Reflections On Behaviorism and Society PDF

:REFLECTIONS
ON
BEHAVIORISM
AND
SOCIETY
B. F. Skinner
P r e n t ic e -H a ll , I n c ., Engletuood CliffSj N.J. 07632

O T H E R C E N T U R Y P S Y C H O L O G Y SERIES T IT L E S
B Y T H E SA M E A U T H O R
Contingencies of Reinforcem ent: A T heoretical Analysis

T h e Behavior of Organisms: A n Experim ental Analysis
T h e Technology of Teaching
Cum ulative Record: A Selection of Papers, 3rd E dition
Verbal Behavior
Library of Congress Cataloging in Publication Data
S k in n e r , B urrh us F r e d e r ic ,
Reflections on behaviorism and society.
Includes bibliographical references and index.
1. Behaviorism. 2. Personality and culture.
I. Title.
B F igg.S ss 301.1 77-28636
ISB N 0-13-770057-1
© 1978 by Prentice-Hall, Inc., Englewood Cliffs, N.J. 07632
A ll rights reserved. No part of this book may be reproduced in any form or

by any means without permission in writing from the publisher.
Printed in the United States of America
10 9 8 7 6 5 4 3 2 1
London
P r e n t i c e - H a l l I n t e r n a t i o n a l , I n c .,
P r e n t i c e - H a l l o f A u s t r a l i a P t y . L im ite d , Sydney
P r e n t i c e - H a l l o f C a n a d a , L t d . , T oronto
P r e n t i c e - H a l l o f I n d i a P r i v a t e L i m i t e d , New Delhi
P r e n t i c e - H a l l o f J a p a n , I n c ., Tokyo
P r e n t i c e - H a l l o f S o u t h e a s t A s i a P t e . L t d . , Singapore
W h i t e h a l l B o o k s L i m i t e d , Wellington, New Zealand
To
Eve renée
Contents
i
\
f
Preface
SOCIETY
1 Human Behavior and Democracy 3
2 Are'W e Free to Have a Future? 16
3 The Ethics of Helping People 33
4 Humanism and Behaviorism 48
5 "Walden Two Revisited 56
THE SCIENCE OF BEHAVIOR

6 The Steep and Thorny W ay to a
Science of Behavior 68
7 Can We Profit from Our Discovery of
Behavioral Science? 83
8 W hy I Am N ot a Cognitive Psychologist 97
9 The Experimental Analysis of Behavior (A History) zr3
viii Contents
III EDUCATION
10 Some Implications of Making Education More Efficient 129
11 The Free and Happy Student 140
ii Designing Higher Education 149
IV A MISCELLANY
13 The Shaping of Phylogénie Behavior 163
14 The Force of Coincidence 169
1j Reflections on Meaning and Structure 1/6
16 Walden (One) and Walden Two 188
17 Freedom and Dignity Revisited 195
18 Freedom at Last, from the Burden of Taxation 199
Acknowledgments 202
Index 205
Preface
T h is is not a book to be read straight through. Most of the papers

were occasional, and the occasions various. For those who prefer an
overview, here is a b rief synopsis:
1. B ehavior m odification is just the technology we need

to promote the face-to-face control of people, by people,
and for people and thus to reduce the scope of the cen
tralized institutions of governm ent and economics.
2. W e are beginning to be seriously concerned about the
future. H ow can people be induced to behave in ways
that take the future into account? T h ere are relevant be
havioral processes, bu t only the most careful planning
w ill enable us to use them to solve our problems.
3. T h e things we call “ good” w ork for our good, or for
the good of others, only w hen they are usefully con
tingent on behavior. W e do not really help others sim ply
by giving them good things.
4. It is a mistake to iden tify hum anism w ith the self-
centered individualism of the existentialists. By identify
ing the role of the environm ent, particularly the social
Preface
environm ent, behaviorism makes it possible to achieve

the goals of hum anism more effectively.
5. W alden Two is more relevant today than when it was
written thirty-two years ago. It describes a m inim ally con
suming, m inim ally polluting, m axim ally socializing cul
ture. It is not a bad start toward restoring the place of
small com m unities in m odern life.
6. T h e m ajor obstacles to progress in a science of be
havior are certain long-standing commitments to an in
ner w orld of the individual, either the m etaphorical
w orld of the m entalistic or cognitive psychologist or the
real bu t not at the m oment relevant w orld of the phys
iologist.
7. T h e effective m anagem ent of hum an behavior is jeo p
ardized when we appeal to feelings and ideas in the ex
planation of behavior. In doing so we neglect useful
environm ental contingencies.
8. C ognitive psychologists engage in a m etaphorical trans
location of the environm ent, m oving contingencies of
reinforcem ent, severally or as a whole, into the supposed
w orld of the m ind. It is assumed that people can some
how adjust to private contingencies more effectively be
cause of the intimacy.
B u t the changes in behavior attributed to the supposed
internal contingencies are due instead to the external
contingencies from w hich they were derived.
9. (A b rief account of the author's part in the evolution
of the experim ental analysis of behavior.)
10. T h e experim ental analysis of behavior has im proved
education by clarifyin g its objectives, suggesting new
practices in classroom management, and introducing in
structional program m ing texts and other materials. As a
result students learn in less time and w ith less effort, but
difficult problems are then raised for the educational es
tablishm ent.
11. Rousseau’s free and happy student appears to be
n either free nor happy. Permissiveness is not the only al
ternative to the aversive control characteristic of educa
tion throughout its history. Other feasible alternatives
m ake it possible to prepare young people for the future
that lies ahead o f them.
12. Teachers w ho leave education to the innate curiosity
of the student in a natural “ learning environm ent” aban
don their role as transmitters of the culture. N ew instruc
Preface
tional practices can restore that essential function. H igher

education is especially resistant to a behavioral technol
ogy, but changes are b eing m ade— as in the personalized
system bf instruction of F. S. Keller.
13. It is possible that instinctive behavior has been
“ shaped” by a process of selection not un like the shaping
of the behavior of the individual, although it required
hundreds of m illions o f years rather than hundreds of
seconds. Recent discoveries in the field of plate tectonics
or “ continental d rift” point to slow changes in the en
vironm ent which m ay have shaped some unusual exam
ples of species behavior.
14. Certain kinds of coincidences are often said to show
an underlying order in the world that has not been rec
ognized by science. It is easy to collect instances, not be
cause they are common, but because they are particularly
likely tq be noted and remembered. Coincidences attract
our attention in part because the relation between re
sponse and consequence in operant conditioning is essen
tially cdincidental.
15. Structuralism in linguistic and literary criticism does
not yield a satisfactory analysis. W e do not have the kinds
of inform ation about the writer needed for a convincing
functional alternative, b u t internal relations am ong the
parts of w hat a person writes sometimes point to relevant
verbal processes.
16. Walden Tw o is not as different from T h o re a u ’s
W alden (One) as critics have claimed. Both books argue
that we should exam ine the way of life into w hich we are
born and if possible replace it w ith a better way. B oth
point to the advantages of sim plifying one’s life, and
neither is escapist in any real sense. You ng people today
are discovering not on ly how they can sim plify their lives
b u t how they can solve a problem that T h o rea u ne
glected— the problem of com m unity.
17. B y predicting that m an w ould deliberately go m ad to
prove that his behavior could not be predicted, Dos
toevsky cut off that last line of escape, because going m ad
was henceforth a predictable reaction. H e thus illustrated
one of the great paradoxes of freedom: any attem pt to
prove that m an is free is probably less productive than
changing our cultural practices in such a way that people
feel freer than they have ever felt before. T h e change,
nevertheless, im plies a certain measure of control.
xii Preface
18. “ State lotteries can become an ideal way in w hich

free, happy, and affluent people support their govern
m ent w ithout paying taxes.”
T h ere is more in each chapter than I have been able to pu t in

a few sentences, as I trust the reader who now turns to parts that
may be of interest w ill discover.
B. F. S k in n e r
REFLECTIONS
ON
BEHAVIORISM
AN D
SOCIETY
PART I
SOCIETY
1 Human Behavior and Democracy

2 Are We Free to Have a Future?
i
3 The Ethics of Helping People
4 Humanism and Behaviorism
5 Walden Two Revisited
I
Human Behavior
and Democracy
i
\
?
. . . T h a t this nation, under God, shall have a new

birth of freedom; and that control of the people, by the
people, and for the people, shall not perish from the
earth.
Abraham Lincoln? N o t exactly. L in coln said “governm ent of

the people,” not “ control,” and there is a difference. T o govern
once meant simply to guide, but the w ord soon acquired a stronger
meaning. Governm ents “ com pel obedience to authority.” In other
words, they treat people aversively— punishing them w hen they
behave badly and relax in g a threat of punishm ent when they be
have well.
W hen they are too aversive, people escape from them or attack
and weaken them w ith violence, terrorism, protests, strikes, boy
cotts, or revolution. T h e y thus impose a kind o f countercontrol
upon the power to punish. Some sort of equilibriu m m ay be
reached, and we then speak of governm ent “ by the consent of the
governed,” where “ consent” marks the lim it beyond w hich an
4 SOCIETY
authority may not com pel obedience. Note that the countercontrol,
like the control, is aversive. T h e presum ed value o f a “ governm ent
by the people” is that when people govern themselves they w ill use-
aversive measures w ith restraint.
B ut w hy should governm ents confine themselves to aversive
control? W h y not use positive reinforcem ent? M any governments
have the means o f doing so; they have the power to provide as
well as punish. O ne answer may be that positive reinforcem ent is
not w ell understood. Its effects are easily overlooked; w e do not feel
the control exerted when our own behavior is positively reinforced.
Aversive action also has a kind of genetic priority. Aggressive reper
toires, as w ell as the capacity to acquire aggressive behavior readily,
have had survival value. It is also easy to learn to treat others
aversively because the results are especially quick. Nevertheless,
negative reinforcem ent and punishm ent have serious disadvantages
w hich deserve attention, particu larly now that dem ocracy as a
philosophy of governm ent is in trouble. T h e re are on ly a few real
democracies in the w orld today, and the demise o f democratic
governm ent is being w idely predicted. Em erging nations tend to
adopt the pattern of obedience to authority, epitom ized in the m ili
tary dictatorship, and m any older nations are m oving in that direc
tion. Sim ply as the aversive countercontrol of the pow er to treat
people aversively, dem ocracy is losing ground. Can we save it, and
preserve and further its achievem ents, by m aking a greater use of
nonaversive measures?
It m ay be argued that som ething of the sort is done in the
welfare state. O ur own governm ent is perhaps as m uch concerned
w ith freedom from w ant as w ith freedom from fear; consider the
services it provides in health, education, and welfare. B ritain and
the Scandinavian countries have gone touch further, o f course, and
co, at least in ^hnc.iy, have the com m unist countries. B u t it is hard
to find positive reinforcem ent in any of this. W elfare states sustain
themselves w ith aversive practices. T h e y acquire the goods they
distribute through taxation (backed by a threat of punishm ent) or
through the coercion of labor, and if they distribute goods “ accord
ing to need” it is largely according to w hether the needy w ill
otherwise protest. T h e w elfare or com m unist state also shows an
unstable equilibriu m between aversive control and countercontrol.
Human Behavior and Democracy 5
Moreover, and this is the im portant point, it does not m ake the
goods it distributes contingent upon the behavior of its citizens. It
does not u'se them as reinforcers but as appeasem ent, to reduce
countercontrolling action. A t best it m oderates certain conditions
that may otherwise lead to punishable behavior, since people are
presumably more likely to behave w ell in a w orld free of poverty, ill
ness, unem ploym ent, and ignorance. B ut even full-fledged welfare
states continue to punish m isbehavior, and strong punitive sanc
tions certainly survive in com m unist countries.
W e cannot avoid the conclusion that som ething that could
contribute to governm ent in the broadest sense is being overlooked.
Positive reinforcem ent, as the term implies, is strengthening. It lacks
both the suppressive and the aggressive effects o f punishm ent, and
it is free of the effects o f negative reinforcem ent that we associate
w ith anxietyi and fear. Positively reinforced behavior is active par
ticipation in life, free of boredom and depression. W hen our be
havior is positively reinforced we say we en joy w hat we are doing;
we call ourselves Jiappy. C ertainly these features o f hum an behavior
should be am ong the goals of any governm ent “ for the people,” but
they are out of reach of governm ents w hich m erely com pel obedi
ence and are, at best, left to chance in w elfare states. C an they be
brought w ithin reach in a democracy?
Let us look at two problem s faced by all forms of governm ent

in the w orld today b u t especially relevant here because they have
been created by what w ould certainly be regarded as the two great
triumphs of a dem ocratic way of life. Significantly, they are also the
products of the basic behavioral processes we are considering. T o
borrow an expression which is perhaps too fam iliar, the ways in
which people react to negative and positive reinforcem ent have led
to the establishment of the rights to life and liberty and to the
pursuit of happiness, respectively. T h e y have also led to trouble. N o
matter how essential to the survival of a species a process m ay once
have been, it can becom e troublesom e or even lethal w hen the
environm ent changes, and this has happened to both positive and
negative reinforcem ent.
T h e processes through w hich, organisms learn to escape from
6 SOCIETY
or avoid various kinds o f physical damage have had an obvious

survival value, but in w hat we call a civilized- environm ent they
become less im portant, and a poin t m ay be reached at w hich they
w ork against survival. For exam ple, a vast technology has been
developed to prevent, reduce, or term inate exhausting labor and
physical damage. It is now dedicated to the production o f the most
trivial conveniences and comfort. N ot only do we not suffer ex
tremes of cold and heat, we keep our buildings w ithin a narrow
range of temperatures. N ot only do we not work to near exhaustion,
we ride escalators rather than clim b stairs and push buttons to open
the windows of our cars. Unless we then devise strenuous and stress
ful substitutes, we find ourselves vulnerable to any strong environ
m ental demand, as well as to stronger people (the archetypal pat
tern of the civilized person versus the barbarian). M oreover, because
the technology cannot be made available to everyone, our trivial
gains mean costly losses for others.
T h e social parallel is far m ore im portant. N o one w ill ques
tion the im portance of the historical struggle for freedom , through
w hich people have escaped from and weakened or destroyed those
w ho have treated them aversively, but this process of establishing
the right to life and liberty has reached the poin t at w hich any
infringem ent upon the free m ovem ent o f the in d ivid u al is chal
lenged. People claim the right to do as they please— to gam ble
away a fortune, risk unnecessary danger by not w earing a seat belt,
die an alcoholic, and consume resources and pollu te the environ
m ent w ithout restriction. Students are to enjoy free and open class
rooms, people w ith problem s are not to be told w hat to do bu t are
to find solutions within themselves, business is to flourish in an
atmosphere of laissez-faire, and the form of behavior most subject
to com plaint by one’s peerc is com plaint itself.
It is perhaps a natural m istake to suppose that the abolition
o f aversive social control leads in the end to this kind of permis
siveness but, like convenience and comfort, small personal freedoms
are purchased at great social cost. Everyone suffers w hen people are
ill-mannered, illiterate, and ignorant, when laws are frequently
broken, when people continue to need help, w hen goods are
u n equally distributed, and w hen so-called victimless crimes prove
to have victims. In short, the w orld has changed, and the processes
through w hich we free ourselves from aversive stim ulation, non
social and social, have begun to w ork against the survival of the
culture and possibly the species.
T h ere has been a com parable miscarriage o f the process of

positive reinforcem ent. Jefferson borrow ed the phrase “ the pursuit
of happiness” from John Locke, bu t Locke had said “ the pursuit of
property.” T h e technology now devoted to the production of rein
forcing goods is far more extensive than that concerned w ith the
avoidance of exhausting labor and physical damage, and unless it
is restrained it w ill soon exhaust the w orld ’s resources. It has
another seribus effect because people differ in the ability to acquire
property and hence in the quantities they possess, and since posses
sion usually makes acquisition easier, differences have become very
great. Positive reinforcem ent has led not only to great w ealth but
to extrem e poverty. W hen the poor become numerous enough or
otherwise pow erful enough to protest, they m ay be given some share
of the wealth, but that leads to further trouble. W elfare— either as
a social measure or as a political philosophy— raises the problem
of the noncontingent reinforcer, to w hich I shall return.
Here, then, are two basic issues faced by all modern govern
ments. Somewhere between freedom and despotism and between
affluence and poverty there are points at w hich personal and social
gains are balanced, but how can those points be reached? T h e most
likely answer shows the traditional preoccupation w ith aversive
control: we 'should enforce the laws, lim it the extent to w hich
people can acquire goods (as by taxing excesses), and m ake people
work for what they get. B u t are there n onpunitive alternatives? C an
we design an environm ent in w hich people w ill treat each other
well, keep the size of the population w ithin bounds, learn to w ork
and w ork productively, preserve and enhance th e reinforcing char
acter o f'th e w orld, explore and analyze that world, lim it the use
of resotiirCes and keep the environm ent safe for future generations,
and do all this because the results are positively reinforcing?
A social environm ent in w hich people thus behave as they

like, rather than as they have, to behave, has been the dream of
many political and social reformers, bu t it is usually called
SOCIETY
“ utopian ” in the pejorative sense of impossible. Nevertheless, we

are already under w ay in developing just such an alternative to
governm ent as the pow er to com pel obedience. A n d it may lead to
som ething that is closer to a governm ent of people by people than
anything yet proposed in the name o f democracy.
People are governed, in the broadest sense, by the w orld in
w hich they live, particularly by their social environments. The
operation of such an environm ent is most obvious in a small
homogeneous group, where behavior injurious to others is punished
and behavior favoring others is reinforced, either by relaxin g a
threat or by presenting goods. As a social environm ent evolves,
supportive practices appear. T h e group classifies behavior as good,
bad, right, and w rong and uses these terms as conditioned rein
forcers in strengthening or suppressing behavior. It describes some
of the more im portant contingencies in the form o f rules, and by
follow ing rules its members conform more quickly and avoid direct
exposure to punitive consequences. Individuals may act to m ain
tain the very contingencies to w hich they conform and when they
do so w ithout supervision, they are said to show self-control or the
possession o f an ethical or m oral sense.\Such a social environm ent
transmits itself as new members o f a group acquire the behavior of
m aintaining the contingencies.
U nfortunately, people govern people in this rather idealistic
sense only w hen everyone has essentially the same power, and this
is almost never the case. Someone emerges as a leader and, un fortu
nately, almost always by exerting a special share o f the power to
com pel obedience. Countercontrol m ay lim it that power, bu t the
result is not a truly egalitarian society. Som ething of the same sort
follows when a group delegates control to representatives, since
delegation can have the same effect as usurpation. Preventing the
misuse of pow er by on e’s own representatives is only a m ilder form
o f the struggle for freedom from tyranny. N either process gu aran
tees a balanced governm ent.
It was once the practice to divide the social environm ent into
three parts: (1) the polity (government in the narrow sense, special
izing in aversive control), (2) the economy (specializing in the p ro
duction and exchange o f reinforcing goods), and (3) the culture, or
all the other contingencies of reinforcem ent m aintained by the
group— in fam ily practices, religious rituals, arts, crafts, and so on,
It is probably im possible to keep these fields apart, and in its

modern use ithe term culture covers them all. A culture is a com
plete social ênvironment, in w hich some contingencies are m ain
tained by individuals and others by institutions. T h e earlier d ivi
sion was useful, however, because culture in the older sense m eant
the social contingencies not m aintained by centralized agencies.
Democracy has a special m eaning when we apply the term to a
culture in that sense.
It is then m ore obvious that control rests w ith the people. A
social environm ent exists only because of w hat people do for and
to other people, and it is never m ore than that even w hen pow er
is usurped by, or delegated to, a special agency, bu t in a culture
in the older sense the control is direct. Concentration o f pow er in
an agency is objectionable not on ly because it is characteristically
misused and jwasted but because it destroys interpersonal contacts.
If I w ork for a com pany m anufacturing shoes and my neighbor for
a company m anufacturing shirts, and if we both earn enough so
that I buy a shirt and he or she a pair of shoes, we have in a sense
produced som ething o f value for each other, bu t there has been no
direct exchange. A special opportunity to reinforce each other's
behavior has been lost. Com panies are no doubt needed for the
efficient production of shoes and shirts, and we must have an econ
omy rather than sim ply a culture in the older sense, bu t som ething
has been given up. Sim ilarly, if I delegate the censure of m y neigh
bor to the police, I am less likely to search for nonpunitive alter
natives than if I act sim ply as a neighbor. In a large group a police
force is no doubt needed and we shall continue to have punitive
governments, b u t the chances of w orkin g out better personal rela
tions are then reduced.
W hen we delegate the control of people to p olitical and
economic institutions, we relinquish the face-to-face control o f an
equitable governm ent o f people by people, and it is a mistake to
suppose that we recapture it by restricting the scope of those to
whom we delegate it. A better strategy is to strengthen face-to-face
control. A social environm ent, or culture, can operate w ithout the
help of usurping or delegated rulers and entrepreneurs, and it is
most clearly a governm ent of people by people w hen it does so.
Something of the sort has been proposed from time to time— for
exam ple, in the p olitical philosophy of anarchy— b u t n othin g could
SOCIETY
better illustrate the failure to find appropriate means than the

public stereotype of the anarchist as a m an w ith a bom b. W e are in
a m uch better position today. W e have begun to understand how
the environm ent, particularly the social environm ent, works, and
we already have some glimpses of how it can be made to w ork
better.
M uch of this has come abou t through the application o f the
experim ental analysis of behavior, or what has come to be called
behavior m odification. O ne cannot use that term today w ithou t
adding a caveat and a definition. I do not m ean the m odification of
behavior by im planted electrodes or psychotropic drugs. I do not
mean Pavlovian conditioning w ith vom it-inducing drugs or electric
shock. B y “ behavior m odification” I mean w hat the term was intro
duced to mean— changing behavior through positive reinforcem ent.
T h e un derlying processes have lon g been know n and occasionally
used, but we now have a better understanding of their role in the
social environm ent and can therefore make significant changes in
the face-to-face control o f people by people.
M any people have had frightening visions of behavior m odifi

cation in the hands of pow erful governments or rich corporations,
but the fact is that the m ajor applications to date have been
precisely at the level of the face-to-face control of people by people
— by teachers who find better ways of w orking w ith students in the
classroom and who use instructional m aterials w hich enable stu
dents to progress as rapidly as possible and w ith a m inim um of
aversive pressure, by attendants in hospitals and homes for psy-
chotics and retardates w ho arrange conditions under w hich those in
their care lead more interesting and dignified lives, by psychothera
pists ;n Cace-to-face consultation w ith those who need help, by
parents w ho discover how to m ake the fam ily a warm er and more
h elpful institution, by em ployers who design incentive systems
under w hich employees not only w ork w ell b u t enjoy w hat they do,
and by individuals who discover how to m anage their own lives
effectively when face to face w ith themselves.
M ore than a hundred books have been published about be
havior modification in the past five years and the rate of pu blica
tion continues to rise. T h ere is no indication that the principles are
being sequestered or m onopolized by individuals or organizations

bent on exploitative control. On the contrary, the basic practices
are finding tjheir w ay into d aily life as part of our culture. It is
difficult to prescribe practices appropriate to a given situation.
Th ere are no general rules w hich w ill perm it us to gloss over
details. But some of the principles com m only observed in the ap p li
cation of an experim ental analysis to d aily life are w orth noting
because they are particularly concerned w ith the governm ent of
people by people. In one form or another they have a long history.
T h e veiy substitution of positive reinforcem ent for aversive

control is, of course, at the heart of the struggle for freedom. A l
though we still have a long way to go, we have moved from slavery
to the paym ent o f wages, from the birch rod to the free school, and
from bedlam to hum ane care of the psychotic and retarded. Positive
reinforcement has a strengthening effect not only upon the behavior
of the individual^ but also upon the culture, by creating a w orld
from which people are not likely to defect and which they are
likely to defend, prom ote, and im prove. A ll those w ho act to
make the physical world m ore b eau tifu l— the ecologists concerned
with natural beauty and the artists, musicians, architects, and
others who create beautiful things— all increase the chances that
living in the w orld w ill be positively reinforced. T hose who use
behavior m odification, properly defined, could be said to be con
cerned w ith preserving and furthering the beauty, o f the social
environment— or, to borrow a phrase from a vanishing culture, to
create more b eautiful people.
A second principle in im proving the control of people by
people is the avoidance of contrived reinforcers. Here, again, there
is a long history. W e all live in a token econom y. M oney was in
vented as a conditioned reinforcer because it has m any advantages:
it is easily given and received; consum ing the uncontrived rein
forcers for w hich it is exchanged can be conveniently postponed;
reinforcing values can be easily compared, and so on. B u t behavior
is most expeditiously shaped and m aintained by its natural conse
quences. T h e behavior of the production line w orker w hich has no
im portant consequence except a w eekly w age suffers in com parison
with the behavior of the craftsman w hich is reinforced by the things
12 SOCIETY
produced. T h e separation of workers from the natural products of

their w ork was, of course, w hat M arx m eant by. “ alienation.” T h ere
is a similar effect when punitive sanctions are delegated to authori
ties, because negative reinforcers like fines or im prisonm ent alienate
citizens from the direct censure of their peers.
T h ere is nothing w rong w ith contrived reinforcers as such.
Teachers and counselors need them to shape and strengthen be
havior w hich the individual w ill find help fu l in the n atural con
tingencies in daily life. B u t contrived reinforcers must be abandoned
before the preparation is complete. T h e student w ho continues to
turn to a teacher has not been successfully taught; the client who
continues to consult a counselor has not been successfuly counseled.
T h e uncontrived reinforcers of the world at large must take over.
T h e practices of industry and governm ent are different. W orkers
must continue to receive the contrived reinforcers called wages, and
citizens to be threatened with the contrived consequences called
punishm ent. A lien ation is then likely to follow .
A third principle is rather similar. B ehavior which consists
of follow ing rules is inferior to behavior shaped by the contin
gencies described by the rules. T h u s, we m ay learn to operate a
piece o f equipm ent by follow ing instructions, bu t w e operate it
sk illfully only when our behavior has been shaped by its effect on
the equipm ent. T h e instructions are soon forgotten. Sim ilarly, by
learning the rules of a culture we are enabled to deal w ith people
effectively, but our behavior w ill be most sensitive to the contin
gencies m aintained “ by the people” when we are directly censured
and commended, and the rules of the culture, like the operating
instructions for a piece o f equipm ent, forgotten. (A fam iliar obser
vation in jurisprudence is that laws survive long after the personal
relationships they describe have changed, and they then m isrepre
sent the prevailing social control.)
A fourth principle is not so w idely recognized. C ontrol of
people by people is likely to be disturbed by “ noncontingent” rein
forcers. M any good things come to us free— from a b o u n tifu l cli
mate, from a run of good luck, from other people w h o give them
to us or allow us to take them w ithout a struggle, or from a store
o f goods we have already accum ulated. W e count ourselves lucky
when these potential reinforcers come our way when we have done
nothing for them, but we should not overlook the dam age they may
do. N oncontingent reinforcers are characteristic of both affluence

and w elfare and have the same troublesom e effects in both. B y re-
ducing the level of deprivation, they preem pt m any possibilities of
reinforcement, and reinforcers o f a lesser b iological significance take
over. T h e results are sometimes productive. W e m ay turn to art,
music, literature, science, or the other great achievements of the
hum an species. M ore often, however, they are stu ltifyin g and waste
ful— as when we turn to alcohol or other drugs, surrender to the
variable-ratio schedules exploited by gam blin g systems, vicariously
live the serious lives o f others in gossip, literature, films, and
spectator sports, or turn to violence as an escape from boredom . A
policy o f “ work not w elfare” m ay solve the problem of the non
contingent reinforcer for the unem ployed, but not for the affluent.
N oncontingent reinforcers keep the group from most fu lly develop
ing the capacities of its members and threaten the strength of the
culture and presum ably its chances of survival.
Still another principle concerns the extent to w hich a culture
prepares its members to m eet its contingencies. A social environ
ment is extraordinarily com plex, and new members o f a group do
not come prepared w ith appropriate behavior. T h e ind ivid u al was
once inducted into a culture by n atural instructional programs, in
the presence of favorable models. T h ese are no longer an im portant
part of grow ing up, and m ore exp licit control is now needed. P ro
grammed sequences of contingencies, in the hands of skillful
teachers and counselors, can lead efficiently to the com plex reper
toires demanded by a social environm ent.
These, then, are some of the principles to be observed in

prom oting the effective control of people by people. James Reston,
writing ih m e New York Tim es, quoted th e-L o n d o n Econom ist on
the contribution Am erica can m ake in its third century. It w ill
depend, the Econom ist said, on how its three m ain institutions
evolve. “ These three m ain institutions are, in reverse order of im
portance, its business corporations, its governm ent, and its m echa
nisms for livin g together” — in other words, the econom y, the polity,
and the culture in the older sense. Perhaps wc m ay leave business
to the economists and governm ent to the p o litical scientists, but to
whom shall we assign the “ mechanisms for livin g together,” w hich
14 SOCIETY
the Econom ist puts at the top of the list? I submit that they are
simply the contingencies w hich define the social environm ent as a
culture and therefore precisely the field of a technology of behavior.
“ Mechanisms for livin g together” compose the w hole field of
social psychology, but that does not mean that we can look to all
social psychologists for help. A pure structuralism makes very little
difference, and developm entalism not much more. T h e measure
ment of feelings and attitudes and other states of m ind is scarcely
a spur to action. Psychologists in general are not distinguished by
any great readiness to act. N ot only do they hesitate to change the
behavior of other people, m any of them strongly oppose any effort
to do so. T h is narrows the field when we are looking for those who
w ill contribute to our third century by im proving our mechanisms
for livin g together.
T h e trouble is that any allusion to the control of hum an be

havior evokes the challenge: who w ill control?— often w ith the im
plication that a technology of behavior w ill n aturally fall into the
hands o f despots. Like all sources of power it could very w ell do
so, especially if those who are not despots refuse to act. B ut the very
threat o f misuse is the best reason for looking as clearly as possible
at how a science o f behavior can w ork “ for the people.” Behavior
modifiers who stop intervening when their w ork is finished are
certainly not classic exam ples of despotic rulers. T ru e , they may
pose a different threat. T h e y are perhaps no more likely to engage
in despotic control than atom ic physicists to conquer the w orld with
nuclear weapons, but w ill they not lend their skills as consultants
to potential despots? A M achiavelli who uses his insights to advise
a prince is perhaps as dangerous as the prince. B ut behavior m odi
fication is prim arily a w ay of m aking people m ore effective, not in
ru lin g others, but in m aintaining and im provin g the social en
vironments in w hich they live.
It is often said that in the end the question is w ho w ill control
the controllers (Qiiis custodiet ipsos custodesf), bu t the issue is not
W ho but What. People act to im prove cultural practices w hen their
social environm ents induce them to do so. C ultures w hich have this
effect and w hich support the relevant sciences are m ore likely to
'fn
.i-M
M
Human Behavior and Democracy *5
solve their problem s and survive. It is an evolving culture, then,

which is most lik ely to control the controller.
U n fo r tu n a te ly , it does not have the same effect on everyone.
Those who act to im prove governm ent of the people by the people
for the people have been selected by special, possibly accidental,
circumstances. Since they have been selected, they are an elite, bu t
they are not the exp lo itin g elite that has given the w ord such bad
connotations. T h e ir task is not to control people b u t to b rin g
people under the control of more effective physical and social en
vironments. T h e y operate upon the environm ent, not upon people.
Physical and biological technologists w ork w ith one part of
that environment, as they construct contingencies affecting hum an
behavior; and they do not in any sense rem ain in control o f the
people whom their achievements affect. Teachers, therapists, and
other behavioral technologists w ork w ith another part of the en
vironm ent— as they construct the contingencies under w hich people
control people. B u t they themselves do not continue to intervene.
W e see this in m iniature in such a field as fam ily counseling. T h e
f • • •
counselor changes certain practices— for exam ple, by teaching the
members of a fam ily to commend each other, rather than criticize
or com plain— b u t the project is not finished u n til the fam ily works
more efficiently as a system w ithout further intervention by the
counselor. N o cultural practice designed through the application
of an experim ental analysis of behavior involves a behavior m odifier
who remains in control. Control rests w ith “ the people.”
T h ere w ill no doubt continue to be governm ental and eco

nomic agencies, organizations, and institutions, for they have their
proper functions, but they should not be given an exclusive fran
chise. A social environm ent functions most successfully for the
individual, the group, and the species if, so far as possible, people
directly control people. T h e design of a social environm ent in
which they do so is one of our most pressing needs. It is quite
clearly a special challenge to psychology as a science of behavior.
Are W e Free
to Have a Future?
It is often pointed out that I have specialized in the behavior

of rats and pigeons, and it is usually im plied that as a result my
judgm ent about people has been warped, bu t at least sixty percent
of w hat I have published has been about hum an behavior. I have
discussed governm ent, religion, psychotherapy, education, language,
incentive systems, art, literature, and m any other hum an things.
A n d so, of course, have thousands of other people, but I do not
believe I have offered my readers just m ore of the same, for that is
where the other 40 percent comes in. In w ritin g about hum an affairs
I have always stressed the im plications o f an experim ental analysis
of behavior w hich was, indeed, first carried out on
lower species, but w hich was eventually extended to hum an subjects
w ith com parable results. Even the w ork w ith other species was rele
vant to hum an affairs, because it revealed the extraordinary role
played by the environm ent in the determ ination of behavior. One
did not need to believe that men and wom en were just like rats and
pigeons to begin to look more closely at the w orld in w hich they
lived . It became clear that certain features o f that w orld had a bear
ing on some long-standing problems. W h a t follows is offered as an
exam ple.
Are We Free to Have a Future? l7
Doomsday prophecies are now a com m onplace of daily life. W e

are continually rem inded that, for all its past triumphs, m ankind
may be headed straight for disaster. Unless som ething is done, and
soon, there w ill be too m any people in the world, and they w ill ever
more rapidly exhaust its resources and pollute its air, land, and
water, until in one last vio len t struggle for what is left, some m ad
man w ill release a stockpile o f nuclear missiles. T h ere are optimists,
of course, who contend that the hum an species, like some other
species, w ill prove to have some built-in mechanism w hich limits
population (a mechanism m ore acceptable than the famine, plague,
and w ar which have served that purpose in the past), that new and
n onpolluting sources of energy w ill be discovered, and that some
kind of world governm ent or possibly the deterrent effect of even
more horrible weapons w ill put an end to war. But the trend is
certainly omipous, and Cassandra, w ho always prophesies disaster,
may again be right. If so, it w ill be for the last time. If she is righ t
now, there w ijl be no m ore prophecies o f any kind.
One of t.he most om inous things about the future is how little
is being done' about it. T h e great m ajority of the people on the
earth do not know that there is a problem , and of those w ho know
very few take any relevant action. A m ajor difficulty is that the
future always seems to conflict w ith the present. It may be obvious
to commuters that their private cars are p ollu tin g the air they
breathe, but a private car is nevertheless m uch more convenient
than public transportation. Energy m ay be in short supply, bu t it
is pleasant to 'h e a t buildings in the w inter and cool them in the
summer so that roughly the same kind of clothing can be worn in
both seasons. Inflation underm ines the future w hich w ould other
wise be provided for by personal savings or social security, bu t higher
wages for labor and higher prices for m anagem ent are m om entarily
rewarding. O verpopulation m ay be a m ajor threat, but people take
pleasure in procreation and pride in children, W ars m ay be inevi
table so long as w ealth is u nevenly distributed, bu t those w ho are
lucky enough to have an undue share n aturally defend it. Physical
and biological technologies are probably pow erful enough to solve
these problems and guarantee a decent future, b u t they w ill do so
only if they are put to use. T h e problem is hum an behavior. H ow
can people be induced to take the future into account? T h a t is a
question to which, I think, an analysis of behavior is relevant.
i8 SOCIETY
W h at does it mean to say that a person “ takes the future into

account” or acts in a given w ay “ because of.” som ething that w ill
happen in thé future? C an anything have an effect before it occurs?
Final causes were soon ruled out of physics and eventually out of
biology, but must we suppose that there is some w ay in w hich they
function in the field o f hum an behavior?
T h e traditional answer is yes. H um an beings, it is said, differ
from physical objects or non-hum an livin g things because they can
think about the future. T h e y can imagine the consequences of their
action. T h e y can act because they predict the future and therefore
know what is going to happen. T h ey can be affected by the mere
idea or concept of a future. T h is is a m entalistic explanation of
hum an behavior, of course, and it has the weakness w hich has
always been the hallm ark of mentalism. T hou ghts, images, know l
edge, ideas, and concepts are no explanation at all until they have
been explained in turn. H ow do people come to think, im agine,
have ideas, or develop concepts about the future? W hat does know
ing about the future mean? Questions of this sort bear directly
on the practical problem . Is it any easier to get people to think
about the future than to get them to act w ith respect to it? In fact,
are not the measures we say we take to change m inds the very
measures we take to change behavior? Even for the m entalist the
pi'oblem is to get people to act as if they were thinking about the
future. A ll we can change are the circumstances in w hich people
live, and we want to change them in such a way that people w ill
behave differently. W e are on safer and m ore prom ising ground if
we stick to the behavior.
Some biological processes are relevant to the problem of final
causes. A lthou gh no future ever has an effect on the present, there
is a sense in w hich livin g things are affected by consequences. A n
..‘’effect o f the future” was first recognized in D arw in ’s principle of
natural selection. A genetic change or m utation does not occur be
cause of any relation to the survival of the species, but if the result- ;
ing trait promotes survival, as it does in a few cases, the m utation
becomes a characteristic of the species. W e say that it enables the J|
species to adapt or adjust to an environm ent, and adaptation and M
-adjustm ent, like survival, point toward a future. M oreover, charac- |f
teristics selected by past events seem designed to have an effect on '1
the future. (T he environm ent must rem ain essentially unchanged |
Are We Free to Have a Future? *9
with respect to the features w hich have played a part in selection.

O nly that future is “ taken into account” w hich resembles the past.)
T h e term “ purpose” shows the change in form ulation re
quired. Before D arw in the purpose of any feature of the hum an
body— say, the hand— seemed to lie in the future. A baby was born
w ith a hand designed to grasp objects in the w orld in w hich it was
to live. T h e theory o f n atural selection m oved the significance of
grasping into the past. A person is born w ith a hand w hich w ill be
effective in his environm ent because his ancestors had hands w hich
were effective in theirs. Procreation is an exclusive characteristic of
living things; and it is the transmission of traits from generation to
generation w hich makes natural selection an apparently creative
principle “ taking the future into account.”
T h e individual organism is also affected by consequences. T h e
process evolved '.through natural selection, b u t it operates on a very
different scale. It was foreshadowed by philosophies of hedonism
and fairly exp licitly stated in T h o rn d ik e ’s Law o f Effect. It has been
most clearly dem onstrated in the experim ental analysis of operant
conditioning. If? a given bit of behavior has a consequence of a
special sort, it is more likely to occur again upon sim ilar occasions.
T h e behavior is said to be strengthened by its consequences, and
consequences having this effect are called reinforcers. For exam ple,
a foraging pigeon brushes aside a leaf lyin g on the ground and in
doing so uncovers a seed; if the seed is reinforcing, the pigeon is
more likely to brush aside sim ilar leaves in the future.
In spite o f the difference in tim e scale, operant conditioning
bears a striking resem blance to natural selection. It builds an
adaptation or adjustm ent to the environm ent. It seems designed to
have an effect. It makes possible a sim ilar disposition of purpose,
m oving it from the future into the past. A ll of this gives behavior
a kind of- orientation toward the future. (As in n atural selection,
the environment must be reasonably stable; behavior w hich is
strengthened under a given set o f circumstances w ill continue to be
effective so long as the circumstances do not greatly change. T h e
process “ takes into account” a future which resembles the past.)
A n d only an im m ediate future. O perant cond itioning w ould
be m axim ally effective if it strengthened behavior w hich actually
produced its consequences. H edonism and the L aw o f Effect seemed
to guarantee this because they both appealed to feelings— the
20 SOCIETY
pleasure and pain or satisfaction and annoyance w hich resulted

from action. But the reinforcers w hich figure in the analysis of
operant behavior are physical things, and they are consequences
sim ply in the sense that they follow behavior. T h e y need not be
produced by it. T h e equipm ent used in the operant laboratory
arranges temporal sequences only; there is no functional connection
betw een a response and its effects. It is easy to show that a rein
forcer which follows a response but has no other relation to it is
effective; what we call superstition is an exam ple.
T h is is a defect, and it must be attributed to the exigencies of
n atural selection. O perant conditioning evolved as a useful process
in w hich behavior was brought under the control o f any conse
quence, functional or not. It was useful because in general any
event w hich followed an action was likely to have been produced
by it. It was not necessary to take into account the reasons w hy a
reinforcer occurred, and it is difficult to see how that could have
been done.
T h e more im m ediate the consequence, the m ore likely it is to
have been produced by the behavior it follows, bu t there are other
reasons why reinforcem ent must be quick. If there is a delay, inter
venin g behavior w ill be affected, possibly more strongly than the
b ehavior responsible for the reinforcer. A n d reinforcem ent must
overlap behavior if we are not to suppose that som ething w hich has
not yet occurred can have an effect. T h e future m ediated by oper
ant conditioning is therefore not very remote.
(A possible exception was once called “ stomach m em ory.” In
a laboratory dem onstration, a rat is made sick a few hours after
eatin g a particular kin d o f food and is then found to show a
weakened preference iox tne food. If the rat is m ade sick through
radiation, no interm ediate activity can be involved. Such a mecha
nism should have great survival value in protecting organisms
against indigestible or poisonous foods. T h e aversive consequence %
is anatom ically linked to ingestion and “ overlaps” it in that sense, J
and for the same reason it need not affect intervening behavior of :|
other kinds. If the evidence is valid, a fairly rem ote future is medi- f|
ated by this mechanism, bu t it is an exception. In general a rein- I
- forcer must be closely contingent on behavior if it is to be effective.) 1
Nevertheless, organisms do behave “ because o f” events w hich ||
take place a long time in the future. A possible connection is made j l
Are We Free to Have a Future? 21
through a different process called respondent conditioning. T h e

process probably evolved because it prepared organisms for unpre
dictable features o f their environm ents. Foodstuffs like sugar and
salt elicit salivation as an early step in digestion, but because sweet
and salty foods vary greatly in appearance, organisms could not
have developed the capacity to salivate appropriately to their mere
appearances, no m atter how im portant such a preparatory saliva
tion m ight be. T h ro u g h conditioning, the visual appearance of a
particular food comes to elicit salivation, w hich is “directed toward
the future”— though again not a very remote future.
Som ething of the sort affects the role o f the stim ulus in oper
ant behavior. Sweet and salty foods reinforce the behavior o f finding
or capturing them, and they do so because organisms inclined to
be so affected were more likely to survive and transmit the inclina
tion. But, again, since foods vary w idely in appearance, a suscepti
bility to reinforcem ent by the appearance of a food could have had
little chance tq evolve. W h at evolved instead was a process in w hich
any occasion upon\which behavior is likely to be reinforced becomes
reinforcing in its own right.
G ood exam ples appear when behavior is only interm ittently
reinforced. In a standard experim ent a hungry pigeon must re
spond, say, 5000 times before a response is reinforced w ith a small
amount of food. It must then respond 5000 times again before
another response is reinforced. Shortly after reinforcem ent the
pigeon could be said to be responding “because o f” an event w hich
lies in the fairly distant future. A ratio of 5000 to one can be m ain
tained for hours, but only after a special program in w hich progress
through the ratio becomes reinforcing.
L on g chains of responses can be b u ilt up by conditioning
reinforcers. In a typical classroom dem onstration a rat executes a
series of perhaps ten different responses, each of w hich is reinforced
by the opportunity to execute the next, u n til a final, usually uncon
ditioned, reinforcer appears. T h e first step seems to be taken “ for
the sake of the last,” w hich lies in the fairly distant future. Some
thing of the same sort occurs, when, for exam ple, a person builds
a shelter. T h e last step brings protection from the weather, but it
can be taken only after earlier stages have been com pleted. As the
shelter is constructed, each step is reinforced by the opportunity to
take another step. (Not all sequences originate in this way, as we
22 SOCIETY
shall see, but once established they usually continue to be supported

