has earlier
received food after pecking at that particular point. But if this is to be applied
in a consistent way, in keeping with Skinner’s position, then this principle
also holds for me as a scientist. I do what I do (in my experiment with the
pigeon) as a consequence of outcomes of similar experimentation earlier. As
a scientist, I do not hold any objective or exclusive position. I am not outside
or above the principles I study. If this understanding is applied consistently,
all claims of representing the ontological truth have to be dropped. Based on
11
Learning RFT
this position, we cannot maintain that “this is the way it really is.” Radical
behaviorists repudiate the notion that the scientist operates from an objective
and neutral position. As mentioned earlier, from the perspective of radical
behaviorism you cannot understand behavior without studying its context.
All behavior takes place within a context. But neither can the context be
studied independent of behavior. This is because the scientist’s attempt to
study something is a behavior as well. After all, the object of our study is
something that we are acting upon, just by studying it. So just as we cannot
understand behavior without context, there is no context available for the
organism without behavior.
This point about the behavior of the scientist is also true in a more general
sense. Stimulus and response (behavior) are codependent and should be con-
sidered together. They make up a single unity (Kantor, 1970). We can separate
them for practical reasons, with a certain aim in mind. And the behavioral
science that Skinner wanted to create has an aim: to predict and influence
behavior. Radical behaviorists are not claiming to be “uncovering reality”;
rather, we maintain that this method, the scientific project of radical behav-
iorism, is a method that works for what we want to do. The pigeon in Skinner’s
experiment could say something similar: “Pecking this spot works when it
comes to getting Skinner to give me food.”
When we radically apply the fundamental principles of behavior that
we have identified, this leads to another important result, one involving the
definition of the term “behavior.” In everyday speech, the word “behavior”
normally refers only to external actions, which can be observed by anyone
else who is present. So how should we regard the things a person does but
that no one except the person himself can observe, things like feeling, remem-
bering, and thinking? Traditionally, these phenomena have been assigned to
another sphere—the psyche—as if they were of a different nature than the
things we can observe. Here too, Skinner called for consistency, maintaining
that there is nothing to indicate that the same principles are not valid for these
phenomena as well (Skinner, 1953, 1974). This means these phenomena are
also behavior, and that they can and should be analyzed according to the same
principles as behavior that is observable by others.
FUNDAMENTAL BEHAVIOR ANALYTIC

PRINCIPLES
What, then, are the fundamental principles demonstrated by and studied
through experimental research that we can use to understand and influence
12
behavior? For a more detailed answer to this question, the reader is referred
to other publications (Catania, 2007; Ramnerö & Törneke, 2008). Still, I
will provide a short summary here, before turning to this book’s main quest,
which is to point out how these principles should be used to shed light upon
the function of human thinking.
The two fundamental principles for behavior analysis are operant and
respondent conditioning. The latter has been described since Pavlov’s well-
known experiment with dogs at the beginning of the twentieth century,
including how their natural reaction of salivation can be influenced through
conditioning. Operant conditioning is the principle of learning that Skinner
investigated and demonstrated via his experiments, so that is where I will
begin.
Operant Conditioning: Learning Through

Consequences
Human actions never take place in a vacuum. There is something pre-
ceding and something following each action. It is among these contextual
factors—those that precede and those that follow—that the behavioral analyst
looks for answers to questions about what governs behavior. If someone, in
a certain context, turns his eyes to me with a specific expression on his face,
I might address him by saying something like “Can I help you?” My utter-
ance is followed by a new occurrence: The person replies. So my action is
followed by a consequence, in this case that someone answers me. The core
principle in operant conditioning is that the consequences following a behav-
ior (a response) influence the probability of the behavior being repeated.
Let’s speculate a bit using two rather different consequences in the everyday
example above. Imagine that what follows upon my utterance “Can I help
you?” is that the person gives me a friendly smile and tells me what he wants.
An alternative would be the response “Mind your own business, you jerk!”
It is hard to say exactly how each of these consequences might influence the
probability of me asking if I could help if someone else were to look at me with
this specific facial expression in the future. What is essential is that earlier
consequences do have an influence. “A burned child shuns the fire” is an old
saying in Swedish, corresponding to “Once bitten, twice shy” in English. A
well-known author has turned this “truth” around in a book entitled Burned
Child Seeks the Fire (Edvardson, 1997). It is not always easy to determine what
kind of behavior to expect based on previous consequences. But the saying
and the author agree on one thing: Previous consequences have an influence.
This is the core of operant learning.
13
Learning RFT
In operant psychology, the different influences of consequences are cate-

gorized based on whether they increase or reduce the probability of an earlier
behavior being repeated. If I get friendly responses to utterances like the one
above and thereafter more often address people who exhibit the same type of
facial expression in similar situations, the friendly responses would be said to
have had a reinforcing effect on this specific behavior of mine. A consequence
that increases the probability of the preceding behavior being repeated is thus
termed reinforcing. In this case the reinforcing consequence is that I receive
something: a friendly response. Something is added. This kind of process is
called positive reinforcement.
A behavior can also be reinforced through a consequence consisting of
something being removed. This is illustrated by the behavior of the person
who responded to my question by saying, “Mind your own business, you jerk!”
Let’s assume that when this is uttered, I become quiet and turn my attention
away from the speaker. I do what I am told to do. This consequence could
increase the probability of the other person repeating the utterance “Mind
your own business” in similar situations in the future. The consequence—
that I became quiet and turned away—has in this case become reinforcing
to this person’s behavior of telling off jerks. This time, though, the reinforce-
ment consists of something being removed, namely, the attention from a jerk.
When a behavior increases because something is removed, it is termed nega-
tive reinforcement.
Distinguishing between positive and negative reinforcement (which are
both processes that increase the probability of a certain behavior) is not always
essential. These two concepts can be said to describe two different sides of the
same thing (Michael, 1975). If the behavior of telling off jerks was reinforced
by me becoming quiet, then one factor is that my annoying questions came
to an end. This is negative reinforcement. But another way of describing the
same thing is to note the condition that resulted: silence, for example, or any-
thing else that was added. That would be positive reinforcement. It is often
convenient to distinguish between positive and negative reinforcement, even
though the difference may not be clear-cut from a theoretical perspective. It
may sometimes be more obvious that something is removed than that some-
thing is added. By speaking about this as negative reinforcement, we clarify
the process. The distinction is often practical in clinical situations, a subject I
will return to in part 3 of the book.
When a consequence reduces the probability of a certain behavior being
repeated, it is known as punishment. If I receive an unfriendly response in
the example above and I subsequently refrain from addressing people in
14
that specific social context or do so less often, then the earlier consequence
has been punishing. Punishment, too, can be separated into positive punish-
ment—in which something has been added—and negative punishment—in
which something has been removed. Remember, however, that there is no
way of determining what is reinforcing versus punishing based on any intrin-
sic quality that signifies the consequence as such. Of course, it is true that
some consequences more often function as reinforcing to people, for example,
certain types of social attention. But this is not always the case. To be recog-
nized and addressed in a kind way is usually reinforcing for human behavior,
but we can all think of situations when this is something we want to avoid.
Likewise, certain consequences usually function in a punishing way, like
being hit, yet this is not always the case. There are situations when being hit
reinforces the behavior that preceded this consequence. A child who encoun-
ters only indifference, despite several actions meant to attract attention, may
repeat a behavior that leads to getting smacked, simply because the smack
involves attention. It is the function of the consequence that provides the defi-
nition. When the consequences increase the probability of a certain behavior,
this is reinforcement; and when the consequences reduce the probability of a
behavior, it is punishment. To recap:
DI F F ER E N T T Y PE S OF CONSEQU E NCE S
Reinforcement: A consequence that increases the likelihood that a
certain behavior will be repeated.
Positive reinforcement is when the consequence is
something that is added.
Negative reinforcement is when the consequence is
something that is taken away.
Punishment: A consequence that decreases the likelihood that a

certain behavior will be repeated.
Positive punishment is when the consequence is some-
thing that is added.
Negative punishment is when the consequence is some-
thing that is taken away.
15
Learning RFT
Before I go on to describe how we can analyze a specific behavior, I must

clarify what it is that we are analyzing. Two behaviors are rarely, if ever, identi-
cal; even if they seem to be, they differ in detail. I can raise my coffee cup in
a number of different ways, and I can address another person in many differ-
ent ways. In behavior analysis, behaviors that are alike in the sense that they
have the same function are said to belong to the same functional class. This
categorization is essential if we are to analyze behavior. When you analyze
a given, historically defined behavior, it is of course this particular behavior
that you are analyzing—for instance, the way I addressed someone in the
above example. At the same time, this is only of interest if it is useful when
analyzing a similar behavior in the future—a behavior that is similar enough
to have the same or almost the same function. Over time, we are primarily
interested in functional classes or categories of behavior. Some such classes of
behavior are narrow or very specific, such as the behavior of a biathlon com-
petitor in erect shooting position when he aims to score a bull’s-eye. Other
classes are broad or contain a wide range of behavior, such as what people do
to avoid painful memories.
ABC
In behavioral analysis, a common way of describing an operant sequence