by some such arrangem ent o f conditioned reinforcers.)
Even when supplem ented by the conditioning o f reinforcers,
operant conditioning w ill not, without help, generate m uch of the
hum an behavior w hich “ takes the future into account.” N o individ
ual could, in a single lifetim e, acquire a very large repertoire in this
way. A farm er plants in the spring “ in order that he m ay harvest in
the fall,” but it is unlikely that anyone ever learned to do so for
that reason alone. A n oth er process comes into play. It involves
other people, who accum ulate and transmit useful behavior.
A basic process, im itation, may be part of the hum an genetic
endowment. Other people have been a stable feature of the hum an
environm ent, and a tendency to behave as others are observed to
behave should have had great survival value; others presum ably
behave as they do for good reason, and by im itating them an indi
vidual can expediently acquire behavior useful for the same reason.
M any species show innate im itative behavior, although its exis
tence in man is still debated. In any case, there are contingencies
of reinforcem ent, rather like those of survival, w hich induce people
to behave as others are behaving. By im itating those whose behavior
has already been shaped by prevailing contingencies, people acquire
appropriate behavior w ith ou t being directly exposed to the con
tingencies themselves. T h e customs and manners of a group seem
to be m aintained by such a process. W ith the help of im itation,
individuals need not construct for themselves the lon g sequences
w hich b rin g their behavior under the control of fairly remote
consequences. T h e y acquire m uch greater repertoires than w ou ld
be possible in a nonsocial environm ent.
T h ere are other arrangements of reinforcers w hich seem to
bring the future more actively into play. G overnm ental practices
supply good examples. T h e reinforcers used to “ keep the peace” are
almost exclusively aversive or punitive; for exam ple, citizens are
fined, flogged, or im prisoned w hen they behave illegally. T h e rein- ^
forcers used to induce citizens to defend a governm ent against its
enemies are also largely aversive; defectors and deserters are im- ’j
prisoned or shot. A system of conditioned positive reinforcers is also |
used, ran gin g from medals to memorials. T h e behavior strengthened J
has consequences w hich reinforce the governm ent for m aintaining
these conditions, but citizens m ay gain indirectly (if less immedi-
ately) from the order and security w hich result. T h e ir behavior is
due prim arily to contrived governm ental contingencies, but it has

consequences'in the possibly distant future w hich w ould be rein
forcing if they occurred sooner. T h e governm ental practice bridges
a temporal gap.
_ R eligious agencies also control their com m unicants w ith con
trived reinforcers both positive and negative. T h e claim ed pow er to
determine extraordinary rewards and punishm ents after death is
used first of all to strengthen the agency, but the com m unicant may
acquire useful practices of self-control, as w ell as the advantages o f
livin g am ong well-behaved people.
Possibly the greatest of all conditioned reinforcers is money.
Worthless in itself, it becomes reinforcing w hen exchanged for
established reinforcers. Industry induces people to work by paying
them. It enjoys relatively im m ediate gains, bu t people in general
may profit fropi the resulting developm ent and production of goods.
Education shows the same pattern. T h e craftsman teaches his ap
prentice because he acquires a useful helper, b u t the apprentice
gains by becomingsa craftsman in his own right. It w ould be difficult
to spot all the reasons w hy parents, peers, employers, religions, and
governments contrive educational contingencies, bu t a distinction
may still be drawn between the advantages gained by those who
teach or pay for teaching and the possibly long-deferred gains of the
learner. Ethical and m oral practices are less conspicuously organized,
but the same pattern prevails. People control each other— govern
ing, teaching, giving incentives— because of im m ediate gains but in
ways which yield possibly long-deferred advantages for all.
T h e consequences w hich lie in the possibly distant future are
often cited to justify practices in governm ent, religion, economics,
education, and ethics. Governm ents may act prim arily to m aintain ~
their power, bu t they seek legitim acy by po in tin g to peace and
security. R eligious agencies appeal to values such as peace of m ind
and compassion. Entrepreneurs justify themselves by po in tin g not
to their profits bu t to the resources they develop and the goods they
make available. A n d when a proposal is made to change a practice,
it is usually supported by p o in tin g to the deferred advantages
rather than the im m ediate gains of those who propose it. N everthe
less, it is quite un likely that the deferred consequences have any
effect as reinforcers. T h e y are, on the contrary, sim ply incidental
by-products.
T h is is not to deny that they serve a different kind o f func-
24 SOCIETY
tion. T h e fact is that cultural practices have evolved in w hich con

tingencies of im m ediate reinforcem ent generate behavior having
rem ote consequences, and this has presum ably happened in part
because the consequences have strengthened the culture, perm itting
it to solve its problem s and hence survive. T h a t the rem ote conse
quences, no m atter how im portant for the culture, are nevertheless |
not havin g any current effect is all too evident when efforts are \
made to take into account a future w hich is not the by-product of
currently reinforced behavior.
W e have, of course, turned to various controlling agencies to
forestall the disasters w hich threaten us. T o reduce pollution, parts
of cities are legally closed to private cars. Special lanes on bridges,
in tunnels, and on highways are reserved for cars w ith a certain
num ber of passengers or for buses. T h e use of energy is taxed. T h e
m anufacture of nondegradable detergents, herbicides, and insecti
cides is prohibited. R eligious and legal sanctions against birth-
control or abortion are eased, and econom ic incentives favoring
large fam ilies are reduced or abolished. C h ildren are taught to
avoid waste, and cam paigns in the mass m edia are designed to have
the same effect on adults. W e are to insist upon returnable bottles
and cloth towels; we are to use recyclable handkerchiefs rather than
tissues.
T h ese measures are obviously taken for the sake o f possibly
long deferred consequences, but it has proved to be difficult to sup
port them w ith im m ediate reinforcers. In fact, in dem ocratic coun
tries few if any institutional sanctions and suasions, designed for
w hatever purposes, are now w orking well. In our own culture, for
exam ple, people do not seem to be as law -abiding as they once were
or as readily disposed to serve in the armed services. T h is does not
mean that they have developed crim inal tendencies or lost their
patriotism; it means that laws are no longer as strictly enforced or
m ilitary service as highly honored. W e impose light punishm ents or
suspend sentences, and in m any states the death penalty has been
abolished. W e no longer shoot deserters, or glorify our heroes. (Only
the returnin g prisoner of w ar is met w ith a brass band playing
“ See, the conquering hero comes!” )
Fewer people now go to church or observe religious practices.
T h is does not mean that they are less devout; it means that, as the
Pope recently pu t it, a veil of silence has been draw n over the Devil.
Few theologians of my acquaintance m ention hellfire, and most of

them speak of heaven w ith a certain embarrassment.
Few people now w ork very hard. T h is does not mean that they
have grown lazy; it means that econom ic incentives are no longer
very effective. In the nineteenth century, it was believed that a
hungry labor force was needed if industry was to prosper, b u t wages
are now exchanged for goods w hich are less acutcly needed than
food. W elfare, like affluence, makes m oney less reinforcing; m edi
care and social security have replaced the threat of the poor farm,"
and even the prospect of a crystal palace is failin g to recruit execu
tive eager beavers.
More and more young people drop out of school and college.
T h is does not mean they have lost their curiosity, their love of
learning, or their desire for an education; it means that educational
contingencies »are no longer very com pelling.
People no longer observe m any of the social graces. T h is does
not mean that; they have become rude or thoughtless; it means that
they are no longer consistently com m ended or punished by their
peers. It has 6ften been rem arked that we no longer com plain. In
fact, the only behavior likely to be punished by one’s peers is the
behavior of com plaint itself.
I have suggested elsewhere that this failure of institutional
and ethical control can be attributed to certain features of the
struggle for freedom. M en and wom en (exhibiting certain standard
features of hum an behavior) escape from dangerous, irritating,
annoying, or uncom fortable things. A m ong such things we must
list the aversive measures used for purposes of control by other
people, organized or unorganized. A person m ay escape from them
by breaking contact— through defection, for exam ple, or apostasy,
truancy, or vagrancy— or by w eakening or destroying their control
ling power— by riots, say, or strikes, boycotts, or revolt. People are
said to govern themselves— electing their own rulers, m aking their
own mystical contact w ith God, sharing in the decisions m ade by the
companies they work for, and so on. It is not surprising that they
should not use strong measures, that they should avoid severe p u n
ishment and the extrem e deprivation needed to m ake a small
reward effective. T h e y may end by destroying all forms of control,
trusting to the hum an genetic endow m ent to survive w ith ou t help.
T h e process can be follow ed in m iniature in education. In
26 SOCIETY
struction was once quite aversive. T h e Egyptians, the Greeks, and

the Rom ans all beat their schoolboys, and m edieval sculpture
portrayed the'schoolm aster w ith the tool of his trade, the birch rod.
C orporal punishm ent is still w ith us. Positive reinforcers in the
form of good marks, grades, prom otion, diplom as, and prizes have
been suggested but only as parts of ineffective contingencies of
reinforcem ent. Rousseau proposed to solve the problem by letting
students study what is naturally reinforcing. T h is is the strategy of
the contem porary “ free school.” As a final step it has been suggested
that schools be abolished, that the whole w orld be converted into
a “ learning environm ent.”
T hose w ho have proposed and effected changes of this kind
have m oved to destroy certain aversive or exploitative features of
the environm ent. As a result people have more often felt free, and
they have also probably enjoyed a greater sense of achievem ent or
worth. B u t we can scarcely overlook the fact that some of the con
trived contingencies under w hich hum an behavior has had im
portant deferred consequences have been destroyed. As a result
people are m ore susceptible to im m ediate consequences. It w ould be
un fair to take the hippy culture as typical of A m erican life today,
b u t it served to point up certain features. Y o u n g people turned in
large num bers to the im m ediate gratifications of drugs and sex, to
forms of art and music w hich can be enjoyed w ith ou t preparation,
and to idleness as an escape from social and econom ic responsibili
ties. In d oin g so they had the support, often m erely im plicit, of
several current philosophies-ôf existentialism w ith its rejection of
both past and future, of phenom enology w ith its concentration on
the experience of the mom ent, and of the structuralism of anthro
pology and linguistics w ith its neglect of causal factors lyin g outside
the topography of behavior itself. T h e y turned to m ystical practices
of the East, surrendering themselves, as one exponent has p u t it, “ to
the awareness of the present m om ent w hile forgettin g the past and
ignoring the future.” H um anistic psychology added its support by
em phasizing self-actualization, the m eaning of w hich is perhaps
clearer in French where actuel means current or contem porary.
T hose w ho are alarm ed by this excessive concern for im m edi
a te gratification are likely to argue that we should restore strong
measures. A ll Com m unist countries and police states have taken this
step, and stronger sanctions are being proposed in Am erica. W e are
urged to make penalties more severe, restoring the death penalty

for certain crimes including the sale of hard drugs, and enforce the
law. T h e m oney people receive should be made more sensitively
contingent on what they do. W elfare paym ents should depend on
useful work. B ut this is not necessarily a way to m ake the future
more effective. Stronger measures are also likely to be proposed for
the sake of im m ediate consequences— for others. Pow erful control
lers are also com mitted to quick effects, and the remoter gains
which sometimes occur as a kind of by-product are by no means
guaranteed. T h e pendulum has swung from despotism through
democracy to anarchy and back again m any times, w ith little or
no change in the future prospects of m ankind. A stable equilibriu m
between control and countercontrol may occur from tim e to time,
but equilibrium w ill not suffice.
W e see th e lim itation of control and countercontrol in the
incentive conditions in most industries. A hundred years ago the
editor of the Scientific Am erican wrote this: “ W e must fairly and
honestly exam ine the conditions of the laboring classes, upon whom
the whole structure of the social organism rests. T h e questions
raised by them and in their behalf can never be adjusted by the
two extremes— those anxious to secure the greatest possible am ount
of pay for the least possible work, and those anxious to obtain the
greatest possible am ount of w ork for the least possible pay.” T h e
opposing contingencies in industry are fairly obvious: em ployers
control their employees w ith reinforcem ent, mostly monetary; the
employees control employers w ith such measures as slowdowns,
strikes, or boycotts. In w hat is called bargaining, conditions are
worked out w hich are acceptable to both parties. T h e trouble is,
they are not good conditions. T h e y do not induce m any people to
work hard or carefully or enjoy w hat they are doing. N o r do they
take into account the consequences for society as a whole, such as
the usefulness of the product, the overall level of em ploym ent, or
the developm ent and conservation of resources.
T h e same lim itations of control and countercontrol are seen
in the other fields we have been exam ining. In governm ent, for
example, a system of checks and balances m ay make for a kin d of
stability but not for the most productive order; and between one
government and another w hat is significantly called a balance of
power yields at best the uneasy equ ilibriu m we call peace.
28 SOCIETY
W e cannot continue to leave the future to the occasionally

beneficial side effects of a strong concern for the present. Som ething
more exp licit must be done. B u t who w ill do it and why? W ho are
to plan for the future and under w hat conditions are they likely to
do so? O ne possibility is that people w ill be m ore concerned for the
future (for whatever reason) if they are less concerned for the
present. Leaders in governm ent, religion, and industry have some
times thought about the future and acted w ith respect to it but
they have usually done so only when their present problems appear
to be solved. It is the successful governm ent or governor who can
afford to become benevolent. Very rich men have stopped using
money to make more m oney and, especially w hen about to die, have
set up foundations w hich are relatively free of present concerns and
can act w ith respect to the future.
O ther candidates for the custodianship of the future are to
be found am ong those who have little or no pow er and hence little
or n othing to gain from the present. In the nineteenth century, the
press emerged as a “ fourth estate.” T h e governm ent, the church,
and the merchants represented the powers exerted through the
police and m ilitary, the m ediation of supernatural sanctions, and
money, respectively. T h e press controlled no com parable rein
forcers; it was lim ited to uncovering and reportin g facts and ex
horting to action, and it enjoyed little or noth in g by way of
im m ediate gain. T h e press was, therefore, concerned w ith the
future, and it criticized the other three estates accordingly. From
time to tim e it has been im portant enough to be suppressed by the
other estates. A press w hich becomes the instrum ent of governm ent,
religion, or economic system can, however, no longer play this role.
Teachers satisfy the same specifications: T h e y have little
power, and teaching has few im m ediate consequences. T h e crafts
man who teaches an apprentice quickly gains the advantage of
having a useful helper, bu t teachers in schools, colleges, and uni- j
versities are affected, if at all, by long-deferred results. Education |
is, indeed, prim arily a preparation for the future; it gives the stu- I
dent current reasons for learning to behave in ways w hich w ill be I
useful later. L ike the press, education serves this function only to
the extent that it is not controlled by the current interests of a i
governm ent, religion, or economic system.
W e should expect that those most lik ely to take the future i
into account w ill have two other qualifications which lie in the
field of science. W hatever the reasons m ay be, people are more likely
to act if th ey have a clear picture of the future. It does not take a
scientist to be aware of changes in population, pollution, dw ind ling
supplies of energy, and so on or to m ake rou gh extrapolations to the
future, but science can do all this more effectively. It can collect
data far beyond the range of personal experience, and it can p ro
ject trends. T h e projections of the C lu b of R om e reported in L im its
to Growth are an exam ple.
Scientists should also be best able to say what can be done.
T h e physical and biological sciences are needed if we are to redesign
our cities to avoid the effects of crowding, to develop new forms of
transportation, and to discover new sources of energy and new
methods of contraception. U nfortunately physical and biological
technology alone cannot guarantee that its solutions w ill be put into
effect. T o solve the m ajor problem we need an effective technology
of behavior. W e need, in short, a new field of specialization— the
design of cultpral^ practices.
Frazier, the protagonist of W alden Tw o, is a kind of arche
type. H e has all the qualifications of the designer of the future. H e
wields none of the pow er to be found in a police force, in the
mediation of supernatural sanctions, or in money. H e has no
personal power; to m ake that clear I gave him what m ight be
called negative charisma. Since his place in the history of W alden
T w o has been deliberately concealed, he gains nothing by w ay of
acclaim as a founder. H e enjoys no special share of the proceeds
of the comm unity. H e is, in short, the com plete nonhero. In him
the present has been almost totally suppressed; the future and its
surrogates have taken com plete control.
T h e specifications of that future were listed in Beyond Free
dom and Dignity. Frazier has tried to construct a world in w hich
“ people live together w ithout quarreling, m aintain themselves by
producing the food, shelter, and clothing they need, enjoy them
selves and contribute to the enjoym ent of others in art, music,
literature, and games, consume only a reasonable part of the
resources of the w orld and add as little as possible to its pollution,
bear no more children than can be decently raised, continue to
explore the w orld around them and discover better ways of dealing
with it, and come to know themselves accurately and, therefore,
3» SOCIETY
manage themselves effectively.” He has done this by constructing a

social environm ent rich in im m ediate reinforcers, so selected that
they strengthen the kinds of behavior w hich m ake a future possible.
A n d the reinforcers are positive. T h a t is w hy the citizens of
W alden T w o feel free. Frazier himself, as the designer of. a culture, is
also under the control of positive consequences, no m atter how
remote. H e has responded to the appeal of U topian rather than
Cassandran predictions— an im portant point. T h rea ten in g predic
tions sometimes spur action (it is perhaps just another sign o f the
weakness of the future that we so often respond only to threats),
but they also induce people to escapc sim ply by tu rnin g to other
things. It is possible that we shall act more consistently w ith respect
to the future when we see the possibility of b u ild in g a better world
rather than merely fending off disaster.
B ut som ething more is needed. W h y should anyone design a
better way of life? T h e answer has been w aitin g for us in the
U topian literature. A n intentional com m m unity emphasizes the
issue of survival. T h e overriding question is: W ill it work? It is not
so obvious that the same question must be asked of every culture.
It is asked, at least im plicitly, by all those who are trying to solve
the problems w hich face our culture today, and it must eventually
be asked about m ankind as a whole. O verpopu lation, pollution, the
exhaustion of resources, nuclear war— these are threats to the sur
vival of the hum an race. W ill the world that m ankind has made
for itself work?
A n d so we come at last to m y title: A re we free to have a
future? P ut commas around “ free” and the question is this: W e who
call ourselves free, are we to have a future? W e call this the free
world and Am erica the land of the free. W e insist that the wars we
fight are fought in the defense of freedom. W e value practices in
governm ent, religion, economics, education, and psychotherapy to
the extent that they prom ote feelings of freedom . T h e question is
w hether practices chosen for that reason have survival value. Are
they to make a m ajor contribution to the future, or w ill practices |
chosen for different reasons by different people— say, the Chinese—
displace them? T h e question once suggested a k in d o f social Dar-
• • •
winism, but cultural practices are no longer confined to any one |
territory, nation, race, religion, or econom ic system. W h at is evolv
Are We Free to Have a Future? 3»
ing is a social environm ent, in w hich the genetic endowm ent of the
hum an species w ill be m axim ally effective.
T h is is-, a test o f freedom in the sense of a test of cultural
practices selected because they m ake people feel free. W e escape
from or destroy aversive control w hen we can do so; that is the
point of the struggle for freedom. W hen we act because we have
been positively reinforced, we feel free and do not try to escape
or countercontrol. T h e m istake is to believe that we are then
actually free. T h is is not a philosophical or theological quibble. O n
the contrary, it is a point of the greatest practical importance.
L et us com pare the lives of you n g people in C h in a and the
U nited States today. W e say that youn g Am ericans are sexually free,
while the Chinese, if we can trust the accounts, observe a strict
moral code. W e say that youn g Am ericans choose their work— or
even not to wbrk at all— w hile the Chinese are assigned to jobs and
work long hours. Y o u n g Am ericans have access to a great variety of
books, movies, theaters, and sports, b u t in C h in a almost all of these
are selected ]}y the governm ent. W e say that young Am ericans
choose where they are to live, while the Chinese have space assigned
to them. Young Am ericans wear w hat they please; the Chinese wear
standard uniforms. It is easy to exaggerate these differences. T h e
Chinese no doubt have some choice, and not all Am ericans are free
to choose their work or where they live. But, even so, the Am ericans
seem to have m uch more freedom. Clearly, they have m any m ore
opportunities; they can do a great m any more kinds of things. B u t
are they really free to choose am ong them? W hy, in fact, do they
wear particular kinds of clothing, live in particular places, go to
particular movies, w ork at particular jobs, or observe a particular
sexual standard? C ertain ly the answer is not as easy as, "because the
government tells them to do so,” bu t that does not m ean that there
is no answer. It is m uch harder to dem onstrate the control exerted
by family, friends, education, religion, work, and so on, b u t it
would be foolish to neglect it.
T h e feeling of freedom is another matter. It depends on the
kinds of consequences responsible for behavior. W hether either
Americans or Chinese feel free depends upon w hy they behave as
they do. If young Chinese are conform ing to their w ay of life b e
cause they w ill be denounced by their fellows and severely punished
3* SOCIETY
if they do not, we may be sure that they do not feel free. In that
case they are doing w hat they have to do. -But if M ao Tse-tung
created a social environm ent rich in positive reinforcers, then they
may be doing what they want to do, and it is quite possible that
they feel freer than Am ericans. Moreover, it is possible that the
reinforcers affecting their behavior have been chosen precisely be
cause of their bearing on the future of the Chinese w ay of life.
Rem ove the commas, and my title is more to the point: A re we
sufficiently free of the present to have a future? O ur extraordinary
com m itm ent to im m ediate gratification has served the species
well. T h e powerful reinforcing effects of drugs like alcohol and
heroin are no doubt accidents, bu t our susceptibilities to reinforce
ment by food, sexual contact, and signs of aggressive damage have
had great survival value. W ith o u t them the species w ould probably
not be here today, but under current conditions they are almost as
nonfunctional as drugs, leading not to survival bu t to obesity and
waste, to overpopulation, and to war, respectively.
N o m atter how free we feel, we are never free of our genetic
endowm ent or of the changes which occur in us d u rin g our lifetim e.
B ut if other aspects of hum an nature, aspects we sum up in the
word intelligence, come into play, we may design a w orld in w hich
our susceptibilities to reinforcem ent w ill be less troublesom e and
in w hich we shall be more likely to behave in ways w hich promise
a future. T h e task can scarcely be overestimated. Happiness is a
dangerous value, and the pursuit of happiness has clearly been too
successful. Like other affluent nations, we must, to coin a horrid
word, “ deaffluentize.” People have done so in the past when pesti
lence and fam ine have deprived them of n atural reinforcers, and
when revolutions in governm ent and religion have changed their
social environments, bu t the power of im m ediate reinforcem ent
continues to reassert itself and w ith ever more threatening conse
quences. T h is could happen once too often. It is possible that the
hum an species w ill be “ consumed by that w hich it was nourished
by.” W e have it in our power to avoid such an ironic fate. T h e
question is whether our culture w ill induce us to do so.
3
The Ethics
of Helping People
(
\
t
W e sometimes act for the good of others. W e feed the hungry,

clothe the naked, and heal the sick. W e say that we care for them,
provide for their needs, do good to them, help them. O u r behavior
often has unforeseen consequences w hich need to be taken into
account.
W e presumably help people in part for reasons that concern
the survival of the species. M aternal behavior is a kind of help
which is either part of an organism ’s genetic equipm ent or w hich
is quickly acquired because of a genetic susceptibility to reinforce
ment; it is obviously im portant for survival. T h e hum an species is,
presumably, more likely to survive if people generally help each
other or are naturally reinforced by signs that they have done so.
Something of the sort m ay contribute to the behavioral disposition
which is part of w hat we call love or compassion.
It is more obvious that we learn to help or do good and that
we learn because of the consequences w hich follow. W e sometimes
help because we find the helplessness of others aversive. W e help
those who help us in return, and we stop doing so when they stop
— when, as we say, they are ungrateful. W e often fail to help those
34 SOCIETY
who are too weak to reciprocate, or to protest effectively when we

fail to help. T h e very young, the aged, the . infirm, the retarded,
and the psychotic are classic examples of people who often have
been not only not helped but positively mistreated.
W e may also help others because in doing so we further the
survival of the group to w hich we belong. A social environm ent
(a “ culture” ) may induce us to give help even though we gain
n othing directly from the advantage for the group. T h u s, we may
be a G ood Sam aritan at some personal sacrifice, and the group
supplies overriding reasons for doing so w ith practices w hich have
been selected sim ply because they have contributed to its survival.
T h e group plays such a role when it steps in to guarantee adequate
care for the very young, the aged, the infirm, the retarded, and
the psychotic. T h ere are few, if any, behavioral processes w hich
provide for such care in the absence of a disposing social environ
ment, w ith the possible exception of such genetic considerations as
the care of the very young.
T h e sanctions arranged by a group are often treated in a dif
ferent way. T h e y are “ justified” as defending ind ivid u al rights, as
guaranteeing that people shall get what they deserve or what is fair
or just. It was perhaps easiest to justify helping those who were
most in need of help, but, in m any cultures, people are now said
to have the right not only to life, liberty, and the pursuit of happi
ness, but to a share in the common wealth. “ T o each according
to his [or her] need” was St. A ugustine’s program before it was
K arl M a rx ’s, and it is still a program, rather than an achievement.
B ut it suggests the extent to w hich groups are now engaged in the
business of m aking sure that their members help each other. T h e
program is not w ithout problem s of an ethical nature. In solving
them, all the consequences of an act of help must be considered.
T h e follow ing discussion with certain possibly relevant be-
.lo ia i processes.
T o begin w ith a very simple example, we may not really help

others by doing things for them. T h is is often the case when they
“are learning to do things for themselves. W e w atch a child tying
a shoelace, grow jittery, and to escape from our jitteriness, we
“ help” the child tie the lace. In doing so, we destroy one chance
The Ethics of Helping People 35
to learn to tie shoelaces. Com enius made the point nearly 400
years ago when he said that “ the more the teacher teaches, the
less the student learns." T h e m etaphor of “ com m unication,” or
the transmission and receipt of inform ation, is defective at just this
point. W e ask students to read a text and assume that they then
know what they have read. Effective com m unication, however,
must provide for the so-called acquisition of knowledge, m eaning,
or inform ation. A traditional m ethod has been to repeat w hat is
said, as in a verbose text. However, new methods in w hich textual
help is progressively w ithdraw n have emerged in the field of pro
grammed instruction. T h e aim is to give as little help as possible
when readers are saying things for themselves.
By giving too m uch help, we postpone the acquisition of effec
tive behavior and perpetuate the need for help. T h e effect is crucial
in the very profession of helping— in counseling and psychotherapy.
Therapists, like teachers, must plan their w ithdraw al from the lives
of their clients. One has most effectively helped others when one
can stop helping them altogether.
M ore serious un an ticipated effects of the good we do to others

often arise because “goods” function as “ reinforcers.” It has long
been known that behavior is affected by certain kinds of conse
quences.! T h a t is w hy rewards and punishm ents are such well-
established social measures. T h e U tilitarians proposed to quantify
consequences in terms of pleasure and pain, for social purposes.
For example, the pleasure enjoyed as the rew ard of unethical or
illegal behavior was to be offset by a corresponding am ount of pain
administered as punishm ent. Both rewards and punishm ents were
regarded as com pensation; and w hen they were fairly balanced,
the ethical account was closed.
T h e form ulation neglected certain contingent relations be
tween behavior and its consequences which were recognized by the
Am erican psychologist, Edw ard L. T h orn d ik e, in his Law of
Effect. By “ effects” he also meant feelings, b u t they were more
than compensation; they strengthened the connection betw een be
havior and the situation in which it occurred. T h e strengthening
effect of reinforcem ent has been an im portant consideration in the
experim ental analysis of operant behavior. E xtrem ely com plex en-
36 SOCIETY
vironm ents are constructed in w hich reinforcing consequences are

contingent upon both behavior and the setting- in w hich it occurs;
and the effect upon the prob ab ility that a given instance of be
havior w ill occur upon a given occasion is analyzed.
T h e fact that strength in the sense of probability of occurrence
is an im portant property of behavior has come to be understood
only very slowly. W ith respect to the present issue, an im portant
point is that strength is not related in any simple way to quantity
of reinforcers and, therefore, not in any simple way to the help we
give or the good we do to others, as these are traditionally eval
uated. W e need to consider the possibility that strength o f behavior
is more im portant than the receipt or possession of goods.
T h ose who are in a position to help others by givin g them

things can use the things as contingent reinforcers. T h is is, of
course, the poin t of behavior m odification. T h e righ t to change
the behavior of others in this w ay has been challenged on ethical
grounds, as we shall see, and C arl Rogers has suggested that the
help given by the therapist (and one could also say teacher or
friend) should be made carefully noncontingent on the behavior
of the recipient. U nfortunately, reinforcers are always tem porally
contingent on some behavior, and they are effective, even though
there is no causal connection. A dventitious reinforcem ents bu ild
superstitions. For example, w hatever people are d oin g just before
rain falls at the end of a drought, they are more likely to do again
in another drought. A n d since the more conspicuous their behavior,
the more effective the adventitious contingencies, a ritu al such as a
rain dance may emerge, and in turn a m yth to exp lain it— for
exam ple, as the propitiation of a giver of rain. T h e grace of G od i
was defined by St. Paul as noncontingent upon works— “ for if by ;
works, then grace is no longer grace,” and Rogers is proposing |
essentially that therapeutic help should have this divine quality.
B u t there are behavioral processes w hich cannot be denied, and |
offerings and sacrifices to the G iver of H elp are an im portant prob- I
lem for the therapist.
U nanticipated consequences w hich follow when we are said to J

give people help can be m uch m ore serious. In an environm ent in |
which such things as food, shelter, and safety are guaranteed as

rights, these things are less likely to serve as reinforcers. T h e re
cipients of boiantiful help are rather in the position of those who
live in a benign clim ate or possess great wealth. T h ey are not
strongly deprived or aversively stim ulated and, hence, not subject
to certain kinds of reinforcem ent. Some im portant forms of be
havior are never acquired or, if they have been acquired, are no
longer exhibited. B u t such people do not sim ply do nothing; in
stead, they come under the control of lesser reinforcers. N o objec
tion is likely to be raised to the classic exam ples found in art,
music, literature, and scientific exploration. Individuals are en
couraged to devote themselves to these fields through the kind of
help called patronage or grants-in-aid. B ut these reinforcing con
sequences are, unfortunately, seldom as im m ediate or as personally
effective as others, w hich have long given the leisure classes a
special character. Sweets rem ain reinforcing to the nonhungry;
alcohol and dsugs have anom alous reinforcing effects; sexual re
inforcement survives
\
because we do not leave satiation to others;
certain special schedules of reinforcem ent (such as those basic to
all gam bling devices) m ake w eak reinforcers effective; and just the
spectacle of other people livin g seriously or dangerously is often
reinforcing, as in films or television.
These are the reinforcers, rather than those of art, music, lit
erature, and science, w hich are more Hkely to be given free play
by any help which preempts the serious business of life, and there
is little to be said for them. Some are stultifying, and none leads
to the fu ll developm ent of the hum an genetic potential. O n e’s
behavior may be reinforced for a lifetim e in these ways and yet
undergo almost no im portant change, and when these alternative
reinforcers lose their pow er or are suppressed by societal rules,
behavior falls to a very low ebb. W e call the child who has b e e n 1
given excessive help “ spoiled,” and the term applies as well to the
adult. iW, s '
Organisms are at least as strongly disposed to take goods away

from others as to supply them in the form of help, particularly
when unmerited, and the disposition may serve as a natural cor
rective to excessive help. (W e are inclined to speak of the feeling
of compassion that accom panies helping others and the feeling of
38 SOCIETY
resentment that accompanies taking goods away from those who

have not worked for them, bu t it is the tendencies.to act w hich are
involved here.) Aggressive behavior offsets or corrects compas
sionate help and may have survival value, for either species or
group, if it leads to a more equitable distribution of goods, but
the question is not who should have how much of what but, rather,
how they are to get what they have.
T h e plight of those whose behavior is not often reinforced—
because others do things for them, or because they have not learned
to do things for themselves, or because they are given the things
their behavior would otherwise be reinforced by— is fam iliar
enough. T rad ition ally, their behavior is attributed to feelings and
states of m ind. Such people are said to lack initiative, to show little
strength of character, to have w eak wills, to lack spiritual strength,
or to have egos that are not w ell developed. T h e y are said to suffer
from abulia (lack of will), acedia (spiritual torpor), apathy (lack
of feeling), or boredom. What they are suffering from is a world in
which their behavior is not positively reinforced.
It is easy to dismiss that statem ent as the idée fixe of a be
havioral analyst, but strength of behavior, in the sense of the prob
ability that behavior w ill occur, is a basic aspect of hum an nature.
It is to be attributed to external contingencies of reinforcem ent,
rather than internal deficiencies. Hence", it is an aspect about which
som ething can be done. Som ething is being done by those who
understand the im portance of contingencies of reinforcem ent.
A good exam ple of the neglect of relevant aspects of the en

vironm ent is to be found in analyses of incentive conditions in
modern industry.1 T h e “ degradation o f labor” is said to have be
gun with the systematic destruction of craft skills. W orkers move
from craft to industrial conditions for m any reasons. W ork is
usually easier and, because a task is divided am ong m any workers,
each share is simpler and can be learned durin g a briefer appren
ticeship. W orkers produce m ore in less tim e and can be paid more.
1 Heilbroner, R. Review of Braverman, Labor and Monopoly Capi

tal: The Degradation of Work in the Twentieth Century. New York
Review of Books, Jan. 23, 1975.
Yet som ething has been lost. M any interpretations have appealed
to feelings and states of m ind: T h e w orker has come to think of
him self as a cog'in a machine; he is no longer the possessor o f the
“ accum ulated knowledge o f the m aterials and processes by w hich
production is accom plished” ; w ork has been reduced to “ a series
of bodily movements entirely devoid of m eaning” ; the w orker is
separated (“ alienated” ) from the product of his labor; and so on.
B ut w hy is this degrading? It is true that work on a production
line is probably faster than the w ork of a craftsman w ithou t a
deadline. Because it has been reduced in scope, it is also neces
sarily m ore repetitious and, hence, likely to yield the “ fatigue of
repeatedly doing the same thin g” (not to be confused w ith physical
exhaustion). Yet the gam bler “ works” fast and repetitiously and
calls his life exciting; and the craftsm an uses machines to save
labor when he can and often works w ith a time-and-motion effi
ciency that an industrial engineer w ould give m uch to duplicate.
T h e im portant difference lies in the contingencies of rein
forcement. It is often supposed that industrial workers work to get a
reward, rather than avoid punishm ent. B u t as M arx and others
have noted, they work because to do anything else w ould be to
lose a standard of livin g m aintained by their wages. T h e y work
under the eye of a supervisor upon whose report their continued
em ploym ent depends. T h e y differ from slaves only in the nature
of the “ punishm ent” they receive for not w orking. T h e y are sub
ject to negative reinforcem ent, a condition obscured by the u n
critical use of the term reward.
T h e craftsm an’s behavior, in contrast, is reinforced at every
stage by those conditioned reinforcers called signs of progress. A
particular task may take a day, a week, a m onth, or a year, but
almost every act produces som ething w hich w ill form part of the
whole and is, therefore, positively reinforcing. It is this condition of
“ nondegrading” work w hich has been destroyed by industrialization,
and some of those concerned with incentive conditions have used
the principles of behavior m odification to restore it.
A sim ilar correction needs to be m ade to offset the unwanted

by-products of help ing others by supplying goods. U nfortunately,
it is difficult to see this and to act accordingly just because our
40 SOCIETY
behavior in helping others is determ ined to such a large extent by

reciprocal reinforcem ent. G iven a choice between receiving some
thing gratis and the opportunity to work to get it, those whom we
help are likely to choose the former, and they w ill therefore more
abundantly reinforce our behavior when we give them things rather
than the opportunity to work for things. It is in the long run that
the advantage of getting, rather than possessing, makes itself felt,
both by them and by us, and w hat happens in the long run does
not often have m uch of an effect. W hat a person is said to deserve
as a right is subject to a sim ilar bias.
It is just at this point that behavior m odification plays a unique
role. T h e term needs careful definition. Behavior has been modified
ever since it was m odifiable— w hich is to say, from the beginning.
B ehavior is modified by the threat of the bully or of the nation
w ith a nuclear stockpile, by incentive tax allowances, by advertising,
by religious rituals, by state lotteries and other gam blin g enter
prises, and, recently, by certain physiological measures and explicit
Pavlovian conditioning. T h e term was introduced, however, to
refervto certain applications of the experim ental analysis of be
havior, particularly through the arrangem ent of contingencies of
positive reinforcem ent. Behavior m odification in that sense helps
people by arranging conditions under w hich they get things rather
than by giving them things. T h a t is its essential feature. A n d for
that very reason, it was inevitable that there w ould be some conflict
w ith traditional views of helping others— especially w ith principles
of what was just or fair or to be defended as the rights of the
individual.
T h e issue first arose w hen behavior m odification was used in

institutional care. In m any cultures, food, shelter, clothing, secu
rity, and possibly privacy have been made available to those who
for any reason cannot otherwise obtain them. Hom es for the very
young, the aged, the infirm, and the retarded, hospitals for psy-
chotics, and prisons are far from a benign world, b u t those w ho
live in them, characteristically, have little reason to w ork for the
basic reinforcers because the reinforcers have been guaranteed as
rights. Most of the alternatives, such as gam bling, sex, alcohol,
and .drugs, are not available (except surreptitiously in prisons). As
a result, such people suffer all the ills of havin g n othin g to do.
The Ethics of Helping People 4i
Troublem aking m ay be unintentionally reinforced, and if possible

they escape, but otherwise we say that their behavior tends to be
marked by boredom, abulia, acedia, and apathy.
B ehavior m odification, properly defined as ‘‘the applied analysis
of behavior,” is precisely what is needed to correct this shortcom ing
of institutional life because it is concerned with establishing effec
tive contingencies of reinforcement. A ctual practices need not be
described here, b u t the behavior m odifier usually begins w ith a
search for available reinforcers and then arranges especially clear-
cut contingencies— as w ith the use of tokens. Contingencies can be
programmed to shape com plex topographies and to bring behavior
under the control o f com plex stim uli. For those w ho w ill eventually
leave the institution, such a program is called educational, thera
peutic, or rehabilitative. For those who must rem ain, the goal is
simply a “ prosthetic” environm ent— an environm ent in w hich peo
ple behave in reasonably effective ways in spite of deficiencies, in
which they talê an active interest in life and begin to do for them
selves w hat the institution previously did for them.
W h eth er'w e are concerned w ith education, therapy, and re
habilitation, or w ith the construction of a prosthetic environm ent,
we need those reinforcers w hich have acquired special pow er in
the evolution of the species. Y et they are the very things supplied
in the act of helping or caring for people— the things guaranteed
as rights. In order to make them contingent on behavior in an in
stitutional setting, we must w ithhold them u n til the behavior oc
curs. T h e ind ivid ual must, therefore, be deprived to some extent
and, consequently, w ill appear to rem ain unhelped or to be denied
certain rights. W e cannot avoid this conflict so long as we continue
to view help as providin g goods rather than as arranging contin
gencies of reinforcem ent.
T h e conflict first came into the open in an attack u pon oper
ant reinforcem ent programs in m ental hospitals. O ne set of pro
posed regulations contained the follow ing:
“D eprivation is never to be used. N o patient is to be de

prived of expected goods and services and ordinary
rights, including the free m ovem ent o f his limbs, that he
had before the program started. In addition deficit re
w arding must be avoided; that is, rewards m ust not con
sist of the restoration of objects or privileges that were
taken away from the patient or that he should have had
42 SOCIETY
to begin with. T h e ban against deficit rew arding includes

the use of tokens to gain or regain such objects or p rivi
leges.” 2 ■ . ..
T h e authors insist that they are concerned w ith the legitim acy
of the rationale for using operant conditioning, bu t it is the rationale
of rights which is at issue. W h y have these things been guaranteed
to the patient? W hat “ should” patients have had to begin with?
T h e mistake is to generalize from those who cannot help themselves
to those who can. For the latter, a m uch more fundam ental right— -
the right to live in a reinforcing environm ent— must be considered.
If the function of an institution is education, therapy, or. rehabilita
tion, all available resources should be used to speed the process, and
the strong reinforcers are u ndoubtedly to be classified as such. For
those who w ill never return to the w orld at large, a strongly rein
forcing environm ent is equally im portant.
U nder proper contingencies, many institutionalized people
can engage in productive work, such as caring for themselves, keep
ing their quarters clean, and w orkin g in laundry, kitchen, or truck
garden. B ut when these things have previously been done by paid
personnel, suspicion falls on the motives of m anagement. Should
residents not be paid the same wages? O ne answer is that they
should unless the contingencies are “ therapeutic,” but that raises
the question of help in only a slightly different form. Residents
are receiving help when their behavior is being reinforced in a
prosthetic environment, though they are not necessarily being
“ cured.” Especially when we consider the economics of institutional
care, can there be any objection to the residents themselves pro
ducing all the goods and services it was once supposed to be
necessary for others to give them?
A t least one state has recognized the issue. A b ill was recently
passed in Iowa w ith the provision that:
“ T h e administrator m ay require of any resident of the

County Care Facility w ith the approval of a physician
reasonable and m oderate labor suited to the resident’s
2 Lucero, R . J., V ail, D . J., and Scherber, J. R e g u la tin g o p e ra n t-

con d ition in g program s. H osp ita l and Com m unity Psychiatry, 1968,
53- 54 -
age and bodily strength. A n y income realized through

the labcir of a resident together w ith the receipts from
operation of the C oun ty Farm if one is m aintained shall
be appropriate for use by the County Care Facility in
such m anner as the B oard o f Supervisors may direct.” 3
T h e constitutionality of the b ill is being questioned.