of events is ABC, and analyzing such a sequence is referred to as carrying out
a functional analysis. The core is B, which stands for “behavior”: Something
is being done. This behavior or response is what we intend to predict and
influence. C stands for “consequence”; we have just seen the significant role it
plays in an operant analysis. Finally, there is A, which stands for “antecedent,”
or “that which precedes.” Even the consequences that govern a certain behav-
ior actually precede the behavior they govern, because they have followed an
earlier, similar behavior. If speaking to another person has previously been
followed by receiving kind attention, the likelihood of this behavior being
repeated may increase. However, the antecedents referred to as A in ABC are
those conditions that are present when a behavior occurs. Within behavioral
analysis, we think of A as having at least two different types of function: dis-
criminative function and motivational function. I will begin by describing
discriminative function.
The fact that my question (“Can I help you?”) was reinforced earlier does
not mean that from then on I will always be asking the same question. This
behavior has been reinforced in a specific context, and it is in this context—or
rather, in similar contexts—that the likelihood of my asking the question again
16
increases. Antecedent refers to precisely this context: the conditions under

which my behavior previously put me in contact with certain consequences
that have become governing. It was under the condition of having a person
look at me with a certain facial expression (A) that I spoke to the person (B)
and as a result of this behavior encountered different consequences (C). This
means that A now has a function established by a certain connection in my
history—a connection between a condition, a behavior, and a consequence.
Now, when I meet a new condition that is similar enough to the one I encoun-
tered earlier, the previous consequence influences my current behavior. This
function of A is termed discriminative, and when this is the function we have
in mind, we speak of the antecedent as a discriminative stimulus.
A discriminative stimulus signals a historical connection between a
behavior and a certain consequence. A specific type of behavior in another
person—a glance or a facial expression, for example—signals to an individual
a historical connection between a certain behavior, like asking a question, and
a certain consequence. You might say that a discriminative antecedent in the
present signals the availability of a certain consequence, simply based on this
historical contiguity. This connection is called a contingency in behavior ana-
lytic language. We also say that a certain consequence needs to be contingent
on a certain behavior in order to exert its function. What we mean is that
there must be a direct connection between the behavior and the consequence.
In order for discriminative antecedents and reinforcing or punishing conse-
quences to have their respective functions in relation to a specific behavior,
they must occur in contiguity with that behavior. A great deal of research
has been carried out in the field of experimental analysis of behavior to try
to describe these connections more explicitly, as well as how they may vary
(Catania, 2007).
The prevailing circumstances (A) can affect the probability of a certain
behavior in another way too. There are circumstances that are not discrimina-
tive; that is, they don’t signal any historical connection between a behavior
and a certain consequence. Although they don’t signal an increased availabil-
ity of a specific consequence, they influence the likelihood of a certain behav-
ior. A classic example is food deprivation (hunger). If my daughter walks past
the kitchen, where I am cooking a meal, and I say, “Dinner is ready,” this
might function as a discriminative stimulus for her. If so, she stops walking
toward the TV room and sits down at the kitchen table. There could be differ-
ent types of learning history behind this, but one possibility is that my words
signal to my daughter a historical connection between the current conditions
and the availability of a certain consequence: that of being served food. This
17
Learning RFT
is a description of a discriminative function.1 Alongside this, however, the

fact that my daughter is either full (after just eating a couple of sandwiches)
or hungry (because she has not eaten since breakfast) may influence whether
she sits down at the table to eat. Her being full versus being hungry does not
indicate the availability of food. The meal is equally available regardless of
whether she is hungry or not. Her hunger is representative of a different func-
tion of conditions that may precede and influence a behavior. This function
is usually called an establishing operation or motivational operation (Michael,
1993). It is a function of A, but not a discriminative function. These are ante-
cedent conditions that influence the reinforcing or punishing effectiveness of
a consequence. In this example, the dinner will be more or less reinforcing
to my daughter depending on her recent history with food—that is, whether
she is hungry or not.
Let’s take a look at the earlier example of asking “Can I help you?” with
this aspect in mind. We could assume that I have a history in which a certain
behavior by another person (facial expression, for example) functions as a dis-
criminative antecedent for asking a question. Let’s imagine two possible sce-
narios: I could be unusually tired because I didn’t sleep well the night before;
or, on the other hand, I may have been missing social contact with others for
some time. Both of these conditions could influence whether I ask my ques-
tion or not, if I indeed encounter a discriminative stimulus for this behavior.
And this is despite the fact that neither my tiredness nor my desire for human
contact indicates any availability of the governing consequence. Instead, the
issue here is that these conditions influence the degree to which the governing
consequence is motivating to me, or how much influencing power the conse-
quence has in this particular situation.
1 As I will go on to describe later, this is a simplification that does not take
into account the difference between verbal and nonverbal discriminative
functions. The example works for the intended point in this case, though.
There are quite a few of these types of simplified examples in this chap-
ter. The alternative would have been to only use examples from organisms
without human language, but this would have affected the text in a negative
way and would hardly have helped the reader’s understanding. The problem
with these types of simplifications, which have been necessary within the
area of behavioral analysis due to its difficulty in handling phenomena like
language and cognition, will be dealt with in detail later in this book. It is
also important to remember that within behavior analysis we are aiming at
usefulness, not necessarily at covering all possible aspects of an event (see
Ramnerö & Törneke, 2008).
18
BA SIC F U NCT IONS OF T H E DI F F ER E N T

PA RTS I N A N A BC A NA LYSIS
A B C
Antecedent Behavior Consequence
Discriminative function Reinforcing function
Something an individual does
Motivational function Punishing function
It is often useful to distinguish between the discriminative and motiva-

tional functions of the conditions that precede and influence operant behav-
ior. But again, it is the degree of usefulness that determines how essential this
distinction is. In practice, it is not always possible to make the distinction, and
other times it may not be important to do so even if the possibility is there.
This leads me to an important point about what ABC is not.
A NOTE ON WHAT ABC IS NOT
When we describe a behavioral sequence in the way I have above and

assign different terms to the different functions, it is easy to be misled in
thinking that we’ve discovered a mechanical chain of events “out there in
reality.” But, as mentioned earlier, this is not what is intended. This type of
discussion is simply a way of speaking and writing about behavior—a way
that is useful for gaining both understanding and influence. In the constant
flow of events we face, we can distinguish between different processes because
it is useful to us to do so. This applies to all humans at all times, even those
engaged in scientific study. Distinguishing between different processes, prin-
ciples, antecedents and consequences, reinforcement and punishment, and
so on is a behavior performed by the behavior analyst. Within this frame of
reference, knowledge is seen as a skill for doing certain things; it is not an
object that you discover or find yourself in possession of. Analyzing behavior
from the perspective of radical behaviorism is not uncovering a hidden reality.
Everything we do is acting on reality, or behaving. This means that our activ-
ity of performing ABC analyses is operant behavior, as well, governed by the
consequences that the behavior analyst has previously met when doing this
19
Learning RFT
very activity. We do what we do because we have a history of connections

between different antecedents, earlier behaviors, and consequences. We do it
because it helps us reach certain goals.
From this position, making claims regarding discovering or understand-
ing “the nature of reality” is a self-contradiction. Being based in functional
contextualism means we refrain from such claims of discovering or knowing
the “truth.” Instead, we adopt a pragmatic truth criterion wherein what is true
is what is serviceable toward a certain aim or goal. This also means that a
science needs to clarify its aims. Nothing works “in general.” If it works, it
works for something specific, for what we are aiming at. The behavior analyst
has a twofold aim or purpose: prediction and influence.
Respondent Conditioning: Learning by Association

Whereas the power of consequences over behavior is the main point
in what we call operant conditioning, respondent conditioning describes the
power of certain antecedents to trigger a reflexive behavior. Put simply, under
a certain circumstance we will react. If the same circumstance, or one much
like it, reoccurs, it provokes the same reaction based entirely on the anteced-
ent. The behavior occurs regardless of earlier consequences following this
reaction.
Again, it is important to remember that this does not uncover mechanical
processes that are actually out there “in reality,” as phenomena in their own
right. The terminology of behaviorism is simply a way of speaking about this
issue, and we employ it because it is useful. It serves our purpose to distin-
guish between operant and respondent in this way. Although these processes
coexist in the web of events we are trying to understand and influence (more
on this below), for the sake of clarity I will isolate what we think of as respon-
dent processes.
There are certain basic reactions that we do not need to learn. They are
already there from the start. Loud noises, a physical blow, rapidly approach-
ing movement toward our eyes, contact with strong heat, and so on—all of
these trigger spontaneous movements in humans, as well as in other animals.
Consequences of our behavior do not seem to affect this substantially. If we
make sure that these spontaneous movements are followed by specific conse-
quences that would typically affect operant behavior, this does not cause the
corresponding change in behavior we might expect. If I encounter a certain
consequence because of putting something in my mouth, this consequence
will probably take on a governing function. Whether what I put in my mouth
tastes good or bad will influence my tendency to put the same thing in my
20
mouth in the future. This is operant behavior, governed by consequences. But

if we were somehow able to make my salivation result in a bad taste, and then
we did something that normally makes me salivate, this would not affect my
salivation to any appreciable degree. Salivation is respondent behavior; it is a
result of antecedents and is not significantly controlled by consequences. A
response that has not been learned is called an unconditioned response, and a
stimulus2 (in this case an antecedent) that triggers such a response is called
an unconditioned stimulus. Typically, salivation is an unconditioned response,
and food is an unconditioned stimulus that triggers it. Some affective reac-
tions, like fear, are also examples of unconditioned responses. These are all
innate reactions that have evolved because of their survival value.
The fact that respondent learning occurs means that experience never-
theless influences both these reactions and their occurrences. This is true of
both salivation and affective reactions, among others. Some external situa-
tions trigger fear without the individual having to learn this (Öhman, 2002).
But in a situation where I feel afraid, other stimuli that are also present may
take on or acquire the same triggering function as the stimulus that originally
triggered fear. If I get assaulted while strolling in the main square of the town
where I live, we can assume that I’d feel some kind of fear during the event.
But if I later take another walk in the square, stimuli that had previously been
neutral—or that had perhaps even triggered positive emotions in me—might
instead provoke fear. The stimuli that have been conditioned in this way
could be the square itself, the aroma outside the restaurant near where I was
assaulted, or any of a number of other details, in themselves irrelevant, such
as the statue standing in the square. As a result of these stimuli being present
at the time of the assault, they are now associated with it. This type of learned
response, called a conditioned response, is an example of how respondent reac-
tions can propagate when neutral stimuli become associated with stimuli that
2 The term “stimulus” is common within behavioral psychology. It denotes

a quality or a phenomenon in the context or environment surrounding a
behavior (Catania, 2007). Thus, both antecedents and consequences can be
called stimuli. A problem with this term is, perhaps due to the experimen-
tal history of the behavioral position, that it can easily be associated with
something small and very clearly defined. This is not a requisite meaning.
It can refer to something very complex and composite and be synonymous
with what, in ordinary terms, we call an event. For this reason, in this book
both words are used, with the latter (“event”) often used to refer to the
more composite meaning.
21
Learning RFT
trigger unconditioned responses. In this way, formerly neutral stimuli become

conditioned stimuli.
In respondent learning, the direct connection between events is impor-
tant. A conditioned stimulus acquires its function by occurring in direct con-
nection with an unconditioned stimulus and its attendant response. These
connections and their variations make up another area that has been the focus
of a lot of research (Catania, 2007).
R E SPON DE N T L E A R N I NG
Unconditioned stimulus Unconditioned response
Assault Fear
Neutral stimulus
The main square in my town
After the above experience this might follow:
Conditioned stimulus Conditioned response

The main square in my town Fear
Operant and Respondent Learning Interact

In many cases, it can be useful to distinguish between operant and
respondent learning. Some processes are more easily understood or influ-
enced based on operant principles, while for others respondent principles are
more useful. However, learning often takes place under the influence of both
principles simultaneously, and through interaction between them. If my son
lives in a different city and I like talking to him on the phone, I may call
him up once in a while. If it turns out that it’s easier to get ahold of him on
Tuesday nights, when I realize it is Tuesday night, I might call him. So far this
is operant learning through positive reinforcement for calling, where Tuesday
night becomes a discriminative stimulus, indicating the increased availability
of a certain consequence: that my son will be there to answer the phone. Now,
let’s say that when I call my son on the phone, a specific melody is played
22
while I wait for him to pick up. After I have encountered this on several occa-
sions of calling, one day I hear the same melody on the radio. I start thinking
about my son, and maybe some emotional reactions, originally occurring due
to my interaction with him, also surface. How did this happen? The answer
is respondent learning. The melody has become a conditioned stimulus, and
thoughts and feelings connected with my son are a conditioned response. It is
easy to see how these reactions can in turn function as antecedents for more
operant behavior. I might, for example, call my son earlier than I would have
if I hadn’t heard this melody on the radio.
Another way in which respondent and operant learning interact is in how
reinforcers and punishers are established. Stimuli that have not previously
functioned as reinforcers can acquire this function by being associated with
things that already function as reinforcers. Status symbols, designer clothes, a
photo of a loved one, or a favorite TV program can all have reinforcing func-
tions. These stimuli have acquired their function by being associated with
other reinforcers; for example, receiving interpersonal attention because of
status symbols. Reinforcers that have their function without being learned—
such as interpersonal attention, food when you are hungry, and warmth when
you are cold—are called unconditioned or primary reinforcers. Reinforcers
that have acquired their function by way of learning are called conditioned or
secondary reinforcers. The corresponding terminology also applies to punish-
ers: They can be unconditioned or primary punishers, or they can acquire their
functions through association with other punishers and therefore be called
conditioned or secondary punishers.
Extinction
Behavior that has been learned does not necessarily last forever. Whether
governed by consequences or by associations, behavior can cease or decrease
following the removal of particular contingencies. We often use the term
extinction for this; operant extinction or respondent extinction, respectively.
Operant extinction occurs when a certain behavior no longer provides that
which has been a reinforcing consequence. If my Tuesday calls to my son
begin to go unanswered, I will probably continue to call him on Tuesdays for
a while. But if I encounter that he never answers on Tuesdays anymore, I’ll
stop calling him on that day. Tuesday is no longer a discriminative stimulus
for the consequence that was previously reinforcing. Tuesday night no longer
signals a historical connection between a specific behavior (calling my son)
and a certain consequence (he answers the phone).
23
Learning RFT
Respondent extinction also involves crucial changes or a discontinuation

of the relations between the different stimuli that lay behind the learning. If,
despite the respondently aroused fear I experience after being assaulted in my
town square, I continue to regularly visit the square, my fear will decrease,
provided that the event (the assault) that originally caused my reaction is not
repeated. New associations between my emotional reactions and the square,
the aroma from the restaurant, and the statue are established. I may experience
other events in the square that awaken positive emotions, so new respondent
learning takes place. That this happens when I walk in the square (operant),
even though I feel afraid initially (respondent), is yet another example of how
operant and respondent learning act together.
That a particular behavior is extinguished does not mean it has been
unlearned. Once I have become afraid in connection with the assault in the
square in my hometown, this fear will most likely be more easily triggered
than it would have been if I had never experienced the assault in the first
place. If my son’s phone does not play the special melody any longer, then the
melody itself may not bring up thoughts of him if I hear it on the radio. And
yet we’ve all experienced how associations like these can suddenly reemerge
after being gone for years. The same is true for operant extinction. Actions we
once learned often lie within reach, even if we have long since ceased to act
in that way. Although I once learned to ride a bike, this behavior might have
been extinguished many years back if the reinforcing consequences ceased.
If I get on a bike again, I might notice that I cannot ride it as well as before;
nevertheless, this behavior is not unlearned.
The fact that learning has these kinds of lasting effects tells us something
important about working for change. I will return to this point when discuss-
ing this book’s main issue: the power of thinking. What we can establish at
this juncture is that the key is to learn new things, rather than trying to extin-
guish what we have already learned.
Generalization
The fact that, in an operant sequence of events, a certain stimulus func-
tions as a discriminative or motivational antecedent, or as a reinforcing or
punishing consequence, does not mean that a new event must be identical to
have the same function. If that were the case, learning would practically be
impossible, since two events are, in fact, never exactly the same. Instead, two
stimuli or two events need only be “similar enough.” If a certain social behav-
ior from my side, like addressing someone with a question, has taken place
under the condition of this person looking at me with a certain expression on
24
his face, and if my asking the question has led to positive reinforcing conse-
quences, then this increases the likelihood that I will speak to someone again
when the conditions are similar; the situation need not be identical. How
similar the conditions must be in order to function in a discriminative way
depends on the individual’s specific learning history. A young child initially
may have learned to say something in an interaction with a parent. This is the
condition under which the behavior has been reinforced. A glance, a facial
expression, or an utterance by a different person does not at this stage con-
stitute an antecedent for the child to interact in a similar way. This anteced-
ent function will, however, gradually spread, becoming generalized. As time
passes, a much broader category of conditions (different people, different
environments, different utterances, and slightly different facial expressions)
might function as antecedents for a certain social behavior.
This does not apply only to antecedents. Events that function as rein-
forcing or punishing consequences are generalized, as well. Friendly behavior
may be encountered in many different forms, but regardless of the differences
it can still have the same reinforcing function. The same goes for the things
that often function as punishers, like being criticized, for example. Another
example of a reinforcing function that can be generalized involves money.
Money is a conditioned or secondary reinforcer. In our experience, it has been
associated with other things that have been reinforcing, giving these pieces
of metal and paper reinforcing functions in themselves. Money has acquired
this function by association with such a large number of other reinforcing
functions that it becomes a generalized reinforcer. For most children, atten-
tion from adults also functions as a generalized reinforcer. In certain con-
texts, however, the same attention can have a punishing function. This shows
us, again, that whether a given stimulus is reinforcing or punishing is not an
intrinsic quality of that stimulus; rather, it can only be understood in context,
in the interaction between an organism and its environment.
Generalization is relevant in connection with respondent learning, as
well. The fact that I was assaulted in the main square of my hometown can
make me feel this fear in a similar square located in a different town if that
other square is similar enough. Or this fear can emerge in a totally different
situation from being in a square; for example, seeing someone in the subway
who is somehow similar enough to the person who assaulted me. Likewise,
hearing that familiar melody on the radio can make me think of my son even
if it is performed by a different singer and in a different version than the one
on his phone.
25
Learning RFT
Discrimination
Discrimination can be said to be the opposite of generalization. Just as the
function of an event can be spread to other events because they are in some
way similar, a function can be restricted to a more specific event in situa-
tions where a similar event doesn’t have this function. Another person’s facial
expression can be an antecedent for a certain social behavior on my part, and
we might assume that generalization has taken place in my learning history
since my first attempts at interacting with other people years ago. There are
quite a few slightly different facial expressions of other people that lead to
more or less the same type of social behavior from me. But what happens if
I want to play poker? In that case, the very small variations that I might be
able to detect in the other players’ expressions become antecedents for quite
different behaviors from my side. A certain glance may make me raise my
stakes, and if another player looks at me in a slightly different way, this can be
an antecedent for my decision to fold. The differences could even be so small
that I’m unable to describe them, yet I still act on them. The fact that some
people are so much better than others at playing poker can partly be assumed
to be connected to a highly trained ability to discriminate when it comes to
other people’s behavior.
Generalization and discrimination, and the balance between them, is
an important part of all kinds of learning, both operant and respondent.
Sometimes we need to catch one very specific signal in a noisy surrounding,
while at other times it is important to act on anything that moves.
WHERE WE GO FROM HERE

This concludes my run-through of important foundations of the psychology
of learning. Let me once again encourage interested readers to seek more
detailed knowledge from other sources. In chapter 2, we will move on to
the issue of how these principles have been used and are used in order to
understand thinking and the power of thinking. We will also take a look at
some difficulties in relation to this—what can be described as inadequacies
in the behavior analytic attempts to deal with this profound type of human
behavior.
26
CHAPTER 2
Thinking and Human
Language
The phenomenon we commonly call “thinking” is obscure to most of us.