T h e so-called rehabilitation of the prisoner raises a special
problem. Prisoners usually undergo very little useful change. T h ey,
have been separated from society for the latter’s protection or as
punishment and are unable to help themselves only because they
have been cut off from the usual means. T h e destructive changes
which follow are well known. Some prom ising results have been
obtained from the application of an experim ental analysis of b e
havior— for exam ple, in a project at the N ational T ra in in g School
for Boys in W ashington, D .C .4 U nfortunately, experim ents of this
sort have been confused w ith efforts to change prisoners w ith drugs
or the more violent forms of aversive conditioning, and protests
against the letter— for exam ple, by the Am erican C ivil Liberties
Union— have been extended w ith ou t warrant to efforts to construct
more sustaining prison environments.
L ike everything else, operant conditioning can be misused.
Management may solve some of its problems by arranging contin
gencies which suppress disruptive behavior and under w hich a
child, a prisoner, or a psychotic m ay simply sit quietly and do little
or nothing all*day long. Even so, this may be better than achieving
the same result through punishm ent, but both solutions m ay be
challenged if nothing further is done. M uch m ore can be done
through the applied analysis of behavior when the problem is
understood.
Some of the same issues arise in the w orld at large, where

helping people takes on a m uch broader m eaning. V ery little has
3 Behavioral Voice § 5 (C en ter for H um an D evelopm en t, D rake U n i

versity, Des M oines, Iowa). (T h e b ill is called “ T h e R ed esig n atio n
o f C ou n ty H om es as C o u n ty C are Facilities,” HF659.)
4 C ohen, H . L., and F ilip czak, J. A new learning environm ent. San
Francisco: Jossey-Bass, 1971.
44 SOCIETY
ever been achieved simply by supplying goods and services. G overn

ments do not help their citizens by giving them order and security
— that is the claim only of the police state; they help them by
arranging environm ents in w hich they behave in orderly and m u
tually supportive ways. T h e y do not defend the rights to life,
liberty, and the pursuit of happiness as things w hich their citizens
possess; they m aintain environm ents in w hich people do not
threaten the lives and political freedom of one another. Schools and
colleges do not give their students inform ation, knowledge, or skills;
they are environments in w hich students acquire inform ed and
skillful behavior. T h e “ good life ” is not a w orld in w hich people
have w hat they need; it is one in w hich the things they need figure
as reinforcers in effective contingencies.
A case history w ill show how easily the basic issue is missed.
A fter the Second W orld W ar, Denm ark entered upon a program
of “ modern reform atory guidanc.e” to raise the standard of livin g
of the Eskimos of G reenland.5 Thousands of construction workers
were sent in to bu ild m odern houses and facilities. B u t the local
industry, fishing, could not support these m aterial standards, and
an annual subsidy of many m illions of dollars w ill now be needed
— indefinitely— for the 50,000 inhabitants. T h e goods supplied are
not contingent on productive behavior, and it is not surprising that
a long-established, cooperative culture has broken down. U nder
the surface, there is said to be “ an alarm ing chaos of hum an frus
tration.” A n antagonistic class society is developing. G ood dyadic
social relations have yielded to drunken brawls.
It means little to say that a high standard of livin g was “ an
artificial creation,” that it can be made natural by giving each per
son a more direct influence in governm ent, or that a “ strategy of
wholeness” is needed. T h e trouble is that certain basic contingen
cies of reinforcem ent have been destroyed. A n d it is difficult to see
how they can ever be restored except by greatly increasing the
behavioral repertoires of the Eskimos or by sharply reducing their
so-called standard of living. It w ill not be enough that the teams
of construction workers are now to be follow ed by teams of social
workers. T h e U n ited States is repeating the experim ent on a small .
I
5 Jensen, B. Human reciprocity: An Arctic exemplification; Ameri- |
can Journal of Orthopsychiatry, 1973, 43, 447-458.
scale on the island of B ikin i, and it w ill be interesting to see

whether the result is the same.
Even in the restricted sense of the applied analysis of behavior,

behavior modification has grown w ith astonishing speed and m uch
of that growth has been uncharted and chaotic. Practitioners have
ranged from scientists high ly skilled in the basic analysis to laym en
applying a few cookbook rules. B ut the accomplishments are too
substantial to be dismissed— am ong them, program m ed instruction
and contingency m anagem ent in the classroom, the design of pros
thetic environments for the retardate and psychotic, personal and
family counseling in ethical self-management, educational environ
ments for juvenile delinquents, and new incentive systems in indus
try. In retrospect, m uch of this often seems to be simply a m atter
of common sense, b u t people have had com m on sense for thou
sands of years, and it has not helped them solve the basic problem .
It has been tdo easy to put possession ahead of acquisition, and
to miss the im portance of strength of behavior and its relation to
contingencies of reinforcem ent. In the classroom, hospital, factory,
prison, home, and the w orld at large, the obvious fact is that some
of the good things in life are in short supply. W e are just begin nin g
to see that a mere shortage is not w hat is causing trouble and that
people w ill not necessarily be helped by increasing the supply.
Behavior m odification through the m anagem ent of contingencies
of reinforcem ent is a special way of helping people just because
it is concerned w ith changing the probability that they w ill behave
in given ways.
For just that reason, it is now under attack. A recent exam ple
is the report of the E rvin Com m ittee, Individual R ights and the
Federal R ole in Behavior M odification, based on a three-year in
vestigation of federal support of a variety of programs. A ccording
to Senator Ervin, “ T h e most serious threat posed by the technology
of behavior m odification is the pow er this technology gives one
man to impose his views and values on another. . . . If our society
is to remain free, one m an must not be em powered to change an
other m an’s personality.” 8 B u t individuals have always had the
6 “ In d ivid u al R ig h ts and the F ederal R o le in B eh av io r M od ifica

tion,” N o. 5270-02620.
46 SOCIETY
power to impose their views on others; the relevant behavioral

processes were not recently invented. One o f . the greatest and cer
tainly the most convenient of all reinforcers is money, and we have
recently seen some extraordinary examples of its misuse. W hy does
the E rvin Com m ittee not consider constitutional safeguards against
the power w hich a person can amass by accum ulating money? W e
have m inim um wage laws and other laws restricting some uses of
money, but we have no m axim um wage laws restricting the extent
to which money can be acquired for use. A n d m oney is only one
of the more conspicuous instruments of control. Possibly, the ex
perim ental analysis of behavior w ill play its greatest role in forcing
an exam ination of all the ways in w hich “ one man can change an
other m an’s personality.”
L ike any other means of control— say, physical force— be
havior modification should be supervised and restrained. T h e con
cept o f the rights of the individual is concerned w ith that problem .
Some traditional principles have emphasized freedom from coer
cive or pun itive control, and they are as badly needed today as they
have ever been. Other traditional principles have emphasized the
possession of goods and services, and here a sweeping revision is
needed. N either a capitalist defense of private property nor a
socialist program of state ownership as a means of equitable dis
tribution takes into account the fu ll scope of relevant behavioral
processes.
It has been suggested that Gross N ational Product should be
subordinated to Gross N ational Happiness in evaluating a culture,
but nothing m uch would be gained if happiness were identified as
a static condition of satisfaction derived from the possession of
goods. Indeed, in that case, there w ould scarcely be a distinction.
T h e greatest good of the greatest num ber may be the greatest bore,
and the U tilitarians lost their case just because they neglected the
reinforcing contingencies w hich b u ild the condition we describe by
saying that we are happy.
T h e intense current interest in ethical, moral, legal, and re
ligious matters is no doubt largely a response to w orsening world
conditions. A burgeoning population forces us to take another
look at birth control, abortion, and selective breeding. Increasing
violence, as in bombings, hijackings, and political kidnappings,
forces us to look again at legal sanctions, possibly reversing a hu
mane trend against capital punishm ent. In addition, however, a

surprising num ber of critical issues have to do w ith w hat is called
helping peopie. “ A id ” is a synonym of “ help,” and foreign aid
raises m any ethical, moral, and legal problem s. In the name of aid,
the U nited States has become one of the Zaharoffs o f the last half
of the twentieth century— one of the great m unitions makers who
were once held in utter contem pt. In the name of aid, we rescue
some of the starving peoples of the w orld w hile allow in g others
to die, and refuse to adm it that we are practicing triage. W ith both
m ilitary and nonm ilitary “ help,” we have nearly destroyed In d o
china. A n d so we begin again to ask to w hat extent the rich nations
of the w orld are to help the poor, or, in domestic affairs, how far a
government should go in increasing the help w hich its rich citizens
must give to its poor? 7
B ut it ist a mistake to turn again to certain earlier principles.
For reasons w hich in themselves illustrate a pow erful behavioral
principle, we have grossly overem phasized the im portance of simple
possession. N either happiness nor the survival of the group depends
on the satisfaction derived from havin g things. A n d the most gen
erous help may fail as ignom iniously as the most aggressive despoli
ation. Som ething else is needed to achieve conditions under w hich
human beings w ill show the productivity, the creativity, and the
strength inherent in their genetic endow m ent and w hich are essen
tial to the survival of the species.
7 Rawls, J. A theory of justice. Harvard University Press, 1971.

Nozick, R. Anarchy, state, and utopia. Basic Books, 1974.
4
Humanism and Behaviorism
T h ere seem to be two ways of knowing, or know ing about,

another person. O ne is associated w ith existentialism , phenom e
nology, and structuralism. It is a m atter of know ing w hat a person
is, or w hat he is like, or what he is com ing to be or becom ing. W e
try to know another person in this sense as we know ourselves. W e
share his feelings through sympathy or empathy. T h ro u g h intuition
we discover his attitudes, intentions, and other states of m ind. W e
com m unicate w ith him in the etym ological sense of m aking ideas
and feelings common to both of us. W e do so more effectively if we
have established good interpersonal relations. T h is is a passive, con
tem plative kind of know ing: If we w ant to predict w hat a person
does or is likely to do, we assume that he, like us, w ill behave
according to what he is; his behavior, like ours, w ill be an expres
sion of his feelings, states o f m ind, intentions, attitudes, and so on.
T h e other way of know ing is a m atter of w hat a person does.
W e can usually observe this as directly as any other phenom enon
in the world; no special kin d of know ing is needed. W e explain
why a person behaves as he does by turning to the environm ent
rather than to inner states or activities. T h e environm ent was effec-
•
48
Humanism and Behaviorism 49
tive during the evolution of the species, and we call the result the
human genetic endowm ent. A m em ber of the species is exposed to
another part-pf that environm ent durin g his lifetim e, and from it
he acquires a repertoire of behavior w hich converts an organism
with a genetic endowm ent into a person. By analyzing these effects
of the environment, we move toward the prediction and control
of behavior.
B ut can this form ulation o f w hat a person does neglect any
available inform ation about w hat he is? T h ere are gaps in time
and space between behavior and the environm ental events to which
it is attributed, and it is natural to try to fill them w ith an account
of the intervening state of the organism. W e do this when we sum
marize a long evolutionary history by speaking of genetic endow
ment. Should we not do the same for a personal history? A n
omniscient physiologist should be able to tell us, for exam ple, how a
person is changed when a bit of his behavior is reinforced, and what
he thus becomes should explain w hy he subsequently behaves in a
different way.. W e, argue in such a manner, for exam ple, w ith
respect to im m unization. W e begin w ith the fact that vaccination
makes it less likely that a person w ill contract a disease at a later
date. W e say that he becomes im m une, and we speak of a state of
immunity, which we then proceed to exam ine. A n om niscient
physiologist should be able to do the same for com parable states in
the field of behavior. H e should also be able to change behavior by
changing the organism directly rather than by changing the en
vironment. Is the existentialist, phenom enologist, or structuralist
not directing his attention precisely to such a m ediating state?
A thoroughgoing dualist w ould say no, because for him w hat
a person observes through introspection and what a physiologist
observes w ith his special techniques are in different universes. B u t
it is a reasonable view that what we feel when we have feelings are
states of our own bodies, and that the states of m ind we perceive
through introspection are other varieties of the same kinds of
things. Can we not, therefore, anticipate the appearance of an
omniscient physiologist and explore the gap between environm ent
and behavior by becom ing more keenly aware of what we are?
It is at this point that a behavioristic analysis of self-knowl
edge becomes most im portant and, unfortunately, is most likely to
be misunderstood. Each o f us possesses a small part of the universe
5» SOCIETY
within his own skin. It is not for that reason different from the rest
of the universe, but it is a private possession; W e have ways of
know ing about it that are denied to others. It is a mistake, however,
to conclude that the intim acy we thus enjoy means a special kind
of understanding. W e are, of course, stim ulated directly by our own
bodies. T h e so-called interoceptive nervous system responds to con
ditions im portant in deprivation and emotion. T h e proprioceptive
system is involved in posture and movement, and w ithou t it we
could scarcely behave in a coordinated way. T h ese tw o systems,
together with the exteroceptive nervous system, are essential to
effective behavior. B ut know ing is m ore than responding to stimuli.
A child responds to the colors of things before he “ knows his colors.”
K now ing requires special contingencies of reinforcem ent that must
be arranged by other people, and the contingencies involvin g pri
vate events are never very precise because other people are not
effectively in contact w ith them. In spite of the intim acy of our own
bodies, we know them less accurately than we know the world
around us. A n d there are, of course, other reasons w hy we know
the private world of others even less precisely.
T h e im portant issue, however, is not precision bu t subject
matter. Just what can be know n w hen we “ know ourselves” ? T h e
three nervous systems just m entioned have evolved under practical
contingencies of survival, most of them nonsocial. (Social contin
gencies im portant for survival must have arisen in such fields as
sexual and m aternal behavior.) T h e y were presum ably the only
systems available when people began to “ know themselves” as the
result of answering questions about their behavior. In answering
such questions as “ Do you see that?” or “ D id you hear that?” or
“ W hat is that?” a person learns to observe his own responses to
stim uli. In answering such questions as “Are you hungry?” or “Are
you afraid?” he learns to observe states of his body related to de
privation and em otional arousal. In answering such questions as J
“ A re you going to go?” or “ D o you intend to go?” or “ D o you feel •.
like going?” or “ A re you inclined to go?” he learns to observe the
strength or probability of his behavior. T h e verbal com m unity asks
such questions because the answers are im portant to it, and in a
sense it thus makes the answers im portant to the person himself.
T h e im portant fact is that such contingencies, social or nonsocial, |
involve nothing more than stim uli or responses; they do not involve |
fH
Humanism and Behaviorism 5i
mediating processes. W e cannot fill the gap between behavior and

the environm ent of w hich it is a function through introspection
because, to p u t'th e m atter in crude physiological terms, we do not
have nerves going to the right places. W e cannot observe the states
and events to which an om niscient physiologist w ould have access.
W hat we feel when we have feelings and w hat we observe through
introspection are nothing m ore than a rather m iscellaneous set of
collateral products or by-products of the environm ental conditions
to which behavior is related. (We do not act because we feel like
acting, for exam ple; we act and feel like acting for a com m on reason
to be sought in our environm ental history.) D o I mean to say that
Plato never discovered the mind? O r that A quinas, Descartes, Locke,
and Kant were preoccupied w ith incidental, often irrelevant by
products of hum an behavior? O r that the m ental laws of physiologi
cal psychologists« like W undt, or the stream of consciousness of
W illiam James, or the mental apparatus of Sigm und Freud have no
useful place in the understanding of hum an behavior? Yes, I do.
A nd I put the m atter strongly because, if we are to solve the prob-
* I
lems that face us in the w orld today, this concern for m ental life
must no longer divert our attention from the environm ental con
ditions of w hich hum an behavior is a function.
B ut why have we attached so m uch im portance to our feelings
and states of mind, to the neglect of the environm ent? T h e answer
seems to lie in the im m ediacy and the saliency of the stim uli. M any
relevant events in our personal history pass w ithou t notice. For one
thing, the behavior to w hich they w ill eventually prove relevant
has not yet occurred and cannot contribute to contingencies that
would lead us to notice them. A n d if we have noticed them, we may
quickly forget. B ut our feelings, “ ideas,” “ felt intentions,” and so
on, often overlap the behavior to w hich they seem related, and they
usually occur in just the place that w ou ld be occupied by a cause
(on the principle of post hoc, ergo propter hoc). For exam ple, we
often feel a state of deprivation or em otion before we act in an
appropriate way. If we say som ething to ourselves before saying it
aloud, what we say aloud seems to be the expression of an inner
thought. A n d if we say som ething aloud w ithout first saying it to
ourselves, it is tem pting to suppose that we m ust be expressing a
nonverbal thought.
T h is apparent causality lodged w ithin the private world
52 SOCIETY
w ithin a skin, together with the organization imposed upon it by

the fact that all its determ ining conditions have occurred in the
history of one person, generates a “ sense of self.” W e feel there is
an “ I ” who knows what he is going to do and does it. Each of us is
aware or conscious of at least one such self, w hich we learn to
m anage more or less effectively.
Since the only selves we know are hum an selves, it is often
said that man is distinguished from other species precisely because
he is aware of him self and participates in the determ ination of his
future. W hat distinguishes the hum an species, however, is the
developm ent of a culture, a social environm ent that contains the
contingencies generating self-knowledge and self-control. It is this
environm ent that has been so lon g neglected by those who have
been concerned w ith the inner determ ination of conduct. T h e
neglect has meant that better practices for b u ild in g self-knowledge
and self-management have been missed.
It is often said that a behavioristic analysis “ dehumanizes
m an.” B ut it merely dispenses w ith a harm ful explanatory fiction.
In doing so it moves m uch more directly toward the goals that
fiction was designed, erroneously, to serve. People understand them
selves and manage themselves m uch more effectively w hen they
understand the relevant contingencies.
Im portant processes in self-management lie in the fields of
ethics and morals, where conflicts between im m ediate and deferred
consequences are considered. O ne of the great achievem ents of a
culture has been to bring rem ote consequences to bear upon the
behavior of the individual. W e m ay design a culture in w hich the
same results w ill be achieved m uch more efficiently by shifting our
attention from ethical problem solving or m oral struggle to the
external contingencies.
W e may move from an inner agent to environm ental determi
nants w ithout neglecting the question of values. It has been argued
that behaviorism is or pretends to be value free, b u t that no value-
free science can properly deal w ith man qua man. W h at is wrong
in the traditional argum ent can be seen in the expression “ value
judgm ent.” A n inner in itiatin g agent is to judge things as good
or bad. B ut a m uch more effective source of values is to be found
in the environm ental contingencies. T h e things people call good,
are positive reinforcers, and they reinforce because of the contin
gencies of survival under w hich the species has evolved. U n til

recently, the ipecies could survive fam ine, pestilence, and other
catastrophes only if its members procreated at every opportunity,
and under such contingencies sexual contact became highly rein
forcing. Sex is not reinforcing because it feels good; it is reinforcing
and feels good for a common phylogenic reason. Some reinforcers
may acquire their power d urin g the life of the individual. Social
goods, such as attention or approval, are created and used to in
duce people to behave in ways that are reinforcing to those who
use them. T h e result m ay be good for the ind ivid u al as w ell as
for others, particularly w hen deferred consequences are m ediated.
T h e values affecting those who are in charge of other people
supply good examples of the im portance of turning from supposed
attributes of an inner man to the contingencies affecting behavior.
There are five classical types of hum an beings who have been mis
treated: the young, the elderly, prisoners, psychotics, and retardates.
Are they m istreated because those who are in charge of them lack
sympathy, compassion, or benevolence, or have no conscience? No,
the im portant fact is that they are unable to retaliate. It is easy
to mistreat any one of these five kinds of people w ithou t being-
mistreated in turn. T h e confrontation in 1972 between Hum anists
and Catholics at the LaFarge Center in N ew Y o rk C ity failed to
make clear that the sources of conscience are not to be found in
psychological realities but in p u n itive sanctions.
A n environm ental analysis has a special advantage in prom ot
ing a kind of value concerned w ith the good of the culture. C u l
tures evolve under special contingencies of survival. A practice
that makes a culture more likely to survive survives w ith the cul
ture. Cultures become more successful in m eeting contingencies of
survival as they induce their members to behave in m ore and more
subtle and com plex ways. (Progress is not inevitable, of course, for
there are extinct cultures as w ell as extinct species.) A n im portant
stage is reached when a culture induces some of its members to be
concerned for its survival, because they may then design m ore effec
tive practices.
Over the years, men and wom en have slow ly and erratically con
structed physical and social environm ents in w hich they have come
closer to fulfillin g or actualizing their potential. T h e y have not
changed themselves (that is a genetic problem w hich has not yet
54 SOCIETY
been solved); they have changed the world in w hich they live. In the '
design of his own culture, man could thus be said to control his
destiny.
I w ould define a hum anist as one of those who, because of the
environm ent to w hich he has been exposed, is concerned for the
future of m ankind. A m ovem ent that calls itself “ hum anistic psy
chology” takes a rather different line. It has been described as “ a
third force” to distinguish it from behaviorism and psychoanalysis;
but “ third ” should not be taken to mean advanced, nor should
“ force” suggest power. Since behaviorism and psychoanalysis both
view hum an behavior as a determ ined system, hum anistic psycholo
gists have emphasized a contrast by defending the autonom y of the
individual. T h e y have insisted that a person can transcend his
environment, that he is more than a causal stage between behavior
and environm ent, that he determines what environm ental forces
w ill act upon him — in a word, that he has free choice. T h e position
is most at home in existentialism , phenom enology, and structural
ism, because the emphasis is on what a person is or is becoming.
M aslow’s expression “ self-actualization” sums it u p nicely: T h e
individual is to fulfill him self— not merely through gratification,
of course, but through “ spiritual growth.”
H um anistic psychologists are not unconcerned abou t the good
of others or even the good of a culture or of m ankind, bu t such a
form ulation is basically selfish. Its developm ent can be traced in the
struggle for political, religious, and econom ic freedom , where a
despotic ruler could be overthrown only by convincing the individ
ual that he was the source of the power used to control him . T h e
strategy has had beneficial results, but it has led to an excessive ,
aggrandizem ent of the individual, w hich m ay lead in turn either
to new forms of tyranny or to chaos. T h e supposed right of the
individual to acquire unlim ited w ealth w hich he is free to use as
he pleases often results in a kind of despotism, and the H indu
concern for personal grow th in spirituality has been accompanied |
by an almost total neglect of the social environm ent.
Better forms of governm ent are not to be foun d in better
rulers, better educational practices in better teachers, better eco- |
nomic systems in more enlightened management, or better therapy |
in more compassionate therapists. N either are they to be foun d in
better citizens, students, workers, or patients. T h e age-old mistake :
is to look for salvation in the character of autonom ous men and

i
women rather than in the social environm ents that have appeared
in the evolution of cultures and that can now be explicitly designed.
By turning from m an qua m an to the external conditions of
which m an’s behavior is a function, it has been possible to design
better practices in the care of psychotics and retardates, in child
care, in education (in both contingency m anagem ent in the class
room and the design of instructional m aterial), in incentive systems
in industry, and in penal institutions. In these and m any other
areas we can now more effectively w ork for the good of the ind ivid
ual, for the greatest good of the greatest num ber, and for the good
of the culture or of m ankind as a whole. Th ese are certainly
hum anistic concerns, and no one w ho calls him self a hum anist can
afford to neglect them. M en and wom en have never faced a greater
threat to the fixture of their species. T h ere is m uch to be done and
done quickly, and n othing less than the active prosecution of a
science of behavior w ill suffice.
\
r
Walden Two Revisited
.■' irrniiiiii
'
III I
T h e early summer of 1945, when I wrote W alden Tw o, was
not a bad tim e for W estern C ivilization. H itler was dead, and one
of the most barbaric regimes in history was com ing to an end. T h e
Depression of the thirties had been forgotten. Com m unism was no
longer a threat, for Russia was a trusted ally. It w ould be another
m onth or two before H iroshim a w ould be the testing ground for a
horrible new weapon. A few cities had a touch of smog bu t no one
worried about the environm ent as a whole. T h ere were w artim e
shortages, bu t industry w ould soon turn again to devoting un
lim ited resources to the fulfdlm ent of unlim ited desires. T h e in
dustrial revolution was said to have stilled the voice of T hom as
R obert M althus.
T h e dissatisfactions w hich led me to w rite W alden Tw o were
personal. I had seen my w ife and her friends struggling to save
themselves from domesticity, w incing as they printed “ housewife”
in those blanks asking for occupation. O u r older daughter had just
finished first grade, and there is nothing like a first child ’s first year ■
in school to turn one’s thoughts to education. W e were soon to leave
M innesota and m ove to Indiana and I had been in search of
'■ I
Walden Two Revisited 57
housing. I would be leaving a group of talented young string

players who had put up w ith my inadequacies at the piano and I
was not sure:I could ever replace them. I had just finished a pro
ductive year on a G uggenheim Fellowship, bu t I had accepted the
chairmanship of a departm ent at In d ian a and was not sure w hen I
would again have time for science or scholarship. W as there not
something to be done about problem s of that sort? W as there not
by any chance som ething a science of behavior could do?
It was probably a good thing that these were small provincial
problems, because I m ight not have had the courage to tackle
bigger ones. In B ehavior of Organisms, published seven years
earlier, I had refused to apply my results outside the laboratory.
“ Let him extrapolate w ho w ill,” I had said. But, of course, I had
speculated about the technology that a science of behavior im plied
and about the {differences it could make. I had recently been takin g
the implications seriously because I had been m eeting once a m onth
with a g ro u p , of philosophers and critics (among them H erbert
Feigl, Alburey.Castell, and R obert Penn W arren) where the control
of hum an behavior had em erged as a central topic.
T h a t all this should come together in a novel about a u topian
community was probably due to the fact that a colleague, A lice F.
Tyler, had sent me a copy of her new book, Freedom’s Ferm ent, a
study of perfectionist movements in Am erica in the nineteenth
century.1 W ith two months to spare before m oving to Indiana, I
decided to write an account of how I thought a group of, say, a
thousand people m ight have solved the problems of their daily lives
with the help of behavioral engineering.
T w o publishers turned Walden Tw o down, and M acm illan
published it only on condition that I w rite an introductory text for
them. These editorial judgm ents were, at the time, quite correct.
One or two distinguished critics took the book seriously, b u t the
public left it alone for a dozen years. T h e n it began to sell, and the
annual sales rose steadily on a com pound interest curve.
T h ere were, I think, two reasons for the awakened interest.
T h e “ behavioral engineering” I had so frequently m entioned in
the book was, at the time, little m ore than science fiction. I had
1 T y le r, A . F. Freedom ’s Ferm ent. M in n eapo lis: U n iv. o f M in nesota

Press, 1944.
I
58 SOCIETY
thought that an experim ental analysis of behavior could be applied

to practical problems, but I had not proved it.. T h e 1950s, however,
saw the beginnings of w hat the public has come to know as be
havior m odification. T h ere were early experim ents on psychotic
and retarded persons, and then on teaching machines and pro
gram m ed instruction, and some of the settings in w hich these
experim ents were conducted were in essence communities. A n d in
the sixties applications to other fields, such as counseling and the
design of incentive systems, came even closer to w hat I had de
scribed in Walden Two. A technology of behavior was no longer
a figment of the im agination. Indeed, to m any people it was alto
gether too real.
B u t there was, I think, a better reason why m ore and more
people began to read the book. T h e w orld was beginning to face
problems of an entirely new order of m agnitude— the exhaustion of
resources, the pollution of the environm ent, overpopulation, and
the possibility of a nuclear holocaust, to m ention only four. Physi
cal and biological technologies could, of course, help. W e could find
new sources of energy and m ake better use of those we had. T h e
world could feed itself by grow ing more nutritious grains and
eating grain rather than meat. M ore reliable methods of contracep
tion could keep the population w ithin bounds. Im pregnable de
fenses could make a nuclear war impossible. B u t that w ould happen
only if hum an behavior changed, and how it could be changed was
still an unanswered question. H ow were people to be induced to
use new forms of energy, to eat grain rather than meat, and to lim it
the size of their families; and how were atom ic stockpiles to be
kept out of the hands of desperate leaders?
From time to tim e policy makers in high places have been
urged to pay more attention to the behavioral sciences. T h e Na- 1
tional Research Council, the operative arm of the N ational A cad
emy of Sciences, made one such proposal a num ber of years ago,
po in tin g out that useful “ insights in policy form ulation ” had been
developed. But it im plied that the chief role of the behavioral
sciences was to collect facts and insisted, possibly to reassure policy |
makers who m ight be alarm ed by the am bitions of scientists, that |
“ knowledge is no substitute for wisdom or com m on sense in m aking .
decisions.” Science w ould get the facts but Congress or the President
w ould m ake the decisions— w ith wisdom and com m on sense.
It is true that when the behavioral sciences have gone beyond

the collection of facts to recom m end courses of action and have
done so by predicting consequences, they have not been too helpful.
Not all economists agree, for exam ple, on how an increase or reduc
tion in taxes or a change in interest rates w ill affect business, prices,
or unem ploym ent, and political scientists are no more likely to
agree on the consequences of domestic or international policies. In
anthropology, sociology, and psychology the preferred form ulations
are those that do not dictate action. A thoroughgoing developmen-
talism, for exam ple, almost denies the possibility of effective action.
A pplied psychology is usually a m ixture of science and common
sense, and Freud regarded therapy as a m inor contribution of
psychoanalysis'.
From thp very beginning the application of an experim ental
analysis of behavior was different. It was doubly concerned w ith
consequences. Behavior could be changed by changing its conse
quences— that» was operant conditioning— but it could be changed
because other -.kinds of consequences would then follow. Psychotic
and retarded persons w ould lead better lives, time and energy
of teachers and students w ould be saved, homes w ould be pleasanter
social environments, people w ould w ork more effectively w hile en
joying w hat they were doing, and so on.
T hese are the kinds of achievements traditionally expected
from wisdom and common sense, but Frazier, the protagonist of
Walden Tw o, insists that they are w ithin reach of a special be
havioral science w hich can take the place of wisdom and common
sense and with happier results. A n d w hat has happened in the past
twenty-five years has increased the plausibility o f his achievem ent—
a com m unity in which the most im portant problem s of daily life,
as well as certain aspects of economics and governm ent, are solved.
Frazier’s critics w ill protest. W h at can we conclude from a
successful com m unity of a thousand people? T r y those principles on
New Y o rk City, say, or on the State D epartm ent and see what
happens. T h e w orld is a vast and com plex space. W h at works for a
small group will be far short of w hat is needed fo,r a nation or the
world as a whole.
Frazier m ight answer by calling W alden T w o a p ilot experi
ment. Industries do not invest in large plants u n til they have tried
a new process on a smaller scale. If we w ant to find out how people
6o SOCIETY
can live together w ithout quarreling, can produce the goods they
need w ithout w orking too hard, or can raise and educate their chil
dren more efficiently, let us start w ith units of m anageable size
before m oving on to larger problems.
But a more cogent answer is this: what is so w onderful about
being big? It is often said that the w orld is suffering from the ills of
bigness, and we now have some clinical examples in our large cities.
M any cities arc probably past the point of good governm ent because
too many things are wrong. Should we not rather ask whether we
need cities? W ith modern systems of com m unication and transpor
tation, businesses do not need to be w ithin w alking or taxicab
distances of each other, and how m any people must one be near in
order to live a happy life? People who flock to cities looking for
jobs and more interesting lives w ill flock back again if jobs and
more interesting lives are to be found where they came from. It has
been suggested that, w ith m odern systems of com m unication, the
A m erica of the future may be sim ply a network of small towns. B u t
should, we not say W alden Twos? A few skeletons of cities may
survive, like the bones of dinosaurs in museums, as the remains of
a passing phase in the evolution of a way of life.
T h e British economist E. F. Schumacher, in his rem arkable
book Small Is B eautiful,2 has discussed the problem s that come from
bigness and has outlined a technology appropriate to systems of
interm ediate size. M any current projects dealing w ith new sources
of energy and new forms of agriculture seem ideally suited to
developm ent by small com m unities. A network of small towns or .
W alden T w os would have its own problems, but the astonishing
fact is that it could m uch more easily solve m any of the crucial
problems facing the w orld today. A lthough a small com m unity does
not bring out ‘‘hum an nature in all its essential goodness” (small
towns have never supported that rom antic dream), it makes it
possible to arrange m ore effective “ contingencies of reinforcem ent”
according to the principles of an applied behavior analysis. W e
need not look too closely at practices derived from such principles
to survey some of those w hich could solve basic problem s in a small
comm unity.
2 Schumacher, E. F. Small Is Beautiful. New York: Harper Torch-

books, 1973.
T o induce people to adapt to new ways of livin g w hich are

less consuming and hence less polluting, we do not need to speak of
frugality or austerity as if we m eant sacrifice. T h ere are contin
gencies of reinforcem ent in w hich people continue to pursue (and
even overtake) happiness w hile consum ing far less than they now
consume. T h e experim ental analysis of behavior has clearly shown
that it is not the quantity of goods that counts (as the law of supply
and dem and suggests) bu t the contingent relation between goods
and behavior. T h a t is why, to the amazement o f the Am erican
tourist, there are people in the w orld w ho are happ ier than we are,
while possessing far less. Inflation is said to be the most serious
problem in the world today. It has been defined, not ineptly, as
spending more than one has. In an experim ental com m unity con
tingencies of reinforcem ent w hich encourage unnecessary spending
can be corrected. As for pollution, small com m unities are optim al
for recycling materials and avoiding wasteful methods of distri
bution. ,
T h e basic research has also shown how im portant it is for
everyone, young and old, wom en and men, not only to receive goods
but to engage in their production. T h a t does m ean that we should
all work like eager beavers according to the Protestant w ork ethic.
T here are many ways of saving labor, but they should not, as
Frazier points out, be used to save laborers and hence to increase
unemploym ent. Sim ply by d ivid in g the total am ount of wages
Americans receive each year by the num ber of people w ho want
jobs, we arrive at a, perfectly reasonable annual w age for everyone.
But that means a reduction in the standard of livin g for many
people, which, as things now stand, is probably impossible. In a
series of small com munities, however, everyone w ould have a job
because work, as w ell as wages, could be divided am ong workers.
And good incentive conditions— for exam ple, those in w hich people
make not money, but the things that m oney buys— do not require
what we call hard work.
If the world is to save any part of its resources for the future,
it must reduce not only consum ption bu t the num ber of consumers.
It should be easy to change the birthrate in an experim ental
community. Parents w ould not need children for econom ic secur
ity, the childless could spend as m uch time w ith children as they
liked, and the com m unity w ould function as a large and affec
62 SOCIETY
tionate fam ily in which everyone w ould play parental and filial
roles. Blood ties w ould then be a m inor issue..
People are more likely to treat each other w ith friendship and
affection if they are not in com petition for personal or professional
status. B u t good personal relations also depend upon im m ediate
signs of commendation or censure, supported perhaps by simple
rules or codes. T h e bigness of a large city is troublesom e precisely
because we meet so many people whom we shall never see again
and whose com m endation or censure is therefore meaningless. T h e
problem cannot really be solved by delegating censure to a police
force and the law courts. T hose w ho have used behavior modifica
tion in fam ily counseling or in institutions know how to arrange
the face-to-face conditions w hich promote interpersonal respect
and love.
W e could solve m any of the problems of delinquency and
crime if we could change the early environm ent of offenders. One
need not be a bleeding heart to argue that m any you n g people
today have simply not been prepared by their homes or school to
lead successful lives w ithin the law or, if prepared, do not have the
chance to do so by getting jobs. Offenders are seldom im proved by
being sent to prison, and judges therefore tend to reduce or suspend
sentences, but crime, unpunished, then increases. W e all know how
early environments can be im proved, and a much-neglected experi
ment reported by Cohen and F ilip cz a k 3 has dem onstrated that
occasional offenders can be rehabilitated.
C hildren are our most valuable resources and they are now
sham efully wasted. W on derful things can be done in the first years
of life, but we leave them to people whose mistakes range all the
way from child abuse to overprotection and the lavishing of affec
tion on the wrong behavior. W e give small children little chance
to develop good relationships w ith their peers or w ith adults, espe
cially in the single-parent home, w hich is on the increase. T h a t is
all changed when children are, from the very first, part o f a larger
com munity.
C ity schools show how m uch harm bigness can do to educa
tion, and education is im portant because it is concerned w ith the
3 Cohen, H. L., and Filipczak, J. A New Learning Environment.

San Francisco: Jossey-Bass, 1971.
transmission and hence the survival of a culture. W e know how to

solve many educational problems w ith program m ed instruction and
good contingency m anagement, saving resources and the time and
effort of teachers and students. Sm all com m unities are ideal settings
for-new kinds of instruction, free from interference by administra
tors, politicians, and organizations of teachers.
In spite of our lip service to freedom, we do very little to
further the developm ent of the individual. H ow m any Am ericans
can say that they are doin g the kinds of things they are best quali
fied to do and most enjoy doing? W hat opportunities have they had
to choose fields related to their talents or to the interests and skills
they acquired in early life? W om en, only ju st begin nin g to be able
to choose not to be housewives, can now discover how hard it is to
choose the right profession when they are young or to change to a
different one later on.
A n d once one is lucky enough to be doing w hat one likes,
what are the chances of being successful? H ow easily can artists,
composers, an<£ writers bring their work to the attention of those
who w ill enjoy it and whose reactions w ill shape behavior in crea
tive ways? Those who know the im portance of contingencies of
reinforcement know how people can be led to discover the things
they do best and the things from w hich they w ill get the greatest
satisfaction.
A lthough sometimes questioned, the survival value of art,
music, literature, games, and other activities not tied to the serious
business of life is clear enough. A culture must positively reinforce
the behavior of those w ho support it and must avoid creating
negative reinforcers from which its members w ill escape through
defection. A w orld w hich has been made beautiful and exciting by
artists, composers, writers, and perform ers is as im portant for sur
vival as one w hich satisfies biological needs.
T h e effective use of leisure is almost com pletely neglected in
modern life. W e boast of our short w orkday and week, b u t w hat we
do w ith the free tim e we have to spend is nothing of w hich we
can be very proud. T h e leisure classes have almost always turned
to alcohol and other drugs, to gam bling, and to w atching other
people lead exhausting or dangerous lives, and we are no exception.
Thanks to television m illions of Am ericans now lead the exciting
and dangerous lives of other people. M any states are legalizing
64 SOCIETY
gam bling and have set up lotteries of their own. A lcohol and drugs
are consumed in ever-increasing quantities. One may spend one’s
life in these ways and be essentially unchanged at the end of it.
These uses of leisure are due to some basic behavioral processes, bu t
the same processes, in a different environment, lead people to
develop their skills and capacities to the fullest possible extent.
A re we quite sure of all this? Perhaps not, but W alden T w o
can help us make sure. Even as part of a larger design, a com m unity
serves as a pilot experim ent. T h e question is sim ply w hether it
works, and one way or the other, the answer is usually clear. W hen
that is the case, we can increase our understanding of hum an be
havior w ith the greatest possible speed. H ere is possibly our best
chance to answer the really im portant questions facing the world
today— questions not about economics or governm ent but about
the daily lives of hum an beings.
Yes, but w hat about economics and government? M ust we
not answer those questions too? I am not sure we must. Consider
the follow ing econom ic propositions. T h e first is from H enry D avid
T h o rea u ’s Walden: by reducing the am ount of goods we consume,
we can reduce the am ount of tim e we spend in unpleasant labor.
T h e second appears to assert just the opposite: we must all con
sume as much as possible so that everyone can have a job. I subm it
that the first is more reasonable, even though the second is defended
by m any people today. Indeed, it m ight be argued that if Am erica
were to convert to a netw ork o f small com m unities, our economy
w ould be wrecked. B u t som ething is wrong w hen it is the system
that must be saved rather than the way of life that the system is
supposed to serve.
B u t what about governm ent? Surely I am not suggesting that
we can get along w ithout a federal government? B u t how m uch of
it is needed? O ne great share of our national budget goes to the
Departm ent of Health, E ducation and W elfare. Health? Educa
tion? W elfare? B u t an experim ental com m unity like W alden T w o
is health, education, and welfare! T h e only reason we have a vast
federal departm ent is that m illions of people find themselves
trapped in overgrown, unw orkable livin g spaces.
Another large share of the budget goes to the D epartm ent of
Defense. A m I suggesting that we can get along w ith ou t that? H ow
can we preserve the peace of the world if we do not possess the
most powerful weapons, together w ith an industry that continues

to develop even more pow erful ones? B u t we have weapons only
because other-, countries have them, and although we feel threatened
by countries w ith com parable m ilitary power, particularly the
Bomb, the real threat m ay be the countries that have next to
nothing. A few highly industrialized nations cannot long continue
to face the rest of the world w hile consum ing and p o llu tin g the
environment as they do. A way o f life in w hich each person used
only a fair share of the resources o f the w orld and yet somehow
enjoyed life w ould be a real step toward world peace. It is a pattern
that could easily be copied, and I was heartened recently w hen
someone from the State D epartm ent called to tell me that he
thought Am erica ought to stop tryin g to export the ‘A m erican way
of life” and export W alden T w os instead. A state defined by re
pressive, formal, legal, social controls based on physical force is not
necessary in the developm ent o f civilization,4 and although such
a state has certainly figured in our own developm ent, we m ay be
ready to move on\to another stage.
Suppose 'we do know w hat is needed for the good life; how
are we to bring it about? In Am erica we almost instinctively move
to change things by political action: we pass laws, we vote for new
leaders. B ut a good m any people are beginning to wonder. T h e y
have lost faith in a democratic process in w hich the so-called w ill of
the people is obviously controlled in undem ocratic ways. A n d there
is always the question w hether a governm ent based on pun itive
sanctions is inappropriate if we are to solve problems nonpuni-
tively.
It has been argued that the solution m ight be socialism, b u t
it has often been pointed out that socialism, like capitalism , is
committed to growth, and hence to overconsum ption and pollution.
Certainly Russia after fifty years is not a model we wish to em ulate.
China may be closer to the solutions I have been talking about,
but a Com m unist revolution in Am erica is hard to im agine. It
would be a bloody affair, and there is always L en in ’s question to be
answered: H ow m uch suffering can one impose upon those now
living for the sake of those who w ill follow? A n d can we be sure
that those who follow w ill be any better off?
4 See Service, Elm an. Origins of the State and Civilization. N ew
York: N orton , 1975.
66 SOCIETY
Fortunately, there is another possibility. A n im portant theme

in Walden Tw o is that political action is to be .avoided. Historians
have stopped w riting abou t wars and conquering heroes and em
pires, and what they have turned to instead, though far less
dramatic, is far more im portant. T h e great cultural revolutions
have not started w ith politics. T h e great men who are said to have
made a difference in hum an affairs— Confucius, Buddha, Jesus, the
scholars and scientists of the R evival of Learning, the leaders of
the Enlightenm ent, M arx— were not political leaders. T h e y did
not change history by ru n n in g for office. W e need not aspire to
their em inence in order to profit from their exam ple. W h at is
needed is not a new p olitical leader or a new kind of governm ent
but further knowledge abou t hum an behavior and new ways of
applying that know ledge to the design of cultural practices.
It is now widely recognized that great changes must be made
in the Am erican way of life. N ot only can we not face the rest of
the world w hile consum ing and p o llu tin g as we do, we cannot for
long face ourselves w hile acknow ledging the violence and chaos in
which we live. T h e choice is clear: either we do n othing and allow
a miserable and probably catastrophic future to overtake us, or we
use our know ledge abou t hum an behavior to create a social
environm ent in w hich we shall live productive and creative lives
and do so w ithout jeop ard izin g the chances that those who follow
us w ill be able to do the same. Som ething like a W alden T w o
would not be a bad start.
PART II
THE SCIENCE
OF
BEHA VIOR
i
\
6 The Steep and Thorny Way to a Science of Behavior
7 Can We Profit from Our Discovery of Behavioral

Science?
8 Why I Am Not a Cognitive Psychologist
9 The Experimental Analysis of Behavior (A History)

6
The Steep and Thorny W ay
to a Science of Behavior
A critic contends that a recent book of m in e 1 does not con

tain anything new, that m uch the same was said m ore than four
centuries ago in theological terms by John C alvin. You w ill not be
surprised, then, to find me com m ending to you the steep and thorny
way to that heaven promised by a science of behavior. B u t I am not
one of those ungracious pastors, of whom O phelia com plained, who
“ recking not their own rede themselves tread the prim rose path of
dalliancc.” N o, I shall rail at dalliance, and in a m anner w orthy, I
hope, of my distinguished predecessor. If I do not thunder or
fulm inate, it is only because we moderns can m ore easily portray a
truly frightening hell. I shall merely allude to the carcinogenic
fallout of a nuclear holocaust. A n d no C alvin ever had better
reason to fear his hell, for I am proceeding on the assumption that
nothing less than a vast im provem ent in our understanding of
hum an behavior w ill prevent the destruction of our way of life or
of m ankind.
W hy has it been so difficult to be scientific about hum an be
1 Skinner, B. F. B eyon d freedom and dignity. N e w Y o rk: A lfre d A .

Knopf, 1971.
68
The Steep and Thorny Way to a Science of Behavior 69
havior? W hy have methods that have been so prodigiously success

fu l almost everywhere else failed so ignom iniously in this one field?
Is it because hum an behavior presents unusual obstacles to a
science? N o doubt it does, b u t I think we are begin nin g to see how
these obstacles may be overcome. T h e problem , I subm it, is digres
sion. W e have been drawn off the straight and narrow path, and
the word diversion serves me well by suggesting not only digression
but dalliance. In this article I analyze some of the diversions pecu
liar to the field of hum an behavior w hich seem to have delayed our
advance toward the better understanding we desperately need.
I must begin by saying w hat I take a science of behavior to be.