There is, of course, always an observer, but only one: the person thinking
her thoughts. Thinking, then, is a phenomenon that is private and cannot be
observed directly by anyone else. Thoughts are included in the phenomena
Skinner termed private events (Skinner, 1953). However, we might critically
remark that thoughts are nevertheless often accessible, like when something
is said and then heard, or is written and then read. If we think about it care-
fully, though, we find that what is accessible isn’t the thought itself; it is what
is said or written. These expressions are often related to the original thought,
but they are hardly identical to it. The experience of thoughts in themselves is
directly available only to the person who is thinking those thoughts. The same
truth applies to recollections, internal bodily sensations, and at least parts of
what we call emotions, all encompassed in Skinner’s term “private events.”
It is important to emphasize that to Skinner, this did not imply that “inter-
nal” phenomena are of a different nature than what we can plainly observe
jointly, or publicly. These phenomena are not enacted in a mental world that is
somehow different from the external world. The distinction is purely in their
degree of accessibility to public observation.
We humans devote a lot of attention to these private events. We talk
about them and assume that they are present in others, even though we are
unable to observe them in anyone but ourselves. We attach great importance
to them in a number of different ways. What is just now taking place is that I,
as an author, am writing a book about some parts of this phenomenon. You,
Learning RFT
as a reader, are devoting your time and attention to understanding what I

am writing, and you may be thinking (if I may dare guess) something along
the lines of “It will be interesting to find out what his thoughts (!) are on
this.” This phenomenon of thinking is also given a prominent position within
modern science’s understanding of human beings, usually under the heading
of “cognition.”
In his interest in private events, Skinner again showed his consistent, or
radical, stance, questioning why we talk about these private events. This is
definitely a very essential question to ask. If we understand why we talk about
this, we will know something about the function these phenomena have for
us. In addition, if we did not learn to talk about things that only the indi-
vidual can note, these phenomena would hardly turn into what they gradually
become to an adult person. The fact that we do things jointly in connection
with these internal phenomena gives them a function for us. In fact, these
phenomena attain a center-stage place in human behavior precisely through
known principles of learning. We learn to talk about private events, and the
way we learn to talk about them, in turn, influences how they evolve. (What
I discuss as “talk” is not limited to producing sounds; for example, a mute
person can learn the same thing through sign language.)
Why would it be important to humans to learn to talk about feelings,
memories, physical sensations, and thoughts? Because this kind of sharing
is valuable to the social community. In Skinner’s words, “It is only when a
person’s private world becomes important to others that it is made important
to him” (Skinner, 1974, p. 35). Expressed in a slightly more technical way, we
learn to talk about the things each of us can observe in only ourselves because
our social environment reinforces that type of behavior. The social commu-
nity reinforces children for speaking based on their own private events. Step-
by-step, these phenomena—which the social community cannot observe in a
direct way—become important to the child, as well, based on the contingen-
cies of reinforcement.
VERBAL BEHAVIOR
Before we take a closer look at how speaking about private events evolves,
we need to reflect on speaking in a more general sense: what Skinner called
verbal behavior. From very early in language training, humans learn to use
combinations of sounds in a way that successively becomes very important
for how we interact. What is functionally crucial in this behavior, according
to Skinner, is the possibility of behavior being reinforced in an indirect way,
by how other individuals act, rather than as a direct result of the speaker’s
28
Thinking and Human Language
actions. This makes it possible for me to be given a desired object without

having to direct my actions toward physically obtaining the object myself. I
can ask for something and receive it by way of another person’s actions. Using
sounds in this way is primarily a social ability, and it is carried out according
to the same fundamental principles for learning that apply to other behavior.
Skinner called this kind of action by a speaker “verbal behavior” and defined
it as a behavior that acquires its effect through the mediation of a listener’s
behavior, the listener having been taught such that her behavior precisely
functions as reinforcing for the behavior of the speaker (Skinner, 1957). If
I reach for an object, the consequence of managing to reach it is reinforcing
for that behavior. Asking to be given the same object is only reinforcing if
it is followed by the listener’s behavior in the form of giving me the object I
wanted.
Someone speaks, and based on this speaking, another individual acts in
a way that makes different reinforcing consequences accessible to the person
who spoke. This also allows the social context to govern the speaker’s behav-
ior in turn, based specifically on the contingencies of reinforcement that are
established. If the listener, when she hears me, gives me what I asked for, the
probability of my repeating this behavior increases. If the listener acts in a
different way, this too will influence my future behavior.
SKINNER’S DESCRIPTION OF
VERBAL OPERANTS
Skinner divided verbal behavior into several primary types: tacts, mands,
echoic behavior, intraverbal behavior, and autoclitic behavior (Skinner, 1957).
When we regard verbal behavior as operant responses, we see it as controlled
by antecedents and consequences, just like any other operant behavior. These
different responses (B in an ABC analysis) are distinguished by the different
relations between the form (what is said or written) and the variable (A and/
or C) that governs the response. The shape the response takes and its relation
to antecedents and consequences is what forms the basis for classification.
Tact
A tact is governed by a preceding stimulus: the stimulus that is being
tacted. An example would be saying “chair” when a chair is present. This
response, the utterance, is a direct result of seeing the chair. When we say
“She is running,” it is governed by the fact that someone (she) is moving her
29
Learning RFT
body in a certain way. We tact our environment in this way because we have a
solid learning history in which tacting has been reinforced. From an early age,
we have experienced reinforcing consequences when, for example, in the pres-
ence of a cow, we have uttered precisely “cow.” If we said “kitty” in the pres-
ence of a cow, there were other consequences. The governing consequences
are primarily of a general and social nature; that is, tacting is followed by
generalized reinforcers. It is tacting we have in mind when in everyday speech
we speak of “describing,” “telling,” “referring to,” and the like. All of these
concepts, however, are highly imperfect for scientific purposes, and thus the
use of the neologism “tact.”
An ideal tact is completely controlled by the stimulus preceding it. In
an everyday context, we would say the statement is objective or corresponds
to the object referred to. This is the type of “pure” tact that we seek in scien-
tific linguistic inquiry. Tacts rarely have this character in normal life, and you
could probably question whether a pure tact is actually possible as anything
other than an ideal, even in scientific settings. Skinner wrote about distorted
or impure tacts (1957), by which he meant tacts that are controlled by other
factors, such as who is listening and how listeners act as a result of a given tact.
What we normally speak of as exaggeration is an example of a distorted tact.
The verbal operant is controlled by that which precedes it, but the size, for
example, is exaggerated, so the tact is distorted. When the “Great Fisherman”
talks about the size of the fish he has caught, his descriptions are governed not
solely by the size of the fish. If they were, the descriptions would be pure tacts.
Instead, they are probably governed by other elements, as well. If a statement
is portrayed as a tact when it actually is governed not at all by what precedes
it, but rather by something completely different, in everyday language we
would call it a lie.
Mand
A mand is verbal behavior controlled by a specific reinforcer, and it speci-
fies this same reinforcer. For example, your saying to a person “Go away!” is
reinforced by earlier experiences of having a person leave after you said this.
Saying “Look at me” specifies its own reinforcer—that is, the listener looks at
the speaker. The typical antecedent (A) of a mand is the presence of a listener
(a discriminative function) and a motivational operation (an establishing
operation) that makes the consequence in question desirable to the speaker.
The motivational operation that precedes “Go away!” is most likely that the
listener’s presence is aversive to the speaker at the moment, whereas with
“Look at me,” something else is going on in the interaction between speaker
30
and listener. Examples of mands are different types of requests and demands,
asking questions, and raising your hand to obtain permission to speak.
It is important to understand that both tacts and mands are defined by
their function, not by sheer topography. The very same word or expression
can have several different functions. This means that topographically identical
linguistic expressions can function both as tacts and as mands, depending on
the relation to antecedents and consequences in the specific situation. When
someone says “the newspaper,” it could be a tact if the governing variable is
an actual newspaper and the utterance is an answer to the question “What
is that on the table?” But “the newspaper” could also be a mand if it serves
as a request that someone hand the newspaper to the speaker. The relation
between what is said and the governing variables determines what type of
verbal behavior is at hand. A statement such as “Those apples are nice” can
seem like a tact but in fact be a mand, if this statement is reinforced by the
fact that the consequence that previously followed upon a similar statement
was that the speaker was given an apple.
Echoic Behavior
In echoic behavior, verbal behavior has an antecedent that is topographi-
cally identical to the response. It is a verbal response that follows a preced-
ing verbal response, echoing or repeating something that has been uttered.
Again, typical reinforcers are generalized social consequences like attention
and other interpersonal processes. Echoic behavior is a core part of early lan-
guage learning. The parent, for example, utters a word and then reinforces
every tendency in the child to repeat it. But echoic behavior continues to exist
as an important verbal behavior throughout life, like when we silently repeat
something we have just heard.
Skinner gave a few more examples of verbal responses that resemble echoic
behavior insofar as they all involve a response that is somehow a reiteration
of the antecedent. He described textual behavior, as well as transcription and
taking dictation. I describe all of these together with echoic behavior based
on this similarity. Textual behavior is saying something that is controlled by
a preceding stimulus in the form of a written text, where a formal correspon-
dence exists between what is written and what is said. Textual behavior, then,
is the response we would normally refer to as reading aloud. Taking dictation
is an inverse process: writing down something that is in formal correspon-
dence with what has been uttered, like when you write down a telephone
number someone just told you. Transcription is writing something wherein
the controlling antecedent is topographically identical with the response. In
31
Learning RFT
everyday language, this is called copying. All of these verbal behaviors are
reinforced by generalized social reinforcers.
Intraverbal Behavior
Intraverbal behavior is also verbal behavior that has other verbal behav-
ior as its antecedent, just as echoic behavior; but in this case there isn’t any
formal correspondence between antecedent and response. In this case, the
relation between the verbal antecedent and intraverbal behavior is arbitrary,
established by social whim. If I say “one, two, three,” and you say “four,” then
your response is intraverbal behavior. If I say “What is casa in English?” and
you say “house,” then this response, too, is intraverbal. Just as for all other
verbal operants except the mand, the important, governing consequences for
intraverbal responses are generalized social reinforcers.
Once again, note that the definitions of these different types of verbal
behavior are functional. What is crucial is the relation between the response
and the antecedents and consequences. Above, I said that the expression “the
newspaper” can be either a tact or a mand. But it could, of course, also be
echoic behavior if it were governed by someone else saying “the newspaper”
and the response is a reiteration of this. It could also be intraverbal behavior.
This would be the case if it were governed by someone else just having said,
“What is another word for ‘the local rag’?”
Autoclitic Behavior
Skinner described one more type of verbal behavior: autoclitic behavior.
This is verbal behavior, or parts of verbal behavior, governed by other verbal
behavior by the speaker and modifying this other behavior. An example of
autoclitic behavior is the word “maybe” in the response “Maybe it’s the news-
paper” when someone asks, “What is that on the table?” The word “maybe”
functions as an autoclitic because it modifies the totality and lets the listener
know something about the position from which the speaker is speaking, in this
case a position of uncertainty. Other autoclitic behavior, which would modify
the tact in a different way, would be a word like “not,” as in answering the
question about what is lying on the table with “not the newspaper.” Autoclitic
behavior can be whole words, as in the examples above, but it can also be
modification of a word, such as adding an “s” to the end when the answer is
“newspapers.” Punctuation, grammatical structure, and syntax are all forms
32
of verbal behavior in the category autoclitic behavior—verbal behavior that is

dependent on or modifies other verbal behavior by the speaker.
The Use of Skinner’s Analysis in This Book

Skinner’s system for analysis of verbal behavior is fairly complex, and
my intention has not been to describe it in detail here. With respect to the
preceding outline of the analysis of verbal behavior, two points are important
from the perspective of this book. The first is that this outline shows us that
a functional analysis of verbal behavior is a feasible task. Verbal behavior is
a human behavior that is governed by antecedents and consequences. The
second point is that this analysis allows me to return to a consideration of
private events. In this, I will primarily use one part of Skinner’s analysis—his
description of the behavior of tacting—to aid in understanding how private
events attain such a central function for us humans.
LEARNING TO TACT PRIVATE EVENTS

What does all of this mean in terms of our initial question about how we learn
to talk about internal phenomena? One thing it shows us is that we learn this
from our social environment, which reinforces certain behavior. For the child,
this environment is initially made up of parents, other guardians, or other
family members. Hence, from a Skinnerian perspective, these members of
our social environment are the ones who train us to appropriately tact events
in our environment. However, these people (and others) do not have direct
access to a child’s private events. When people in the child’s social environ-
ment reinforce tacting and what is to be tacted is an external phenomenon,
like an object, a person, or an external action, the social environment does
have access to whatever is to be tacted. The connection between stimulus and
tact is obvious. Establishing a connection to a reinforcer is rather easy in this
case. The child who is learning to tact a teddy bear can see the teddy bear
in front of her eyes, and the same goes for the parent. When the child says
“teddy” or something similar, the parent can reinforce the behavior. What the
child observes internally, however, is not accessible to the social environment
in the same way. This will make it more difficult for the environment to rein-
force her tacting of such internal phenomena. Skinner described several ways
in which people in the social setting can reinforce tacting of private events,
despite this difficulty (Skinner, 1945).
33
Learning RFT
One way we learn to tact private events is when others are able to observe
phenomena that are parallel to what is being tacted. If others are able to
observe a flush or a swelling in a child’s skin, they can assume that the child
experiences pain. In such a situation, the parent can reinforce verbal behavior
that tacts the assumed pain, such as if the child says “ouch” or “hurts.” The
child’s verbal behavior is followed by reinforcing consequences, established
by the surrounding social environment. Other common, and commonly
accessible, phenomena are events that we know normally produce feelings.
One example would be when someone in the child’s environment behaves
in an aggressive manner, and we afterward ask the child, “Did that make you
feel scared?” Having learned to tact our own private world, we use our own
experiences and assume that the child experiences something similar to what
we experience. However, a parent’s emotions may not always correspond with
those of her child, which leaves open the possibility of problematic tacting of
emotions on the part of the child.
A similar type of learning occurs when observable responses by the indi-
vidual are parallel to the private phenomenon that is tacted. Sounds, facial
expressions, and certain movements are observable responses that commonly
parallel an individual’s private events. Experiences of pain, anger, interest,
and joy are often accompanied by other behavior. Children shy away, draw
closer, cast glances, and act in a number of different ways that are visible
to others in their social environment. The fact that these different forms of
observable behavior vary in accordance with private events makes it possible
for the environment to establish contingencies of reinforcement that, from
then on, influence the child’s growing ability to talk about the things that
only she can observe.
Another way of learning to talk about private events occurs when the
child first learns to talk about the things she does that are observable both to
herself and to others around her. She then gradually goes on to learn to talk
about similar actions of her own that are observable only to herself. This is
especially relevant to what we commonly call thoughts or thinking, and the
same learning path is also relevant to what we normally call memories.
A child carries out a number of actions that she gradually learns to talk
about. She walks, waves, eats, watches the dog, stands still, plays, gets dressed,
and so on. No child can tact her own action without first having performed
it. (She can, however, utter a word that others use to describe something that
is performed without performing it herself. In this case, she’s echoing some-
thing she heard someone else utter.) In order to tact her own actions, the
following sequence is required: The child does something that is observable
to the child herself and to her social environment. When this action is carried
34
out, or immediately thereafter, the environment reinforces a certain verbal

behavior by the child. Most likely, the child first repeats after others (echoes),
then gradually transitions to tacting. The child’s own action then becomes a
discriminative stimulus for the verbal behavior.
Here’s an example: Petra is kicking a ball. When she does this, her father
says, “Kick! Look, you’re kicking the ball!” The word “kick” is repeated, and
if Petra says something that resembles “kick,” her father reinforces this utter-
ance. In time, the very action of kicking becomes a discriminative stimulus
for Petra to say “kick.” Petra has learned to tact her own kicking. Over time,
similar learning occurs for other actions: waving, running, playing, and so
on.
Alongside this process, the child will, of course, experience aspects of
these behaviors that are private, that is, only accessible to the child herself;
for example, what her arm feels like when she waves to someone. This will
contribute to an increase in her ability to tact private events. She learns to tact
her own actions, and because these actions vary in accordance with private
events, her ability to discriminate private events gradually grows, along with
her ability to tact them.
Skinner described yet another way in which we learn to tact private
events, but before we go on to that, let’s first take a closer look at tacting of
the type of private events that are especially important as far as this book is
concerned. This has to do with how we learn to think, how we learn to talk
about our thinking, and how, in time, we learn to think about our thinking.
Learning to Tact Thoughts

An action that is initially accessible both to the person who performs it
and to the social environment can gradually change (Skinner used the term
“weaken”) so that it is no longer accessible to others, although it stays acces-
sible to the person performing it. One simple example is how we go about
learning to count. First, children count aloud. That is the only way this behav-
ior can be reinforced by the environment. As time passes, the child begins
to experience that counting aloud isn’t reinforced in all situations; in fact,
sometimes it is punished. People may find it annoying to have a child count-
ing out loud, so the reinforcement ceases. The child then starts to count more
quietly and realizes that she can count quietly enough that no one else can
hear it. She may still be moving her lips but not be making sounds. A child
who is sitting quietly with a counting book and moving her lips in a concen-
trated way may also receive reinforcing social consequences for this behavior,
35
Learning RFT
and—presto!—she has learned to count silently, to “count in her mind.”