It is, I assume, part of biology. T h e organism that behaves is the
organism that brkath.es, digests, conceives, gestates, and so on. As
such, the behaving organism w ill eventually be described and ex
plained by the anatom ist and physiologist. As far as behavior is
concerned, they \\5ill give us an account of the genetic endow m ent
of the species and tell how that endowm ent changes during the life
time of the individual and why, as a result, the individ u al then
responds in a given way on a given occasion. Despite rem arkable
progress, we are still a long way from a satisfactory account in such
terms. W e know som ething about the chem ical and electrical effects
of the nervous system and the location o f m any of its functions, but
the events that actually underlie a single instance of behavior— as a
pigeon picks up a stick to b u ild a nest, or a child a block to com plete
a tower, or a scientist a pen to w rite a paper— are still far out o f reach.
Fortunately, we need not w ait for further progress of that sort.
W e can analyze a given instance of behavior in its relation to the
current setting and to antecedent events in the history of the species
and of the individual. T h u s, we do not need an exp licit account
of the anatomy and physiology of genetic endowm ent in order to
describe the behavior, or the behavioral processes, characteristic of
a species, or to speculate about the contingencies o f survival u nder
which they m ight have evolved, as the ethologists have convincingly
demonstrated. N or do we need to consider anatom y and physiology
in order to see how the behavior of the individual is changed by
his exposure to contingencies o f reinforcem ent during his lifetim e
and how as a result he behaves in a given way on a given occasion.
7° T H E SCIENCE OF BEHAVIOR
I must confess to a predilection here for my own specialty, the ex

perim ental analysis of behavior, w hich is a quite -explicit investiga
tion of the effects upon individual organisms of extrem ely com plex
and subtle contingencies o f reinforcem ent.
T h ere w ill be certain tem poral gaps in such an analysis. T h e
behavior and the conditions of w hich it is a function do not occur
in close tem poral or spatial proxim ity, and we must w ait for physi
ology to make the connection. W h en it does so, it w ill not invali
date the behavioral account (indeed, its assignment could be said
to be specified by that account), nor w ill it m ake its terms and
principles any the less useful. A science of behavior w ill be needed
for both theoretical and practical purposes even when the behaving
organism is fully understood at another level, just as m uch of
chemistry remains useful even though a detailed account of a
single instance may be given at the level of m olecular or atomic
forces. Such, then, is the science of behavior from w hich I suggest
we have been diverted— by several kinds of dalliance to w hich I
now turn.
Very little biology is handicapped by the fact that the biologist

is him self a specimen of the thing he is studying, b u t that part of
the science w ith which we are here concerned has not been so
fortunate. W e seem to have a kind of inside inform ation about
our behavior. It may be true that the environm ent shapes and con
trols our behavior as it shapes and controls the behavior of other
species— b u t we have feelings about it. A n d w hat a diversion they
have proved to be. Our loves, our fears, our feelings about war,
crime, poverty, and God— these are all basic, if not ultim ate, con
cerns. A n d we are as m uch concerned about the feelings of others.
M any of the great themes o f m ythology have been abou t feelings—
of the victim on his way to sacrifice or of the w arrior going forth
to battle. W e read what poets tell us about their feelings, and we
share the feelings of characters in plays and novels. W e follow
regimens and take drugs to alter our feelings. W e become sophisti
cated about them in, say, the m anner of La Rochefoucauld, noting
that jealousy thrives on doubt, or that the clemency of a ru ler is a
m ixture of vanity, laziness, and fear. A n d along w ith some psychia
The Steep and Thorny Way to a Science of Behavior I1
trists we may even try to establish an independent science of feel

ings in the intrapsychic life of the m ind or personality.
A n d do 'feelings not have some bearing on our form ulation of
a science of behavior? D o we not strike because we are angry and
play music because we feel like listening? A n d if so, are our feelings
not to be added to those antecedent events of w hich behavior is a
function? T h is is not the place to answer such questions in detail,
but I must at least suggest the kind of answer that may be given.
W illiam James questioned the causal order: Perhaps we do not
strike because we are angry but feel angry because we strike. T h a t
does not brin g us back to the environm ent, however, although
James and others were on the right track. W h at we feel are condi
tions of our bodies, most of them closely associated w ith behavior
and with the circumstances in w hich we behave. W e both strike and
i
feel angry for al common reason, and that reason lies in the environ
ment. In short, the bodily conditions we feel are collateral products
of our genetic and environm ental histories. T h e y have no exp lana
tory force; they-, are sim ply additional facts to be taken into account.
Feelings enjoy an enormous advantage over genetic and en
vironm ental histories. T h e y are warm, salient, and dem anding,
where facts about the environm ent are easily overlooked. M oreover,
they are immediately related to behavior, being collateral products
of the same causes, and have therefore com m anded more attention
than the causes themselves, w hich are often rath er remote. In doing
so, they have proved to be one of the most fascinating attractions
along the path of dalliance.
A m uch more im portant diversion has, for m ore than 2,000

years, m ade any move toward a science of behavior particularly
difficult. T h e environm ent acts upon an organism at the surface of
its body, but when the body is our own, we seem to observe its
progress beyond that point; for exam ple, we seem to see the real
world become experience, a physical presentation becom e a sensation
or a percept. Indeed, this second stage may be all we see. R eality
may be m erely an inference and, according to some authorities, a
bad one. W h at is im portant m ay not be the physical w orld on the
far side of the skin but w hat that w orld means to us on this side.
72 T H E SCIENCE OF BEHAVIOR
N ot only do we seem to see the environm ent on its way in, we

seem to see behavior on its way out. W e observe certain early stages
— wishes, intentions, ideas, and acts of w ill— before they have, as
we say, found expression in behavior. A n d as for our environm ental
history, that can also be viewed and reviewed inside the skin, for
we have tucked it all away in the storehouse of our memory. A gain
this is not the place to present an alternative account, b u t several
points need to be made. T h e behavioristic objection is not prim arily
to the m etaphysical nature of m ind stuff. I welcome the view, clearly
gain in g in favor among psychologists and physiologists and by no
means a stranger to philosophy, that what we introspectively ob
serve, as well as feel, are states of our bodies. B u t I am not w illin g
to give introspection m uch of a toehold even so, for there are two
im portant reasons why we do not discrim inate precisely am ong our
feelings and states of m ind and hence why there are m any different
philosophies and psychologies.
In the first place, the world w ithin the skin is private. O nly
the person whose skin it is can m ake certain kinds of contact w ith
it. W e m ight expect that the resulting intim acy should m ake for
greater clarity, but there is a difficulty. T h e privacy interferes w ith
the very process of com ing to know. T h e verbal com m unity w hich
teaches us to make distinctions am ong things in the w orld around
us lacks the inform ation it needs to teach us to distinguish events
in our private world. For exam ple, it cannot teach us the difference
between diffidence and embarrassment as readily or as accurately
as that between red and blue or sweet and sour.
Second, the self-observation that leads to introspective know l
edge is lim ited by anatomy. It arose very late in the evolution of
the species because it is only when a person begins to be asked
about his behavior and abou t why he behaves as he does that he
becomes conscious of him self in this sense. Self-knowledge depends
on language and in fact on language of a rather advanced kind, bu t
when questions of this sort first began to be asked, the only nervous
systems available in answering them were those that had evolved for
entirely different reasons. T h ey had proved useful in the internal
economy of the organism, in the coordination of m ovem ent, and in
operating upon the environm ent, b u t there was no reason why
they should be suitable in supplying inform ation abou t those very
extensive systems that m ediate behavior. T o pu t it crudely, intro
spection cannot be very relevant or com prehensive because the

human organism does not have nerves go in g to the right places.
One other problem concerns the nature and location of the
knower. T h e organism itself lies, so to speak, betw een the environ
ment that acts upon it and the environm ent it acts upon, b u t w hat
lies between those inner stages-^-between, for exam ple, experience
and will? From w hat vantage point do we watch stim uli on their
way into the storehouse of memory or behavior on its way out to
physical expression? T h e observing agent, the knower, seems to
contract to som ething very small in the m iddle of things.
In the form ulation of a science w ith w hich I began, it is the
organism as a whole that behaves. It acts in and upon a physical
world, and it can be induced by a verbal environm ent to respond to
some of its o\yn activities. T h e events observed as the life of the
mind, like feelings, are collateral products, w hich have been made
the basis of many elaborate metaphors. T h e philosopher at his desk
asking him selfiw hat he really knows, about him self or the world,
will quite n aturally begin w ith his experiences, his acts of w ill, and
his memory, but the effort to understand the m ind from that van
tage point, beginning w ith P lato’s supposed discovery, has been one
of the great diversions which have delayed an analysis o f the role
of the environment.
It did not, of course, take inside inform ation to induce people

to direct their attention to w hat is going on inside the behaving
organism. W e almost instinctively look inside a system to see how
it works. W e do this w ith clocks, as w ith livin g systems. It is stan
dard practice in m uch of biology. Some early efforts to understand
and explain behavior in this way have been described by Onians
in his classic Origins of European T h o u g h t .2 It must have been the
slaughterhouse and the battlefield that gave man his first know l
edge of anatomy and physiology. T h e various functions assigned to
parts of the organism were not usually those that had been observed
introspectively. If Onians is right, the phrenes were the lungs, in ti
mately associated w ith breathing and hence, so the Greeks said,
2 O nians, R . D . T h e origins of European thought. C am bridge,

En glan d: U n iversity Press, 1951.
74 TH E SCIENCE OF BEHAVIOR
w ith thought and, of course, w ith life and death. T h e phrenes were
the seat of thumps, a vital principle whose nature is not now clearly
understood, and possibly of ideas, in the active sense of H om eric
Greek. (By the time an idea had become an object of quiet con
tem plation, interest seems to have been lost in its location.) Later,
the various fluids of the body, the humors, were associated w ith
dispositions, and the eye and the ear w ith sense data. I like to
im agine the consternation of that pioneer who first analyzed the
optics of the eyeball and realized that the im age on the retina was
upside down!
Observation of a behaving system from w ithin began in
earnest with the discovery of reflexes, but the reflex arc was not
only not the seat of m ental action, it was taken to be a usurper, the
spinal reflexes replacing the Riickenmark.seele or soul of the spinal
cord, for example. T h e reflex arc was essentially an anatom ical
concept, and the physiology rem ained largely im aginary for a long
time. M any years ago I suggested that the letters C N S could be said
to stand, not for the central nervous system, b u t for the conceptual
nervous system. I had in m ind the great physiologists Sir Charles
Sherrington and Ivan Petrovich Pavlov. In his epoch-m aking In te
grative Action of the Ne'rvous System, Sherrington 3 had analyzed
the role of the synapse, listing perhaps a dozen characteristic prop
erties. I pointed out that he had never seen a synapse in action and
that all the properties assigned to it were inferred from the behavior
of his preparations. Pavlov had offered his researches as evidence of
the activities of the cerebral cortex though he had never observed
the cortex in action bu t had merely inferred its processes from the
behavior of his experim ental animals. B ut Sherrington, Pavlov, and
many others were m oving in the direction of an instrum ental ap
proach, and the physiologist is now, of course, studying the nervous
system directly.
T h e conceptual nervous system has been taken over by other
disciplines— by inform ation theory, cybernetics, systems analyses,
m athem atical models, and cognitive psychology. T h e hypothetical
structures they describe do not depend on confirm ation by direct
observation of the nervous system, for that lies too far in the future
3 Sherrington, C. S. Integrative action of the nervous system. N ew

H aven, C onn.: Y a le U n iversity Press, 1906.
to be of interest. T h e y are to be justified by their internal consis

tency and the '.successful prediction of selected facts, presum ably not
the facts from; which the constructions were inferred.
T hese disciplines are concerned w ith how the brain or the
mind must work if the hum an organism is to behave as it does.
T h ey offer a sort of therm odynam ics of behavior w ith ou t reference
to m olecular action. T h e com puter w ith its apparent sim ulation of
Man T h in k in g supplies the dom inant analogy. It is not a question
of the physiology of the com puter— how it is w ired or w hat type of
storage it uses— but of its behavioral characteristics. A com puter
takes in inform ation as an organism receives stim uli and processes
it according to an in b u ilt program as an organism is said to do
according to its genetic endowm ent. It encodes the inform ation,
converting it to a form it can handle, as the organism converts
visual, auditory, and other stim uli into nerve impulses. L ik e its
human analogue it stores the encoded inform ation in a memory,
tagged to facilitate retrieval. It uses what it has stored to process
inform ation as received, as a person is said to use prior experience
to interpret incom ing stim uli, and later to perform various opera
tions— in short, to compute. Finally, it makes decisions and behaves:
It prints out.
T h ere is nothin g new about any of this. T h e same things
were done thousands of years ago w ith clay tiles. T h e overseer or
tax collector kept a record o f bags of grain, the num ber, quality,
and kind being m arked appropriately. T h e tiles were stored in lots
as marked, additional tiles were grouped appropriately, the records
were eventually retrieved and com putations made, and a summary
account was issued. T h e m achine is m uch swifter, and it is so con
structed that hum an participation is needed only before and after
the operation. T h e speed is a clear advantage, b u t the apparent
autonomy has caused trouble. It has seemed to m ean that the mode
of operation of a com puter resembles that of a person. P eople do
make physical records w hich they store and retrieve and use in
solving problems, but it does not follow that they do anythin g of
the sort in the m ind. If there were some exclusively subjective
achievement, the argum ent for the so-called higher m ental processes
would be stronger, b u t as far as I know, none has been dem on
strated. T ru e, we say that the m athem atician sometimes in tu itively
solves a problem and only later, if at all, reduces it to the steps of
a proof, and in doing so he seems to differ greatly from those who

proceed step by step, bu t the differences could well be in the evi
dence of w hat has happened, and it would not be very satisfactory
to define thought sim ply as u nexplained behavior.
Again, it would be foolish of me to try to develop an alter
native account in the space available. W hat I have said about the
introspectively observed m ind applies as well to the m ind that is
constructed from observations of the behavior of others. T h e acces
sibility of stored memories, for exam ple, can be interpreted as the
probability of acquired behaviors, w ith no loss in the adequacy of
the treatment of the facts, and w ith a very considerable gain in the
assim ilation of this difficult field w ith other parts of hum an be
havior.
I have said that m uch of biology looks inside a livin g system

for an explanation of how it works. B ut that is not true of all of
biology. Sir Charles B ell could w rite a book on the hand as evidence
of design. T h e hand was evidence; the design lay elsewhere. D arw in
found the design, too, but in a different place. H e could catalog
the creatures he discovered on the voyage of the Beagle in terms
of their form or structure, and he could classify barnacles for years
in the same way, but he looked beyond structure for the principle
of natural selection. It was the relation of the organism to the
environm ent that m attered in evolution. A n d it is the relation to
environm ent that is of prim ary concern in the analysis of behavior.
Hence, it is not enough to confine oneself to organization or struc
ture, even of the most penetrating kind. T h a t is the mistake of most
of phenomenology, existentialism , and the structuralism of anthro
pology and linguistics. W hen the im portant thin g is a relation to
the environm ent, as in the phylogeny and ontogeny of behavior, the
fascination with an inner system becomes a simple digression.
W e have not advanced more rapidly to the m ethods and in
struments needed in the study of behavior precisely because of the
diverting preoccupation w ith a supposed or real inner life. It is
true that the introspective psychologist and the m odel builder have
investigated environments, b u t they have done so only to throw
some light on the internal events in which they are interested. T h ey
are no doubt well-intentioned helpmates, b u t they have often
simply misled those who undertake the study of the organism as a

behaving system in its own right. Even when helpful, an observed
or hypothetical inner determ iner is no explanation of behavior
until it has itself been explained, and the fascination w ith an inner
life has allayed curiosity about the further steps to be taken.
I can hear my critics: “ D o you really mean to say that all
those who have inquired into the hum an m ind, from Plato and
A ristotle through the Rom ans and Scholastics, to Bacon and
Hobbes, to Locke and the other British empiricists, to John Stuart
M ill, and to all those who began to call themselves psychologists—
that they have all been wasting their time?” W ell, not all of their
time, fortunately. Forget their purely psychological speculations,
and they were still rem arkable people. T h e y w ould have been even
more remarkahle, in my opinion, if they could have forgotten that
speculation themselves. T h e y were careful observers of hum an
behavior, but the intu itive wisdom they acquired from their contact
w ith real people was flawed by their theories.
It is easiej: to Nmake the point in the field of m edicine. U n til
the present century very little was know n about bod ily processes in
health and disease from w hich useful therapeutic practices could be
derived. Y et it should have been w orthw hile to call in a physician.
Physicians saw many ill people and should have acquired a kind of
wisdom, unanalyzed perhaps b u t still of value in prescribing simple
treatments: T h e history of m edicine, however, is largely the history
of barbaric practices— bloodlettings, cuppings, poultices, purgations,
violent emetics— w hich m uch of the tim e must have been harm ful.
My point is that these measures were not suggested by the intuitive
wisdom acquired from fam iliarity w ith illness; they were suggested
by theories, theories about what was going on inside an ill person.
Theories of the m ind have had a sim ilar effect, less dramatic, per
haps, but quite possibly far more dam aging. T h e men I have m en
tioned made im portant contributions in governm ent, religion,
ethics, economics, and many other fields. T h e y could do so w ith an
intuitive wisdom acquired from experience. B u t philosophy and
psychology have had their bleedings, cuppings, and purgations too,
and they have obscured simple wisdom. T h e y have diverted wise
people from a path that w ould have led more directly to an even
tual science of behavior. Plato w ould have m ade far more progress
toward the good life if he could have forgotten those shadows

on the w all of his cave.
Still another kind of concern for the self distracts us from

the program I have outlined. It has to do w ith the individual, not
as an object of self-knowledge, bu t as an agent, an initiator, a
creator. I have developed this theme in Beyond Freedom and D ig
nity. W e are more likely to give a person credit for w hat he does if it
is not obvious that it can be attributed to his physical or social
environm ent, and we are likely to feel that truly great achievements
must be inexplicable. T h e more derivative a w ork of art, the less
creative; the more conspicuous the personal gain, the less heroic an
act of sacrifice. T o obey a well-enforced law is not to show civic
virtue. W e see a concern for the aggrandizem ent of the individual,
for the m axim izing of credit due him, in the self-actualization of so-
called hum anistic psychology, in some versions of existentialism , in
Eastern mysticism and certain forms of Christian mysticism in which
a person is taught to reject the world in order to free him self for
union w ith a divine principle or with God, as well as in the simple
structuralism that looks to the organization of behavior rather than to
the antecedent events responsible for that organization. T h e difficulty
is that if the credit due a person is infringed by evidences of the condi
tions of w hich his behavior is a function, then a scientific analysis ap
pears to be an attack on hum an worth or dignity. Its task is to explain
the hitherto inexplicable and hence to reduce any supposed inner
contribution w hich has served in lieu of explanation. F reud m oved in
this direction in explainin g creative art, and it is no longer just the
cynic who traces heroism and m artyrdom to pow erful indoctrina
tion. T h e culm inating achievem ent of the hum an species has been
said to be the evolution of man as a moral anim al, b u t a simpler
view is that it has been the evolution of cultures in w hich people
behave m orally although they have undergone no in n er change of
character.
Even more traum atic has been the supposed attack on free
dom. H istorically, the struggle for freedom has been an escape from
physical restraint and from behavioral restraints exerted through
punishm ent and exploitative measures of other kinds. T h e individ
ual has been freed from features of his environm ent arranged by
governm ental and religious agencies and by those who possess great
wealth. T h e Success of that struggle, though it is not yet complete,
is one of m a a ’s great achievements, and no sensible person w ould
challenge it. U nfortunately, one of its by-products has been the
slogan that “ all control of hum an behavior is w rong and must be
resisted.” N oth in g in the circumstances under w hich m an has
struggled for freedom justifies this extension o f the attack on con
trolling measures, and we should have to abandon all of the advan
tages of a well-developed culture if we were to relinquish all
practices involvin g the control of hum an behavior. Y et new tech
niques in education, psychotherapy, incentive systems, penology,
and the design of daily life are currently subject to attack because
they are said to threaten personal freedom , and I can testify that
the attack can be fairly violent.
T h e extent to which a person is free or responsible for his
achievements is not an issue to be decided by rigorous proof, but
I subm it th a t,w h a t we call the behavior of the hum an organism
is no more free th^n its digestion, gestation, im m unization, or any
other physiological process. Because it involves the environm ent in
many subtle ways it is much more com plex, and its lawfulness is,
therefore, much harder to demonstrate. B u t a scientific analysis
moves in that direction, and we can already throw some light on
traditional topics, such as free w ill or creativity, w hich is m ore
helpful than traditional accounts, and I believe that further progress
is imminent.
T h e issue is, of course, determinism. Slightly more than 100
years ago, in a famous paper, C laude Bernard raised w ith respect
to physiology the issue w hich now stands before us in the b e
havioral sciences. T h e almost insurm ountable obstacle to the appli
cation of scientific method in biology was, he said, the belief in
“vital spontaneity.” His contem porary, Louis Pasteur, was responsi
ble for a dram atic test of the theory of spontaneous generation, and
I suggest that the spontaneous generation of behavior in the guise of
ideas and acts of w ill is now at the stage of the spontaneous genera
tion of life in the form of maggots and microorganisms 100 years
ago.
T h e practical problem in continuin g the struggle for freedom
and dignity is not to destroy controlling forces b u t to change them,
to create a w orld in w hich people w ill achieve far more than they
w
have ever achieved before in art, music, literature, science, tech

nology, and above all the enjoym ent of life. It could be a world in
w hich people feel freer than they have ever felt before, because they
w ill not be under aversive control. In b uildin g such a world, we
shall need all the help a science of behavior can give us. T o misread
the theme of the struggle for freedom and dignity and to relinquish
all efforts to control w ould be a tragic mistake.
B ut it is a mistake that may very well be made. O ur concern
for the individual as a creative agent is not dalliance; it is clearly
an obstacle rather than a diversion, for ancient fears are not easily
allayed. A shift in emphasis from the individual to the environ
ment, particularly to the social environm ent, is rem iniscent of vari
ous forms of totalitarian statism. It is easy to turn from w hat may
seem like an inevitable m ovem ent in that direction and to take
one’s chances w ith libertarianism . B u t much remains to be analyzed
in that position. For exam ple, we may distinguish between liberty
and license by holding to the right to do as we please provided we
do not infringe upon sim ilar rights of others, but in doing so we
conceal or disguise the pu b lic sanctions represented by private
rights. R ights and duties, like a m oral or ethical sense, are examples
of hypothetical internalized environm ental sanctions.
In the long run, the aggrandizem ent of the individual jeop
ardizes the future of the species and the culture. In effect, it in
fringes the so-called rights of billions of people still to be born, in
whose interests only the weakest of sanctions are now m aintained.
W e are beginning to realize the m agnitude of the problem of
brin ging hum an behavior under the control of a projected future,
and we are already suffering from the fact that we have come very
late to recognize that m ankind w ill have a future on ly if it designs
a viable way of life. I wish I could share the optim ism of both
D arw in and H erbert Spencer that the course of evolution is neces
sarily toward perfection. It appears, on the contrary, that that
course must be corrected from tim e to time. B u t if the intelligent
behavior that corrects it is also a product of evolution, then per
haps they were right after all. B u t it could be a near thing.
Perhaps it is now clear w hat I mean by diversions and

obstacles. T h e science I am discussing is the investigation of the
relation between behavior and the environm ent— on the one hand,
the environment in which the species evolved and w hich is responsi
ble for the facts investigated by the ethologists and, on the other
hand, the environment in w hich the individual lives and in re
sponse to which at any moment he behaves. W e have been diverted
from, and blocked in, our inquiries into the relations between
behavior and those environments by an absorbing interest in the
organism itself. W e have been m isled by the almost instinctive
tendency to look inside any system to see how it works, a tendency
doubly pow erful in the case of behavior because of the apparent
inside inform ation supplied by feelings and introspectively observed
states. O ur only recourse is to leave that subject to the physiologist,
who has, or w ill have, the only appropriate instruments and
methods. W e have also been encouraged to move in a centripetal
direction because the discovery of controlling forces in the environ
ment has seemed to reduce the credit due us for our achievements
and to suggest ,that the struggle for freedom has not been as fu lly
successful as we had im agined. W e are not yet ready to accept the
fact that the task is to change, not people, bu t rather the w orld in
which they live.
W e shall be less reluctant to abandon these diversions and to

attack these obstacles, as we come to understand the possibility of
a different approach. T h e role of the environm ent in hum an affairs
has not, of course, gone unnoticed. H istorians and biographers have
acknowledged influences on hum an conduct, and literature has
made the same point again and again. T h e E nlightenm ent advanced
the cause of the ind ivid ual by im provin g the w orld in w hich he
lived— the Encyclopedia of D iderot and D ’Alem bert was designed
to further changes of that sort, and by the nineteenth century the
controlling force of the environm ent was clearly recognized. Ben-
tham and M arx have been called behaviorists, although for them
the environment determ ined behavior only after first determ ining
consciousness, and this was an un fortunate qualification because the
assumption of a m ediating state clouded the relation between the
terminal events.
T h e role of the environm ent has become clearer in the present
century. Its selective action in evolution has been exam ined by
8s TH E SCIENCE OF BEHAVIOR
the ethologists, and a sim ilar selective action during the life of the
individual is the subject of the experim ental analysis of behavior.
In the current laboratory, very com plex environm ents are con
structed and their effects on behavior studied. I believe this work
offers consoling reassurance to those who are reluctant to abandon
traditional form ulations. U nfortunately, it is not w ell known out
side the field. Its practical uses are, however, begin nin g to attract
attention. Techniques derived from the analysis have proved useful
in other parts of biology— for exam ple, physiology and psycho
pharm acology— and have already led to the im proved design of
cultural practices, in program m ed instructional materials, contin
gency m anagement in the classroom, behavioral m odification in
psychotherapy and penology, and m any other fields.
M uch remains to be done, and it w ill be done more rapidly
when the role of the environm ent takes its proper place in com
petition w ith the apparent evidences of an inner life. As D iderot
put it, nearly 200 years ago, “ U nfortunately it is easier and shorter
to consult oneself than it is to consult nature. T h u s the reason is
inclined to dwell w ithin itself.” B u t the problem s we face are not
to be found in men and wom en bu t in the w orld in w hich they live,
especially in those social environm ents we call cultures. It is an
im portant and prom ising shift in emphasis because, un like the
remote fastness of the so-called hum an spirit, the environm ent is
w ithin reach and we are learning how to change it.
A n d so I return to the role that has been assigned to me as
a kind of twentieth-century C alvin, calling on you to forsake the
primrose path of total individualism , of self-actualization, self
adoration, and self-love, and to turn instead to the construction of
that heaven on earth w hich is, I believe, w ithin reach of the
methods of science. I wish to testify that, once you are used to it,
the way is not so steep or thorny after all.
Can W e Profit from
Our Discovery
of Behavioral Science?
i
i
\
i
Many things are happening today that seem com pletely sense
less, irrational, insane. T h e population of many countries has been
allowed to reach a point at w hich two or three bad harvests w ill
mean death by starvation for tens or even hundreds of m illions
of people. T h e U n ited States and Russia spend a staggering part
of their incomes on' the production of m ilitary systems w hich every
one hopes w ill never be used and w ill therefore prove to be a total
waste. O ur supplies of energy and many critical m aterials are surely
running out, but we have done very little to curtail current or
future use. T h e environm ent grows steadily less habitable.
People have always been thoughtless and short-sighted, b u t
can we continue to excuse ourselves by saying so? T h e hum an
species has emerged trium phant in a long com petition w ith other
species. Its members can acquire behavioral repertoires of a unique
and extraordinary com plexity. V erbal behavior was perhaps its
greatest achievem ent, and it led to the social environm ents w hich
have produced art, literature, religion, law, and science. W ith the
technologies of physics and biology the species has solved problem s
of fantastic difficulty. Yet w ith respect to its own behavior some-
thing always seems to go wrong. It is easy to understand why the

question should be asked: “ W hen shall we have the behavioral
science and technology we need to solve our problems?"
I believe that that is the w rong question and that we should
be asking: “ W hy do we not use the behavioral science we already
have?” Consider the position of an agricultural specialist visiting a
developing country. He secs farmers plantin g varieties of grain
w hich are not best suited to the soil, rainfall, or clim ate or the most
resistant to disease. He sees them using too little fertilizer, or
fertilizer of the wrong kind. T h e y are cultivating and harvesting
w ith prim itive equipm ent, and processing and storing food in
wasteful ways. If they then ask him , “ W hen shall we have the
agricultural science we need to m ake better use of our land?” must
he not reply, “ W hy are you not using the science w hich already
exists?”
T h e re could be m any answers. Special seed, fertilizer, m a
chinery, and storage space are costly. If m oney is available, those
who have it must be convinced that spending it w ill bring results.
N ew methods often throw people ou t of work and take control out
of the hands of those who have profited from the old. B u t there
is a special kind of explanation that is more im portant. W e have all
heard stories of third w orld farmers who change to new methods
w hile they are being dem onstrated only to change back as soon as
the would-be reform er leaves. T h e stories may be apocryphal, but
they are easy to believe because people do persist in doing things
as they have always done them and the entrenched ways do post
pone or block any advance toward som ething better.
W e have no reason to feel superior to those w ho reject better
methods of agriculture, for we are doing m uch the same thing with
respect to behavioral science. T h e parallel w ith agriculture breaks
down, because I cannot point to any part of the w orld today in
which a behavioral technology flourishes, b u t recent advances in a
science of behavior have led to substantial achievem ents in the
m anagem ent of hum an behavior in such special fields as govern
ment, industry, schools and colleges, institutions for the care of
psychotic and retarded people, and personal and fam ily counseling.
I shall not review this work, or try to indicate how extensive it is.
I simply w ant to ask why it is not more w idely accepted iri the
solution of our problems.
Can We Profit from Our Discovery of Behavioral Science? 85
M oney is again relevant, not because we are poor bu t because

changes have consequences w hich are essentially economic. A n d
again it is true th at those who make the decisions to spend money—
for exam ple, in education— are often unaw are of w hat can be done.
But the m ain.obstacle is, again, the entrenchm ent o f old practices—
in this case of old ways of thin kin g about hum an behavior. A n ti
quated theories ingrained in our language and our culture stand in
the way of prom ising scientific alternatives.
It is not so m uch the com plexity of hum an behavior that
causes trouble as the traditional practice of looking for explanations
inside the behaving person. People are said to act as they do because
of their feelings, their states of m ind, their intentions, purposes,
and plans. T h ey act because they w ill to act. T h e ir roles in society
are reflections of inner selves or personalities.
A science 6f behavior must look elsewhere. It turns to the
environm ent— the environm ent that has produced the genetic en
dowment o f thet species through n atural selection and that now
shapes and m aintains the repertoire of the individ u al through
another selective'process called operant conditioning. By analyzing
these two roles of the environm ent we can begin to understand
behavior and, by changing the environm ent, to m odify it.
Evidence of the pow erful control exerted by the environm ent
is obtained only through rather subtle scientific practices and is by
no means as im m ediate or as obvious as the evidence that seems to
support the traditional view. H ow we feel— or, more precisely, how
our bodies feel to us— is a salient part of the situations in w hich
we take action, and we are therefore likely to regard the feelings as
the causes of action. A n d since we usually assume that when others
act as we do, they feel as we feel, it is not surprising that w hen we
want them to act in a given way, we try to make them feel as we
feel w hen doing so. T h e evidence is convincing because it is fa
miliar, but twenty-five hundred years of speculation about it have
not led to a convincing m entalistic (or, to use a vogue word, cog
nitive) theory, and our practical failures are obvious enough in our
present difficulties.
L et us look at some exam ples in w hich this abiding concern
for an inn er explanation has diverted attention from environm ental
measures w hich m ight have brought us closer to solving our prob
lems. A feelin g or state of m ind fam iliar to everyone is confidence.
T h e term is useful in daily com m unication. As a behaviorist I do

not blush to say that I am at this mom ent possessed of a num ber of
different feelings of confidence, and I shall list a few in order of
degree. I have complete confidence that this chair and desk w ill hold
me as I write. I have a fair degree of confidence that the words I
am w riting w ill eventually reach readers. I have some confidence
that a num ber of those who start reading w ill finish the paper,
and just a touch of confidence that some of them w ill come
to behave in a slightly different w ay because of w hat they read.
I thus report certain conditions of my own body. B ut I hasten
to point out that the degree of my confidence is related to the
extent of my past successes and failures. Sim ilar desks and chairs
have always held me. Sim ilar w riting has usually been published.
A good m any readers finish my papers, and readers sometimes
change their behavior because of w hat I write. I am w riting a paper
at my desk at this m om ent because I have done m uch the same
kind of thing w ith some success on fairly sim ilar occasions. For the
same reasons I have certain feelings of confidence. B u t I am w riting
because of the consequences, not because of the feelings. My feelings
and my behavior are collateral products of my personal history.
T h e point is im portant when the word is used in discussing
practical affairs. T h e B u lletin of the Royal Society of A rts in
London recently reported the remarks of a speaker who had dis
cussed the appearance of the British countryside. H e had told his
audience that “ the key to the survival of our present landscape lay
in the word ‘confidence’— w ith out w hich people w ould not plant
trees.” In the past year or two, he said, “ confidence had been com
pletely destroyed.” But the im portant fact was sim ply that people
no longer planted trees. W hy did they not do so? It is not difficult
to point to relevant facts. People move about a great deal these
days, and when they do so they never w atch a tree they have
planted grow to m aturity. T rees are likely to be w antonly destroyed
as new roads are put through and properties broken up for housing
developments. M ore people now live in cities where the governm ent
plants the trees. Surely changes of this sort are the real key to the
survival of a landscape. T h e y m ake it less likely that people will
plant trees, but why should we say that they first destroy confidence
in tree planting? Confidence is not a key to anything.
In the U nited States there is a Conference Board w hich

reports on “ Consum er Confidence in the Econom y.” A recent
decline in its index was called the “ first significant sag in consumer
spirits since 1974.” W h at the board had discovered was sim ply
that,Am ericans were p lann ing to buy fewer cars, television sets, and
so on. W hy attribute this to a decline in “ confidence” let alone
“sagging spirits” ? T h e board m ade the helpful suggestion that “ the
major culprit may be inflation .” People plan to buy less when
their money does not go as far, and when, as a collateral effect,
they also lack a feeling of confidence. A move to “ restore confidence
in the econom y” is really a m ove to encourage buying. T h e govern
ment may, for example, give people more m oney— say, by reducing
taxes— and they are then more likely to buy things.
In speaking of plan tin g trees or buyin g television sets, the
term confidences may sim ply refer to the likelihood that a person
w ill act in a particular way. Perhaps the feeling is not really m eant
to be an explanation. B ut w hen the behavior cannot be so easily
specified, feelings assume a m ore im portant role. Confidence begins
to be treated as 'if it really were a cause, and that is when in q u iry
into valid explanations is obscured.
A year or two ago, in a newsweekly, D avid E. L ilien th al
discussed “ the prevailing Am erican m ood,” which he said “ has
become negative and fearful.” It is “ a mood of self-doubt and fear
which paralyzes the very will-to-act,” and, he added, it is the will-to-
act that alone can “ remove the causes of fear and lack of confi
dence.” W h at Am erica needed was m ore confidence, and L ilien th al
offered the Tennessee V alley A u th ority as an exam ple that proved
his point. In the early thirties the soil in the valley of the T en n es
see R iver had lost its fertility, the forests had been almost destroyed,
the land was eroding, and nothin g m uch could be done. People
were idle and poor. A fter the dams were built, electric power and
fertilizer were available, and the people turned to new m ethods of
agriculture and restored the land. T h e ir incomes rose, and the
valley became green. L ilien th a l attributes this highly desirable
change to “ restored self-confidence.” B u t surely it was the dams
and their products that m ade the difference! People began to do
things they could not do before, and, being successful, they no
doubt felt confident. (Lilienthal cited Iran as another exam ple, and
insisted that what was happening there was due to a restoration

of a national faith and self-confidence, but a connection w ith the
price of oil is certainly not to be overlooked.)
W hen we speak of a n ation ’s confidence in itself, the behavior
at issue is m uch more complex than planting trees or b u yin g cars,
and we are therefore m uch more likely to give confidence a power
o f its own. B u t if we are content to say that all Am erica needs today
is a new spirit of confidence, we shall neglect the things that can
actually be done to bring about the changes we desire.
To take confidence one step further, Senator D aniel P.
M oynihan recently considered “ the contention by some intellectuals
here and abroad that democracy is declining as a force in world
affairs.” W orld leaders seem to be turning less and less often to
dem ocratic processes to solve their problems. T h e London Tim es
made m uch the same point in a recent issue: “ T h e great democra
cies are themselves partly losing faith in democracy . . . the U nited
States, w hich fifteen years ago was still the dom inant w orld power,
is now only one of the super-powers, the stronger but in im por
tant ways the less self-confident of the tw o” (emphasis added). And
a national poll reports that pu b lic confidence in our institutions—
inclu d ing the Congress, the executive branch of the federal govern
ment, and corporations— has sunk to its lowest point since the poll
began to be taken. A n d so, Senator M oynihan says, “ W e ask our
selves . . . whether hope is not fading and w hether it w ill not
continue to fade unless there is some restoration of confidence and
w ill on our part.”
B ut how can one restore confidence in the dem ocratic process?
T h e President cannot ask drug companies to prepare m illions of
injections o f confidence— like so m any inoculations for influenza.
W e want people to discuss issues, work for candidates, and go to
the polls, bu t they w ill do so, not because they have confidence in
the dem ocratic process, but because certain kinds of consequences
then follow . Potential voters stay hom e when they have nothing
to show for their trouble. V o tin g for the candidate who lost the
1972 presidential election was not, as we say, reinforced, nor, as it
turned out, was voting for N ixon, the candidate w ho won. T h e
behavior o f w orking for, giving to, and voting for a candidate for
President underw ent a well-known process called, if I may be
technical, extinction. T o reverse it we need to m ake sure that the

political behavior of the citizen w ill again have reinforcing conse
quences. T h e dem ocratic process works when it makes a difference
whether people participate or not.
Confidence is only one of hundreds of terms referring to
feelings or states of m ind w hich come to us n aturally and con
veniently in daily discussions of hum an behavior bu t w hich by their
very nature are inim ical to a scientific approach. Perhaps no single
instance causes serious trouble, because more productive measures
are taken when im portant consequences are at stake, b u t the gen
eral practice of thinking about behavior in this w ay discourages
inquiry into the role of the environm ent, and explains, I believe,
why we are still in such trouble.
T h e difficulty is com pounded by the fact that behavioral
scientists themsehies are often not free of the old tendencies. M any
established concepts suffer from the same shortcomings. In its
psychological use^ the term attitude is an exam ple. It m ay have
begun as a metaphor,\ Since things usually fall in the direction in
which they are leaning, we say that a person leans tow ard one p oliti
cal candidate or another. T o be inclined to do som ething is also
to be in an attitude slightly off the perpendicular. B u t at some
point the term began to refer to an internal state. A distinguished
economist, Sir A rth u r Lewis, tells us that “ econom ic grow th de
pends on attitudes to work, to wealth, to thrift, to havin g children,
to invention, to strangers, to adventure.” H e could just as well
have said that economic grow th depends on w hether people work,
acquire wealth, save, have children, invent things, get along w ith
strangers, and explore the world. Should we then say that by “ atti
tude” he simply means the likelihood that people w ill behave in
these ways? No, Sir A rth u r forgoes that chance of exculpation, for
he continues, “ and all these attitudes flow from deep springs in the
human m ind.” For him attitudes are more than probabilities of
action; they are m ental forces. People work, save, spend, bear
children, and so on because of their attitudes. If that is so, we can
account for economic grow th only by looking into the hum an m ind,
but we should then look away from the external circumstances
about w hich something can be done— the econom ic conditions
under which people work, acquire goods, and save, and the social
conditions under which they have families and treat each other
w ell or badly. B y speaking of attitudes we draw attention away
from the role of the environm ent in economic growth.
Another feeling or state of m ind with a secure place in social
science is alienation. W orkers on a production line com plain of
being unhappy, they often stay away from work, and they fre
quently strike or quit. T h ey are said to do so because, am ong other
things, they are alienated. Should we not then study what alienation
feels like? T h e contribution of one authority in the field has been
described as follows: “ [He] does not deny that the causes of aliena
tion lie elsewhere, outside of the individual; either in the environ
ment, or in the relation individual-environm ent. B ut he does well
to insist on alienation itself as a subjective state of an individual,
to be distinguished sharply from alienating social structures.
H avin g made this distinction, it can then become a m atter of argu
ment whether one should concentrate prim arily on alienation itself
[as a subjective state] or on alienating conditions in the social
structure; on the phenom enon itself or on its causes.” B u t the
all-im portant effect is then ignored. T h e problem arises because
certain social structures lead people to behave in certain ways; they
may also generate feelings, but that is a collateral effect.
Feelings and states of m ind m ay usurp this role in the causa
tion of behavior m ainly because we respond to our own bodies
w hile we are responding to the w orld around us, bu t there are
other reasons. A s Freud pointed out, we often act w ithout having
relevant feelings, in which case we should look for other causes.
B u t Freud is probably responsible for the fact that we look in the
w rong place— deep in a person’s m ental life. H e made a great deal
of the depth of psychoanalysis, as linguists do of deep structures,
and this sense of probing makes a behavioral analysis seem super
ficial and the appeal to feelings especially profound. A discussion
of the hum an rights m ovem ent in Russia contains the follow ing
question: “ C an the consciousness of the Soviet citizenry, and of the
Soviet bureaucracy, be brought to the point where the one de
mands, and the other provides, the ru le of law?” W h at is “ conscious
ness” doing there? W hy not say simply, “ C an Soviet citizens be
induced to dem and, and the Soviet bureaucracy to provide, the rule
of law?” W ith the term consciousness the w riter alludes to some
thin g beyond or beneath the behavior itself. A n d indeed something
beyond the behavior does need to be taken into account— b u t in

certain governm ental and social systems, rather then deeply rooted
feelings. :;
I recently received a letter that began: “ H ave you ever
thought about the great reservoir of feeling against w ar that exists
throughout the world? It is being-wasted! B ottled up, w ith n othing
to show for it in terms of progress towards real peace . . . let us
release this great reservoir of feelin g.” W e all know what the
w riter is talking about. W e read the newspapers and watch news
reports on television and are m oved to act, bu t we find there is
little we can do. It does seem as if som ething havin g to do w ith
war is indeed “ bottled up,” and that it w ould be w onderful if we,
and m illions like us, could pull the cork. B u t it w ould be action
not feelings that poured out. W e should have found som ething to
do to prevent War. B ut w ould m y correspondent have bothered to
write simply to say: “A re you aware of how many people in the
world w ould dq som ething to stop war if they could? L et us give
them a chance?’.’ T h a t kind of statement cries out for the response
“ B u t how?” If anything can be done, it w ill be done not by re
leasing feelings but by specifying the steps to be taken to bu ild a
peaceful world. Pent-up feelings and floodgates w aiting to be
opened are pow erful metaphors, but they do not tell us w hat to do.
A research docum ent published by the International Peace
Research Association elaborates upon a fam ous statem ent issued by
U N E S C O m any years ago: “ W ars begin in the minds of men; hence
it is in the minds of men that the defense of peace must be con
structed.” B u t how are we to m ove into the minds of m en (and
presum ably women) and of w hat are we to construct the defenses
of peace? T h e relevant facts lie in the outside world. W ars begin in
many places and for m any reasons— overcrowding, com petition for
world trade, border disputes, concentrations of m ilitary power,
racial and national claims and counterclaims, the un equal distribu
tion of w ealth . . . we know at least some of the things to be done
about matters of that sort. C all it a “ superficial” analysis if you
will, but to turn instead to the minds o f men, no m atter how deeply
im planted they may be, is to abandon any hope of solution.
It is not only a spurious profundity that promotes a men-
talistic account. C ertain advantages are gained from its weaknesses.
T hose who must make im portant decisions suffer w hen things go
wrong. T h ey are held responsible for their action in the sense that
they w ill be punished if they fail. A convenient way to avoid
punishm ent is to call for a change of m ind rather than action.
Thus, to say that Am erica needs m ore of the kind of confidence
attributed to the Tennessee V alley Project is safer than to say that
Am erica needs more projects of the same kind. N o one objects
to a call for confidence, but a proposal to build more dams and
plants may be harshly received. Sim ilarly, those who call for more
confidence in the dem ocratic process are not necessarily ready to
support changes in the conduct of elections, in the m ethods of
financing candidates, in lobbying practices, or in any of the other
conditions w hich underm ine a dem ocratic system. A n d m any of
those who call for new attitudes toward work, thrift, and the fam ily
may hesitate to advocate the social changes w hich w ould effectively
induce people to work, save, and have more or fewer children. It
is safe to call for changes in feelings and states of m ind precisely
because n othing w ill ever happen for w hich one can be held
responsible.
Feelings play a different and possibly more destructive role
when they are taken, not as causes, bu t as values, not as preceding
behavior but follow ing it. N utritious food is essential to the sur
vival of the individual; is it not therefore extrem ely im portant that
it taste good? Sexual behavior is essential to the survival of the
species; is it not extrem ely im portant that sexual contact feel good?
B u t the im portant thin g for the ind ivid u al and the species is not
how things taste or feel but whether they are reinforcing— that is,
whether they strengthen the behavior upon w hich they are con
tingent. Susceptibilities to reinforcem ent have presum ably evolved
because of their survival value. W hen, through a m utation, an
organism ’s behavior is more strongly reinforced by nutritious food
or sexual contact, the organism is m ore likely to get the food it
needs and to have offspring. T h e increased susceptibility to rein
forcem ent is then contributed to the species. T h e im portant thing
is that the susceptibility should survive. T h e feelings are incidental.
T h e same thin g is true of the social reinforcers, w hich are
more likely to be called values. People are said to treat each other
in ways which express compassion and love and w hich inspire
gratitude, bu t the im portant thing is the contribution to the func
tioning of the social environm ent or culture. T h e behavior we call
ethical makes a group function m ore effectively. T h e feelings or

states of m ind associated w ith it are collateral products.
Happiness is a feeling often taken as a value. W e struggle
to achieve happiness, but an inquiry into how it feels to be happy
may not help. O n the contrary, it m ay make the struggle less
successful. W e often feel happy w hen we behave in ways leading
to the possession of goods, and we then m istakenly take the posses
sion to be the cause of the feeling. W e m ake the same mistake when
we act to make others happy by givin g them good things. W h ole
philosophies of governm ent have been tested on the theory that if
goods are distributed “ to each according to his need” people w ill
be happy. B ut happiness is the accom panim ent of successful action
rather than of w hat the action brings. It is characteristic of getting
rather than possessing. Possession leads to happiness only when it
makes further action possible. W hether or not people are happy is
of great political significance, but a subjective measure of the
quality of life w ill'do little m ore than tell us whether a given change
should be made. *t x
Another feeling said to be im portant as a value is freedom —
one of the great hum an goals and possibly the most im portant issue
in the world today. W e gain little from analyzing how it feels to
be free; the im portant things are the conditions under w hich we
do so. W hen we escape from punishm ent or the threat of punish
ment, we say that we are free, and feel free, b u t we are still under
the control of othe^ parts of the environm ent, particularly the kinds
of consequences called positive reinforcem ent. It is a kin d of control
under w hich we do, as we say, what we want to do. W e do not try
to escape from a world in w hich such control prevails; on the con
trary, we work to preserve it. Nevertheless, as I argue in Beyond
Freedom and D ignity, the m isunderstanding of freedom that comes
from a concentration on feelings has actually perpetuated pun itive
practices and slowed the advance tow ard a w orld in w hich people
will feel freer than ever before.
B ut what about the sheer enjoym ent of life— the pleasurable
feelings we get from, say, art, music, and literature? Surely the
important things here are feelings. B u t we must not neglect the
things we do in enjoying life. W e listen to music, look at a picture,
read a book, and these are modes o f behaving. W e stop listening
and looking when we do not, as we say, enjoy the consequences. W e
call some of the things we continue to look at or listen to beautiful,

but we could as well say reinforcing. W e listen-again and again to
music we find beautiful, we look again and again at a picture we
find beautiful. A study w hich confined itself to beautiful things, or
to how beautiful things made us feel, w ould not tell us how to
explore the world in order to find or create more of them. N or
would it tell us why we are less likely to defect from a group and
more likely to defend it and im prove it if m any things about it are
reinforcing in the ways we call beautiful.
A n d so, in general, we enjoy life and call the w orld beautiful
and ourselves free and happy when our behavior leads to an
abundance of good things. No structural account of the things
themselves or any analysis of the feelings w hich arise w hen be
havior is strengthened by them w ill help us make life m ore enjoy
able and ourselves freer and happier. W e must turn instead to the
contingent relations between behavior and its consequences.
Such a program does not rob people of their feelings. It
simply puts feelings in their proper place, and in doing so moves
more rapidly to the kin d of environm ent in w hich they can be
enjoyed. In refusing to accept feelings and states of m ind as causes,
we do not make the behavior that is said to follow from them any
the less im portant. Instead we make it possible to deal w ith be
havior more successfully.
It is clear that the behavioral sciences have not yet fulfilled

their promise. T h ere are economists who question w hether there is
a science of economics, and if we can judge by international strate
gies in the w orld today, governm ents m ake little use of political
science. Anthropologists, sociologists, and social psychologists grow
increasingly uneasy about their fields. (One w riter has said that
sociology is suffering from a “ crisis of confidence!” ) In most of
these fields there is no shortage of facts, and efforts are continually
made to discover m eaningful relations am ong them, m athem atical
or otherwise. W hat is missing is a coherent theory of human
behavior.
T h e fau lt lies, I am arguing, w ith a surviving m entalism . T h e
sooner we abandon explanations of behavior in terms of feelings
and states of m ind the sooner we shall turn to the genetic and
environmental conditions of w hich behavior is a function. Enough
is already kn'own about those conditions to assure reasonable suc
cess in the interpretation, prediction, and control of hum an be
havior. A refusal to take advantage o f w hat is w ithin reach could
mean the difference between the survival and the destruction of our
civilization or even the species.
Th ere are those who w ill say that such a cause is surely in
commensurate w ith such an effect. A m entalistic philosophy is
rather inoffensive, and it need not seriously handicap practical peo
ple. A m I not exaggerating its importance? B u t there must be some
reason why we are not m aking the technological advances in the
m anagement of hum an behavior w hich are so obvious in other
fields, and the, reason could be our lingerin g com m itm ent to the
individual as ah in itiatin g agent. It is of the very nature of hum an
behavior that seemingly trivial causes have profound effects, and
there is a historical exam ple w hich I am inclined to take seriously.
I am not a historian nor do I usually trust arguments based upon
history but in this instance the evidence is, I think, persuasive.
From the fifth century B .C . to about 1400 a .d . C hina was as
advanced in physical technology as any part of the world. T h e
recent exhibition of early Chinese pottery and ceramic and bronze
sculptures sent around the w orld by the Chinese governm ent shows
an art and a technology fully equal to those of the Greeks of the same
period. A com parable position was m aintained for nearly two
thousand years. T h e n three great Chinese inventions— the com
pass, gunpowder, and m oveable type— brought about extraordinary
change. B ut not in China! G unpow der was of little practical use
because Chinese m ilitary activities were cerem onial and largely under
the control of astrologers. L on g sea voyages were forbidden, and
coastwise shipping gained little from the compass. T h e Chinese
system of notation, w ith its thousands of characters, could not take
advantage of m oveable type. It was the W est w hich seized upon
these three great Chinese inventions and exploited them w ith
extraordinary results. W ith the compass the W est explored the
world and w ith gunpowder conquered it. M oveable type and the
printing press brought the revival of learning and the spread of
Western thought. A n d w hile all this was happening, C h in a re
m ained a m edieval society.1 C ertain rather inoffensive cultural prac

tices had deprived it of the benefit of its own discoveries.
Is it possible that som ething of the same sort is happening
again, and that this time W estern culture w ill suffer from essentially
ceremonial, astrological, and geom antic practices? W ill China, for
tunately untouched by the G reek “ discovery of the m ind,” now take
over the behavioral equivalents of compass, gunpowder, and move-
able type and dom inate a new era? O r is it not too late? C an we
begin at last to profit from our discovery of behavioral science and
use it to share in the solution of the problems facing the world
today? T h a t is the question.
1 M cN eill, W illiam . T h e rise of the west. C hicago: U n iversity of