Provided that this behavior results in consequences that are functionally rein-
forcing, the behavior will continue. Since the ability to count silently is very
useful to attain a number of things that people want, this remains a human
ability. The same sequence is easy to see in the process that leads to our ability
to read. First we do it aloud, and after a while we begin to do it silently. Still,
a person who has been able to read silently for many years will often read
aloud again if what she is reading is particularly difficult, like a complicated,
unfamiliar word, even if she is speaking only to herself. And when we want
our thoughts to be particularly effective, like when performing a demanding
task, it is common to “think aloud”: “Go for it, Niklas. You can do it!”
The preceding examples show how verbal behavior, such as counting or
reading, starts out as a public event, where it is accessible to a social environ-
ment that can reinforce it, and then gradually turns private. A very large part
of what we normally call “thinking” develops similarly. First we learn to talk
about our own behavior based on the contingencies of reinforcement that
have been established through the social setting; then, gradually, the public
parts of the behavior wane while the private aspects remain.
I should perhaps point out that this is not a complete description of every-
thing included in the terms “thoughts” and “thinking.” It would be impossible
to give such a description, for many reasons, including the ambiguity sur-
rounding what should actually be included in the terms. They can be defined
in different ways, and they are everyday concepts, not scientific terms. What I
am maintaining, however, is that the process described above is fundamental
to what we usually call thinking and thoughts. Also, this description is rel-
evant to the aims of behavior analysis, that is, to predicting and influencing
what we as human beings do, or how we behave.
Several factors contribute to the persistence of private verbal behavior,
and to it becoming increasingly independent of those external actions in
which it originated. Thinking about running is, in some ways, similar to actu-
ally running. It involves a number of internal stimuli identical to those that
exist when you are in fact running. This is particularly evident when it comes
to the kind of thinking that we call imagining, visualizing, or thinking of.
When I think of my mother, I see something in my mind’s eye that is very
much like what I see when she is standing in front of me. The reality of such
a connection between thinking and external behavior corresponds well with
what we now know about the functions in the human body, and particularly
the nervous system. Thinking about looking at someone is done largely with
the same parts of the brain as actually looking at someone. If you first learn to
play a piece of music and then hear this music played, motor neurons are active
36
while you simply listen (Lahav, Saltzman, & Schlaug, 2007). Thinking about
running involves the same motor centers of the brain as running (Jeannerod,
1994; Kosslyn, Ganis, & Thompson, 2001). This means that thinking about
doing something is an action that in many respects resembles actually doing
what we are thinking about. At the same time, this private action has great
advantages. We can perform an action in the concealed setting of our imagi-
nation without facing many of the consequences that the external action
would involve. Doing things privately—thinking of doing them—can thus
be a way of testing and practicing. As we all know, this type of behavior plays
a significant role in what we often call problem solving. We try things in our
thoughts, and then we perform them more entirely, or else we refrain from
doing them. It is easy to see that this possibility is likely to have increased our
species’ ability to survive.
The path of learning I have just described is thus:
1. We do something.
2. We learn to talk about what we are doing, which in the above termi-
nology means we are tacting our own behavior.
3. We learn to speak without uttering any words; that is, we think.
Once we are doing this, this very behavior becomes something we can
tact. Perhaps you, as a reader, just noticed that you were thinking about some-
thing other than what this text says. In that case, you have something new to
tact: “I was just thinking…” Therefore, when we talk about the fact that we
are thinking, this can lead to thinking about the fact that we are thinking. (In
cognitive theory, this is often referred to as metacognition.)
Thus far, I have described two main ways in which we learn to talk about
private events. The first way is when people in the environment observe
phenomena that are parallel with private events in the individual, and they
use these parallel phenomena to reinforce the verbal behavior for which the
private events gradually become discriminative stimuli. The second way is
when behavior that is initially accessible to the social environment gradually
becomes private, through this behavior as a whole being punished or extin-
guished in certain contexts, while at the same time a part of the behavior—
doing the same thing, only silently—leads to reinforcing consequences. The
behavior as such can thereafter be tacted by the individual, just as other
private events are. Before we turn to the question of why the social environ-
ment places such importance upon teaching each new individual this type
of skill, let’s look at a third possible way of learning the ability to tact private
events.
37
Learning RFT
Learning to Tact Private Events Through

Generalization
Once tacting of private events becomes an established ability, generaliza-
tion allows us to further develop this kind of verbal behavior. A private stimu-
lus can have certain features that resemble other phenomena, either external
or internal. Tacting something that is available only to the individual in
question by using a resemblance to something that is accessible to others is a
behavior that is often reinforced. One example is a child who has been sitting
on the toilet for a long time and then gets up and says, “My legs are prickly like
soda pop.” The child has discriminated some sort of resemblance between the
experience of drinking a carbonated drink and what she experiences in her
legs in connection with having spent a long time sitting on the toilet. She uses
what we could call a metaphor or an analogy to tact the similarity she experi-
ences. The fact that a great deal of the language we use to describe private
events consists of metaphors taken from events in our external environment
illustrates this path for learning. When we use metaphors, we are making use
of the similarity between different phenomena; thus, metaphors can contrib-
ute to generalization. Language like “simmering with anger,” “feeling low,”
“being filled with peace,” or “being overwhelmed” are all typical examples.
WHY IS THIS BEHAVIOR REINFORCED?

In the introduction to this chapter, I quoted Skinner, who wrote, “It is only
when a person’s private world becomes important to others that it is made
important to him” (Skinner, 1974, p. 35). Why, then, is a person’s private
world important to the social environment, and what is the point in teaching
a new individual to talk about this private world? The answer is that an indi-
vidual’s private world contains things that are useful in social interaction. We
are social animals, and social interaction is fundamental to our survival.
What an individual feels in a certain situation communicates a lot about
her history in connection with similar situations. For example, if someone says
she is hungry, this is a brief way of describing her history in regard to the avail-
ability of food. It also lets us know something about her state at present and
what is essential to her in the current situation, and it let us know something
about her inclination to act in the near future. It communicates that soon she
will probably eat something or act in some other way in relation to food. The
same goes for verbal behavior that describes anger, weariness, joy, and so on.
If someone says she is depressed, others receive a condensed description that
38
may be highly relevant to their own actions and to their ability to anticipate
that person’s further actions. The point is not that these internal phenomena
are some kind of autonomous force; rather, the significance of such expres-
sions lies in how the private events they describe are related to the individual’s
learning history.
Once the ability to tact private behavior is established, this ability also
becomes valuable to the individual. To quote Skinner once more, “A person
who has been ‘made aware of himself ’ by the questions he has been asked is
in a better position to predict and control his own behavior” (Skinner, 1974,
p. 35). Being able to foresee and control one’s own behavior naturally implies
an increased ability to achieve things that are desirable to oneself.
Before I conclude this passage about the relationship between think-
ing and verbal behavior, I want to underscore, once again, that this is not a
description of every aspect that could be included in the concepts of “think-
ing” and “thoughts.” It could be argued that even before this learned ability of
silent verbal behavior is established, as described above, there is some sort of
rudimentary behavior in the child that might be called “thinking” (Vygotsky,
1986). It might also be argued that something like this is present in other
animals besides humans, in one way or another. How we view this argument
depends on what we include in the concept of “thinking.” In any case, we
know very little about this possible rudimentary capacity and what function
it has. And regardless of this, something novel and revolutionary happens as
the child’s verbal behavior shifts from being solely public to also becoming
private.
THE OBSCURITY OF THE INTERNAL

WORLD STILL REMAINS
Despite the fact that we learn to talk about private events, it remains more dif-
ficult for us to talk about these events than about phenomena that are observ-
able by others. All through life, it will be easier to describe to others how we
go about driving a car or painting a picture than it will be to describe how
we do our thinking or how we feel when we are sad. All of this is logical as
seen from the preceding analysis. In situations where we have learned to talk
about our pain, anger, or fascination, at least parts of our behavior haven’t
been observable to anyone but ourselves. This means we’ve had relatively
limited support from the environment in learning to talk about these things
compared to phenomena that others can more easily observe. Therefore, our
ability to discriminate and speak about these private phenomena will forever
39
Learning RFT
be less developed and less precise than it is for other, more observable areas
in our field of experience.
A SPEAKER AND A LISTENER