C hicago Press, 1963.
J
8
W hy I Am Not
a Cognitive Psychologist
!
t
T h e variables of w hich hum an behavior is a function lie in

the environment. W e distinguish between (1) the selective action
of that environm ent during the evolution of the species, (2) its
effect in shaping and m ain tainin g the repertoire o f behavior w hich
converts each member of the species into a person, and (3) its role
as the occasion upon which behavior occurs. C ogn itive psychologists
study these relations between organism and environm ent, b u t they
seldom deal w ith them directly. Instead they invent internal surro
gates w hich become the subject m atter of their science.
T a ke, for exam ple, the so-called process of association. In
P avlov’s experim ent a hungry dog hears a bell and is then fed. If
this happens m any times, the dog begins to salivate when it hears
the bell. T h e standard m entalistic explanation is that the dog
“ associates” the bell w ith the food. B u t it was P avlov who associated
them! “ Associate” means to jo in or unite. T h e dog merely begins
to salivate upon hearing the bell. W e have no evidence that it does
so because of an internal surrogate of the contingencies.
In the “ association of ideas” the ideas are internal replicas of
stim uli to w hich I shall return. If we have eaten lemons, we may
97
taste lemon upon seeing a lem on or see a lemon upon tasting lemon
juice, but we do not do this because we associate the flavor w ith
the appearance. T h e y are associated in the lemon. “ W ord associ
ations” are at least correctly named. If we say “ hom e” when some
one says “ house,” it is not because we associate the two words bu t
because they are associated in daily English usage. C ognitive asso
ciation is an invention. Even if it were real, it w ould go no further
toward an explanation than the external contingencies upon which
it is modeled.
A nother exam ple is abstraction. Consider a simple experi
ment. A hungry pigeon can peck any one of a num ber of panels
bearing the names of colors— “ w hite,” “ red,” “ blu e,” and so on,
and the pecks are reinforced w ith small amounts of food. A n y one
of a num ber of objects— blocks, books, flowers, toy animals, and
so on— can be seen in an adjacent space. T h e follow ing contin
gencies are then arranged: whenever the object is white, no m atter
w hat its shape or size, pecking only the panel m arked “ w h ite” is
reinforced; w henever the object is red, pecking only the panel
marked “ red” is reinforced; and so on. U nder these conditions the
pigeon eventually pecks the panel m arked “ w hite” when the object
is white, the panel m arked “ red” w hen the object is red, and so on.
C hildren are taught to name colors w ith sim ilar contingencies, and
we all possess com parable repertoires sustained by the reinforcing
practices of our verbal environments.
B ut what is said to be going on in the mind? K arl P o p p e r 1
has put a classical issue this way: “ W e can say either that (1) the
universal term “ w h ite” is a label attached to a set of things, or that
(2) we collect the set because they share an intrinsic property of
“ whiteness.” Popper says the distinction is im portant; natural
scientists may take the first position but social scientists must take
the second. M ust we say, then, that the pigeon is either attaching
a universal term to a set of things or collecting a set of things
because they share an intrinsic property? Clearly, it is the experi
m enter not the pigeon w ho “ attaches” the w hite key to the white
objects displayed and who collects the set of objects on w hich a
single reinforcing event is made contingent. Should we not simply
1
1 P opper, K. Poverty o f historicism. Lo n do n , 1957. i
Why I Am Not a Cognitive Psychologist 99
attribute the behavior to the experim ental contingencies? A n d if

so, why not for children or ourselves? B ehavior comes under the
control of stim uli under certain contingencies of reinforcem ent.
Special contingencies m aintained by verbal com m unities produce
“ abstractions.” W e do attach physical labels to physical things and
we collect physical objects according to labeled properties, but
com parable cognitive processes are inventions which, even if real,
w ould be no closer to an explanation than the external contin
gencies.
Another cognitive account of the same data w ould assert that
a person, if not a pigeon, forms an abstract idea or develops a
concept of color. T h e developm ent of concepts is an especially
popular cognitive field. (T h e h orticultural m etaphor minimizes
contributions from the environm ent. W e may hasten the grow th of
the m ind but w*e are no more responsible for its final character than
farmers for the character of the fruits and vegetables they so care
fully nourish.) *C olor vision is part of the genetic endowm ent of
most people, and it develops or grows in a physiological sense,
possibly to some extent after birth. Nevertheless, most stim uli
acquire control because of their place in contingencies of rein
forcement. As the contingencies becom e more com plex, they shape
and m aintain m ore com plex behavior. It is the environm ent that
develops, not a m ental or cognitive possession.
A passage from a recent discussion of the developm ent o f
sexual identity in a child m ight be translated as follows: “ T h e
child forms a concept based upon w hat it has observed and been
told of w hat it means to be a boy or girl.” (A child ’s behavior is
affected by w hat it has observed and been told about being a boy
or girl.) “ T h is concept is oversim plified, exaggerated, and stereo
typed.” (T he contingencies affecting the behavior are simplified
and exaggerated and involve stereotyped behavior on the part of
parents and others.) “ As the child develops cognitively, its concepts,
and consequently its activities, becom e m ore sophisticated and
realistic.” (As the child grows older, the contingencies become more
subtle and more closely related to the actual sex of the child.)
Children do not go around form ing concepts of their sexual identity
and “ consequently” behaving in special ways; they slowly change
their behavior as people change the ways in w hich they treat them
lOO T H E SCIENCE OF BEHAVIOR
because of their sex. Behavior changes because the contingencies

change, not because a mental entity called a concept- develops.
M any m entalistic or cognitive terms refer not only to. con

tingencies but to the behavior they generate. T erm s like “ m ind,”
“ w ill,” and “ thought” are often sim ply synonyms of “ behavior.”
A historian writes: “ W hat may be called a stagnation of thought
prevailed, as though the m ind, exhausted after b u ild in g up the
spiritual fabric of the M iddle Ages, had sunk into inertia.” Exhaus
tion is a plausible metaphor when a q u iet period follows an active
one, but it was behavior that becam e stagnant and inert, pre
sum ably because the contingencies changed. C ertain social condi
tions (“ the spiritual fabric of the M iddle Ages” ) made people active.
A second set of conditions, possibly produced by the very behavior
generated by the first, made them m uch less so. T o understand what
actually happened we should have to discover why the contingencies
changed, not w hy thought became stagnant or inert.
Behavior is internalized as m ental life when it is too slight to
be observed by others— when, as we say, it is covert. A w riter has
pointed out that “ the conductor o f an orchestra m aintains a cer
tain even beat according to an internal rhythm , and he can divide
that beat in h alf again and again w ith an accuracy rivalin g any
mechanical instrum ent.” B u t is there an internal rhythm? Beating
time is behavior. Parts of the body often serve as pendulum s useful
in determ ining speed, as when the am ateur m usician beats time
w ith a foot or the rock player w ith the w hole body, but other well-
timed behavior must be learned. T h e conductor beats time steadily
because he has learned to do so under rather exacting contingencies
of reinforcem ent. T h e behavior may be reduced in scale u n til it is
no longer visible to others. It is still sensed by the conductor, but
it is a sense of behavior not of time. T h e history of “ m an’s develop
m ent of a sense of time” over the centuries is not a m atter of
cognitive growth bu t of the invention of clocks, calendars, and ways
of keeping records— in other words, of an environm ent that “ keeps
tim e.” *
W hen a historian reports that in a given period “ a wealthy,
brilliant, and traditional governing class lost its w ill,” he is re
porting sim ply that it stopped acting like a wealthy, brillian t, and
traditional governing class. Deeper changes are suggested by the

term “ w ill” but they are not identified. T h e y could not have been
changes in particular people, since the period lasted more than one
lifetim e. W hat changed were presum ably the conditions affecting
the behavior of members of the class. Perhaps they lost their money;
perhaps com peting classes becam e m ore pow erful.
Feelings, or the bodily conditions we feel, are com m only taken
as the causes of behavior. W e go for a w alk “ because we feel like
going.” It is surprising how often the fu tility of such an explanation
is recognized. A distinguished biologist, C. H . W ad din gton ,2 re
view ing a book by T in b ergen , writes as follows:
It is not clear how far he T in b ergen w ould go along

w ith the argum ent of one of the most perceptive critical
discussions ofj ethology by Suzanne Langer, w ho argues
that each step* in a com plex structure of behavior is con
trolled, not by a hierarchical set of neural centers, but
by the im mediate feelings of the anim al. T h e anim al,
she claims, does the n ext thin g in the sequence, not to
bring about a' useful goal, or even as a m ove toward an
enjoyable consummation, b u t because it actually feels
like doing it at the moment.
Evidently W addington him self goes along partw ay w ith this “ per
ceptive view .”
B ut suppose Langer is right. Suppose anim als sim ply do w hat
they feel like doing? W h a t is the next step in exp lain in g their
behavior? Clearly, a science of anim al behavior m ust be replaced
or supplemented by a science of anim al feelings. It w ould be as
extensive as the science of behavior because there w ould presum ably
be a feeling for each act. B u t feelings are harder to id en tify and
describe than the behavior attributed to them, and we should have
abandoned an objective subject m atter in favor of one of dubious
status, accessible only through necessarily defective channels of
introspection. T h e contingencies w ould be the same. T h e feelings
and the behavior w ould have the same causes.
A British statesman recently asserted that the key to crime in
the streets was “ frustration.” Yoxmg people m ug and rob because
they feel frustrated. B u t w hy do they feel frustrated? One reason
2 W ad d in gto n , C. H ., N ew York R eview , F ebru ary 3, 1974.

may be that many of them are unem ployed, either because they do
not have the education needed to get jobs or because jobs are not
available. T o solve the problem of street crime, therefore, we must
change the schools and the economy. But what role is played in all
this by frustration? Is it the case that when one cannot get a job
one feels frustrated and that w hen one feels frustrated one mugs
and robs, or is it simply the case that when one cannot earn money,
one is more likely to steal it— and possibly to experience a bodily
condition called frustration?
Since many of the events w hich must be taken into account in
explaining behavior are associated w ith bodily states that can be
felt, what is felt may serve as a clue to the contingencies. B u t the
feelings are not the contingencies and cannot replace them as
causes.
By its very nature operant behavior encourages the invention

of m ental or cognitive processes said to initiate action. In a reflex,
conditioned or unconditioned, there is a conspicuous prior cause.
Som ething triggers the response. B u t behavior that has been posi
tively reinforced occurs upon occasions which, though predisposing,
are never com pelling. T h e behavior seems to start up suddenly,
w ithout advance notice, as if spontaneously generated. H ence the
invention of such cognitive entities as intention, purpose, or will.
T h e same issues were debated w ith respect to the theory of evolu
tion and for the same reason: selection is a special causal m ode not
easily observed. Because controlling circumstances w hich lie in an
organism ’s history of reinforcem ent are obscure, the m ental sur
rogate gets its chance. U nder positive reinforcem ent we do, as we
say, w hat we are free to do; hence the notion of free w ill as an
initiatin g condition. (I think it was Jonathan Edwards who said
that we believe in free w ill because we know about our behavior
but not about its causes.)
W hen we do not know why people do one thing rather than
another, we say that they “ choose” or “ make decisions.” Choosing
originally m eant exam ining, scrutinizing, or testing. Etym ologically,
deciding means cutting off other possibilities, m oving in a direction
from which there is no return. Choosing and deciding are thus
conspicuous forms of behavior, b u t cognitive psychologists have
nevertheless invented internal surrogates. A n atole R a p a p o rt3 puts

it this way: ‘A subject in a psychological experim ent is offered a
choice am ong alternatives and selects one alternative over others.”
W hen this happens, he says, “ com m on sense suggests that he is
guided by a preference.” Com m on sense does indeed suggest it, and
so do cognitive psychologists, but w here and w hat is a preference?
Is it anything more than a tendency to do one thing rather than
another? W hen we cannot tell whence the w ind com eth and w hither
it goeth, we say that it “ blow eth where it listeth,” and com m on
sense, if not cognitive psychology, thus credits it w ith a preference.
(List, by the way, is an exam ple of a term w ith a physical referent
used to refer to a mental process. It means, of course, to lean— as in
the list of a ship. A n d since things usually fall in the direction in
which they arq leaning, we say that people lean toward a candidate
in an election hs a rough way of predicting how they w ill vote. T h e
same m etaphor is found in “ in clin ation ” ; we are “ inclined ” to vote
for X. B u t it does not follow that we have internal leanings and
inclinations which affect our behavior.)
“ In ten tion ” is a rather sim ilar term w hich once m eant stretch
ing. T h e cognitive version is a critical issue in current linguistics.
Must the inten tion of the speaker be taken into account? In an
operant analysis verbal behavior is determ ined by the consequences
w hich follow in a given verbal environm ent, and consequences are
what cognitive psychologists are really talkin g about w hen they
speak of intentions. A ll operant behavior “ stretches tow ard” a
future even though the only consequences responsible for its
strength have already occurred. I go to a d rin kin g foun tain “ w ith
the intention of getting a drink of w ater” in the sense that I go
because in the past I have got a drink when I have done so. (I may
go for the first time, follow ing directions, b u t that is not an excep
tion; it is an exam ple of rule-governed behavior, of which m ore
later.)
So m uch for the cognitive internalization of contingencies of

reinforcem ent and the invention o f cognitive causes o f behavior.
3 R a p ap o rt, A . E xp erim en tal games and their uses in psychology.

G en eral L e a rn in g Press, 1973.
Far more dam aging to an effective analysis is the internalization of

the environment. T h e Greeks invented the m ind to explain how
the real world could be known. For them, to know m eant to be
acquainted with, to be intim ate with. T h e term cognition itself is
related to coitus, as in the biblical sense in which a man is said
to know a woman. H avin g no adequate physics of light and sound
nor any chemistry of taste and odor, the Greeks could not under
stand how a world outside the body, possibly some distance away,
could be known. T h ere must be internal copies. Hence cognitive
surrogates of the real world.
T h e distinction between reality and conscious experience has
been made so often that it now seems self-evident. Fred A ttn e a v e 4
has recently written that “ the statement that the world as we know
it is a representation is, I think, a truism— there is really no way in
which it can be wrong.” B u t there are at least two ways, depending
upon the meaning. If the statement means that we can know only
representations of the outside world, it is a “ truism ” on ly if we are
not our bodies but inhabitants located somewhere inside. O ur
bodies are in contact w ith the real world and can respond to it
directly, but if we are tucked away up in the head, we must be
content w ith representations.
A nother possible m eaning is that know ing is the very process
of constructing m ental copies of real things, bu t if that is the case
how do we know the copies? D o we make copies of them? A n d is
that regress infinite?
Some cognitive psychologists recognize that know ing is action
but try to make the poin t by appealing to another m ental surro
gate. Knowledge is said to be “ a system of propositions.” According
to one writer, “ when we use the word ‘see’ we refer to a bridge
between a pattern of sensory stim ulation and know ledge w hich is
propositional.” B u t “ propositional” is sim ply a laundered version
of “ behavioral,” and the “ brid ge” is between stim uli and behavior
and was built when the stim uli were part of the contingencies.
Representational theories of knowledge are m odeled on prac
tical behavior. W e do m ake copies of things. W e construct repre
sentational works of art, because looking at them is reinforced in
4 Attneave, F., American Psychologist, July 1974.

m uch the same way as looking at w hat they represent. W e m ake

maps, because! our behavior in follow in g them is reinforced w hen
we arrive at our destination in the m apped territory. B u t are there
internal surrogates? W hen we daydream , do we first construct
copies of reinforcing episodes w hich we then watch, or do we sim ply
see things once again? A n d when we learn to get abou t in a given
territory, do we construct cognitive maps w hich we then follow or
do we follow the territory? If we follow a cognitive m ap, must we
learn to do so, and will that require a m ap of the map? T h ere is no
evidence of the mental construction of images to be looked at or
maps to be followed. T h e body responds to the world, at the point
of contact; m aking copies w ould be a waste of time.
Know ledge is a key term in cognitive theory, and it covers a

good deal of ground. It is often contrasted w ith perception. W e are
said to be able ,to see that there are three dots on a card bu t only to
know that there are thirteen after counting them, even though
counting is a form of behavior. A fter noting that one spiral can be
seen to be continuous but that another can be discovered to be so
only by tracing, Bela Julesz 5 has said that “ any visual task that
cannot be perform ed spontaneously, w ith ou t effort or deliberation,
can be regarded as a cognitive task rather than as a perceptual one,”
though all the steps in that exam ple are also clearly behavioral.
“ K n ow ing how to do som ething” is an internal surrogate of
behavior in its relation to contingencies. A child learns to ride a
bicycle and is then said to possess know ledge of how to ride. T h e
child’s behavior has been changed by the contingencies of rein
forcement m aintained by bicycles; the child has not taken possession
of the contingencies.
T o speak of know ing about things is also to construct an in
ternal surrogate of contingencies. W e w atch a football game and
are then said to possess know ledge of w hat happened. W e read a
book and are said to know w hat it is about. T h e game and the
book are somehow “ represented” in our minds: we are “ in posses
sion of certain facts.” B ut the evidence is sim ply that we can describe
5 Julesz, B., Scientific A m erican, A p r il 1975.

io6 TH E SCIENCE OF BEHAVIOR
what happened at the game and report what the book was about.
O ur behavior has been changed, bu t there is no evidence that we
have acquired knowledge. T o be “ in possession of the facts” is not
to contain the facts w ithin ourselves but to have been affected by
them.
Possession of knowledge im plies storage, a field in w hich
cognitive psychologists have constructed a great m any m ental sur
rogates of behavior. T h e organism is said to take in and store the
environment, possibly in some processed form. L et us suppose that
a young girl saw a picture yesterday and when asked to describe it
today, does so. W hat has happened? A traditional answer w ould
run som ething like this: when she saw the picture yesterday the
girl form ed a copy in her m ind (which, in fact, was really all she
saw). She encoded it in a suitable form and stored it in her memory,
where it rem ained u n til today. W hen asked to describe the picture
today, she searched her memory, retrieved the encoded copy, and
converted it into som ething like the original picture, w hich she
then looked at and described. T h e account is m odeled on the
physical storage of m emoranda. W e make copies and other records,
and respond to them. B ut do we do anything of the sort in our
minds?
If anything is “ stored,” it is behavior. W e speak of the “ ac
quisition” of behavior, but in w hat form is it possessed? W here is
behavior w hen an organism is not behaving? W here at the present
moment, and in w hat form, is the behavior I exh ib it when I am
listening to music, eating my dinner, talking w ith a friend, taking
an early m orning walk, or scratching an itch? A cognitive psycholo
gist has said that verbal behavior is stored as “ lexical memories.”
V erbal behavior often leaves pu b lic records w hich can be stored in
files and libraries, and the m etaphor of storage is therefore par
ticularly plausible. B u t is the expression any m ore helpful than
saying that my behavior in eating my dinner is stored as prandial
memories, or scratching an itch as a prurient memory? T h e ob
served facts are simple enough: I have acquired a repertoire of
behavior, parts of w hich I display upon appropriate occasions. T h e
m etaphor o f storage and retrieval goes w ell beyond those facts.
T h e com puter, together w ith inform ation theory as designed
to deal Avith physical systems, has made the m etaphor of input-
Why I Am Not a Cognitive Psychologist
storage-retrieval-output fashionable. T h e struggle to make m achines

that think like people has had the effect of supporting theories in
which people: thin k like machines. M in d has recently been defined
as “ the system of organizations and structures ascribed to an in d i
vidual that processess inputs . . . and provides ou tpu t to the vari
ous subsystems and the w orld.” B u t organizations and structures of
what? (T he m etaphor gains power from the w ay in w hich it
disposes of troublesome problems. B y speaking of in p u t one can
forget all the travail of sensory-psychology and physiology; by speak
ing of output one can forget all the problems of reportin g and
analyzing action; and by speaking of the storage and retrieval of
inform ation one can avoid all the difficult problem s of how or
ganisms are indeed changed by contact w ith their environm ents
and how those, changes survive.)
Sensory d&ta are often said to be stored as images, m uch like
the images said to represent the real world. Once inside, they are
moved about fqr cognitive purposes. T h ere is a fam iliar experim ent
on color generalization in w hich a pigeon pecks at a disk of, say,
green light, thé behavior being reinforced on a variable interval
schedule. W hen a stable rate of responding develops, no fu rth er
reinforcements are given, and the color of the disk is changed. T h e
pigeon responds to another color at a rate w hich depends upon
how much it differs from the original; rather sim ilar colors evoke
fairly high rates, very different colors low rates. A cognitive psy
chologist m ight explain the m atter in this way: T h e pigeon takes in
a new color (as “ in p u t” ), retrieves the original color from memory,
where it has been stored in some processed form, puts the two
colored images side by side so that they may be easily compared,
and after evaluating the difference, responds at the appropriate
rate. B ut what advantage is gained by m oving from a pigeon that
responds to different colors on a disk to an inner pigeon that re
sponds to colored images in its mind? T h e sim ple fact is that because
of a known history of reinforcem ent, different colors control diff
erent rates.
T h e cognitive m etaphor is based upon behavior in the real
world. W e store samples of m aterial and retrieve and com pare them
with other samples. W e com pare them in the literal sense of p u ttin g
them side by side to m ake differences m ore obvious. A n d we re
io8 T H E SCIENCE OF BEHAVIOR
spond to different things in different ways. B u t that is all. T h e

whole field of the processing o f inform ation can be reform ulated as
changes in the control exerted by stim uli.
T h e storage of practical knowledge raises another problem.

W hen I learn, say, to take apart the rings of a puzzle, it seems
unlikely that I store my knowledge of how to do so as a copy of the
puzzle or of the contingencies the puzzle m aintains for those trying
to solve it. Instead cognitive theory holds that I store a rule. Rules
are widely used as m ental surrogates of behavior, in part because
they can be memorized and hence “ possessed,” bu t there is an
im portant difference between rules and the contingencies they
describe. R ules can be internalized in the sense that we can say
them to ourselves, but in doing so we do not internalize the
contingencies.
I m ay learn to solve the puzzle in either of two ways. I may
move the rings about until I h it upon a response that separates
them. T h e behavior w ill be strengthened, and if I do the same
thin g a num ber of times, I w ill eventually be able to take the rings
apart quickly. M y behavior has been shaped and m aintained by its
effects on the rings. I may, on the other hand, sim ply follow printed
directions supplied with the puzzle. T h e directions describe be
havior that separates the rings, and if I have already learned to
follow directions, I can avoid the possibly long process of having
my behavior shaped by the contingencies.
D irections are rules. L ike advice, warnings, m axim s, proverbs,
and governm ental and scientific laws, they are extrem ely im portant
parts of a culture, enabling people to profit from the experience of
others. T h ose who have acquired behavior through exposure to
contingencies describe the contingencies, and others then circum
vent exposure by behaving in the ways described. B u t cognitive
psychologists contend that som ething of the same sort happens in
ternally w hen people learn directly from the contingencies. T h ey
are said to discover rules w hich they themselves then follow. But
rules are not in the contingencies, nor must they be “ know n ” by
those who acquire behavior under exposure to them. (W e are lucky
that this should be so, since rules are verbal products w hich arose
very late in the evolution of the species.)
T h e distinction between rules and contingencies is currently

im portant in the field of verbal behavior. C hildren learn to speak
through contact w ith verbal com m unities, possibly w ithou t instruc
tion. Some verbal responses are effective and others not, and over
a period of time more and more effective behavior is shaped and
m aintained. T h e contingencies havin g this effect can be analyzed.
A verbal response “ means” som ething in the sense that the speaker
is under the control of particular circumstances; a verbal stim ulus
“means” som ething in the sense that the listener responds to it in
particular ways. T h e verbal com m unity m aintains contingencies of
such a nature that responses m ade upon particular occasions serve
as useful stim uli to listeners who then behave appropriately to the
occasions.
M ore com plex relations am ong the behaviors of speaker and
listener fall w ithin the fields of syntax and grammar. U n til the time
of the Greeks, no one seems to have know n that there were rules of
grammar, although people spoke gram m atically in the sense that
they behaved effectively under the contingencies m aintained by
verbal com m unities, as children today learn to talk w ith ou t being
given rules to follow. But cognitive psychologists insist that speakers
and listeners must discover rules for themselves. One authority,
indeed, has defined speaking as “ engaging in a rule-governed form
of intentional behavior.” B u t there is no evidence that rules play
any part in the behavior of the ordinary speaker. By using a dic
tionary and a gram m ar we m ay compose acceptable sentences in a
language we do not otherwise speak, and we may occasionally con
sult a dictionary or a grammar in speaking our own language, bu t
even so we seldom speak by applying rules. W e speak because our
behavior is shaped and m aintained by the practices of a verbal
community.
H avin g m oved the environm ent inside the head in the form of
conscious experience and behavior in the form o f intention, w ill,
and choice, and having stored the effects of contingencies of rein
forcement as knowledge and rules, cognitive psychologists put them
all together to compose an internal sim ulacrum of the organism, a
kind of doppelganger, not unlike the classical hom unculus, whose
behavior is the subject of w hat Piaget and others have called “ sub
jective behaviorism .” T h e m ental apparatus studied by cognitive

psychology is simply a rather crude version .of contingencies of
reinforcem ent and their effects.
Every so-called cognitive process has a physical model. W e
associate things by p u ttin g them together. W e store m em oranda
and retrieve them for later use. W e compare things by pu ttin g
them side by side to emphasize differences. W e discriminate things
one from another by separating them and treating them in differ
ent ways. W e identify objects by isolating them from confusing
surrotindings. W e abstract sets of items from com plex arrays. W e
describe contingencies of reinforcem ent in rules. T h ese are the
actions of real persons. It is only in the fanciful w orld of an inner
person that they become m ental processes.
T h e very speed w ith w hich cognitive processes are invented
to explain behavior should arouse our suspicions. M olière made a
joke of a m edical exam ple more than three hundred years ago: “ I
am asked by the learned doctors for the cause and reason why
opium puts one to sleep, to w hich I reply that there is in it a
soporific virtue the nature of w hich is to lu ll the senses.” M olière’s
candidate could have cited evidence from introspection, invoking
a collateral effect of the drug, by saying: ‘‘T o w hich I reply that
opium makes one feel sleepy.” B u t the soporific virtu e itself is a
sheer invention, and it is not w ith ou t current parallels.
A conference was recently held in Europe on the subject of
scientific creativity. A report published in Science 6 begins by p o in t
ing out that more than ninety percent of scientific innovation has
been accomplished by fewer than ten percent of all scientists. T h e
next sentence m ight be paraphrased in this way: “ I am asked by
the learned doctors for the cause and reason why this should be so,
to w hich I reply that it is because only a few scientists possess
creativity.” Similarly, “ I am asked by the learned doctors for the
cause and reason w hy children learn to talk w ith great speed, to
which I reply that it is because they possess linguistic com petence.”
M olière’s audiences laughed.
C ogn itive psychologists have two answers to the charge that

the m ental apparatus is a m etaphor or construct. O n e is that cogni
6 Science, Ju n e 21, 1974. ,

tive processes are known through introspection. Do not all thin kin g
persons know that they think? A n d if behaviorists say they do not,
are they n o t:eith er confessing a low order of m entality or acting
in bad faith for the sake of their position? N o one doubts that
behavior involves internal processes; the question is how w ell they
can be know n through introspection. As I have argued elsewhere,
self-knowledge, consciousness, or awareness became possible only
when the species acquired verbal behavior, and that was very late
in its history. T h e only nervous systems then available had evolved
for other purposes and did not make contact w ith the more im por
tant physiological activities. T h ose who see themselves th in kin g see
little more than their perceptual and m otor behavior, overt and
covert. T h ey could be said t o ' observe the results of “ cognitive
processes” but not the processes themselves— a “ stream of conscious
ness” but not iwhat causes the stream ing, the “ im age of a lem on”
but not the act of associating appearance w ith flavor, their use of
an abstract terpi but not the process of abstraction, a name recalled
but not its retrieval from memory, and so on. W e do not, through
introspection, observe the physiological processes through w hich
behavior is shaped and m aintained by contingencies of reinforce
ment.
B ut physiologists observe them and cognitive psychologists
point to resemblances which suggest that they and the physiologists
are talking about the same things. T h e very fact that cognitive
processes are going on inside the organism suggests that the cogni
tive account is closer to physiology than the contingencies of rein
forcem ent studied by those who analyze behavior. B u t if cognitive
processes are simply m odeled upon the environm ental contingen
cies, the fact that they are assigned to space inside the skin does not
bring them closer to a physiological account. O n the contrary, the
fascination w ith an im agined inner life has led to a neglect of the
observed facts. T h e cognitive constructs give physiologists a mis
leading account of w hat they w ill find inside.
In summary, then, I am not a cognitive psychologist for

several reasons. I see no evidence of an inner w orld of m ental life
relative either to an analysis of behavior as a function of environ
mental forces or to the physiology of the nervous system. T h e re
spective sciences of behavior and physiology w ill m ove forward

most rapidly if their domains are correctly defined and analyzed.
I am equally concerned w ith practical consequences. T h e
appeal to cognitive states and processes is a diversion w hich could
well be responsible for m uch of our failure to solve our problems.
W e need to change our behavior and we can do so only by
changing our physical and social environments. W e choose the
w rong path at the very start when w'C suppose that our goal is to
change the “ minds and hearts of m en and w om en” rather than the
w orld in w hich they live.
I
9
The Experimental Analysis
of Behavior (A History)
i
t
\
*
I was drawn to psychology and particularly to behaviorism

by some papers which B ertrand Russell published in the D ial in
the 1920’s and which led me to his book Philosophy 1 (called in
England A n O utline of Philosophy), the first section of w hich con
tains a much more sophisticated discussion of several epistemologi-
cal issues raised by behaviorism than anythin g of John B. W atson’s.
N aturally I turned to W atson himself, bu t at the tim e only to his
popular Behaviorism.2 I bought P avlo v’s Conditioned R eflexes 3
shortly after it appeared, and w hen I came to H arvard for graduate
study in psychology I took a course w hich covered not only condi
tioned reflexes but the postural and locom otor reflexes of M agnus
and the spinal reflexes reported in Sherrington’s Integrative A ction
of the Nervous System . 4 T h e course was taught by H udson Hoag-
land in the D epartm ent of G eneral Physiology, the head of which,
1 Russell, B. Philosophy. N e w York: W . W . N orton , 1927.

2 W atson, J. B. B ehaviorism . N ew Y o rk: W . W . N o rton , 1924.
8 Pavlov, I. P. C o n d itio n ed reflexes. O x fo rd U n iversity Press, 1927.
4 Sherrington, C. S. Integrative action of the nervous system. N ew
H aven: Y a le U n iversity Press, 1906.
X13
114 t h e s c ie n c e o f b e h a v io r
W . J. Crozier, had worked w ith Jacques Loeb and was studying

tropisms. I continued to prefer the reflex to the tropism, but I
accepted L oeb ’s and Crozier’s dedication to the organism as a whole
and the latter’s contem pt for m edical school “ organ physiology.”
Nevertheless, in the D epartm ent of Physiology at the M edical
School I later worked w ith H allow ell Davis and w ith A lexander
Forbes, who had been in E ngland w ith A d rian and was using
Sherrington’s torsion-wire m yograph to study the reflex control of
movement.
B y the end of my first year at H arvard I was analyzing the
behavior of an “ organism as a w hole” under soundproofed condi
tions like those described by Pavlov. In one experim ent I quietly
released a rat into a small dark tunnel from w hich it could emerge
into a well-lighted space, and w ith m oving pen on a m oving strip
of paper I recorded its exploratory progress, as w ell as its retreat
into the tunnel when I made a slight noise. Some o f my rats had
babies, and in their early squirm ings I thought I saw some of the
postural reflexes stereoscopically illustrated in M agnus’s Korper-
stellung,5 and I began to study them. I m ounted a light platform
on tight wires and am plified its forward-and-backward movement
w ith an arm w riting on a smoked drum. I could put a small rat on
the platform and record the trem or of its leg muscles when I pulled
it gently by the tail, as well as the sudden forw ard leap w ith which
it often reacted to this stim ulation.
I decided to do som ething of the sort w ith an adult rat. I
b u ilt a very light runw ay about eight feet long, the lengthwise
vibration of w hich I could also am plify and record on a smoked
drum , and I induced a rat to run along it by givin g it food at the
end. W hen it was halfw ay along, I would make a slight noise and
record the way in which it came to a sudden stop by the effect on
the runway. I planned to watch changes as the rat adapted to the
noise; possibly I could condition another stim ulus to elicit the
same response. M y records looked a little like those m ade by a
torsion-wire m yograph, bu t they reported the behavior of the or
ganism as a whole.
T h is was all pretty m uch in the tradition of reflex physiology,
b u t quite by accident som ething happened w h ich dram atically
5 M agnus, R . Korperstellung. B erlin : Springer, 1924.

The Experimental Analysis of Behavior (A History)
changed the direction of my research. In my apparatus the rat went

down a back alley to the other end of the apparatus before m aking
its recorded run, and I noticed that it did not im m ediately start to
do so after being fed. I began to tim e the delays and foun d that
they changed in an orderly way. H ere was a process, som ething like
the processes of conditioning and extin ction in P avlov’s work,
where the details o f the act of running, like those of salivation, were
not the most im portant thing.
I have described elsewhere 6 the series of steps through w hich
I simplified my apparatus u n til the rat sim ply pushed open the
door of a small bin to get a piece of food. U n der controlled condi
tions and with pellets of food which took some time to chew, I
found that the rate of eating was a function of the q uan tity of
food already eaten. T h e title of my first experim ental paper, “ On
the Conditions iof E licitation of C ertain E atin g Reflexes,” 7 shows
that I was still applying the concept of the reflex to the behavior
of the organismi as a whole.
Pushing open & door was conditioned behavior, but in order
to study the process of conditioning, 1 needed a more clearly defined
act. I chose pushing down a horizontal bar m ounted as a lever.
W hen the rat pressed the lever, a pellet of food was released into
a tray. T h e arrangem ent wras, of course, close to that w ith w hich
T h orn d ike had dem onstrated his Law of Effect, and in my first
paper I called my apparatus a “ problem b ox,” b u t the results were
quite different. T h o rn d ik e’s cat learned by droppin g out unsuccess
ful bits of behavior un til little or nothing rem ained b u t the suc
cessful response. N oth in g of the sort happened in my experim ent.
Pavlov’s emphasis on the control of conditions had led me to take
certain steps to avoid disturbing my rat. I gave it plenty of tim e
to recover from being p u t into the apparatus by enclosing it first in
a special com partm ent from w hich I later quietly released it. I left
it in the apparatus a long tim e so that it could become thoroughly
accustomed to being there, and I repeatedly operated the food dis
penser until the rat was no longer disturbed by the noise and ate
6 Skinner, B. F. A case history in scientific m ethod. Am erican

Psychologist, 1956, 1 1 , 221-233.
7 Skinner, B. F. O n the conditions o f elicitatio n of certain eatin g
reflexes. Proceedings of the N a tion a l Academ y of Sciences, 1930, 16,
435 - 438 -
n6 T H E SCIENCE OF BEHAVIOR
as soon as food appeared. A ll this was done when the lever was
resting in its lowest position and hence before pressing it could be
conditioned. T h e effect was to rem ove all the unsuccessful be
havior which had composed the learning process in T h o rn d ik e’s
experim ent. M any of my rats began to respond at a high rate as
soon as they had depressed the lever and obtained only one piece
of food.
C on ditionin g was certainly not the mere survival of a success
ful response; it was an increase in rate of responding, or in what
I called reflex strength. T h orn d ik e had said that the cat’s successful
behavior was “ stamped in,” but his evidence was an increasing
priority over other behavior w hich was being “ stamped out.” T h e
difference in interpretation became clearer when I disconnected the
food dispenser and found that the behavior underw ent extinction.
As R. S. W oodw orth 8 later pointed out, T h orn d ik e never investi
gated the extinction of problem -box behavior.
T h o u g h rate of responding was not one of Sherrington’s
measures of reflex strength, it em erged as the most im portant one
in my experim ent. Its significance was clarified by the fact that I
recorded the rat’s behavior in a cum ulative curve; one could read
the rate directly as the slope of the curve and see at a glance how it
changed over a considerable period of time. R ate proved to be a
particularly useful measure when I turned from the acquisition of
behavior to its m aintenance, in the study of schedules o£ interm it
tent reinforcem ent. T heoretically, it was im portant because it was
relevant to the central question: w hat is the probability that an
organism w ill engage in a particular form of behavior at a particu
lar time?
I was nevertheless slow in appreciating the im portance of the
concept of strength of response. For exam ple, I did not immediately
shift from “ condition” to “ reinforce,” although the latter term
emphasizes the strengthening of behavior. I did not use “ reinforce”
at all in my first report of the arrangem ent of lever and food dis
penser, and my first designation for interm ittent reinforcem ent was
“ periodic reconditioning.”
Strength or probability of response fitted com fortably into
/
8 W oodw orth , R . S. Contem porary schools of psychology. N e w York:
R o n a ld Press, 1951.
The Experimental Analysis of Behavior (A History) “ 7
the form ulation of a science of behavior proposed in my thesis.

Russell was again responsible for a central point. Somewhere he
had said that “;reflex” in psychology had the same status as “ force”
in physics. I knew w hat that m eant because I had read Ernst M ach’s
Science of M echanics,9 the works of H enri Poincare on scientific
method, and B ridgm an’s Logic of-M odern Physics.10 M y thesis was
an operational analysis of the reflex. I insisted that the w ord should
be defined sim ply as an observed correlation of stimulus and
response. Sherrington’s synapse was a mere inference w hich could
not be used to explain the facts from w hich it was inferred. T h u s,
a stimulus m ight grow less and less effective as a response was
repeatedly elicited, b u t it did not explain anything to attribute
this to “ reflex fatigue.” Eventually the physiologist w ould discover
a change in the nervous system, b u t so far as the behavioral facts
were concerned,! the only identifiable explanation was the repeated
elicitation. In my thesis 11 I asserted that in the intact organism
“ conditioning, lem otion,’ and ‘d rive’ so far as they concern be
havior were essentially to be regarded as changes in reflex strength,”
and I offered m y experim ents on “ drive” and conditioning as
examples.
One needed to refer not on ly to a stimulus and a response
but to conditions which changed the relation between them. I
called these conditions “ third variables,” and represented matters
with a simple equation:
R = f (S, A)
where A represented any condition affecting reflex strength, such as

the deprivation w ith w hich I identified “ drive” in the experim ental
part of my thesis.
T h e summer after I got my degree, Edward C. T o lm a n was
0 M ach, E. T h e science of m echanics. C hicago: O p e n C ou rt, 1893.