An important aspect of verbal behavior is that when it occurs, there is always
a speaker and a listener. This is particularly obvious when the verbal behav-
ior occurs with another person present, but there is a listener even when
someone is thinking; it’s just that in this case the speaker and the listener
are the same person. The latter situation is of special interest in this book,
but we will commence with what comes first in every person’s history: the
speaker as one person, and the listener as another. Skinner restricted the
concept “verbal behavior” to the speaker. In his definition, it is the speaker
who behaves verbally. The listener reinforces this behavior, but according to
Skinner, this action should be understood as cohering with the same prin-
ciples that apply to other forms of reinforcement. Skinner saw no reason
to deem the listener’s behavior verbal (Skinner, 1957). The listener can, of
course, also speak in her turn, and that would be a verbal act, but the listening
itself is something Skinner mostly left out of his analysis of verbal behavior
(S. C. Hayes, 1991; Schlinger, 2008). Still, it is quite apparent that, in many
cases, what Skinner called verbal behavior has complex effects on the listener.
When someone says, “Please wait outside, and I’ll be right with you,” this is a
typical example of what Skinner called verbal behavior. The point, of course,
is to influence the listener, making her go outside to wait. If private events
in the form of thoughts are powerful, they must be exercised in relation to a
listener, even though in this case the listener and the speaker are one and the
same. All of this gives us good reason to investigate Skinner’s thoughts on the
listener’s behavior. He discussed listening as a separate category of behavior
and referred to it as rule-governed behavior.
RULE-GOVERNED BEHAVIOR
“Please wait outside, and I’ll be right with you” is a verbal statement that is
easy to analyze using operant psychology’s basic formula, ABC. If, as a result
of this statement, the listener goes outside to wait for the person who made the
statement to join her, we could analyze it as follows: The statement functions
as an antecedent (A) for the behavior of going outside (B) in order for the
speaker to join the listener outside (C). According to Skinner, the antecedent
40
is functioning as a rule because it specifies a behavior and a consequence

(Skinner, 1966, 1974). He emphasized the difference between behavior that
is governed by such rules and behavior that is governed by direct experiences
of consequences that have occurred, which he called contingency-shaped
behavior. Here is a classical example of these two types of behavior, set forth
by Skinner:
The difference between rule-following and contingency-shaped
behavior is obvious when instances are pretty clearly one or the other.
The behavior of a baseball outfielder catching a fly ball bears certain
resemblances to the behavior of the commander of a ship taking
part in the recovery of a reentering satellite. Both move about on
a surface in a direction and with a speed designed to bring them, if
possible, under a falling object at the moment it reaches the surface.
Both respond to recent stimulation from the position, direction, and
speed of the object, and they both take into account effects of gravity
and friction. The behavior of the baseball player, however, has been
almost entirely shaped by contingencies of reinforcement, whereas
the commander is simply obeying rules derived from the available
information and from analogous situations. (Skinner, 1966, pp. 241-
242)
Skinner went on to note that the sea captain can, by collecting many sat-
ellites, possibly step away from some of the rules, so that, in time, his behavior
may gradually also become more shaped by contingencies. But rule-governed
behavior has advantages, including that the captain does not necessarily need
personal experience in order to succeed in his task.
It is easy to see that a great deal of human behavior is, in fact, rule-
governed. “Don’t drive through a red light.” “Don’t kick a man who’s down.”
“Work hard on your studies, and you’ll have a better chance of getting a good
job.” Some of the rules specify behavior and consequences that the individual
probably has direct experiences of, as well, like “Put more clothes on, or you’ll
be cold.” Other rules specify both behavior and consequences that the indi-
vidual hasn’t experienced, like “Stop abusing alcohol, or you’ll die.”
When we’re learning to follow rules, they are given to us by people in
our environment. But gradually we also learn to set up rules for ourselves
wherein we become both the speaker and the listener: “Try again, and you’ll
do better!” “Leave it for now, and try again later.” “Don’t say a thing; that will
just make it worse.” Most of us have heard others say things like these to us,
and we have also directed these sorts of statements to ourselves so that we
might follow our own rules. And in precisely this formulating and following
41
Learning RFT
of self-directed rules lies much of what we commonly refer to as the power of

thinking, which is the main theme of this book.
A DIFFICULT QUESTION
To utter a rule or an instruction is, by Skinner’s definition, verbal behavior.
However, as mentioned above, rule-governed behavior—following a rule—is
not necessarily verbal behavior. If someone says, “Wait outside, and I’ll
be right with you,” and I reply in turn, that would be verbal behavior. But
if I simply go outside to wait, then, according to Skinner, it is not. Rather,
it would be a result of previously experienced consequences and should be
understood in the same way we understand contingency-shaped behavior. But
this raises a difficult question: How is it, then, that we humans act with a view
toward the future, toward things we haven’t previously experienced, and that
we do this as a result of something that has been said or thought? How can we
understand the effect verbal behavior has on listeners in everyday situations
like the following? Let’s assume someone tells you, “Tomorrow, when you
hear someone honk five times, go outside and I’ll be there.” Then, the next
day, you go outside when you hear someone honk five times, even though you
have never previously encountered any reinforcing consequences for doing so
in that kind of situation. Or, for a slightly longer-term example, if a colleague
behaves disagreeably, you may think, “The next time she acts like that, I’m
going to give her a piece of my mind.” Then, three weeks later when your col-
league does something similar, you do exactly that. The mechanisms at work
become even more interesting in the extreme long term, when we humans do
things in the present that seem to be governed by how we think things are
going to be after we’re dead, whether it has to do with prospects of going to
heaven, our children’s financial future, or the desire to “finally find peace.”
A more technical way of expressing the same question is to base it on
Skinner’s definition of verbal behavior that functions as rules or instructions.
He wrote that this behavior specifies behavior and consequences (Skinner,
1966). That leaves us with the question of how a verbal behavior now can
specify behavior and consequences that are not taking place in the present
and that the individual has not earlier experienced. And how are we able to
carry out new actions in order to achieve such consequences? How do we
manage to stop smoking based on the assertion “Stop smoking, or you’ll have
a high risk of getting lung cancer,” when the consequences we face in doing
this are primarily of a punishing type? The consequences that result could
include short-term effects like withdrawal symptoms or the loss of pleasant
42
company during smoke breaks, whereas “not getting cancer” is very abstract
and temporally distant. How does this work?
Skinner answered this question by referring to “a long history of verbal
conditioning” (Skinner, 1957, p. 360), but he never pinpointed how to
describe such a potential learning history. Remember that a fundamental
principle for how both antecedents and consequences acquire their govern-
ing functions for behavior is that they are contingent on the behavior they
influence. Experimental behavior analysis sees the direct contiguity between
stimuli as absolutely crucial, for operant as well as respondent conditioning.
Skinner distinguished between this and rule-governed behavior, maintaining
that a complex learning history in one way or another bridges this dividing
line. But what would such a history look like? That is a question he never
answered. Early on, he mentioned the possibility of human language involv-
ing something more than the principles of operant and respondent condition-
ing, which he had accounted for (Skinner, 1938), but he later abandoned this
alternative.
Several leading behavioral analysts, like Michael (1986), Parrott (1987),
and Schlinger (1990), have brought this issue up, along with the observation
that a convincing answer has long been overdue within behavior analysis.
THE COGNITIVE ANSWER AND ITS

PROBLEM, AS SEEN FROM WITHIN
BEHAVIOR ANALYSIS
A common answer to the above problem about how people go about relating
to things and events that are not present is what could be summarized as
cognitive theories or information processing theory. There are many variants
on this approach, which has historical roots extending back millennia (Ellis,
1989). In their modern form, these approaches have in common the assump-
tion that humans are equipped with some type of internal structure for car-
rying information. Common designations for these internal structures are
“schemas” and “mental representations.” These structures are influenced and
transformed by external events and are credited with playing a crucial role in
governing human behavior. External stimuli or events give rise to an internal
processing activity in humans, and this internal processing helps explain our
behavior in a way that isn’t possible by means of analyzing the external events
in themselves. Causes of human behavior, especially more complex human
behavior, are therefore primarily sought in these internal processes.
43
Learning RFT
Within behavior analysis, on the other hand, the guiding principle is that
the causes of different behaviors are things that occur in the context of these
behaviors. Causes are processes outside the actual behavior, and therefore
they are accessible to direct influence, at least in principle. This is a pragmatic
approach, adopted because it supports the aim of behavior analysis, which
is not just to achieve prediction, but also to achieve influence. In this light,
assumed internal structures, like schemas, are problematic. After all, such
structures are not accessible to direct influence. They are merely assumed,
and they are not available for contact in time and space. All we can contact is
their effects: the phenomena they are assumed to cause. If something cannot
be contacted in time and space, then it also cannot be influenced in a direct
way. In behavior analysis, these theories based on assumed internal structures
and mental representations are seen as a historical remnant of prescientific
discussions that included the soul or the psyche (Skinner, 1963).
Arguments for internal structures as causes of behavior follow the
same pattern as everyday expressions we use when, for example, we say that
someone is eating “because she is hungry.” How do we actually know that
someone is hungry? This is simply a conclusion arrived at based on the per-
son’s behavior, which is what we can observe. She is acting in a certain way in
relation to food. The expression “being hungry” only summarizes a number
of behaviors and phenomena that we can observe or contact in some other
way. Some of these phenomena can be contacted by everyone present, such as
seeing the person eating food or hearing her talk about it. Other phenomena
are available for contact only to the person who experiences them, like the
feeling in her stomach. But the assessment “She is hungry” is nothing more
than these phenomena taken together. If these phenomena are not present,
then “hunger” as an internal object disappears. Of course, this way of express-
ing ourselves is often linguistically convenient. It is, to use one of the terms
I have accounted for in this chapter, an example of how we tact our own or
someone else’s behavior. But the fact that “she is hungry” does not suffice as
a scientific explanation for why that person is eating—at least not in behavior
analysis. The expression “she is hungry” is simply a summary of the behavior
one wishes to explain. In order to answer the question of why someone is
eating, it is necessary to examine the context of the behavior. You have to
search among the antecedents and consequences surrounding the behavior
of eating. The causes must be sought both in the present and in the historical
context.
Many modern cognitive theories or information processing theories
use neurobiology in their explanatory models (Siegel, 1999) and see differ-
ent brain structures and the activity taking place within them as causes of
behavior. Although seemingly more scientific, this is much the same as the
44
assumptions underlying hypothetical structures like schemas. And even