10 Bridgm an, P. W . T h e logic of m'odern physics. N e w Y o rk: M ac
m illan, 1928.
11 Skinner, B. F. T h e concept o f the reflex in the description o f
behavior. T hesis. H arvard U n ive rsity Library, C am bridge, M assa
chusetts. (Part O n e rep rin ted in B. F. Skinner, C um ulative record
[3rd ed.]. N ew Y o rk: A p pleton -C en tu ry-C rofts, 1972, alon g w ith 47
other papers.)
n 8 TH E SCIENCE OF BEHAVIOR
teaching at H arvard, and I saw a great deal of him. I expounded '

my operational position at length and the relevance of third vari
ables in determ ining reflex strength. T o lm a n ’s book Purposive B e
havior in Anim als and, M en 12 was then in press, and in it he speaks
of “ independent variables” bu t only as such things as genetic
endowm ent or an in itiatin g physiological state. T h re e years later he
published a p a p e r 13 containing the equation:
B = f (S, H, T , P)
in which B stood for behavior, as my R stood for response, S for

“ the environm ental stimulus setup” (my S), H for heredity, T for
“specific past training” (my “ conditioning” ), and P for “ a releasing
internal condition of appetite or aversion” (my “ drive” ). Wood-
worth later pointed out that these equations were similar. T h ere
was, however, an im portant difference: w hat I had called a “ third
variable” T o lm a n called “ intervening.” For me the observable
operations in conditioning, drive, and em otion lay outside the
organism, bu t T olm an p u t them inside, as replacem ents for, if not
simply redefinitions of, m ental processes, and that is where they 7':
still are in cognitive psychology today. Ironically, the arrangement
is m uch closer than m ine to the traditional reflex arc.
A lth o u gh rate of responding, in the absence of identifiable
stim ulation, had no parallel in Sherrington or Pavlov, I continued
to talk about reflexes. I assumed that some features of the lever
were functioning as stim uli w hich elicited the response of pressing
the lever. B u t I was unhappy about this, and I began to look more
closely at the role of the stimulus. I reinforced pressing the lever
when a light was on but not when it was off and found that in
the dark the behavior underw ent extinction. T u r n in g on the light
then appeared to elicit the response, but the history behind that |
effect could not be ignored. T h e light was not eliciting the be
havior; it was functioning as a variable affecting its rate, and it :| r
derived its power to do so from the differential reinforcem ent with :|
w hich it had been correlated.
12 T o lm a n , E. C. Purposive behavior in anim als and men. New

Y o rk : C entury, 1932. 1 ' ■„
13 T o lm a n , E. C . P h ilosoph y versus im m ediate experien ce. Philos
ophy o f Science, 1935, 2, 356-380.
The Experimental Analysis of Behavior (A History) 119
In the summer of 1934 I subm itted two papers for publication

in separate efforts to revise the concept of the reflex. In “ T h e
Generic N atiire of Stim ulus and Response” 14 I argued that neither
a stimulus nor a response could be isolated by surgical or other
means and that the best clue to a useful u n it was the orderliness
of the changes in its strength as a function of “ third variables.” In
“ T w o types of conditioned R eflex and a Pseudo-type” lfl I distin
guished between Pavlovian and what I w ou ld later call operant
conditioning. Q uite apart from any internal process, a clear dif
ference could be pointed out in the contingent relations among
stimuli, responses, and reinforcem ent.
I was forced to look more closely at the role of the stimulus
when Konorski and M ille r 16 replied to the latter paper by describ
ing an experim ent they had perform ed in the late twenties which
they felt anticipated my own. T h ey had shocked the paw of a dog
and given it food when it flexed its leg. E ventually the leg flexed
even though the paw was not shocked. I replied that true reflexes
seldom have the kinds of consequences w hich lead to operant
conditioning. Shock m ay be one way of in d u cin g a hungry dog to
flex its leg so that the response can be reinforced w ith food, bu t it
is an unusual one, and an eliciting stim ulus can in fact seldom be
identified. (As to priority, T h o rn d ik e was, of course, ahead of us
all by more than a quarter of a century.)
In my r e p ly 17 I used the term “ operan t” for the first time
and applied “ respondent” to the P avlovian case. It w ould have been
the right time to abandon “ reflex,” but I was still strongly under
the control of Sherrington, Magnus, and Pavlov, and I continued
to hold to the term doggedly when I wrote T h e Behavior of Or
ganisms (1938).18 It took me several years to break free of m y own
stimulus control in the field of operant behavior. From this point
14 Skinner, B. F. T h e generic nature o f the concepts o f stim ulus and

response. Journal of G eneral Psychology, 1935, 12, 40-65.
15 Skinner, B. F. T w o types of con d ition ed reflex and a pseudo type.
Journal of G eneral Psychology, 1935, 12, 66-77.
16 K onorski, J. and M iller, S. O n two types o f con dition ed reflex.
Journal of G eneral Psychology, 1937, 16, 264-272.
17 Skinner, B. F. T w o types o f con dition ed reflex: A rep ly to K o n o r
ski and M iller. Journal of G eneral Psychology, 1937, 16, 272-279.
18 Skinner, B. F. T h e behavior of organisms. N e w York: A p pleton -
C entury, 1938.
on, however, I was clearly no longer a stimulus-response psycholo

gist.
T h e lack of an identifiable eliciting stimulus in operant be
havior raises a practical problem : we must w ait for behavior to
appear before we can reinforce it. W e thus start w ith m uch less
control than in respondent conditioning. Moreover, there is a great
deal of com plex behavior for w hich we shall certainly w ait in vain,
since it w ill never occur spontaneously. In hum an behavior there
are many ways of “ prim ing” an operant response (that is, evoking
it for the first time in order to reinforce it), and one of them is
also available in lower organisms: com plex behavior can be
“ shaped” through a series of successive approxim ations. T o re
inforce pressing a lever w ith great force, for exam ple, we cannot
simply wait for a very forceful response, but we can differentially
reinforce the more forceful of the responses which do occur, w ith
the result that the mean force increases.
I used a sim ilar program m ing of contingencies of reinforce
ment to shape com plex topography in a dem onstration (reported
in the Behavior of Organisms) in w hich a rat pulled a chain to
release a marble, picked up the m arble, carried it across the cage,
and dropped it into a tube. T h e term inal behavior was shaped by |
a succession of slight changes in the apparatus. Later my colleagues
I
and I discovered that we could avoid the time-consuming process
of altering the apparatus by constructing program m ed contingen
cies w hile reinforcing directly by hand.
I soon tried the procedure on a hum an subject— our nine-
month-old daughter. I was h old ing her on my lap one evening when
I turned on a table lam p beside the chair. She looked up and
smiled, and I decided to see w hether I could use the light as a
reinforcer. I w aited for a slight m ovem ent of her left hand and
turned on the light for a m om ent. Alm ost im m ediately she moved
her hand again, and again I reinforced. I began to w ait for bigger
movements, and w ithin a short time she was liftin g her arm in a
w ide arc— “ to turn on the ligh t.”
I was w riting Walden T iu o 19 at the time, and the book is
often cited as an essay in behavioral engineering, b u t I believe it
contains no exam ple of the exp licit use of a contrived reinforcer.
19 Skinner, B. F. W alden Tw o. N e w York: M acm illan , 1948.

T h e com m unity functions through positive reinforcem ent, bu t the

contingencies are in the natural and social environments. T h ey
have been carefully designed, but there is no continuin g interven
tion by a reinforcing agent. T h e only contrived contingencies are
Pavlovian: children are “ desensitized” to frustration and other de
structive emotions by being exposed to situations of carefully
graded intensity.
I began to analyze the contingencies of reinforcem ent to be
found in existing cultures in an undergraduate course at H arvard
in the spring of 1949. Science and H um an Behavior (1953) 20 was
w ritten as a text for that course, and in it I considered practices in
such fields as government, religion, economics, education, psycho
therapy, self-control, and social behavior— and all from an operant
point of view.
Practical demonstrations soon followed. A graduate student at
Indiana, Paul Fuller, had reinforced arm-raising in a twenty-year-
old hum an organism w hich had never before “ shown any sign of
intelligence,” and-in 1953 I set up a small laboratory to study oper
ant behavior in a few backward patients in a m ental hospital.
Ogden R. Lindsley took over that project and found that psychotics
could be brought under the control of contingencies of reinforce
ment if the contingencies were clear-cut and carefully programmed.
Ayllon, Azrin, and many others subsequently used operant condi
tioning in both management and therapy to im prove the lives of
psychotic and retarded people.
A t the U niversity of Pittsburgh in the spring of 1954 I gave
a paper called “ T h e Science o f L earning and the A rt of T ea ch in g ” 21
and demonstrated a m achine designed to teach arithm etic, using an
instructional program. A year or two later I designed the teaching
machines which were used in m y undergraduate course at Harvard,
and my colleague, James G. H olland, and I wrote the program m ed
materials eventually published as T h e Analysis of Behavior (1961).22
T h e subsequent history of program m ed instruction and, on a
20 Skinner, B. F. Science and human behavior. N ew York: M acm il-

lan, 1953.
21 Skinner, B. F. T h e science of le arn in g and the art o f teaching.
Harvard Educational R eview , 1954, 24, 86-97.
22 H o llan d , J. G . and S kinner, B. F. T h e analysis of behavior. N ew
York: M cG raw -H ill, 1961.
122 THE SCIENCE OF BEHAVIOR
:
'
broader scale, of what has come to be called applied behavior
analysis or behavior m odification is too well known to need further
review here.
M eanwhile, the experim ental analysis of operant behavior
was expanding rapidly as m any new laboratories were set up.
Charles B. Ferster and I enjoyed a very profitable five-year collabo
____________________________________________________________________________________________ ______ ____________________________

ration. M any of our experim ents were designed to discover whether
the perform ance characteristic of a schedule could be explained by
the conditions prevailing at the m oment of reinforcem ent, includ
ing the recent history of responding, but adm inistrative exigencies
drew our collaboration to a close before we had reached a sound
form ulation, and we settled for the publication of a kind of atlas
showing characteristic perform ances under a w ide range of sched
ules (Schedules of R e in fo rcem en t23). T h e subsequent developm ent
of the field can be traced in the Journal of the Experim ental Anal
ysis of Behavior, which was founded in 1958.
Several special themes have threaded their way through this
history, and some of them call for comment.
Verbal Behavior. I began to explore the subject in the middle

thirties.. T h e greater part of a manuscript was w ritten w ith the help
of a Guggenheim Fellowship in 1944-4.5, from w hich the W illiam
James Lectures at H arvard in 1947 were taken. A sabbatical term
in the spring of 1955 enabled me to finish most of a book, which
appeared in 1957 as Verbal Behavior . 2 i It will, I believe, prove to
be my most im portant work. It has not been understood by linguists
or psycholinguists, in part because it requires a technical under
standing of an operant analysis, bu t in part because linguists and
psycholinguists are prim arily concerned w ith the listener— with
w hat words mean to those who hear them, and w ith w hat kinds of
sentences are judged gram m atical or ungram m atical. T h e very
concept of com m unication— w hether of ideas, meanings, or infor
m ation— emphasizes transmission to a listener. So far as I am con
cerned, however, very little of the behavior of the listener is worth
distinguishing as verbal.
23 Ferster, C . B. and Skinner, B . F. Schedules o f reinforcem ent. New

York: A pp leton -C en tu ry-C rofts, 1957.
24 Skinner, B. F. Verbal behavior. N ew Y o rk: A ppleton-C entury-
C rofts, 1957.
In Verbal Behavior verbal operants are classified by reference

to the contingencies of reinforcem ent m ain tain ed by a verbal
community. T h e classification is an alternative to the “ moods” of
the gram m arian and the “ intentions” of the cognitive psychologist.
W hen these verbal operants come together under m ultiple causa
tion, the effect may be productive if it contributes, say, to style and
wit, but destructive if it leads to distortion and fragm entation.
Speakers m anipulate their ow n verbal b ehavior in order to control
or qualify the responses of listeners, and gram m ar and syntax are
“ autoclitic” techniques of this sort, as are m any other practices in
sustained composition. A technology of verb a l self-management
emerges which is useful both in “ discovering w hat one has to say”
and in restricting the range of controlling variables— emphasizing,
for example, the kinds of variable (characteristic of logic and
science) most likely to lead to effective practical action or the kinds
found to be more productive o f poetry or fiction.
I
T h e Nervous^ System. M y thesis was a sort of declaration of
independence from the nervous system, and t restated the position
in T h e Behavior of Organisms. It is not, I th in k, anti-physiological.
Various physiological states and processes interven e between the
operations performed upon an organism and the resulting behavior.
T h ey can be studied with appropriate techniques and there is no
question of their im portance. A science of ibehavior has its own
facts, however, ( and they are too often obscured w hen they are
converted into hasty inferences about the nervous system. I w ould
still say, as I said in T h e Behavior of Organisms, that no physiologi
cal fact has told us anything about behavior th a t we did not already
know, though we have been told a great d eal abou t the relations
between the two fields. T h e h elpful relation is the other way round:
a behavioral analysis defines the task of the physiologist. O perant
theory and practice now have an im portant p la ce in the physiologi
cal laboratory.
Psychopharmacology. A t M innesota, W - T. H eron and I

studied the effects o f a few fam iliar drugs on o p eran t behavior, and
in the early fifties, Dr. Peter Dews of the D ep artm en t of Pharm a
cology at the H arvard M edical School, becam-e associated w ith my
laboratory and co-workers. A t about the san*e tim e m any of the
ethical drug companies set up operant laboratories, some of which

contributed to the present arm am entarium of. behavior-m odifying
drugs. O perant techniques are now w idely used in the field, as well
as in the study of drug addiction and related m edical problems.
Ethology. Ethologists often assert that their w ork is neglected

by behaviorists, but W atson’s first experim ents were ethological,
and so were mine. T h e process of operant conditioning itself is
part of the genetic equipm ent of the organism, and I have argued
that reinforcers are effective, not because they reduce current drives M
(a w idely held view), but because susceptibilities to reinforcem ent
have had survival value. Species-specific behavior m ay disrupt
operant behavior, but the reverse is also true.
In Science and H um an Behavior I pointed ou t that contin
gencies of survival in natural selection resembled contingencies of
reinforcem ent in operant conditioning. B oth involve selection by
consequences, a process which, in a work in progress, I argue to be
particularly relevant to the question of whether hum an behavior
can indeed take the future into account. Phylogenic contingencies
which could have shaped and m aintained, say, im itative behavior
resemble the contingencies of reinforcem ent w hich shape similar
behavior in the individual, but one repertoire does not evolve from
the other. A n experim ent on im printing has shown how an operant
analysis may clarify field observations and correct conclusions
drawn from them: the young duckling does not inh erit the behavior
of follow ing its m other or an im printed object; it acquires the
behavior because of an innate susceptibility of reinforcem ent from
being close.
A Theory of Knowledge. I came to behaviorism , as I have

said, because of its bearing on epistemology, and I have not been
disappointed. I am, of course, a radical rather than a m ethodologi
cal behaviorist. I do not believe that there is a w orld of m entation
or subjective experience that is being, or must be ignored. One
feels various states and processes w ithin one’s body, b u t these are
collateral products of one’s genetic and personal histories. N o crea
tive or in itiatin g function is to be assigned to them. Introspection
does not perm it us to m ake any substantial contrib u tion to physi
ology, because “ we do not have nerves going to the righ t places.”
C ognitive psychologists m ake the m istake of internalizin g environ-
mental contingencies— as in speaking of the storage of sensory con

tacts w ith the environm ent in the form of m emories which are
retrieved and responded to again at some later date. T h ere is a
sense in which one knows the world, b u t one does not possess
knowledge; one behaves because of one’s exposure to a com plex and
subtle genetic and environm ental history. As I argued in a final
chapter in Verbal Behavior, thin kin g is sim ply behavin g and may
be analyzed as such. In A b o u t Behaviorism ,25 I attem pted to make a
comprehensive statement of the behaviorist’s position as I under
stood it forty-six years after I first entered the field.
Designing a Culture. W alden Tw o was an early essay in the

design of a culture. It was fiction, bu t I described a supporting
science and technology in Science and H um an Behavior. I was
made aware cjf a basic issue when W alden Two was im m ediately
attacked as a threat to freedom. Its protagonist was said to have
m anipulated the lives of people and to have made an unw arranted
use of his own value system. I discussed the issue in a paper called
“ Freedom and' the Control of M en” in 1955 20 and in a debate w ith
C arl Rogers in 1956.27 T h e control of behavior becam e especially
critical with the rise of an applied behavioral analysis in the 1960’s,
and I returned to the issue in Beyond Freedom and Dignity in
1971.28 U nfortunately that title led m any people to believe that I
was opposed to freedom and dignity. I did, indeed, argue that
people are not in any scientific sense free or responsible for their
achievements, but I was concerned w ith identifying and prom oting
the conditions under which they feel free and w orthy. I had no
quarrel w ith the historical struggle to free people from aversive
control or from punitive restrictions on the pursuit of happiness,
and I proposed that that struggle be continued by sh iftin g to prac
tices which employed positive reinforcem ent, bu t I argued that
25 Skinner, B. F. A b o u t behaviorism. N ew York: A lfre d A . K n op f,

1974 -
26 Skinner, B. F. Freedom and the con trol o f m en. A m erican Scholar,
W in ter 1955-56, 25, 47-65.
27 Rogers, C. R . and Skinner, B. F. Some issues con cern in g the con
trol o f hum an behavior: A sym posium . Science, 1956, 124, 1057-1066.
28 Skinner, B. F. B eyon d freedom and dignity. N ew Y o rk: A lfre d A.
K n op f, 1971.
126 T H E SCIENCE OF BEH AVIOR
certain aspects of the traditional concepts stood in the way. For

exam ple, to make sure that individuals receive credit i o r their
actions, certain punitive practices have actually been perpetuated.
I believe that a scientific form ulation of hum an behavior can help
us m axim ize feelings of freedom and dignity.
T h ere is a further goal: w hat lies beyond freedom a n d dignity
is the survival of the species, and the issues I first discussed in
Walden Two have become m uch more pressing as the thffeat of a
catastrophic future becomes clearer. U nfortunately, we m ove only
slowly toward effective action. A question com m only askeci is this:
when shall we have the behavioral science we need to solve our
problems? I believe that the real "question is this: when sh all we be
able to use the behavioral science we already have? M o re and
better science would be helpful, but far more effective (decisions
w ould be made in every field of hum an affairs if those wlho made
them were aware of what we already know.
PART III
EDUCATION
\
\
i
Q Some Implications of Maying Education More

Efficient
1 The Free and Happy Student
2 Designing Higher Education
I
i
IO
Some Implications
of Making Education
More Efficient
T h ere is little doubt that education is in trouble. It faces

many different kinds of problems, for w hich m any different kinds of
solutions w ill have to be found. One of them is economic. T h e
educational assignm ent grows steadily greater. For exam ple, chil
dren are to start school at an earlier age, special classes are to be
arranged for 'exceptional children, students are to be adm itted
to college w ith fewer qualifications, and new fields are to be cov
ered to im prove the relevance of education. T hese changes come
at a time w hen costs are rising sharply. T u itio n and taxes continue
to rise, teachers are asked to take on more work and cut down on
outside activities, and m any schools, particularly parochial schools,
are closing.
One solution to the economic problem is simply to make in
struction m ore efficient. If we could teach, say, twice as m uch in the
same time and w ith the same effort, our present staff and facilities
would suffice to teach more students, teach each one more, allow
for a wider range of abilities, and cover m ore fields, w hile at the
same time h old in g sm aller classes, giving teachers a more reasonable
workday and m ore pay, and getting more support from the public
129
13° E D U C A T IO N
by giving more in return for its money. Alm ost any other enter
prise would try to solve an economic problem in that way. It
would see whether its practices could not be made more efficient.
B u t teachers and school administrators seldom look in that direc
tion. Why?
Past experience w ith research in the field of learning may be

at fault. T h e learning curves obtained w ith mazes and the for
getting curves obtained w ith memory drums have never given the
teacher any real help. E ducational psychologists soon turned from
basic research on the processes underlying teaching to the measure
ment of their effects— a change exem plified in the personal history
of Edward L. T hornd ike.
Part of the explanation m ay be a distaste for pedagogy or edu
cational method. Can teaching really be taught? Does not any rea
sonably intelligent person already know what is needed? A teacher
must attract the students’ attention and keep them interested, but
he w ill have learned how to do this in his daily life. T h e neglect of
pedagogy is seen in current books which tell us how to im prove
our schools. One needs no special vocabulary or any scientific
knowledge to read the contributions of men like John Holt, Jona
than Kozol, Paul Goodm an, and Charles Silberm an. Even the teach
ing aids which are most com m only m entioned (audiovisual devices
and television, for exam ple) sim ply clo w hat people do and make no
use of a more technical analysis of basic processes. Teaching, in
other words, is regarded not as a special skill b u t as an art in deal
ing w ith people. T h e only problem is to find those who practice it
well.
M any educators have gone a step further. T h e y show no interest
in m aking teaching more efficient because they do not believe in
teaching. In the classical expression, the teacher cannot teach, he
can only help the student learn, and he cannot help much. Carl
Rogers has recently said that in his opinion teaching is a “ vastly
overrated function.” 1 “ Free schools” (for exam ple, Summerhill)
and many experim ental colleges boast of how little teaching ac-
1 Rogers, C. Freedom to learn. C olum bus, O h io: M errill, 1969.

Some Im plications of M aking E ducation M ore Efficient
tually goes on, and Ivan Illich 2 has com pleted the reductio ad
absurdum by calling for the dcschooling of society. It w ill be
enough simply to make the world a “ livable learn ing environm ent.”
T h ere is often a note of despair in these proposals. W e have tried
so hard and failed so miserably; there must be a better way.
T h e way that is most often suggested goes back to Jean
Jacques Rousseau. W e are to let the child learn in school as he
learns in the w orld at large— through a n atural love of learning, a
natural curiosity. Let him know the joy of discovery. T h e proposal
is especially appealing in contrast w ith what goes on in the joyless
punitive schools which have so lon g characterized education. It is
also attractive because it seems to raise no problems. T h e real w orld
is conveniently at hand and it does not need to be made to work.
B ut Rousseau’s proposal has been tried, episodically at least, for
two hundred iyears and that is presum ably tim e enough to dem on
strate its feasibility. W hy, then, are we still at the stage of m aking
proposals? W{iy is it that the average life of an experim ental free
school is said .to be som ething on the order of eighteen months? It
is true that new proposals in education, as elsewhere, are not likely
to be well supported, and that the great changes which need to be
made in established practices can be made only slowly. Nevertheless,
more progress should have been made.
A more likely explanation is that the real world is not an
effective teacher. Children do not learn m uch from the natural en
vironment. T h e feral child, the child said to have been raised by
wolves, or one said to have m atured alone in a benevolent environ
ment, is about all we have to show for unaided natural curiosity or
a love of learning. A physical environm ent breeds awkward, dan
gerous, and superstitious behavior, and a social environm ent breeds
hostile as well as friendly behavior, selfish as w ell as generous be
havior. W hat seem like successful dem onstrations of “ free” class
rooms must be attributed to unanalyzed skills in dealing w ith peo
ple, and the difficulty is that because they have not been analyzed,
they cannot be transmitted. T h ere has been no accum ulation of
better ways of teaching w ithin R ousseau’s program . O n the con
trary, apparent successes have usually m eant a contraction in the
2 Illich , I. D eschoolin g society. N e w Y ork: H a rp e r & R ow , 1971.

132 EDUCATION
educational assignment. Less and less is taught, by definition, as

learning is left to the natural environm ent, but to that we must add
that less and less is learned. T h e extent to w hich we have accepted
this consequence is suggested by current proposals to reduce if not
abolish “ compulsory education.”
Education is an im portant function of a culture— possibly in
the long run its most im portant or only function. A culture, as a
social environment, must transmit itself to its new members. Some
transmission occurs when new members learn from those w ith
whom they are in contact, w ith or w ithout inform al instruction;
but transmission on a scale needed to m ake people m axim ally
effective needs a carefully designed system.
A lack of confidence in any effort to im prove teaching is espe
cially crucial because there are m any other reasons why new prac
tices are not likely to be adopted. So far as most administrators and
teachers are concerned, an im provem ent in teaching w ill demand
troublesome changes, in w hich m uch is to be lost and not much to
be gained. N o penalty is imposed if an adm inistrator or teacher
overlooks a better way of teaching, and current inefficiencies can
be justified by arguing that the task is too difficult, that there are
too many students, that facilities are inadequate, and that social
and racial problems are insurm ountable. If this is not enough, it is
possible to fall back upon that argum ent w hich has always been
used to exonerate bad teaching; it is the student who fails the
course, not the teacher or school.
It is a hopeful sign that administrators and teachers are be
ginning to be held accountable for their work. T h is has always been
true in other professions. A doctor may not cure every patient, but
if he cures few or none, he fails as a doctor. A lawyer m ay not win
every case, but if he wins few or none, he w ill not last long as a
lawyer. T h e salesman may not make every sale, bu t if he makes few
or none, he fails as a salesman. A n d of course artisans have always
been judged in terms of the qu ality of their work. W h y should the
teacher not also be held accountable for the results of his teaching?
T h e commonest answer is that the results cannot be evaluated.
T h e y are not as obvious as a cure, a favorable verdict, a sale, or
a jo b well done. Some specialists in educational measurement, in
Some Implications of Making Education More Efficient 133
a surprising reversal of an earlier position, have been quick to agree.

For more than half a century we have been told that measures of
ability and achievement are reliable and valid. B u t suddenly they
have become socially relevant: intelligence has taken on racial over
tones, and achievem ent has been tied to the accountability of
teachers. Some authorities are therefore beating a retreat. H enry S.
Dyer, vice-president of the E ducational T estin g Service, has char
acterized tests of intelligence and of grade equivalency as “ m on
strosities” and has said that the developm ent of tests w hich could
be used to hold teachers accountable for their w ork w ould be an
enterprise of the order o f m agnitude of the atom ic bom b.3 B u t
students have been held accountable w ith the same kinds of tests for
decades; they have been adm itted to college according to m easured
abilities and , prom oted and graduated according to m easured
achievements. *(That we should not hold students accountable, that
we should admit them to college regardless of m easured abilities
and give no «examinations to measure their achievem ent is, of
course, part o£ the ^philosophy of a n atural learning environm ent.)
It is no doubt easier to measure some effects of teaching than
others. H ow well a student has learned to read is more obvious than
what he has learned in, say, social studies. B u t education w ould be
in serious trouble if we could not tell w hether a student has learned
anything in social studies. B oth the teacher and the student need
evidence of progress. One source of trouble is the traditional prac
tice of defining the goals of education in terms of m ental processes.
If the teacher is to “ transm it know ledge,” “ cultivate skills,” “ evoke
ideas,” or “ change attitudes,” neither he nor the student is likely
to have any clear evidence that a change has occurred. A n other
source of trouble is that the very large repertoire acquired in a
course cannot be reliably sampled in a brief exam ination. W e shall
see how these problems can be solved in other ways.
Teachers and administrators are likely to reject any proposal
that they be held accountable, m ainly for econom ic reasons. T h ere
is always the danger that a teacher w ho is not very efficient w ill be
fired or paid less than one w ho is conspicuously successful. B u t only
3 N ew York T im es, M arch 23, 1971.

*34 EDUCATION
when the administrator or teacher is held accountable w ill he search

most actively for better ways of teaching.
It w ould be unfair to say that teachers and administrators do
little to make teaching more effective only because they lack eco
nomic incentives. A better explanation is that they do not know
w hat to do. It has always been supposed that the principal source of
technical knowledge in education is classroom experience. T h e
young teacher learns how to teach either by teaching or by em ulat
ing someone who has learned how to teach by teaching. T h e possi
bility that technical help m ay come from outside the profession
is seldom recognized. T h ere was a com parable stage in the history of
medicine. M edical practices were once entirely the product of the
experience of physicians, bu t most physicians accept the fact that
advances in m edicine w ill now come from the scientific laboratory.
A change in provenance is inevitable as soon as a relevant science
appears, and that stage has now been reached in education.
W h at has come to be called the “ experim ental analysis of be

havior” has already given ,rise to an effective technology of teaching,
although it is not yet w idely know n or used. T h ree contributions
may be noted. One has to do w ith the teacher’s assignment. It has
long been supposed that the task of the teacher is to im part infor
m ation, train the m ind, help the student grasp relations, teach him
to appreciate literature, art and music, encourage creativity, and
change his attitudes (for exam ple, toward racial problems). B ut the
teacher does not act upon the m ind or its faculties, or upon traits of
character or personality. H e acts upon the behavior of the student,
and he does so by changing the verbal or nonverbal environm ent
in w hich the student lives.
It is not always easy to redefine the goals of teaching. In par
ticular, an analysis of the so-called higher m ental processes may be
quite complex. B ut progress has already been made, and it has
given the teacher a clearer conception of his assignment and better
evidence of the extent to w hich it has been fulfilled. It has also
made it less likely that he w ill seek exoneration for failure by in
venting mental objectives w hich he can claim to have achieved. He
is less likely to discount the fact that a child cannot read by arguing
that he is acquiring reading readiness or an interest in reading, or

that a student who cannot solve problem s in arithm etic is neverthe
less learn ing'to understand m athem atics or acquirin g a love for it.
A second contribution of the experim ental analysis of behavior

has to do with classroom management. W hy does a student come to
school, behave w ell in class, pay attention, apply him self to his asr
signments, answer questions, and so forth? So far as traditional prac
tice is concerned, the answer is simple: to avoid the consequences
of not doing so. It is now clear that m any of the disciplinary prob
lems faced by teachers (truancy, vandalism , and apathy) are the
by-products of a lon g history of aversive control, w hich has not
yet come to an end. T h e experim ental analysis of behavior has sug
gested powerful alternatives through the use of positively reinforc
ing consequences. T o pu t it roughly, the student can be given posi
tive reasons f6r doin g the kinds of things w hich w ill advance his
education. T h e lawman speaks of these as rewards and may object
to new classroom practices as bribery, but to do so is to m isunder
stand the whole science of contingency management. W h at is im
portant is not only the rew arding things a student gets b u t the
ways in which they are contingent upon his behavior. T h e pow er
of contingency m anagem ent in the classroom is w ell established—
though, again, it is not yet w idely used.
T h e experim ental analysis of behavior has m ade a third contri

bution to education in the design of instructional m aterials— both
in the m aterial itself and in modes of presentation. T echn iqu es of
shaping com plex behavior through a program of progressive ap
proxim ation emerged from the operant laboratory, particularly in
the extension of basic principles to the analysis of verbal behavior.
T h e m ain features of a good program are well known: the student is
asked to proceed in small steps and to master each step before m ov
ing on to the next. M aterial is so designed that correct responses
are highly probable, and progress through a program may be all
that is needed to keep the student at work. A good program im parts
an extensive repertoire in a very efficient way.
136 EDUCATION
T h e personalized system of instruction (PSI) designed by F. S.

K e lle r 4 brings these contributions together in the redesign of
courses in colleges and graduate schools. T h e basic elements have
been described by K eller as follows:
(1) the go-at-your-own-pace feature which permits a stu

dent to move through a course o f study at a speed com
mensurate w ith his ability and other demands upon his
time; (2) the unit-perfection requirem ent for advance,
which lets the student go ahead to any m aterial only
after dem onstrating mastery of that w hich precedes it;
(3) the use of lectures and demonstrations as vehicles of
m otivation rather than sources of critical inform ation;
(4) the related stress upon the w ritten w ord in teacher
com m unication; and, finally, (5) the use o f [a] proctor
(student), w hich permits repeated testing, immediate
scoring, almost unavoidable tutoring, and a m arked en
hancem ent of his personal social aspect o f the educa
tional process (p. 13).
T h e PSI system is spreading rapidly throughout colleges and uni

versities, and there is no reason w hy it cannot be adapted to high
school and the lower grades.
T h e definition of objectives in behavioral terms, the design of
effective classroom contingencies, and the program m ing of instruc
tional materials may be all that is needed to solve many current
problems in education. O perant conditioning is a m atter of both
“ acquisition” and “ m otivation,” and signs of progress through a
program are for most students a highly reinforcing consequence.
Individualized treatment removes the greatest source of inefficiency
in traditional instruction— the requirem ent that large num bers of
students advance at the same speed, w hich is almost necessarily the
w rong speed for most of them. B u t perhaps the most im portant re
sult is that there is no need for final exam inations. In a well-designed
course of instruction, the behavior a student is acqu irin g is obvious
because he uses it in pursuing the course. A glance ahead shows him
w hat he does not yet know; a glance backward shows him w hat he
4 Keller, F. S. Neglected rewards in the educational process. Proceed

ings of the T w enty-T h ird A n n u a l M eetin g of the Am erican C onfer
ence of Academ ic Deans. Los Angeles, 1967. Pp. 9-22.
has learned. Both student and teacher can see w hat has been done
w ithout trying to sample large repertoires. Such instruction is
rather like teaching a m anual skill or a sport. T h e golf instructor
does not give his student a final exam ination, m easuring the length
of ten drives from a tee, the distances from the pin in ten approaches
from a sand trap, and scoring the num ber of successes in ten long
putts and ten short putts, and then assign a grade showing how
well his student has learned to play. Each step in a program may
be considered an exam ination because the student responds and
his response is evaluated. In the K eller system b rief tests are taken
to determine mastery at each level, bu t this is very different from
trying to measure all that a student has learned at the end of a
course.
Im pending exam inations have well-known em otional effects
due, in part, to the feared risk of inaccurate sampling. A dm inistra
tors and teachers also, faced w ith accountability, are now beginning
to show these effects and for the same reasons. B u t a well-designed
course of instruction solves the problem for both students and
teachers. T h e course itself is the exam ination. If the student is to re
ceive a grade, it w ill indicate only how far he has advanced. It is not
necessary to determ ine the degree to w hich the materials of the
whole course are retained, since most of them must have been
retained in order to finish the course. T h e critic may com plain that
retention is not being measured, bu t a final exam ination does not
measure it successfully, and it encourages practices, such as last-
m inute cramming, w hich actually interfere w ith the retention the
exam ination is designed to guarantee.
H ow much im provem ent is to be expected? Is it fair to say
that what is now taught could be taught in h alf the time and
w ith half the effort on the part of student and teacher? Anyone
who has worked through a well-designed program of instruction
(in a subject w ith which he was not u n til then fam iliar), anyone
who has seen a high school class under good contingency m anage
ment, or anyone who has talked w ith or read the reports of stu
dents in a personalized system of instruction w ill be inclined to say
yes. Comparisons w ith so-called control groups in set experim ents
are not very helpful. T h e com parison should be w ith what now pre
vails in our schools and colleges. T h ere are no doubt other ways in
which teaching can be made more effective, b u t the practices de
138 EDUCATION
rived from an experim ental analysis of behavior already have shown

great promise.
W e shall not be in the clear, however, until other problem s

have been solved. Sim ply to change from a system in w hich large
numbers of students progress at the same rate to a truly in d ivid u
alized mode of instruction may mean drastic changes in the architec
ture of schools, in the roles of supervisors and teachers, and in daily
routines. “ M ore efficient instruction” should mean, if it means any
thing, that students w ill learn more rapidly, bu t if the first-grade
teacher also teaches what has been reserved for the second grade,
w hat is the second-grade teacher to teach? T h e recent history of
education in Am erica has been m arked by a postponem ent of in
struction— for exam ple, u n til students are “ ready” — bu t the trend
may now be reversed. A classical exam ple is the course in logic de
signed by Professor Laym an A llen of Yale U niversity L aw School.
T h e course worked so w ell w ith law students that it was tried in
college, and it w orked so w ell there that it was tried in high school.
A t last report it was being taught in the sixth grade. W h at happens
to a standard curriculum when changes of that m agnitude become
possible?
A reasonable answer m ight be that students w ill be taught a
great deal more during the same period of instruction. B u t it may
be tem pting, instead, to term inate education at an earlier age and
this raises other problems. W h at happens to em ploym ent figures if
large numbers of young people are turned loose on the jo b market
at an earlier age? (T h e term inal age in B ritain has recently been
raised one year, in part, it is said, to solve such a problem .)
Im proved instruction w ill also affect the em ploym ent of
teachers. In dividualized instruction could m ean a return to the
tutorial practices w hich existed before there were schools in the
present sense, and that m ay m ean that more teachers w ill be
needed. B ut tutorial instruction was not feasible for large numbers
and is clearly out of the reach of present educational systems. Hence
the search for new types of m aterials suitable for self-instruction
and for devices w hich evaluate students’ responses to such materials.
T h e K eller system takes advantage of the fact that one learns most
effectively when teaching, that individualized instruction m ay be
furthered by letting students teach each other. T h ese solutions seem

to suggest that it w ill eventually be possible to dispense w ith teach
ers. B ut any'increase in efficiency brings added educational objec
tives w ithin reach, m any of w hich dem and personal attention.
A loss in personal contact between student and teacher is not
necessarily a disadvantage in some fields of learning. T h e student
does not need a person to tell him w hether he has correctly trans
lated a particular sentence or solved a problem . T h e “ approval”
offered by a teacher differs from the confirm ation to be found in
program m ed materials, but it is not a “ natu ral” consequence of
behaving correctly and may, in fact, cause trouble. A n d before
regretting a loss in personal contact, we should look at the kinds
which now prevail in classrooms. W h en large num bers of students
are taught at the same time, few of them acquire effective verbal
behavior, oral’, or written. In m ultiple-choice exam inations and in
some kinds of program m ed materials, students merely check sen
tences which have been composed by others. T h e y have no chance
to learn to compose sentences themselves. Program m ed m aterials
can teach effective composition, bu t the flexibility characteristic
of social discourse calls for a teacher as an essential figure because
verbal exchange is almost necessarily individualized. W e may see
a revival of the art of speaking and w riting, and it w ill be im portant
because it involves much of the art of thinking.
Im proved efficiency in education makes time available for a
greater emphasis on personal exchange between teacher and student.
In addition, the teacher remains an essential figure in follow ing the
progress of a student and advising him on different courses of ac
tion. T hese new demands w ill require new kinds of training, and
some direct contact w ith the experim ental analysis of behavior may
be needed if the teacher is to take advantage of available behavioral
technology. T h e im portant thing is that more efficient practices
w ill give the teacher far greater pow er in fu lfillin g a far more ex
plicit assignment, and that should m ean a vast im provem ent in the
status of the teaching profession.
II
The Free
and Happy Student
His name is Emile. H e was born in the m iddle of the eigh

teenth century in the first flush of the modern concern for personal
freedom. His father was Jean-Jacques Rousseau, bu t he has had
many foster parents, am ong them Pestalozzi, Froebel, and Montes-
sori, down to A. S. N eill and Ivan Illich. H e is an ideal student.
Full of goodw ill toward his teachers and his peers, he needs no
discipline. H e studies because he is naturally curious. H e learns
things because they interest him.
U nfortunately, he is im aginary. H e was quite explicitly so
with Rousseau, who put his own children in an orphanage and pre
ferred to say how he would teach his fictional hero; b u t the modern
version of the free and happy student to be foun d in books by Paul
Goodm an, John H olt, Jonathan Kozol, or Charles Silberm an is also
im aginary. Occasionally a real exam ple seems to turn up. T h ere are
teachers w ho would be successful in dealing w ith people anywhere
— as statesmen, therapists, businessmen, or friends— and there are
students who scarcely need to be taught, and together they some
times seem to bring Em ile to life. A n d unfortunately they do so
140
The Free and Happy Student 141
just often enough to sustain the old dream. B ut Em ile is a w ill-o’-

the-wisp, who has led m any teachers into a conception of their role
w hich could ‘prove disastrous.
T h e student who has been taught as if he were E m ile is, h ow
ever, almost too painfully real. It has taken a long time for him to
make his appearance. C hildren were first made free and happy in
kindergarten, where there seemed to be no danger in freedom, and
for a long time they were found nowhere else, because the rigid
discipline of the grade schools blocked progress. B u t eventually
they broke through— m oving from kindergarten into grade school,
taking over grade after grade, m oving into secondary and on into
college and, very recently, into graduate school. Step by step they
have insisted upon their rights, ju stifying their demands w ith the
slogans that philosophers of education have supplied. If sitting
in rows restricts personal freedom , unscrew the seats. If order can
be m aintained only through coercion, let chaos reign. If one cannot
be really free while w orrying about exam inations and grades, down
with exam ination^ and grades! T h e w hole establishm ent is now
awash w ith free and happy students.
If they are what R ousseau’s Em ile w ould really have been

like, we must confess to some disappointm ent. T h e E m ile we know
doesn’t work very hard. “ C uriosity” is evidently a m oderate sort of
thing. H ard work is frow ned upon because it im plies a “ w ork
ethic,” which h’as som ething to do w ith discipline.
T h e Emile we know doesn’t learn very much. H is “ interests”
are evidently of lim ited scope. Subjects that do not appeal to him
he calls irrelevant. (We should not be surprised at this since
Rousseau’s Emile, like the boys in Sum m erhill, never got past the
stage of a knowledgeable craftsman.) H e m ay defend him self by
questioning the value of knowledge. K now ledge is always in flux,
so why bother to acquire any particular stage of it? It w ill be
enough to remain curious and interested. In any case the life of
feeling and emotion is to be preferred to the life of intellect; let us
be governed by the heart rather than the head.
T h e Emile we know doesn’t think very clearly. H e has had
little or no chance to learn to think logically or scientifically and is
E D U C A T IO N
easily taken in by the mystical and the superstitious. Reason is

irrelevant to feeling and emotion.
And, alas, the Em ile we know doesn’t seem particularly happy.
H e doesn’t like his education any m ore than his predecessors liked
theirs. Indeed, he seems to like it less. H e is m uch more inclined to
play truant (big cities have given up enforcing truancy laws), and he
drops out as soon as he legally can, or a little sooner. If he goes to
college, he probably takes a year off at some time in his four-year
program. A n d after that his dissatisfaction takes the form of anti-
intellectualism and a refusal to support education.
Are there offsetting advantages? Is the free and happy student
less aggressive, kinder, more loving? Certainly not toward the
schools and teachers that have set him free, as increasing vandalism
and personal attacks on teachers seem to show. N or is he particu
larly well disposed toward his peers. H e seems perfectly at home in
a world of unprecedented domestic violence.
Is he perhaps more creative? T rad itio n al practices were said
to suppress individuality; w hat kind of ind ivid u ality has now
emerged? Free and happy students are certainly different from the
students of a generation ago, bu t they are not very different from
each other. T h e ir own culture is a severely regim ented one, and
their creative works— in art, music, and literature— are confined to
prim itive and elemental materials. T h e y have very little to be
creative w ith, for they have never taken the trouble to explore the
fields in which they are now to be front-runners.
Is the free and happy student at least more effective as a
citizen? Is he a better person? T h e evidence is not very reassuring.
H avin g dropped out of school, he is likely to drop out of life too.
It w ould be unfair to let the hippie culture represent youn g people
today, but it does serve to clarify an extreme. T h e members of that
culture do not accept responsibility for their own lives; they sponge
on the contributions of those who have not yet been m ade free and
happy— who have gone to m edical school and becom e doctors, or
who have become the farmers who raise the food or the workers who
produce the goods they consume.
These are no doubt overstatements. T h in gs are not that bad,
nor is education to be blam ed for all the trouble. Nevertheless, there
is a trend in a well-defined direction, and it is particularly clear in
education. O u r failure to create a truly free and happy student is

sym ptomatic of a more general problem .
.W hat we may call the struggle for freedom in the W estern

world can be analyzed as a struggle to escape from or avoid pu n itive
or coercive treatment. It is characteristic of the hum an species to
act in such a way as to reduce or terminate irritating, painful, or
dangerous stimuli, and the struggle for freedom has been directed
toward those who w ould control others w ith stim uli of that sort.
Education has had a long and sham eful part in the history of that
struggle. T h e Egyptians, Greeks, and Romans all w hipped their
students. M edieval sculpture showed the carpenter with his ham m er
and the schoolmaster with the tool of his trade too, and it was the
cane or rod. W ei are not yet in the clear. C orporal punishm ent is
still used in many schools, and there are calls for its return where
it has been abandoned.
A system in. w hich students study prim arily to avoid the conse
quences of not studying is neither hum ane nor very productive. Its
by-products include truancy, vandalism , and apathy. A n y effort to
elim inate punishm ent in education is certainly com m endable. W e
ourselves act to escape from aversive control, and our students
should escape from it too. T h e y should study because they w ant to,
because they like to, because they are interested in w hat they are
doing. T h e mistake— a classical m istake in the literature of freedom
— is to suppose that they w ill do so as soon as we stop punishing
them. Students are not literally free when they have been freed
from their teachers. T h e y then sim ply come under the control of
other conditions, and we m ust look at those conditions and their
effects if we are to im prove teaching.
T hose who have attacked the “ servility” of students, as Mon-
tessori called it, have often p u t their faith in the possibility that
young people w ill learn w hat they need to know from the “ w orld
of things,” w hich includes the w orld of people who are not teachers.
Montessori saw possibly useful behavior being suppressed by school
room discipline. C ould it not be salvaged? A n d could the environ
ment of the schoolroom not be changed so that other useful be
havior would occur? C ould the teacher not sim ply gu ide the
144 EDUCATION
student’s natural development? O r could he not accelerate it by

teasing out behavior w hich would occur naturally bu t not so
quickly if he did not help? In other words, could we not bring the
real w orld into the classroom, as John Dewey put it, or destroy the
classroom and turn the student over to the real world, as Ivan Illich
has recommended. A ll these possibilities can be presented in an
attractive light, but they neglect two vital points:
(a) N o one learns very m uch from the real world w ithou t help.
T h e only evidence we have of what can be learned from a non
social world has been supplied by those w ild boys said to have been
raised w ithout contact w ith other members of their own species.
M uch more can be learned w ithout form al instruction in a social
world, but not w ithout a good deal of teaching, even so. Form al
education has made a tremendous difference in the extent of the
skills and knowledge w hich can be acquired by a person in a single
lifetim e.
(b) A much more im portant principle is that the real world
teaches only, what is relevant to the present; it makes no explicit
preparation for the future. T hose who w ould m inim ize teaching
have contended that no preparation is needed, that the student w ill
follow a natural line of developm ent and move into the future in
the norm al course of events. W e should be content, as C arl Rogers
has put it, to trust
the insatiable curiosity w hich drives the adolescent boy

to absorb everything he can see or hear or read about
gasoline engines in order to im prove the efficiency and
speed of his “ hot rod.” I am talking about the student
who says, “ I am discovering, draw ing in from the out
side, and m aking that w hich is drawn in a real part of
m e.” I am talking about my learning in w hich the ex
perience of the learner progresses along the line; “ N o,
no, that’s not w hat I w ant” ; “ W ait! T h is is closer to
w hat I ’m interested in, what I need.” “ Ah, here it is!
N ow I ’m grasping and com prehending w hat I need and
w hat I w ant to know !” 1
Rogers is recom m ending a total com m itm ent to the present mo
ment, or at best to an im m ediate future.
1 Rogers, C. Freedom to learn. Columbus, Ohio: Merrill, 1969.

B ut it has always been the task of form al education to set up

behavior which would prove useful or enjoyable later in the stu
dent’s life. P unitive m ethods had at least the m erit of providing
current reasons for learning things that w ould be rew arding in the
future. W e object to the punitive reasons, bu t we should not forget
their function in m aking the future im portant.
It is not enough to give the student advice— to explain that
he will have a future, and that xo enjoy him self and be more
successful in it he must acquire certain skills and know ledge now.
Mere advice is ineffective because it is n ot supported by current
rewards. T h e p o sitive. consequences that generate a useful be
havioral repertoire need not be any more exp licitly relevant to the
future than were the punitive consequences of the past. T h e student
needs current reasons, positive or negative, bu t only the educational
policy maker who supplies them need take the fu tu re into account.
It follows that m any instructional arrangements seem “ contrived,”
but there is nqthing w rong w ith that. It is the teacher’s function to
contrive conditions under which students learn. T h e ir relevance
to a future usefulness need not be obvious.
It is a difficult assignment. T h e conditions the teacher ar
ranges must be powerful enough to com pete w ith those under
which the student tends to behave in distracting ways. In what has
come to be called “ contingency m anagem ent in the classroom”
tokens are sometimes used as rewards or reinforcers. T h e y become
reinforcing when they are exchanged for reinforcers that are already
effective. T h ere is no “ natural” relation between w hat is learned
and what is received. T h e token is simply a reinforcer that can be
made clearly contingent upon behavior. T o straighten out a wholly
disrupted classroom something as obvious as a token economy may
be needed, but less conspicuous contingencies— as in a credit-point
system, perhaps, or possibly in the long run m erely expressions of
approval on the part of teacher or peer— m ay take over.
T h e teacher can often make the change from punishm ent to
positive reinforcem ent in a surprisingly sim ple way— by responding
to the student’s successes rather than his failures. T eachers have too
often supposed that their role is to point out w hat students are
doing wrong, but pointin g to w hat they arc doing right w ill often
make an enormous difference in the atmosphere of a classroom
and in the efficiency of instruction. Program m ed m aterials are help
146 EDUCATION
ful in bringing about these changes, because they increase the

frequency with w hich the student enjoys the, satisfaction of being
right, and they supply a valuable intrinsic reward in providing a
clear indication of progress. A good program makes a step in the
direction of competence almost as conspicuous as a token.
Programmed instruction is perhaps most successful in attack
ing punitive methods by allow ing the student to move at his own
pace. T h e slow student is released from the punishm ent which
inevitably follows when he is forced to move on to m aterial for
which he is not ready, and the fast student escapes the boredom of
being forced to go too slow. T hese principles have recently been
extended to college education, w ith dram atic results, in the Keller
system of personalized instruction.2
T h ere is little doubt that a student can be given nonpunitive

reasons for acquiring behavior that w ill become useful or otherwise
reinforcing at some later date. H e can be prepared for the future.
B ut w hat is that future? W h o is to say w hat the student should
learn? Those who have sponsored the free and happy student have
argued that it is the student him self who should say. His current
interests should be the source of an effective educational policy.
C ertainly they w ill reflect his idiosyncrasies, and that is good, but
how much can he know about the world in w hich he w ill eventually
play a part? T h e things he is “ naturally” curious about are of
current and often tem porary interest. H ow m any things must he
possess besides his “ hot rod” to provide the insatiable curiosity
relevant to, say, a course in physics?
It must be adm itted that the teacher is not always in a better
position. A gain and again education has gone ou t of date as teachers
have continued to teach subjects w hich were no longer relevant at
any time in the student’s life. Teachers often teach sim ply w hat they
know. (M uch of what is taught in private schools is determ ined by
what the available teachers can teach.) Teachers tend to teach what
they can teach easily. T h e ir current interests, like those of students,
may not be a reliable guide.
2 PSI New sletter, O cto b er 1972 (published by the C e n ter for Per
sonalized Instruction, G eorgetow n U niversity).
Nevertheless, in recognizing the mistakes that have been m ade

in the past in specifying what students are to learn, we do not
absolve ourselves from the responsibility of setting educational
policy. W e should say, we should be w illing to say, w hat we believe
students w ill need to know, taking the ind ivid u al student into ac
count wherever possible, but otherwise m aking our best prediction
with respect to students in general. V alu e judgm ents of this sort are
not as hard to make as is often argued. Suppose we undertake to
prepare the student to produce his share of the goods he w ill con
sume and the services he w ill use, to get on w ell w ith his fellows,
and to enjoy his life. In doing so are we im posing our values on
someone else? N o, we are merely choosing a set o f specifications
which, so far as we can tell, w ill at some time in the future prove
valuable to the student and his culture. W h o is any more likely to
be right? *
T h e natural, logical outcom e of the struggle for personal free
dom in education is that the teacher should im prove his control of
the student rafh e rxthan abandon it. T h e free school is no school at
all. Its philosophy signalizes the abdication of the teacher. T h e
teacher who understands his assignment and is fam iliar w ith the
behavioral processes needed to fulfill it can have students who
not only feel free and happy w hile they are being taught but w ho
will continue to feel free and happy when their form al education
comes to an end. T h e y w ill do so because they w ill be successful in
their work (haying acquired useful productive repertoires), because
they w ill get on well w ith their fellows (having learned to u nder
stand themselves and others), because they w ill enjoy w hat they do
(having acquired the necessary know ledge and skills), and because
they w ill from time to time m ake an occasional creative contribu
tion toward an even more effective and enjoyable way of life.
Possibly the most im portant consequence is that the teacher w ill
then feel free and happy too.
W e must choose today between Cassandran and U topian prog
nostications. A re we to work to avoid disaster or to achieve a better
world? A gain, it is a question of punishm ent or reward. M ust we
act because we are frightened, or are there positive reasons for
changing our cultural practices? T h e issue goes far beyond educa
tion, but it is one w ith respect to w hich education has m uch to
offer. T o escape from or avoid disaster, people are likely to turn to
148 EDUCATION
the punitive measures of a police state. T o w ork for a better world,

they may turn instead to the positive methods of education. W hen
it finds its most effective methods, education w ill be almost uniquely
relevant to the task of setting up and m aintaining a better way of
life.
I
IZ
Designing
Higher Education
T h e principal function of education is to transmit the culture

— to enable new members of a group to profit from what others have
already learned. It follows that the principal task of the student is
to learn what others already know.
T hese are not popular contentions. T h e y do not seem com
patible w ith & conception of the teacher as one who helps the
student discover the world for himself, or who stimulates a n atural
curiosity, or who guides intellectual, em otional, or m oral develop
ment, or who makes the learning process more m eaningful. O n the
contrary, teaching as the transmission of what is already know n is
often openly attacked as im posing the teacher’s values on the stu
dent, as intervening in a natural process of growth, and as under
m ining the freedom and dignity of the student. T h e teacher is
enjoined to avoid telling what he knows and to look for m eaningful
interchanges of other kinds.
If we ask for supporting evidence of the benefits w hich follow ,
we are likely to be told that the effects of these alternative practices
are not measurable. M easurem ent is appropriate only to the trans
mission of w hat is already known. W e must not expect to quantify
149
150 EDUCATION
the extent to which a teacher helps a student discover the world

for himself, or arouses his curiosity, or guides his development, or
makes learning m eaningful. Q uantifiable evidence may well ind i
cate that new ways of teaching have been less effective (the student
has indeed learned less of w hat is already known), b u t he has under
gone other, more im portant, changes, the evidence of w hich is
necessarily qualitative. In fact, exam inations mean little and should
be abolished.
T h e qualitative evidence is not always reassuring, however.
Practices designed to replace the transmission of w hat is already
know n have had some unanticipated effects w hich can scarcely be
said to recommend them. For exam ple—
1. T h e student who from kindergarten through college has

been commended by teachers who are on the alert for signs of dis
covery is likely to have an exaggerated notion of his powers of
discovery and of how much he has actually learned.
2. W hen tradition is dismissed as restrictive and creative
efforts therefore start from scratch, works of art, music, and litera
ture are necessarily, in a quite literal sense, prim itive.
3. T h ose who have been left to learn how to think by them
selves are quite unable to discover all the techniques w hich have
accum ulated over the centuries. W ith respect to intellectual self-
management, they are therefore almost defenseless against bad logic,
superstition, mystical nonsense, and dem agoguery. T h e ir ethical
self-management is similarly defective. It is not surprising that some
of them should have tried to make a virtue of irrationality.
4. M any cultural practices have prevailed because they sup
port behavior w hich “ takes the future into account” : they give
people current reasons for behaving in ways w hich have important
though possibly remote consequences. T hose w ho confine them
selves to matters of current relevance lack this support and are
forced to be, in the strictest sense, existentialists; they have not dis
covered the past and have no reason for behavin g effectively with
respect to a future.
T h e picture is not so bleak as I have painted it, because no

teacher merely aids discovery, or stimulates curiosity, or guides de-
Designing H ig h er Ed u cation
velopm cnt, or makes learning m eaningful. Students do read books,

enter into inform ative discussions, listen to music, see works of art,
and thus learn about w hat others have said and done and how they
have said and done it. Some transmission of the culture has gone on,
as it were, under the counter. M any contem porary educational
philosophers nevertheless seem dedicated to m inim izing transmis
sion, and I believe that they have done so for ideological reasons
largely unrelated to education.
It is easy to be m isled by w hat I have called the Idols of the
School.1 T h e Idol of the G ood T each er is the belief that w hat a
good teacher can do any teacher can do, and the Idol of the G ood
Student is the b elief that what a good student can learn any student
can learn. For reasons w hich are still beyond analysis, teachers are
sometimes extraordinarily effective even when their students are
not outstanding, and students sometimes learn a great deal even
w ithout the help of good teachers. A com bination of good teacher
and good student m ay have almost m iraculous results. N othin g
much needs to be done about higher education when these con
ditions prevail, but we must not forget the vast numbers of ordinary
teachers who cannot profit from the selection of good students, or
the vast numbers of ordinary students who clo not have good
teachers. For them, effective educational practices must be designed.
A n effective design must be based upon an understanding of
behavioral processes. T h e basic questions are these: W h y should
anyone teach,? and why should anyone learn? T h e y are questions
about hum an behavior, and recent advances in the analysis of be
havior are helpful in answering them.
Education was once m ainly punitive. T h e rod, the cane, and
the dunce cap were the schoolm aster’s tools. T h e student learned or
suffered the consequences. T h e pattern is still often defended, even
by those who were once caned or rid icu led (“ It was good for m el” ),
and som ething can indeed be said for it. U n der punitive sanctions
many students acquire techniques of self-control w hich carry over
into higher education. T h is was true when universities were little
more than libraries, and when only those who did not need to be
taught acquired an education. W h en universities began to teach,
1 Skinner, B . F. T h e technology of teaching. N ew Y o rk: A p p leton -

C entury-C rofts, 1969. Pp . 1 1 1 - 1 1 2 .
152 EDUCATION
punitive sanctions were added, and it is still true that most college'
students, whatever their professed am bitions or long-term goals, go
to lectures and read textbooks largely to avoid the consequences
of not doing so. L et those who disagree look at the evidence to be
found in the average student’s response to an occasional relaxation
of sanctions (an unexpected holiday or reduced assignment, for
example) or at the anxiety characteristically associated w ith exam
inations. (And let those who still disagree beware of the Idols of
the School!)
Aversive control is not easily justified in a dem ocratic society,
however, and there are m any other reasons why hum ane efforts
have been made to find alternatives. Teachers have naturally pre
ferred that their students should learn w ithout being coerced and
that they should even enjoy their studies. T h e learning that occurs
in daily life seems to show these features. W hy not bring the real
world into the classroom and throw away the birch rod? Arrange
conditions under which the student can do what he wants or likes
to do rather than what he has to do.
T h is is undoubtedly a step in the right direction, b u t it has
been misunderstood and misrepresented. T h e punitive conditions
are contrived by teachers, bu t nonpunitive conditions in the real
world are natural. W h at began as a change from coercion to posi
tive inducem ent seemed to emerge as a change in the role of the
teacher. T h e teacher could find things to interest the student; he
could guide his developm ent; he could be part of his natural social
environment; but he could not teach. T h e real world w ou ld do
the teaching. T h e teacher could only help the student learn.
W as this not an unexpected gain? W ou ld the real w orld not
be more likely to produce naturally effective behavior? A n appeal
to nature is always com pelling, and it is still a strong theme in
educational philosophy. It appears to challenge the notion of teach
ing as the transmission of w hat others already know. W h at the
student learns from contact w ith the real w orld is jeopardized when
the teacher interferes or meddles w ith the natural process. There
must be no intervention.
U nfortunately, the real w orld cannot bear the strain w hich is
“ thus imposed upon it. N ot m uch can be learned from it in one
short lifetime. T h e natural environm ent has more variety than a
badly designed classroom, bu t it is nevertheless still repetitious, and
Designing Higher Education 153
personal contact with it is lim ited in scope. W hen teachers aban

doned aversive practices, they lost control. T h e y discarded some
subjects arijd postponed others u n til the students were “ ready” for
them. T h ere is a sim ilar effect in higher education. Coverage is re
duced when the exp licit transmission of know ledge is m inim ized.
T h e relevant world also lacks-scope.
W hat is learned from real life is also faulty. It is w ell known
that the first effect of a natural environm ent in b u ild in g athletic or
artistic skills often proves troublesome at later stages; the coach or
teacher must suppress early natural forms of responding if a final
performance is to be perfected. T h ere are intellectual parallels.
O utstanding achievements have, no doubt, often been the culm ina
tion of a natural process of developm ent, b u t it is because the
instances are so rare that we call them outstanding, and it is perhaps
just because’.they are the exception that we attribute them to genius
rather than to outside help.
The physical environment, teaches awkward behavior as
readily as skilled; the social environm ent teaches aggression and
com petition as readily as good w ill and cooperation; in both worlds,
adventitious consequences breed superstition. T h e real world is
strongly punitive, and we are as likely to escape from m any parts of it
as we are to play hooky from school. A ll these features have parallels
in the world of books. T h e student m ay not learn m uch if he reads
only books of current interest; w hat he learns may be useless or
untrue; and bookish know ledge m ay be largely adventitious. C er
tainly, vast numbers of students learn to stay away from books al
together as soon as they are free to do so.
T u rn in g students over to the things w hich currently interest
them does not solve the m otivational problem s that arise in dis
pensing w ith overtly punitive methods. It is essentially the aban
donment of teaching. T h ere is a m uch more prom ising way to
remove punitive techniques from the classroom— in clu d in g the
easily concealed aversive features of college instruction— w ithout
abandoning the transmission of the culture or the com m unication
of what has already been learned by others. T h e alternative is sur
prisingly simple, and I have no doubt that the historian can find
many early statements. T h e problem is not, however, to state an
alternative, but to put it into general practice, and we are only now
in a position to do so. R ecent advances in our understanding of
154 EDUCATION
hum an behavior supply not only the means, bu t the confidence

needed to make significant changes.
W e need to replace contrived punitive conditions w ith con
trived alternatives, rather than w ith the natural alternatives to be
found in the current “relevant” environment. T h e alternative to
punishm ent is w hat the laym an calls reward, but lay usage has long
obscured an im portant detail. Behavior is indeed m odified by its
consequences, whether rew arding or punishing, but the im portant
thing is the way in which a consequence is contingent upon behavior.
T h e prescriptions of the U tilitarians never worked because they
emphasized the consequences (pleasure and pain), w hile neglecting
the contingent relations. U nder w hat conditions and at w hat m o
ment is an act follow ed by a pleasurable or painfu l consequence?
T h e experim ental analysis of behavior is concerned w ith the
contingent relations w hich prevail am ong three things— the situa
tion in w hich behavior occurs, the behavior itself, and its rew arding
or reinforcing consequences. Extrem ely com plex and subtle con
tingencies are set up in the laboratory and their effects studied. T h e
results have suggested alternatives to both the punitive sanctions of
traditional education and practices in which teaching is turned over
to the real world.
A simple contingency of reinforcem ent in prim ary or secondary
education— such as a token economy or a credit-point system in a
classroom— may seem like n othing more than reward in the tradi
tional sense. T h is is true because the reinforcers have been m ade as
conspicuous as possible in order to make them more clearly con
tingent upon behavior. T o bring a disrupted classroom under con
trol or to replace a punitive environm ent as qu ickly as possible,
simple and conspicuous consequences may be needed. B u t higher
education clearly calls for som ething else.
T h e traditional instrum ent through w hich one person benefits
from the experience of another is a book, and w hat is called a text
book is designed to work, as expeditious a change as possible. B ut
why should a student read a book or study a textbook? T h e possi
bility is worth considering that his behavior in doing so is a simple
function of the clarity and frequency of the reinforcing conse
quences. B ut w hat are those consequences? T rad ition ally, he dis
covers the extent to which he has understood what he has read from
the grade he receives on a test, bu t a grade is not contingent on
Designing Higher Education *55
behavior in an effective way. Unless the m aterial is itself reinforcing

because it is currently of interest, w hich it cannot always be, he w ill
presum ably read, if at all, only to avoid the aversive consequences
of a low grade. M uch more im m ediate positive consequences need
to be contrived.
T h e traditional concept of reward suggests som ething ex
trinsic to the behavior itself, and it is true that we could rew ard
correct responses to passages in a book with, for exam ple, m oney—
with, say, the remission of parts of a fee paid at the beginning of a
course. B u t nothing so crude is needed. A student is presum ably in
college to “ get an education,” and progress in doing so is itself
reinforcing. T h e only requirem ent is that progress should be con
spicuous. A student w ill continue to read a book if there is evidence
that he is undergoing a significant change, that he is increasingly
better able to db and say things, that he is progressing toward the
com pletion of the book or the course of w hich it is a part or the
curriculum of which the course is a part. For students whose be
havior is not thus reinforced, other reinforcers must be found, but,
in general, students can be induced to read— attentively and w ith
pleasure— by m aking sure that the consequences are im mediate,
clear cut, and frequent. A n d what holds for reading holds for other
parts of the educational assignment. T h e necessary conditions can
be most easily met if—
1. T h e student moves at his own pace. Differences in the rate

at which students work may be genetic or environm ental and are
probably both. O f importance, however, is not the source of the
differences, but the solution of the problem s they raise. A student
who is forced to go too fast misses m any reinforcing consequences
and, indeed, misses more and more of them as he falls further and
further behind. A student who could m ove faster bu t is held back
is not receiving reinforcements w hich lie w ith in his range. T h e
principle of individual pacing is as applicable to graduate instruc
tion as to teaching in the first grade.
2. T h e student should not m erely “ soak up inform ation” ;

he should respond, and his responses should be im m ediately evalu
ated so that successful responses w ill be reinforced.
3. T h e student should move through the m aterial in such a

156 EDUCATION
way that what he has just learned helps him to take the next step.
Signs of increasing power are im portant reinforcers. Reinforcem ent
w ill be m axim ized if he masters each stage before m oving on.
T h ere are readers who w ill strongly resent these references to

behavioral processes in a discussion of higher education, and, u n
fortunately, only those who have had some firsthand experience
with operant conditioning w ill be easily persuaded of their rele
vance. B ut it is no longer a matter of theory. Instructional systems
which observe these principles have been designed and tested and
have been conspicuously successful in inducing students to study
energetically, carefully, and w ith pleasure. W h at is learned need
not be relevant to their current life; it can therefore be selected to
be relevant to their futures. U nder these conditions students are less
likely to move to escape from education— not only in m inor ways
such as being inattentive or forgetful, but by taking time off or
dropping out of school or college altogether.
G ood programmed instruction observes these principles, and
a good program is a dram atic dem onstration of their power. It is
an extraordinarily useful device in acquiring knowledge of a new
field in an expeditious way. B u t the exigencies of contem porary
higher education often call for a more thoroughgoing restructuring
of practice, and the Personalized System Instruction originated by
Fred S. K eller is an outstanding exam ple of w hat can be done. A
recent report evaluating the system w ith respect to science teaching
describes the procedure as follows:
A student beginning a K eller course finds that the course

work is divided into topics or units. In a sim ple case, the
content o f the units m ay correspond to chapters of the
course text. A t the start of a course, the student receives
a printed study guide to direct his w ork on the first unit.
A lthough study guides vary, a typical one introduces the
unit, states objectives, suggests study procedures, and lists
study questions. T h e student may w ork anywhere— in
cluding the classroom— to achieve the objectives outlined
in the study guide.
Before m oving on to the second un it in the sequence, the
student must demonstrate his mastery of the first by per
fect or near-perfect perform ance on a short exam ination.
He is exam ined on the u n it only when he feels ade

quately prepared, and he is not penalized for failure to
pass first, second, or later exam ination. W hen the
student demonstrates mastery of the first unit, he is given
the study guide for the next unit. H e thus moves through
the course at his own pace. H e m ay meet all course re
quirements before the term is h a lf through or he m ay
require more than a term for com pleting the course.2
Thousands of courses are now being given on the K eller Plan

in the U nited States, South Am erica, and elsewhere. Adjustm ents
to local conditions may have to be m ade concerning scheduling,
grading, and the logistics of m aterial, space, and record-keeping,
but the essential features can be preserved. Such a course covers
standard material, but covers it m uch m ore thoroughly w hile avoid
ing most, if^not all, of the aversive features of traditional practices.
A t any given time, the student knows where he stands— w hat he has
done and still has to do— and the same inform ation is available to
the instructor arid useful to him for different reasons.
Such a course may be constantly im proved, since, as in other
kinds of instructional programs and in contrast w ith traditional
texts and lectures, weak points can be easily spotted and corrected.
O verall modes of presentation w ill no doubt continue to be im
proved as experience dictates, but it has already been established
that the so-called m otivational problem in higher education can be
solved in this, way. An environm ent can be constructed in w hich the
student has abundant reasons for studying, and for m astering m a
terial, even when it is not currently relevant to his personal p rob
lems or interests. It is therefore possible for education to enable
new members of a group to profit from what other members have
already learned— and in a highly expeditious way.
1 suspect that to some these w ill appear to be words of a
philistine. B ut the position seems to me not only defensible, bu t
inescapable. M ust we not ask educators and philosophers of educa
tion to state as clearly as possible the observable differences betw een
students who have had and who have not had a “ higher education” ?
2 K u lik , J., K u lik , C., and C arm ichael, K. T h e K eller P lan in science
teaching. Science, 1974, 183, 381-383. (Also see T h e K eller Plan
H a n db ook by Fred S. K e lle r and J. G ilm ou r Sherm an, pu blish ed in
1974 by W . A . B en jam in, Inc., M en lo Park, Cal.)
i 58 EDUCATION
A n d should they not say how education is to convert one into the
other? I subm it that the purposes and goals o f education most often
set forth in traditional discussions have suggested useful practices,
but have masked an unwillingness to be specific about these basic
issues.
O f course it is im portant to stimulate the student’s “ natural
curiosity,” but curiosity is of little avail if the student looks only at
the world about him. If in our efforts to stim ulate curiosity we
sacrifice the transmission of w hat other curious people have already
discovered, we deny the student access to an immense w orld lying
beyond his immediate reach.
O f course it is im portant that the student be creative and
im aginative, but if, in m aking absolutely sure that he is not being
im itative, we ignore or conceal the creative achievem ents of others,
we deny him the chance to play a role in a creative process reaching
far beyond his own lifetim e. T h e creative achievements of the past
have come from men who for the most part, as N ew ton said of
himself, “ stood on the shoulders of giants.” It is no service to the
student to insist that he stand w ith his feet firm ly planted on solid
ground.
Sooner or later a discussion of the goals of education turns to
ethics and morals, and it is precisely here that the appeal to a nat
ural process o f growth is most damaging. T h a t part o f a culture
which unquestionably demands transmission is its ethical and
moral practices. People are not ethical or m oral by nature, nor do
they simply grow ethical or moral. It is the ethical and moral
sanctions m aintained by other members of a group w hich induce
them to behave in ethical and m oral ways. T o leave ethical and
moral behavior to the n atural endowm ent of the in d ivid u al and a
natural process of growth is to promote ethical and m oral chaos.
W e must accept that a culture imposes its ethical and m oral stan
dards upon its members. It can do nothing else.
In more general terms, we must also accept that in trans
m itting a culture, education imposes w hat has already been learned
by others upon its students. T o a considerable extent it must decide
in advance what a student is to learn. C urrent philosophies of edu
cation spring in part from an unwillingness to take on this respon
sibility.'E ducational policy-makers are u n w illin g to specify what is
worth knowing, and once again they leave the decision to the
student. B ut the student is in no position to specify w hat w ill

ultim ately be useful. T h is is obvious w hen he is beginning the study
of a large field, such as biology or physics, but it is equally true as
he enters upon a broader “ education for life.” Personal idiosyncra
sies certainly need to be taken into account, and a program w hich
yielded regim entation and uniform ity w ould be bad educational
design, but there are sources of regim entation and u n iform ity in
programs which leave decisions to students, too. W e should not be
misled because students are readier than we are to accept the
responsibility of designing their programs, because their readiness
may spring only from current satisfactions.
Those who let students themselves decide what they are to
study and leave teaching to the physical, social, and textual envi
ronments are essentially abdicating as teachers. T h e y betray students
who already1, care about their future, and fail to help those who
have never had any reason to care. It is possible that education
w ill someday be held responsible for the m illions of youn g people
who are now not only not w ell prepared for the future, b u t not
even sure that they have one.
. li PART IV
A MISCELLANY
The Shaping of Phylogenic Behavior
The Force of Coincidence
Reflections on Meaning and Structure
Walden {One) and Walden Two
Freedom and Dignity Revisited
Freedom at Last, From the Burden of Taxation

J 3
The Shaping
of Phylogenic Behavior
i
t
\
A n unusual topography of operant behavior can often be

shaped by m aking the contingencies of reinforcem ent increasingly
more com plex. In a simple dem onstration a box is divided into two
parts by a low wall, and a hungry rat is placed on one side and food
on the other. T h e rat possesses an in itial repertoire of responses
(clim bing and* jum ping), some of w hich take it over the wall and
are reinforced by food. As a result, responses having the required
topography are strengthened and soon occur on later occasions. If
the w all is then made slightly higher, only some of these responses
will be successful, but they w ill begin to occur more frequently, and
as a result new topographies of response w ill appear, w hich w ill
meet even more dem anding contingencies w hen the height of the
w all is again increased. If the height is not increased too rapidly
(if some responses are always successful), a very energetic and skill
ful repertoire w ill result. T h e rat w ill eventually go over a w all
that it w ould never cross if it had not been exposed to such a
program.
A sim ilar result could be obtained by genetic selection. Rats
which most readily or successfully crossed a w all of a given height
163
164 A MISCELLANY
could be selected for breeding to produce a population among

which some members would be more likely to cross a higher wall;
and the process could then be repeated. T w o strains of hooded rats
used in some early genetic experiments at the old Bussey Institute
at H arvard U niversity could be separated instantly by putting
them one at a time in a shallow box; members of one strain quickly
escaped, w hile members of the other strain rem ained in the box
indefinitely. From the observed facts alone, it was impossible to tell
whether the difference was due to operant conditioning, follow ing
some such procedure as just described, or a genetic characteristic.
W hen ontogenic shaping can be ruled out, it is standard practice
to infer that genetic selection has been responsible for an observed
difference of this sort.
T h e behavior of hom ing to a fixed site raises similar ques
tions. A n organism can be taught to hom e through operant con
d itioning by repeatedly placing it in positions from which re
turning to a fixed site is reinforced. T h e field may be progressively
enlarged, subject only to the limits imposed by the available
time and the locom otive capacities of the organism. A parallel pro
cess in which contingencies of survival replace contingencies of rein
forcement is usually inferred in order to explain, for example,
the behavior of bats in leaving and returning to a cave. As the size
of the original group increased, those bats w hich w ent farther and
farther afield to obtain food and returned successfully were pre
sumably more likely to survive and breed and transmit the behavior.
Contingencies of reinforcem ent w hich shape ontogenic be
havior can be arranged and studied in the laboratory. Most of the
contingencies of survival responsible for phylogenic behavior ob
served in the field are merely inferred. B u t some evidence of
environm ental conditions which probably changed in such a way
as to shape com plex phylogenic behavior has emerged in connection
with the theory of continental d rift and the spreading of the sea
floor.
A n exam ple w hich has recently attracted attention is the be
havior of the green turtle (Chelonia mydas), w hich feeds in the
water pastures along the coast of Brazil and swims m ore than a
thousand miles to breed on Ascension Island. T h e journey takes
several weeks and demonstrates rem arkable n avigational skills, since
a five-mile target must be h it after about one thousand miles of
The Shaping of Phylogénie Behavior 165
travel in the open sea. As Carr 1 has argued, it is hard to im agine

that behavior as com plex as this could have evolved through natural
selection under the present circumstances. In 1964, Fraser 2 pointed
out that 150 m illion years ago the turtles “ would have only a nar
row arm of the sea to negotiate; for since the ancient latitudes of
R io Doce and the projection of Ascension on the A frican coast are
in precise agreement, the nesting ground was lying there just across
the water, a mere hundred miles or so from its land-based home
am ong the elephant grass of east B razil.”
T h e case is not quite that simple, however, as C arr and
Colem an 3 have recently pointed out. Ascension Island is a rela
tively late mem ber of a chain of volcanic islands w hich have ap
peared as the sea floor spread. T h e turtles may first have gone to
islands close {o Brazil, but these slowly submerged. T h e y presum
ably then w eht on to more distant islands in the same general
direction, of w hich Ascension was the last to appear. T h e fact
remains that the behavior of feeding along the shores of Brazil and
swim m ing to a breeding ground relatively safe from predators met
progressively m ore dem anding conditions as the distances increased,
either continuously or in a step-wise fashion.
A n other program of contingencies o f survival resulting from
the spreading of the sea floor may explain the behavior of A m eri
can and European freshwater eels, w hich appear to have a com m on
spawning ground in the Sargasso Sea. It was A lfred W egener, the
father of the theory of continental drift, who noted the relevance
of this fact to his theory, in the fourth edition of his Origin of con
tinents and oceans.4 T h e point was suggested to him , as early as
1922, by J. Schmidt. Schmidt’s early research5 showed that the
European eel (Anguilla anguilla) breeds in an area northeast and
north of the W est Indies. T h e young eels in the larval stage (called
1 Carr, A . A d a p tiv e aspects o f the scheduled travel o f Chelonia. In

Animal orientation and navigation. Corvallis: O regon State U n iv e r
sity Press, 1966.
2 Fraser, R. Understanding the earth. N ew York: P en gu in, 1964.
3 Carr, A . and C olem an, P. J. Seafloor spreadin g theory and the
odyssey of the green turtle. Nature, 1974, 249, 128-130.
4 W egen er, A . The origin of continents and oceans (4th ed.). 1929.
(English translation, N ew York: D over, 1966).
5 Schmidt, J. Nature, 1923, 1 1 1 , 51-54 .
166 A MISCELLANY
leptocephali) are small and leaflike in appearance. T h e y rise to the

surface and w ith the help of the G u lf Stream move toward Europe.
Year-old larvae are to be found in the m id-Atlantic and as far east
as the Azores; two-year-old larvae are found on the shores of
E urope and in the M editerranean; and after three years they un
dergo metamorphosis and appear as elvers in freshwater streams,
where they mature. Years later the m ature eels return to the place
in which they were hatched. As the sea floor spread, the spawning
ground moved m uch farther from European rivers than from
Am erican, and W egener credits H . Osterwald w ith realizing that
“ the gradual drift of this ocean basin plus Am erica away from
E urope” explains the fact that the larval stage of the European eel
is three years while that of the Am erican is only one.
In 1969 the present author pointed out the possible relevance
of the spreading of the sea floor to the shaping of the phylogenic
behavior of the eel.6 T h e behavior is truly rem arkable. From the
breeding ground to the m outh of the N ile, for exam ple, a young
eel in the larval stage travels, as Schmidt pointed out, about 6,000
miles in a period of three years. Ocean currents w ill explain only
part of this m igration. T h e m ature eel makes a return journey
of the same distance against the current. As w ith the green turtle,
it is hard to believe that their extraordinary behavior could have
arisen from natural selection under present environm ental condi
tions. B ut if the distances were at first short, and if they increased
no more than a few inches each generation, as the theory of con
tinental drift implies, then some members of each generation could
have satisfied the new contingencies and bred to transmit the be
havior.
In 1948 W oJfson7 argued that “ continental drift was the
stimulus for the evolution of the more highly developed forms of
m igration [of birds].” He stated his hypothesis in four steps:
(a) Before the advent of continental drift m any birds were
perform ing short flights between breeding and feeding areas.
(b) W ith the onset of d rift these areas diverged slowly.
0 Skinner, B. F. C ontingencies of reinforcem ent: A theoretical anal

ysis. N ew York: Appleton -C en tury-C rofts, 1969. (See C hap . 7).
7 W olfson, A. B ird m igration and the con cept o f con tin en tal drift.
Science, 1948, 108, 23-30.
(c) T h e birds continued their use of these areas because of

their well-developed hom ing instincts.
(d) As 'the distances increased, only those individuals that had
the necessary sources of energy for the flight survived.
W olfson pointed to the fact that, for exam ple, the A rctic tern,
w hich breeds in northern N orth Am erica and m igrates to the A n t
arctic, first flies eastward across the A tlan tic to E urope and then
southward along the African coast. T h e eastward journey m ay at
first have been very short, bu t as the continents separated, successive
generations w ould have flown slightly greater distances and what
seems now like a nonadaptive flight pattern is thus explained.
N orth A tlan tic salmon show a pattern w hich is the reverse of
that of the eel; they breed in freshwater rivers bu t live most of their
adult lives in the ocean. T h e y also show a long east-west m igration
that could haVe been shaped by the separation of the continents.
A ccording to O rr,8 ‘‘Am erican A tlan tic salmon have been taken on
the W est Coast of Greenland, and those from Sweden have also
been recorded-there, a distance of almost 3,000 miles from the hom e
streams of both.” A t one time the distances were very m uch shorter,
and navigation must have been scarcely more than a m atter of
m oving along a coastline.
A different effect of the plate tectonics related to continental
drift may have produced contingencies o f survival w hich shaped
certain features of the behavior of salmon on the W est Coast.
C olum bia R iv er salmon, for exam ple, breed in shallow, relatively
still fresh water, with a gravel base, but spend most of their lives in
the Pacific Ocean, particularly in the G u lf of Alaska. A t the present
time the m ature salmon reach suitable breeding grounds only by
fighting their way up through treacherous rapids and waterfalls. A t
an earlier stage in the developm ent of the river, however, suitable
breeding conditions may have been found close to shore. A t that
time the salmon need not have possessed any of the extraordinarily
skillful and pow erful behavior w ith w hich they now conquer the
hazardous flow of the river. As the river m atured, suitable breeding
grounds should have receded from the shore, setting up a program
of contingencies of survival w hich shaped the present behavior,
8 O rr, R . T . A nim als in m igration. N ew Y o rk: M acm illan , 1970.

168 A MISCELLANY
each generation being required to meet contingencies only very

slightly more difficult than the preceding.
Phylogenic shaping of behavior is plausible, of course, only
if the species, or at least some recognizable earlier form, was in
existence when the geological change occurred. T h e green turtle,
the eel, and m igrating birds seem to meet this requirem ent, and
Professor P. J. Colem an (personal com m unication) has pointed out
that “ the W estern seaboard is a geologically young entity and a
developm ent of the west flow ing rivers is certainly w ithin the time
span of the salmon group as it is recognized today.”
A continuous shaping process, in both ontogenic and phylo
genic behavior, has been only slowly recognized. W egener and
Schmidt pointed to the breeding practices of the eel m ainly to sup
port the theory of continental drift. W olfson emphasized the selec
tion of birds capable of flying greater distances and pointed out that
the path of flight was more significant than the extent in support
in g an explanation in terms of continental drift, but he assumed
that birds follow ed such a path because of a hom ing instinct. Fraser
drew on conclusions from the lengthening voyage of the green
turtle, but C a r r 9 emphasized the im portance of evidence “ bearing
on probable paleolithic conditions at the time of origin, or of refine
m ent, of each pattern [of island-finding behavior]” (italics added).
C arr argued that “ any female with the urge and capacity to go out
to the island w ill contribute m ore genes to reinforce the island
breeding pattern in the race. Each generation, more turtles go out
to the island simply because their genotype was m ade more
prevalent by the island-seeking tendency of the preceding genera
tion. T h e island-seeking m igration is thus a successful evolutionary
venture, and has become the established regim en for the popula
tion.”
It is this genotype w hich changed under selection as the con
tinents moved apart. C arr and Colem an speak of the repetitive ex
tension of previous travel paths and suggest that “ the process of
racial learning is of the repetitive, stepping-stone type, w hich re
quires no radical change in behavior at any p oint.”
It is possible to be a little more specific. T h e shaping of
/ , ‘."-32
9 Op. cit.
phylogenic behavior, like the shaping of oniogenic behavior, in

volves at least three things:
1. Behavior comes under the control of new stim uli. In the

phylogenic case this could involve extensive changes in sense organs,
but it also involves the developm ent of particular forms of stimulus
control which, as in the ontogenic case, do not require changes in
sensitivity. Shaping usually involves a shift in the range of con
trolling stimuli. W hen a stim ulus w hich is not central to an exist
ing range acquires special controlling pow er (either through genetic
selection or operant reinforcem ent), a new range emerges in w hich
new stimuli are now for the first time effective and can be
strengthened to produce a range differing still further from the
original. T h u s, iif an unusual visual pattern becomes particularly
effective, new patterns even more unusual begin to exert control
and are subject to further selection as the contingencies change.
A distinction is often made in discussing hom ing behavior or
migration between “ know ing where to go” and “ know ing how to
find the way,” but the concept of know ledge causes trouble. T h e
organism begins to m igrate or home by responding to current stim
uli; some of its responses m ay produce other stim uli w hich then
take over. In follow ing a long path the organism may be under the
control of a succession of discrete stim uli or such a sustained
stimulus as that used in celestial navigation. (T h e present argum ent
does not, unfortunately, throw any light on the stimuli w hich
actually exert control, although the search for such stim uli may be
aided by a consideration of the probable requirements in earlier
stages of the shaping process).
2. T h e topography of behavior changes. In phylogenic be

havior this may involve elaborate changes in the strength and m ode
of operation of effectors, but it also involves changes in the effective
topography of a stable system, as in the ontogenic case. Shaping
usually involves a shift in the range of effective topographies. W hen
an unusual form of response is strengthened by genetic selection or
operant conditioning, new forms m ay appear for the first tim e
which can in turn be strengthened as the contingencies change.
3. Easily overlooked is a third effect of shaping— a mainte-
A MISCELLANY
nance of, or an increase in, the probability that behavior having a

given topography and under the control of given stimuli w ill
actually occur. T h e effect on probability is due, in phylogenic be
havior, to the selection of genotypes and, in ontogenic behavior, to
operant conditioning. T h e tendency to behave in a given way upon
a given occasion has been attributed to instinct in the phylogenic
case and habit in the ontogenic case. In both it has been associated
w ith the concept of purpose, and in ontogenic behavior w ith expec
tation or intention. Concepts of this sort add nothin g to the ob
served facts, and they cause trouble because, by appearing to refer
to inner determiners of behavior, they often serve as substitutes for
the further explanation w hich w ill eventually be provided by
physiology.
It is not necessary to refer to underlying structures or func
tions in order to study the way in w hich an organism inherits a
tendency to behave in a given way in the presence of given stimuli,
but the physiology w ill, it is hoped, eventually be understood. W hat
evolves is an organism as a physical system, and it is such an
organism that is m odified by operant conditioning. W e do not
know whether the physiological changes which occur in the shaping
of phylogenic behavior are similar to those w hich occur in operant
conditioning. Certainly, there are vast differences in the conditions
under which the two processes occur. It is not im possible, however,
that operant conditioning, itself an evolved feature of an organism,
should have utilized a physiological system w hich had already been
developed in natural selection.
H
i
. \
In the grade school that I attended as a child, a single teacher

taught two grades in the same room. W h ile one class recited, the
other w orked on its assignments. O ne day in third grade, w hen
my teacher was talkin g w ith the other class, I raised my hand, waved
it w ild ly to attract her attention, and said “ I was reading the word
‘m id d le’ just when you said it.” B oth classes laughed. I had been
impressed by the coincidence, but I should have been impressed
by the fact that I was impressed.
T h e current revolt against reason and science has m ade m uch
of psychic phenom ena, such as precognition in dreams and extra
sensory perception, and various tran scen dental,,states of conscious
ness. In his book T h e R oots of Coincidence 1 A rth u r Koestler has
discussed another k in d of evidence said to be neglected by the
scientific establishment: things happen w hich cannot be explained
“ by the laws of chance.” A fter the book was published m any people
wrote to him to report additional strange coincidences, and a
1 K oestler, A . T h e roots of coincidence. N ew York: R a n d om H ouse,

*972-
17 1
172 A MISCELLANY
second volum e is, I believe, to appear containing further data. T h e

evidence cuts both ways. It shows that there are.m any coincidences
w hich are hard to explain, bu t it also shows that coincidences attract
an unusual am ount of attention and are long remembered.
Coincidence is the heart of operant conditioning. A response
is strengthened by certain kinds of consequences, but not neces
sarily because they are actually produced by it. Indeed, it is quite
unlikely that a behavioral process could have evolved w hich took
into account the manner in which a response produces an effect.
T h ere are too m any reasons why consequences follow behavior, and
they depend on features of the environm ent w hich are too unstable
to play any part in natural selection. B ut since an event which
follows another event is likely enough to have been caused by it,
coincidence suffices.
T h a t solution to the problem of causality is not, however, free
of trouble. It means that behavior may be strengthened by merely
adventitious consequences, and such behavior is not likely to be
useful. V uln erability to coincidence must have increased as the
process of operant conditioning accelerated, and when a single
instance of response-and-consequence began to w ork a significant
change, various kinds of superstitious behavior were inevitable. T h e
m ore “ intelligen t” the organism, the more likely it was to be super
stitious. Moreover, superstitious behavior is often self-perpetuating
and even self-enhancing. Recovery from a self-limited illness, for
example, reinforces any therapeutic action a person may take, and
since he is therefore more likely to take it again when ill, it is likely
to be adventitiously reinforced again and hence further strength
ened.
T h e fact that two basic types of superstitious behavior are
com m only observed in such an “ u n in telligen t” organism as a
pigeon 2’3 suggests that superstition must have been very widespread
before corrective measures were developed. Such measures are, of
course, now common. W hen a response appears to have had an
unlikely consequence, a fairly characteristic m ove is to repeat it
2 Skinner, B. F. ‘Superstition ’ in the pigeon. Journal of Experim ental

Psychology, 1948, 38, 168-172.
3 M orse, W . H . and Skinner, B. F. A second type o f superstition in
the pigeon. Am erican Journal of Psychology, 1957, 70, 308-311.
immediately. If the same consequence follows, the response is fu r

ther strengthened. (Using essentially a synonym for “reinforced,”
we say it is “ confirmed.” ) If the same response does not follow (as
is likely to be the case if the first was adventitious), the acquired
strength is lost through extinction and subsequent behavior is then
“ in better touch w ith reality.”
It is possible that people learn to test the causal efficacy of
their behavior sim ply because they are then m ore likely to be con
sistently reinforced, but more com plex tests of the significance of
consequences are usually acquired from others. Someone must
devise each test for the first time, bu t no one person could devise
m any of them w ithin a single lifetim e. M ost people probably learn
even the simplest measures from others.
T h is is all part of the field of self-knowledge and self-manage
ment, and it isialmost w holly a social product. It is only when other
people ask “ W hy did you do that?” that we begin to exam ine the
contingencies responsible for our behavior. As a simple operant
we open a w indow because we then get fresh air, but it is only
when someone asks “ W hy?” that we describe the relation between
our behavior and its consequences, as by saying “ I opened the w in
dow because fresh air then came in .” T h e sequence is taken to be
sufficient evidence. R ecent pow er failures turned up a n um ber o f
stories of people who described them as adventitious consequences
in the same way. A small boy w alkin g along the street striking trees
and picket fences w ith a stick happens to strike an electric ligh t
pole, or a housewife happens to p lu g in an iron, just as all the
lights in the city go out, and both may report, and under certain
social circumstances may insist, that they have caused the trouble.
T h e physical facts which explain w hy fresh air comes in
through an open w indow or w hy plu ggin g in an iron blows a fuse
or blacks out a city have nothing to do w ith personal action. A s the
history of the idea of causality abundantly demonstrates, one thing
has often been said to cause another simply because it precedes it—
as the operant paradigm seems to imply. A very simple exam ple,
involving spatial features, is seen in the k in d of settings studied by
M ichotte.* W hen one black spot m oving on a white field ap
proaches another and the other moves away ju st as contact is m ade,
4 M ich otte, A . L a p ercep tion de la causalité. Paris, 1946.