though brain structures are obviously available for contact in time and space,
the basic, pragmatic objection remains: Brain structures, or what occurs in
them, are not external to the behavior they are said to cause. They are, in fact,
a part of the same behavior. If I lift my hand, events are taking place in my arm,
my shoulder, my aorta, my brain, and more. But all of these phenomena are a
part of my action of lifting my hand. In behavior analysis, behavior is defined
as an action performed by the organism as a whole, and a part of the action
cannot explain the action in its entirety. The behavior of “lifting my hand” is
an action performed by me as an entire organism, and what takes place in my
brain cannot suffice as the cause for my action, any more than what happens
in my arm, my shoulder, or my aorta. All of these are contributing elements
and therefore are parts of the action. And in the behavior analytic approach,
the cause of the act cannot be a component of the action itself; causes must be
sought in events that precede and/or follow the action. They are to be found
in the action’s context. Again, this is because behavior analysis seeks causes
that are accessible to direct intervention, based on our fundamental goal of
influencing behavior.
This means that to a behavior analyst, cognitive theories, with their
explanations based on internal structures or mental representations, are a
dead-end street. (For a more thorough account of the scientific-theoretical
approaches within behavior analysis, see Moore, 2008, and Wilson, 2001.) If
mentally “representing” is an important part of a sequence of human behav-
ior, the behavior analytic question would be How are we to understand this
presumed human activity of “representing”? This gives us even more human
behavior to explain, which is then done by performing analyses of events in
the action’s context. Referring to mental representations does not, as seen
from the behavior analytic perspective, present us with any useful answers
to questions concerning the causes of human behavior. This is, at best, only
a description of more human behavior that then needs to be explained. At
worst, this model places the causes of behavior in an assumed internal world
that is not accessible to scientific psychological analysis. That said, neurobio-
logical research—and its different theories about how what takes place in the
brain plays a role in the actions of human beings—is of course an important
field of research in its own right. But such research cannot answer the ques-
tions within behavior analysis about how we can understand and influence
behavior using psychological interventions.
Cognitive theories are multifaceted, and they vary among themselves. My
intention here has not been to do an exhaustive comparison between behavior
analysis and any of these approaches, or to account for how such approaches
could succeed in solving the behavior analyst’s dilemma, explained above.
45
Learning RFT
There is extensive literature on this topic available for interested readers (e.g.,
Andersson, 2005; Solso et al., 2005). My purpose has simply been to point
out, from a behavior analytic perspective, what unites these other perspec-
tives, and what makes them unacceptable as alternatives.
The fact that cognitive and behavior analytic approaches are very differ-
ent from each other does not eliminate their points of contact in connection
with the phenomena described. (I will return to this topic later in this book.)
Neither do the differences imply that a dialogue between these approaches is
useless. On the contrary, there are writers who argue that we are currently at
a point where such a dialogue could be productive (Overskeid, 2008).
BEHAVIOR ANALYSIS AND HUMAN

THINKING—A BRIEF SUMMARY
An examination, from a behavior analytic perspective, of what we normally
call thoughts or thinking starts out from the fact that we do indeed speak
about this phenomenon, which means that we have learned to speak in this
way. This has taken place in accordance with the same principles that govern
other human behavior, that is, through operant and respondent conditioning.
The way we have learned to speak—aloud to begin with, and then silently—has
shaped the phenomenon we are to analyze. Large parts of this phenomenon—
parts that are crucial to our purpose—are, then, what Skinner called verbal
behavior. This means the behavior of thinking is a part of a greater class of
behavior; it is one part of verbal behavior. Verbal behavior is primarily a social
action. It occurs as a result of a listener being present, and it is reinforced
by the listener’s actions. At the same time, this behavior can have dramatic
effects on the listener’s behavior. Verbal behavior can have immediate conse-
quences when it is uttered, as when someone gives me something I have asked
for. But its effects can also be greatly delayed, as when I travel abroad to visit
a certain place because of something my parents told me twenty years earlier.
The latter has been called rule-governed behavior in behavior analysis ever
since Skinner first used the term. When we first learn this type of behavior,
we are listeners and someone else is the speaker. Someone else supplies the
rule, and the listener learns to act on it. But the same behavior can occur after
we become both speaker (silently) and listener, within ourselves. We learn to
formulate rules, which we then follow. Human behavior is often influenced
by self-generated rules.
Behavior analysis has faced significant problems in explaining the effects
of verbal behavior on the listener’s behavior, and not only in regard to effects
46
that are delayed in time. A verbal statement can have short-term consequences
that are hard to explain based on the classic principles. The statement “Hide
behind the statue; the two guys in brown jackets are out to get you” can have a
rapid effect on the listener’s behavior. This is true even if the listener has never
had any unpleasant experiences connected with people in brown jackets, and
even if she has never been assaulted or hidden behind a statue when threat-
ened. Similar effects can follow upon self-generated rules: rules that can only
be apprehended by the person following them. There have been proposals for
an explanation of these effects through respondent-like associative learning
of some kind (Parrott, 1984). However, these explanations seem far-fetched
when it comes to more complex linguistic behavior, and they haven’t led to
any research-related progress connected to behavior analysis’s second objec-
tive: influencing behavior (S. C. Hayes, Barnes-Holmes, & Roche, 2001).
A REDUNDANT QUESTION?
My main quest in writing this book is to describe theoretical conclusions
based on new experimental findings concerning human language and cogni-
tion. I also want to show how we can use those conclusions to solve behavior
analysis’s dilemma concerning verbal and rule-governed behavior. And all
of this is done with the purpose of increasing our understanding of complex
human behavior and our ability to influence it. This knowledge can then be
applied to the types of problems that make people seek psychological help,
and therefore can help provide clinical tools in working for change. But before
exploring this important issue, let me bring your attention to the question of
whether working to understand and influence the function thinking has for
humans is, in fact, clinically relevant. In chapter 3, I’ll examine whether this
has any significance in working for change.
47

1 RFT Learning

Uploaded by

Copyright:

Available Formats

1 RFT Learning

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 RFT Learning

Uploaded by

Copyright:

Available Formats

US $49.

A Clear and Comprehensive Guide

Foreword writer DERMOT BARNES-HOLMES, PH.D., is foundation professor of psychol-

Afterword writer STEVEN C. HAYES, PH.D., is University of Nevada Foundation Professor

Context Press Foreword by Dermot Barnes-Holmes, Ph.D.

“Until now, explanations of relational frame theory have remained largely

Distributed in Canada by Raincoast Books

Copyright © 2010 by by Niklas Törneke

New Harbinger Publications, Inc.

All Rights Reserved

PDF ISBN: 978-1-57224-908-0

The Library of Congress has cataloged the print edition as:

Clarity, simplicity, depth.

A PRAGMATIC THEORY OF HUMAN

to provide a well-developed technical account of the almost infinite novelty

Dermot Barnes-Holmes is foundation Professor of Psychology at the National

colleagues on rule-governed behavior was in my awareness, but I never saw it

As a psychotherapist, I am a child of my time. I grew up in a psychodynamic

dialectical behavior therapy (DBT), which revealed a strong influence from

intention. I am aware that the psychodynamic and cognitive theories that

described following the experimental work are applied to different types of

THE CHARACTER OF THE BOOK

Radical behaviorism is the philosophical basis on which B. F. Skinner built

to understand and influence behavior, we need to study its function—that is,

BEHAVIORISM AND ITS PREMISES

Is it possible to understand humans based on an understanding of animals?

WHAT IS RADICAL ABOUT RADICAL

FUNDAMENTAL BEHAVIOR ANALYTIC

Operant Conditioning: Learning Through

In operant psychology, the different influences of consequences are cate-

Punishment: A consequence that decreases the likelihood that a

Before I go on to describe how we can analyze a specific behavior, I must

In behavioral analysis, a common way of describing an operant sequence

increases. Antecedent refers to precisely this context: the conditions under

is a description of a discriminative function.1 Alongside this, however, the

BA SIC F U NCT IONS OF T H E DI F F ER E N T

Discriminative function Reinforcing function

Something an individual does

Motivational function Punishing function

It is often useful to distinguish between the discriminative and motiva-

A NOTE ON WHAT ABC IS NOT

When we describe a behavioral sequence in the way I have above and

very activity. We do what we do because we have a history of connections

Respondent Conditioning: Learning by Association

mouth in the future. This is operant behavior, governed by consequences. But

2 The term “stimulus” is common within behavioral psychology. It denotes

trigger unconditioned responses. In this way, formerly neutral stimuli become

The main square in my town

After the above experience this might follow:

Conditioned stimulus Conditioned response

Operant and Respondent Learning Interact

Respondent extinction also involves crucial changes or a discontinuation

WHERE WE GO FROM HERE

The phenomenon we commonly call “thinking” is obscure to most of us.

as a reader, are devoting your time and attention to understanding what I

actions. This makes it possible for me to be given a desired object without

of verbal behavior in the category autoclitic behavior—verbal behavior that is

The Use of Skinner’s Analysis in This Book

LEARNING TO TACT PRIVATE EVENTS

2 The term “stimulus” is common within behavioral psychology. It denotes