174 A MISCELLANY
the first is said to cause the second to move. T h e first spot “ strikes”
the second as one billiard ball strikes another. A n d if we convert
spots into livin g things a whole new realm of causality seems to
open. I once made some small “ turtles” for a child by pasting
M exican ju m pin g beans on small squares of paper w ith the corners
bent down as legs. T h e turtles moved about on a plate of glass as
the beans “ jum ped.” W hen one turtle moved toward another just
as the other moved away, the child im m ediately reported that the
second turtle had been frightened.
W e gain from analyzing the contingencies w hich affect our
behavior— using scientific and statistical methods— in part because
we reduce our vulnerability to merely incidental cases, and our
gains lead us to continue to do so when the contingencies are super
stitious. M any myths appear to represent this function. A n y be
havior executed just before it begins to rain is strengthened if rain
is reinforcing, as it is at the end of a severe drought. A n d because
the more conspicuous the behavior, the more effective the coinci
dence, an elaborate ritual such as a rain dance may evolve. In an
area in w hich drought is self-limited, people are likely to begin to
dance near the end of a drought— when the probability of “ rein
forcem ent” is particularly high— and such a superstition is therefore
self-perpetuating and even self-enhancing. A person who is asked
w hy he dances may simply reply that rain then soon follows, bu t if
he is asked why dancing produces rain, he may answer by generalizing
from instances in w hich sim ilar consequences are not adventitious.
Social contingencies offer the richest sources, and the dance may be
interpreted as a form of asking for rain or pleasing and hence
appeasing someone who is w ithholding rain.
W e dismiss rain dancing as a form of superstition because the
adventitious nature of the consequences can be dem onstrated “ sta
tistically,” but we continue to be fascinated by coincidences which
are “ inexplicable according to the laws of chance.” T h is is likely
to be the case so long as we forget that the w orld in w hich we live
is an extrem ely com plex sample space, in w hich it is doubtful
whether there are any “ laws” of chance w hich apply to m any of
the single events occurring in it. Coincidences are certainly to be
expected, and the sheer num ber m ay be felt to b u ild u p a case for
a force or agent w hich is metaphysical, supernatural, or at least not
part of the current corpus of science. B u t the mere accum ulation of
instances has less to do w ith probability than w ith the striking force
of coincidence.
It is a rare person w ho picks up a hand o f thirteen spades at
bridge and views it as no less likely to occur than any of the other
hands he has picked up durin g his experience as a player, or who
enjoys a'ru n of luck at roulette w ithout calling it his lucky day or
acknowledging his debt to Lady Luck, or who when an honest coin
has come heads twenty-five times in a row w ill not then be more
likely to bet on tails. T h e genetic endowm ent responsible for our
behavioral processes cannot fu lly protect us from the whim s of
chance, and the statistical and scientific measures we devise to bring
our behavior under the more effective control of nature are not
adequate for the extraordinarily com plex sample space in w hich we
live. Science has not ignored some un derlying order; it has not yet
devised ways of «protecting us against spurious evidences o f order.
Reflections on Meaning
and Structure
T h ’expense of spirit in a waste of shame

Is lust in action; and till action, lust
Is perjur’d, m urd’rous, bloody, fu ll o f blame,
Savage, extreme, rude, cruel, not to trust;
E njoy’d no sooner, bu t despised straight;
Past reason hunted; and no sooner had,
Past reason hated as a swallow’d bait
O n purpose laid to make the taker mad:
M ad in pursuit, and in possession so;
Had, having, and in quest to have, extreme;
A bliss in proof, and p rov’d, a very woe;
Before, a joy propos’d; behind, a dream
A ll this the w orld well knows; yet none knows well
T o shun the heaven that leads men to this hell.
In their detailed analysis of the structure o f Shakespeare’s

Sonnet 129, Jakobson and J o n es1 note that “ and very wo instead of
a very wo is an obvious misprint, under the assim ilative influences
of the antecedent and in the same line and in the first two lines
of the same quatrain.” H avin g w ritten (or set in type) the word and
three times in three lines, the poet (or typesetter) w rote (or set) it
again although the m eaning called for a. B ut w hat abou t the third
of these four ands? C an we be sure that w ithout the first two it
w ould not have been but, say, or yet? T h e evidence is clearer in the
1 Jakobson, R ., and Jones, L . G . Shakespeare’s verbal art in th’

exp en ce of spirit. T h e H ague: M ou ton, 1970.
Reflections on Meaning and Structure 177
fourth instance because a n d is a mistake, but there are presumably

reasons w hy words are written when they are not mistakes.
A n assimilative influence may bear on less than a w hole word,
and the fact that w o rimes w ith so is an exam ple. Here again the
evidence is better when it explains a blemish. Jakobson and Jones
quote J. M. Robertson to the effect that “ collapse recurs w hen a
very w o fades into a d re a m e for the rim e’s sake” and Edw ard
H ubler to the effect that “ the anticlim actic position of n o t to tru st
is ow ing entirely to the need for a rim e.” B u t have we any reason
to suppose that assimilative influences are not at work in rimes
which do not show collapse or anticlim ax? In Sonnet 90, for ex
ample, in the line A n d o th e r str a in s o f w o, w h ic h n ow se e m w o
the second w o may be attributed in part to the first and to an
earlier occurrence in rim ing position in the second quatrain. In a
prose version som ething closer to u n b e a r a b le m ight have been more
to the point.
T h e m eaning of other words must also be taken into account.
It is not surprising that one who has been speaking of perjury,
murder, and madness should say w o, either through w ord associa
tion or as the effect of a common subject m atter. B ut m eaning
raises special problems. Structure has the enorm ous advantage of
being accessible. T h e formal properties of Sonnet 129 are not all
im m ediately obvious, as Jakobson and Jones have convincingly
shown, but once pointed out they are there for everyone to see.
B ut what are meanings, and where are they to be found? A dic
tionary does not give the meanings o f words, it gives other words
having the same meanings. T h e m eaning of a poem is sim ilarly
elusive. W hen a person tells us w hat a poem means to him , he
merely tells us how a m eaning m ight be otherwise expressed. Sup
pose he paraphrases Sonnet 129 in some such w ay as this: “ Sexual
behavior is both rewarded and punished, and w hen we enage in it
because of the rewards, we subject ourselves to the punishments. N o
one knows w hat to do about it.” T h is is not one m eaning o f the
sonnet; it is only one other way of saying w hat the sonnet says.
T o get closer to m eaning we should have to look at the cir
cumstances under which the sonnet was w ritten. W e cannot do that
with Sonnet 129. W e are lim ited to m aking a few guesses about
what m ight have happened to Shakespeare to induce him to write
as he did. It has often been pointed out that the sonnet is bitter.
178 A MISCELLANY
W hat could have been so bad about sex? Tem porary impotence
(“ . . . p a ssio n e n d in g , d o th th e purpose lo s e ” ) is scarcely bad
enough. Social, legal, or religious sanctions m ay have been “ blouddy
full of blam e” and could have led Shakespeare to “ dispise” himself,
but they are scarcely perjured or murderous. Perhaps the best guess
is the pox, but we shall probably never know. Fortunately, so far
as the present point is concerned, it does not matter. Assume any
plausible set of circumstances; how could they have given rise to a
sonnet?
W e gain nothing from supposing that the sonnet first came
into existence in some preverbal form, that circumstances gave rise
to an idea in Shakespeare’s m ind w hich he then put into words. If
we begin in that way, we must explain how circumstances give rise
to ideas, and that is m uch more difficult than explainin g how they
give rise to verbal behavior.2 Certain events in Shakespeare’s life
induced him to em it two opposed and seemingly incom patible sets
of responses w ith respect to sex. T h e sets are epitom ized by h ea ven
and h e ll. W hen lust is heaven it is a b liss and a jo y , and it is then
h u n te d and p u r s u e d . W hen lust is hell, it is uncouth (e x tr e a m e ,
ru d e), deceptive (p e r ju r e d , n o t to tru st), costly (a w aste, an e x p en se),
dem eaning ( a thing of s h a m e , fu ll of b la m e), and violent (cruel,
savage, b lo u d d y , m u rd ro u s), and it is then d is p is e d and h a te d . T h e
two sets of responses are not really incom patible, because lust is one
thing or the other depending on the time. H eaven comes first and
hell follows, and this tem poral aspect of the circumstances evokes
several pairs of terms (in a c tio n — till a c tio n , n o s o o n e r — str a ig h t, in
p u r s u it, — in p o ssessio n , h a d — h a v in g , and b e fo r e — b e h in d .)
T h ese key expressions, w hich can be thus arranged in thematic
groups, may be close to the “ prim ordial” verbal m aterial from
which the sonnet was composed. (T hey were not all necessarily
available when the poet began to write, since associative and assimi
lative influences could have generated other m aterial as the writing
proceeded.) T h e y are far from being a sonnet, and there is much
about them that any im agined set of circumstances w ill not easily
explain. C ertainly many other responses could have been evoked.
W hy this particular selection of synonyms? A n d w hat determined
2 Skinner, B. F. Verbal behavior. N ew Y o rk : A ppleton-Century-

C rofts, 1957.
the attribution of “ parts of speech?” A setting which gave rise to a

w aste o f sh a m e could easily have evoked a s h a m e fu l w aste, or s h a m e
fu lly wastedy, or a sh a m e a n d a w aste. A n d what about the order in
which the responses occur? N ine adjectives or adjectival phrases
were strung together in the first quatrain: p e r ju r d , m u r d r o u s,
b lo u d d y , f u l l o f b la m e , sa vage, e x tr e a m e , r u d e , c r u e l, n o t to trust.
W hy that one order, when 362,879 other orders were possible? A n d
which pair of terms was to go w ith w hich indicator of time? W h y
not b e fo r e in jo y d , b e h in d d is p ise d ? O r in jo y d in p r o o f, d is p is e d
w h en p r o v d ? A n d what was to be done about assertion? (Very little
was in fact done in Sonnet 129: only the two is’ s in the first quatrain
and the two k n o w ’ s in the couplet assert anything. It is this lack of
assertion rather than the lack of “ logical organization” of w hich
John Crowe Ransom com plains that keeps 129 from being “ a true
sonnet.” ) i
I have argued (using Sonnet 129 as an exam ple; see note 2)
that verbal m aterial is w orked over in this way to sharpen and im
prove the effect on the reader, possibly on the poet him self as his
own reader. T h e point here is sim ply that at every stage the m aterial
w ill necessarily have form or structure. Even unw orked verbal be
havior has form al properties, though possibly only as by-products.
A t that stage it is possible that a poet’s philosophy of com position
could be expressed in the words of the Duchess: “ T a k e care of the
sense and the sounds w ill take care of themselves.” B u t there are
other sources of form or structure.
One is to be found in certain prior specifications. Shakespeare
intended a sonnet, and it did not turn to an ode. B ut w hat is the
role of “ intention” ? H ow does a prior specification work? One effect
may be severely restrictive. O n ly topics of a certain size are available
to the w riter of a sonnet, and they must be developed w ith phrases
of a lim ited length; only words or sequences of words w hich fit the
iambic m eter can be used; the rim e scheme must be respected; and
so on. Shakespeare did not suffer m uch from these restrictions. From
an extraordinarily rich vocabulary suitable words were available,
although some of them puzzled his contemporaries and still puzzle
us today. H e fudged his gram m ar (Jakobson and Jones point out
that n o t to tru st and T o s h u n th e h e a v en “ seem even to transgress
the gram m atical standard of the Elizabethan time” ). H e punctuated
ambiguously, so that his sentence structure was often unclear, bu t
i8o A MISCELLANY
possibly therefore more effective. R id in g and G ra v e s3 and Empson 4

have paid particular attention to this device. H e was occasionally
illogical (why till action, when the rest of the sonnet makes it clear
that lust is perjurd, murdrous, and so on after action?).
W e have some evidence of his success in fitting form to subject
matter. H e packed his lines w ith different quantities of meaning,
and to some extent according to the position of a line in the sonnet.
W hat m ight be called the density of m eaning in each of the fourteen
lines in the first hundred sonnets was determ ined in the follow ing
way. T o escape from preconceptions about density, a rather me
chanical m ethod of scanning was adopted. Each line was first
scanned strictly as iambic pentam eter: “ Bite to the blood and burn
info the bone.” W hen the accent fell on a preposition, auxiliary,
possessive pronoun, article, copula, or such an ending as -ness, -ing,
-ance, or -ment, it was shifted forw ard or backw ard whenever pos
sible to an adjacent syllable not classified as above and not already
included in the scanning; otherwise it was om itted. In the example
the accent on to was shifted backw ard to bite, b u t the accent on
to in into could not be shifted because there was no appropriate
adjacent syllable and was therefore elim inated. A line of four
accented syllables remained. T h is stage usually yielded fou r or five
stressed syllables per line. Each line was then exam ined for syllables
not yet included which were parts of nouns, verbs, adjectives, and
adverbs, and these were added. T h e result was a scansion which
took account of practically every im portant syllable. T h e method
was arbitrary b u t yielded a plausible reading in almost every case.
T h e average numbers of stressed syllables in each of the four
teen lines in the hundred sonnets were then determ ined. T h e results
are shown in Figure 1, which should be read as if it were the right-
hand profile of a sonnet as usually printed, except that length of
line is due to “ density of m eaning” rather than letters and spaces
(the couplet not being indented). T h e average for the three qua
trains taken together is almost exactly five syllables. T h e second line
in each quatrain is, however, conspicuously short. Density rises
toward the end of the sonnet, and both lines in the couplet are
3 R id in g , L. and Graves, R . A survey of m odernist poetry. New

York, 1928.
4 Em pson, W . Seven types of ambiguity. L o n d o n , 1930.
5.08
First 4.99
qu a tra in
5.02
5.08
4.99
Second 484
quatrain
5.01
4.96
5.02
Third 486
quatrain
1 5.13
5.18
Couplet
F igure 1 Average Densities of L in es in O ne H undred Shakespeare

Sonnets
T h e right-hand mean profile, where length of line represents density

of m eaning rather than num ber of letters and spaces. T h e second
line in each quatrain tends to be short. B oth lines of the couplet
contain many m eaningful syllables.
denser than any line in the quatrains. T h e last line is particularly

so, as if Shakespeare had fallen behind in m aking his point and
was forced to pack the line tightly. T h e fact that the couplet often
makes a point of its own and has little room in w hich to do so is
probably relevant.
W hy should a poet subm it to the restrictions im posed by a
prior specification of form or structure? W hy w rite sonnets rather
than maxims, aphorisms, letters, or short essays? W h at is gained
from dancing in chains? It has been argued that early literature had
the form of poetry because it was m ore easily rem em bered by those
who recited it, and w hat is easily m em orized is likely to be quoted.
182 A MISCELLANY
A philistine m ight say that a poet has an eye on public relations. He

is concerned with putting his points across and chooses a memor
able, quotable form. A philistine m ight also say that the structure
of a poem makes what is said particularly convincing. A re we not
more likely to assent to the conclusion of a syllogism if it is ex
pressed in a meter which has been established by the premises and
if it ends in a satisfying rime? T h in gs have come out right; why
ask whether they are true or false? B u t there are, of course, better
reasons, m any of which lie above and beyond prose meanings,
although they are not exclusively matters of form or structure.
Form al properties which are not the result of a prior specifi
cation arise during the w riting o f a poem from the form al and
thematic “ influences” m entioned at the beginning of this paper.
W ord play is an example. A pun is necessarily a com bination of
structure and meaning, and even a vague double-entendre may be
related to structure. In what form can one write an opening line
that means not only, as a French tran slation 5 has it, L ’ esprit
dispersé dans un abîme de honte, b u t also, if Jakobson and Jones
are right, ejaculation or the “ spendings” of nineteenth-century por
nography?
Some evidence of the mode of action of an “ influence” is
available. Form and m eaning are both involved in alliteration and
assonance. T hese properties are usually avoided in prose (we re
w rite pour éviter les assonances) bu t are accepted in poetry. A n
excessive predilection for alliteration may have the effect of a prior
specification and impose restrictions w hich are not always success
fully evaded, but a moderate use is condoned and valued. It need
not be “ intentional.” T h e dom inant “ influence” is form al rather
than thematic. A fter em itting a response having a given sound, the
poet is somewhat more likely to em it another response havin g th a t
sound. T h e result is a structural feature which lends itself, in some
degree, to objective analysis.
Shakespeare’s sonnets contain m any alliterative lines. T o what
extent do they show an alliterative tendency? I have reported an
attem pt to answer that question.6 T h e stressed syllables in the first
5 Jouve, Jean Pierre. M ercure de France, M ay 1, 1955. '

6 Skinner, B. F. T h e a llitera tio n in Shakespeare’s sonnets: A study
in literary behavior. Psychological R ecord > 1939, 5, 186-192.
hundred sonnets were determ ined in the m anner described above,

their initial sounds were exam ined, and lines containing no instances
of a given sound, or one, two, three, or fou r instances were counted.
T h e results were com pared w ith the numbers of lines to be expected
from chance, as calculated w ith a binom ial expansion. T h e conclu
sions of that study were summarized as follows:
L in e s c o n ta in in g jo u r lik e in itia l c o n s o n a n ts.

(Ex.: £orne on the bier w ith w hite and bristly beard.)
O f these lines there are only eight more than w ould be
expected from chance, and four of these are due to the
repetition of the same w ord or words. N ot more than
once in twenty-five sonnets (350 lines) does Shakespeare
lengthen a series of three like consonants into four, ex
cept when he repeats a word.
L in e s c o n ta in in g th r e e lik e in it ia l co n s o n a n ts.
(Ex.: Save that my soul's im aginary sight.)
O f these lines there are thirty-three too many, bu t twenty-
nine of these\are due to repetition of the same word.
O n ly four are, therefore, “ pu re” alliteration. E xcept
when he repeated a w hole word, Shakespeare changed
a line of two like consonants into one of three not oftener
than once in twenty-five sonnets.
L in e s c o n ta in in g tw o lik e in itia l co n so n a n ts.
T h ere are ninety-two excess lines of this sort, bu t the
correction for repetition gives a sh o rta g e of approxi
m ately forty lines. A llo w in g for eight lines extended to
contain three or four occurrences, we may say that once
in about every three sonnets Shakespeare d is ca r d e d a
w ord because its initial consonant had already been used.
Jakobson and Jones note the presence of this kind of struc

tural feature in Sonnet 129: “ Each line displays a conspicuous
alliteration or repetition of sound sequences and entire morphemes
or words.” B u t can we be sure that roughly the same alliteration
w ould not have occurred if Shakespeare had drawn his words out
of a hat? T h e result is not a statistical artifact. A sim ilar study of
W ordsworth showed, as one m ight expect, that he discarded m any
of the alliterative words w hich must have turned up as he wrote.
In a poet like Swinburne on the other hand, alliteration is statisti
cally conspicuous.
184 A MISCELLANY
In a study of Sw inburne’s alliteration 7 the initial consonants

in the stressed syllables of 500 lines of Atalanta in Calydon were
exam ined. Instances were counted in which a sound was followed
by the same sound in the next syllable, in the next syllable but one,
in the next syllable but two, and so on. These observed frequencies
were converted into percentages of the expected frequencies calcu
lated from the total num ber of sounds. (No correction was made for
the repetition of whole words.) T h e results are shown in Figure 2.
W hen Swinburne uses a stressed initial sound, he shows a strong
tendency to use it again in the next syllable, a slightly weaker
tendency to use it in the next b u t one, and so on, the tendency
rem aining statistically significant for four syllables. T h e open circles
show insignificant differences.
Figure 2 also shows a sim ilar tabulation for Shakespeare. If
there is any alliteration in these hundred sonnets, it is confined to
successive syllables, and even there it is largely a m atter of repeated
whole words. (Some instances of repetition follow from Putten-
ham “ redoubles” or “ translacers,” in which a word or root at the
end of a line is repeated at the beginning of the next line. T h ere
are at least six of these in the first hundred sonnets.)
W ritin g under the control of prior specifications must be
called “ intentional.” O nly passages are allowed to stand w hich have
the effect of fulfilling the conditions of a contract. Nevertheless, the
first person who wrote three quatrains and added a couplet, all on a
single theme, did not “ inten d” to write an English sonnet. If he
found the result pleasing, however, he may have w ritten other
poems w ith similar structural properties, w hich at some point must
have begun to act as a set of rules: to produce a particular kind of
literacy effect write three quatrains and add a couplet, all in iam bic
pentameter. T h e structural features w hich result from form al and
them atic processes are not basically intentional (that is, they are
not introduced by the w riter because of their effects), but if the
effects are pleasing, the w riter may take steps to give these processes
greater play.
W here should we place the structural properties pointed out
7 Skinner, B. F. A q u an titative estimate of certain types o f sound-

p a ttern in g in poetry. Am erican Journal o f Psychology, 1941, 54,
64-79.
In te rv e n in g syllables
1
Figure 2 T h e Alliterative Spans of Shakespeare and Swinburne
T h e point at zero for Sw inburne should be read as follows: “In

500 lines of Atalanta in C a ly d o n the num ber of successive stressed
syllables beginning w ith the same sound is 154 percent of the num
ber expected from chance.” T h e percentage declines but remains
significant even w hen syllables are separated by three intervening
syllables. T h e figure for Shakespeare is about 124 percent, but
much of this is due to the repetition of w hole words.
by Jakobson and Jones? A re they “ negligible accidents governed by

the rule of chance,” 8 are they generated by form al and them atic
verbal processes, or are they the fulfillm ent of prior specifications?
T h e extent to w hich the features of Sonnet 129 are to be fou n d in
the other sonnets is relevant. In a purely physical sense every sonnet
has a center, and one moves toward it in reading the first h alf and
away from it in reading the second half; the first seven lines in
every sonnet are therefore centripetal and the last centrifugal. T h e
term inal couplet is necessarily “ asymmetrically contrasted” w ith the
non-terminal quatrains. B u t other features are quite idiosyncratic.
8 Jakobson, R . S u b lim in al verbal p a ttern in g in poetry. Studies in

General and Oriental Linguistics, T o k y o , 1970 (Q uoted by R ichard s
in reference 9).
A M IS C E L L A N Y
O f how many of the sonnets can it be said that “ the odd strophes in
contradistinction to the inner ones abound in substantives and ad
jectives” ? O r th a t'“ the outer strophes carry a higher syntactic rank
than the inner ones” ? O r that “ the anterior strophes show an in
ternal alteration of definite and indefinite articles” ? O r that “ the
term inal couplet opposes concrete and prim ary nouns to the ab
stract and/or deverbative nouns of the quatrains” ? O r that “ each
of the six initial lines displays a gram m atical parallelism of its two
hem istichs” ?
Idiosyncratic or not, accidental or not, the features are there,
and we should perhaps turn from the conditions w hich may have
produced them to their effect on the reader. Jakobson and Jones
insist that this “ am azing external and internal structuralism [is]
palpable to any responsive and unprejudiced reader” ; bu t Richards
certainly comes closer to the truth when he says that Sonnet 129
“ is now shown to have a degree of exactly describable structural
order which— could it have been pointed out to them in such pre
cise unchallengeable detail— w ould certainly have thrown Shake
speare him self along w ith his most intent and adm iring readers into
deeply wondering astonishment,” 9 and Jakobson has referred to
“ sublim inal structure,” as if it were out of reach of direct observa
tion, and to “ deep structure,” as if it could be reached only through
a penetrating analysis. C ertainly the reader need not be aware of
the structural features of a poem in order to enjoy it. T h e effect of
music on a listener is due to its structure, since there is nothin g else
- to have an effect, but few listeners— even those who are “ most intent
and adm iring”— know anything about the structure of music and
can see it only w ith difficulty when it is pointed out.
T h e visibility of structure is particularly im portant to the
writer, who is his own first and most im portant reader. A w riter
accepts some of the verbal responses w hich occur to him and rejects
others. H e puts those he accepts into some kind of effective order,
he adds gram m atical tags, he asserts or denies the result, and so on.
T o do this he must see w hat he has written— the simple physical
structure of his verbal behavior. Moreover, he m ay learn to write
in given ways because w hat he sees pleases him . R ichards has sug-
f ’. ' V
9 R ichards, I. A. Jakobson ’s Shakespeare: T h e su b lim in al structures
of a sonnet. T im es Literary Supplem ent, M ay 28, 1970.
gested “ traceable linkages” between Jakobson’s work and recent

genetics, and I have raised the question of a different genetic linkage
elsewhere.10 T h e effect on the reader— particularly on the w riter as
reader— is im portant because a poem evolves under a k in d of
natural selection. A ll behavior is intim ately affected by its conse
quences, and just as the conditions of selection are more im portant
in the evolution of a species than the m utations, so the selective
action of a pleasing effect is more im portant than the m eaningful
sources of the responses selected. Pleasing responses survive as a
poem evolves.
W hether or not the structure of a poem is “ sublim in al” has
a bearing on this issue. (T he linguist’s “ deep structure,” like F reud’s
“ depth psychology,” is a spatial m etaphor w hich serves several
functions. It is useful in referring to the visib ility of behavioral
processes and fiheir effects and to the role played by visibility in the
determ ination of behavior. It should not, of course, be used to sug
gest that an analysis is profound rather than superficial.) Richards
has pointed to. a Useful distinction between two kinds of know ing.
Shakespeare, to put it crudely, “ knew how” to w rite Sonnet 129: but
how m uch did he “ know about” his behavior in doin g so? H e must
have known about prior specifications and the extent to w hich what
he was w riting fulfilled them. A fourth quatrain m ight have given
him useful extra space, but he did not add one. H e kept to iam bic
pentameter. H e need not have been aware of a d ouble entendre, or
any other kind of word play, at the time it occurred— as anyone
w ho has made a Freudian slip can testify— but he m ay have “ seen” it
after the fact and allowed it to stand if it pleased him. H e need not
have been aware of associative or assim ilative influences or resulting
features such as alliteration.
H e need not have know n about the greater part of the struc
tural features pointed out by Jakobson and Jones. T h e y could have
played no part in the production of the prim ordial m aterial (the
“ m utations” ), and they are not likely to have played a part in the
elaboration or selection of features as the poem evolved.
10 Skinner, B. F. A lecture on “having a poem.” In B. F. Skinner,

Cumulative record (3rd ed.). New York: Appleton-Century-Crofts,
1972. Pp. 345-355.
Walden (One)
and Walden Two
First, my credentials. I am not a T h oreau scholar, but I claim

to be an am ateur in the original sense of a lover. It was not love at
first sight. I read excerpts from Walden in a course in Am erican
Literature at H am ilton College, bu t they were not “ relevant.” In
those days we joined fraternities and played golf, we could not have
cared less how the country was run, and as for the rest of the world
we learned about that from the N ational Geographic.
W hen I came to H arvard for graduate study, I becam e inter
ested ,.^ N ew England and its history, and I discovered W alden
Pond. I had a bicycle and would ride out to the pond to swim, not
where the bathing houses are now, but in the cove near the site of
T h o rea u ’s hut. T h e bottom was m uddy in those days, and as I
w alked abou t in shallow water, I knew w hat T h o rea u m eant by his
riparian or allu vial walks. I began to read T h oreau . I took an
interest in the site; I used to go out in the late autum n to clean up
after the picnickers.
- H aw thorne said that T h oreau made people feel gu ilty about
their possessions, and I know w hat he meant. W h en I got m y doc
tor’s degree, my fam ily gave me a car, but I felt gu ilty about it and
Walden (One) and Walden Two
bought a copy of Walden to keep in the car to take the curse off.
I made good use of it. I am almost always on time for appointm ents,
and as O scar'•W ild e once pointed out, “ Promptness is the thief of
time.” Walden is an excellent book to pick up for occasional read
ing; even if you have time for only a few sentences, they are w on
derful sentences. It does not m uch m atter what preceded or w ill
follow.
W hen I met the girl I was to marry, I took her on ou r first
date to W alden. W e had just bought a chess set in one of the
shops on Beacon H ill, and on the shores of the Pond she taught me
to play chess.
I moved on to the other works of T h o rea u when I bough t a
leather-bound eleven-volume R iverside E dition. It was not com
plete, of course, and for m any years I turned to O dell Shepard’s
T h e Heart o f'T horea u’s Journals for additional reading. I analyzed
a rather long quotation from that collection in my book Verbal
Behavior. I also bought T h o re a u ’s translation of T h e Transmigra
tion of the Seven Brahmins. A n d, oh yes, I own a T h o rea u pencil—
not made by Thoreau himself, I am sure, but by his fam ily. I
bought it at Goodspeed’s and assume it is genuine, though I can
im agine that before long someone w ill begin to m anufacture T h o
reau pencils again.
I hope this is enough to establish m y status as an am ateur. It
may not, however, quiet the em otion some of you m ay have felt at
my outrageous^ title. H ow could I have the nerve to put a O ne
after W alden, even in parentheses, and set it alongside m y own
U topian novel, Walden Two? If you foun d that disturbing, you
were in good company. W hen the book appeared in 1948, L ife
m agazine published a bitter editorial, denouncing it on just those
grounds. W alden Two was called “ an entirely presum ptuous title.”
“ In spirit Walden Two is as m uch like T h o rea u ’s original W alden
as a Quonset hut is like a com fortable and properly proportioned
Cape C od house.” Further along, my book was described as “ such a
trium ph of mortmain, or the dead hand . . . as has not been en
visioned since the days of Sparta . . . If Dr. Skinner wants to
im agine such a utopia, that is his privilege. B u t w hat should really
be held against him is the egregious liberty he has taken w ith the
title of H enry D avid T h o rea u ’s original W alden. For the truth of
the m atter is that T h o rea u ’s book is profou ndly anti-utopian; it
i go A MISCELLANY
does not belong in the long line of antiseptic literature that began
with P lato’s R epublic. Far from trying to escape into a ‘brave new
w orld,’ Thoreau , the cosmic bum, set out resolutely to make the
best of what he could find right around home. W here Samuel Butler
traveled to Nowhere for his Erewhon, where Edward Bellam y
marched ahead to the year 20000 a .d . for his L ookin g Backward,
T h oreau set up housekeeping by the edge of a duck pond outside
of his native village. As E lliot Paul has said, he ‘got away from it
all’ by m oving just a little farther from town than a good golfer
could drive a ball. T h e lum ber for T h o re a u ’s cabin was taken from
a shanty that had belonged to James Collins, an Irishm an who had
worked on the Fitchburg Railroad; the beans that T h oreau hoed
and ate were Yankee beans, grown in recalcitrant N ew England
soil.” L ife ’s com plaint was summarized in this way; “ Books like
Walden Tw o, then, are a slur upon a name, a corruption of an
impulse. A ll Thoreauists w ill properly resent them, and if Dr.
Skinner comes around w ith any of his advice the good Thoreauist
will, like Diogenes when confronted w ith the proffered largesse of
the M acedonian king, tell the author of W alden Tw o to stand from
between him and the free rays of the sun.”
A few corrections, please. I subm it that T h oreau would have
settled for a Quonset hut. H e discussed the “ necessaries” of habita
tion (we should call them the necessities), and he designed his livin g
quarters to satisfy them. T h e well-proportioned Cape C od house is
far from what he wanted. It is m uch more like the kind of house
which, T h oreau pointed out contem ptuously, cost the Concord
farmer fifteen years of his life. If James Collins had left behind a
small Quonset hut, I ’m sure T h o rea u w ould have been glad to
move it into the woods near W alden Pond.
N or is the com m unity described in Walden Tw o “ getting
away from it all.” It is one point of the book that you can have a
better life here and now. You don’t need to go to a Shangri-La be
hind high m ountains, or to a new A tlantis on some hitherto
undiscovered island, or move about in time to a distant past or
future. You can have the kind of life you w ant in the present
setting.
L ife also called T h oreau perhaps the greatest exponent of the
Yankee virtue of “ use it up and m ake it do,” and that is another
point in Walden Two. As T h oreau said, you don’t own things;
\
Walden (One) and Walden Two
things own you. In Walden Tw o every effort is made to reduce the

things needed for “ the good life.” I did n ’t realize it at the time, b u t
there is a bonus. W alden Tw o is not only m inim ally consuming, it
is m inim ally polluting.
- T h ere is no gadgetry in W alden Tw o— no computers, no
tricky technical equipm ent. It’s a simple life, rather rem iniscent o f
an English country house in the nineteenth century, bu t w ithout
the servant problem . T h ere is technology in W alden T w o, but it is
concerned w ith hum an behavior, w ith producing pleasant, effective
personal relations— in daily life, in education, and in the produc
tion of goods.
I submit that T h oreau was a utopist in a basic sense. If you
do not like the way of life that is offered you, sim ply b u ild a better
one. T h e difference is that W alden (One)— if you w ill perm it me to
call it that for Clarity’s sake— was a utopia for one. T h o rea u was no
herm it (he could w alk into C on cord— to the post office or the
lyceum— whenever he felt like it), bu t he never came to grips w ith
the problems ■yvhich arise w hen people must interact w ith each
other. Walden Tw o is an experim ent in the design of a social envir
onment.
A n d that brings me to the issue of freedom. T h e editorial in
L ife was contemptuous of “ ‘cond itioning’ for a ‘freedom ’ planned
long in advance according to the rigid specifications of a gang of
hierarchs. In the argot of 1948, in W alden One there was sim ply
freedom, period.” B ut w hat m ade it possible for T h o rea u to be
free? O nly an extraordinary set of circumstances. In the w orld in
w hich he lived he was not com pelled to do m uch of anything. H e
was free to do the things he w anted to do— to be a “ self-appointed
spectator at a snow storm ,” to anticipate nature, to begin an adven
ture in life starting w ith a vacation from toil. H e could do these
things by w alking away from C on cord and squatting on the shores
of W alden Pond. B ut how m any people can do that today?
It is easy to contrast a w orld in w hich people are controlled
by other people w ith a w orld in w hich they seem free. Freedom
from control was the dream of Jean-Jacques Rousseau, nearly a
hundred years before T horeau . B y T h o re a u ’s time the dream had
seemed to come true in a successful struggle for political and reli
gious freedom. T h oreau was opposed to political and religious
despots, to armies, and to pu n itive education. H e was opposed to
192 A MISCELLANY
punitive labor— not just slavery (to which, of course, he was actively
opposed) b u t the slavery of the worker who commits him self to a
trade or a way of life. Like M arx, who made the point at about
the same time, T h oreau was opposed to wage slavery as well as the
slavery w hich depended upon physical force. T h e person who works
for wages is avoiding, not a flogging, bu t the loss of a standard of
living. T h a t is easy to demonstrate in a factory, and M arx blam ed
wage slavery on capital, but the principle holds for the personal
entrepreneur— say, a farmer. A m an may own his farm and still be
a slave to it. H e must plant at a certain season, and if the weather
is bad w ithin a very short season. T h ere is no way out; he w ill lose
the whole thin g if he doesn’t plant. If he has cows, there are certain
times of day when they must be m ilked. His day is paced; he can
not do as he pleases, he must do things when he doesn’t feel like
doing them. As a result, T h oreau said, the farm er plows the better
part of him self into the soil as compost. A n y possession exacts its
toil. Luxuries are a hindrance to the good life. O nly leisure w ill
show what a man is really like.
For T h oreau the alternative to the punitive sanctions of daily
life seemed to be personal freedom. T h e feeling of freedom is asso
ciated w ith doing the things a person wants to do. B u t w hy does he
want to do them? T h oreau never had to ask. H e could also neglect
other requirem ents of the good life. H ow m any people today have
the ethical training w hich gave T h oreau an interest in doing things?
His fellows thought him lazy, but he knew that you “ could not kill
time w ithout dam aging eternity.” H e employed himself, but he did
it because of his education and the ethic he had received from his
culture.
He also had the benefit of the perfectionist spirit w hich was
blow ing across the land in those days. T h e founding of Am erica
was a un iqu e event in the history of the world. Here was a nation
w hich seemed to be explicitly designed in advance. Its success in
duced Am ericans to set u p smaller versions of designed ways of
life. More than two hundred intentional com m unities were founded
in the U n ited States in the nineteenth century. Perfectionist activi
ties declined at the turn of the century, but they are begin nin g to
return, and the change is reflected in the publishing history of
Walden Tw o. In the first fourteen years, the book sold only ten
Walden (One) and Walden Two *93
thousand copies; last year it sold a quarter of a m illion. Som ething

had happened in the interim . T h e w orld has come round to the
necessity of doing som ething about the ways in w hich people live,
and the initiative is being taken by youn g people. T h ey understand
what T horeau m eant when he said, “ I have yet to hear the first
syllable of valuable or even earnest advice from m y seniors,” (U nfor
tunately for T h oreau , he was over thirty when he said it.)
L ike T horeau, young people today are m uch less concerned
with the purely physical conditions under w hich they live. L ike
him they avoid aversive labor, in part by cu ttin g down on w hat they
consume. T h e y refuse to w ork hard for things w hich are not
essential— clothing, for exam ple. W h at T h o rea u called the neces
saries of clothing are conspicuous in H arvard Square today. T h o
reau pointed put that a citizen of Concord— Emerson, say— w ould
rather walk dcKvn the street w ith a broken leg than w ith a broken
pant leg. Y oung people today do not m ind w earing patches— they
even sew patches on where there are no holes, just to prove their
point. L ike T horeau , they are arguing that “ L ife is an experim ent
largely untried.” T h e ir communes are a step in the direction of new
social structures. I have just read the m anuscript of a charm ing
book describing an experim ental com m unity in V irg in ia that is
patterned after Walden Txuo. It is perhaps even closer to Walden
(One).
T horeau clearly stated what m ust becom e the dom inant prin
ciple in the im jnediate future of the world: we must cut dow n on
the consum ption of resources. It is quite im possible for our level of
affhience to prevail in all parts of the world. Im agine a billion
Chinese scooting around in a third o f a b illion cars on hundreds
of m illions of m iles of superhighways. If I m ay use a horrible
neologism, the rich nations must “ deaffluentize.” W e m ust learn
how this can be done, but T h o re a u ’s advice is still sound: the good
life is to be reached by deliberate planning.
In my contract with the publishers of W alden Tw o, the book
was called T h e Sun is a M orning Star. T h e publishers rejected that
title because another “ star” book had recently been published. T h e
phrase is from W alden, of course, and I w orked it into the book
after the title was changed. W hen the narrator makes his decision
to go back and jo in W alden T w o , he buys a copy of W alden, and
194 A MISCELLANY
as he starts his long w alk back, he reads that w onderful final para
graph: “ I do not say that John or Jonathan w ill realize all this;
but such is the character of that m orrow which mere lapse of time
can never m ake to dawn. T h e light w hich puts out our eyes is
darkness to us. O nly that day dawns to w hich we are awake. T h ere
is more day to dawn. T h e sun is b u t a morning-star.”
I
1 7
Freedom and Dignity

Revisited
i
1
. \
In a famous passage in N otes From the Underground D ostoev

ski insisted that man w ill never adm it that his behavior can be
predicted and controlled. H e w ill "create destruction and chaos to
gain his point. A n d if all this could in turn be analyzed and
prevented by predicting that it w ould occur, then m an w ould d elib
erately go mad to prove his point.” Dostoevski was him self m aking
a prediction, of course, and it had the curious effect of cutting off
this last avenue of escape, since henceforth even deliberately going
mad could be said to have been predicted.
My critics have, nevertheless, seemed bent on proving that he
was right. M any of them have shown a taste for destruction and
chaos, some of it not far short of madness. T h ey have resorted to
highly em otional terms, and a kind of hysterical blindness seems to
have prevented some of them from reading what I actually wrote.
A n author who has been so w idely m isunderstood w ill naturally
value Dostoevski’s explanation.
M y argum ent was surely simple enough. I was not discussing

a philosophical entity called freedom but rather the behavior of
195
196 A MISCELLANY
those who struggle to be free. It is part of the hum an genetic en

dowm ent that when a person acts in such a way as to reduce “ aver-
sive” (e.g., potentially dangerous) stimuli, he is more likely to do
so again. T hus, when other people attem pt to control him through
a threat of punishm ent, he learns to escape from them or attack
them in order to weaken them. W hen he succeeds, he feels free, and
the struggle ceases. B u t is he really free? T o say w ith John Stuart
M ill, that “ liberty consists in doing what one desires” is to neglect
the determiners of desires. T h ere are certain kinds of control under
w hich people feel perfectly free. T h e point has been made before,
but I was offering some further evidence recently acquired in the
experim ental analysis of operant conditioning.
Such an interpretation is not metaphysics: it is a m atter of
iden tifyin g certain processes in an im portant field of hum an be
havior. It does not— it cannot— lead to the suppression of any free
dom we have ever enjoyed. O n the contrary, it suggests that there
are ways in w hich we could all feel freer than ever before. For
exam ple, in spite of our supposed love of freedom, most of our
practices in governm ent, education, psychotherapy, and industry are
still heavily punitive. People behave in given ways to avoid the
consequences of not doing so. Perhaps this means sim ply that the
struggle for freedom has not yet been finished, but I have argued
that the continuing use of punishm ent is, on the contrary, an un
wanted by-product of that struggle. W e refuse to accept nonpunitive
practices because they make it too clear that control is being
exerted. W hen we punish bad behavior, we can give the individual
credit for behaving well, but if we arrange conditions under which
he “ desires” to behave well, the conditions must get the credit.
I neglected to point out that under punitive practices we even
justify behaving badly. Fortunately, this has now been done for me
by the film “ A Clockw ork O range.” W ritin g in T h e New York
Review , Christopher Ricks argues that aversion therapy takes the
protagonist A lex “ beyond freedom and d ign ity,” and he quotes
A n thon y Burgess (author of the novel) in defense of the film.
“ W hat my, and K u b rick’s [director of the film] parable tries to state
is that it is preferable to have a world of violence undertaken in
fu ll awareness— violence chosen as an act of w ill— than a world
conditioned to be good or harmless.” Ricks says that I am one of
the few who w ould contest that statement. I hope there are far
Freedom and Dignity Revisited
197
more than a few. T h e film misrepresents the issue because the

“ therapy” that makes A lex good is brutally conspicuous w hile the
conditioning that lies behind his “ acts o f w ill undertaken in fu ll
awareness” is easily missed.
T h e struggle for freedom has not reduced or elim inated con

trol; it has m erely corrected it. B ut w hat is good control, and who
is to exert it? Either my answers to these questions have been un
forgivably obscure or m any of my critics have not reached the last
chapters of m y book. T h e question W h o w ill control? is not to be
answered w ith a proper name or by describing a kind of person
(e.g., a benevolent dictator) or his qualifications (e.g., a behavioral
engineer). T o do so is to make the m istake of looking at the person
rather than \at the environm ent w hich determines his behavior.
T h e struggle for freedom has m oved slowly, and alas erratically,
toward a culture in w hich controlling power is less and less likely
to fall into the hands of individuals or groups who use it tyran
nically. W e have tried to construct such a culture by exerting coun
tercontrol over those who misuse power. C ountercontrol is certainly
effective, but it leads at best to a kin d of uneasy equilibrium . T h e
next step can be taken only through the exp licit design o f a culture
w hich goes beyond the im m ediate interests of controller and coun
tercontroller.
Design for what? T h ere is only one answer: the survival of

the culture and of m ankind. Survival is a difficult value (compared,
say, w ith life, liberty, or the pursuit of happiness) because it is hard
to predict the conditions a culture must meet, and we are only
beginning to understand how to produce the behavior needed to
meet them. M oreover, we are likely to reject survival as a value
because it suggests com petition w ith other cultures, as in social
Darwinism , in w hich aggressive behavior is aggrandized. B u t other
contingencies of survival are im portant, and the value of coopera
tive, supportive behavior can easily be demonstrated.
M ust individ u al freedoms be “ sacrificed” for the sake of the
culture? Most of my critics contend that I am saying so, b u t the
answer depends on how people are induced to w ork for the good
A MISCELLANY
of their culture. If they do so under a threat of punishm ent, then

freedom (from such a threat) is sacrificed, bu t if they are induced
to do so through positive reinforcement, their sense of freedom is
enhanced. You ng Chinese wear plain clothing, live in crowded
quarters, eat simple diets, observe a rather puritanical sexual code
and work long hours— all for the greater glory of China. Are they
sacrificing freedom? T h e y are if they are under aversive control, if
they behave as they do because they w ill be denounced by their
fellows when they behave otherwise. B ut if M ao succeeded in
m aking signs of progress toward a greater C hina positively rein
forcing, then it is possible that they feel freer, and happier, than
most young Americans.
M isunderstanding no doubt arises from the word “ control.”
Dostoevski used the m etaphor of a piano key: strike it and it re
sponds w ith a given tone. T h e m etaphor was appropriate to the
early reflexology of Dostoevski’s time, which P avlov’s conditioned
reflexes did little to change. But in operant conditioning a stimulus
merely alters the likelihood that a response w ill be emitted. Good
examples are to be found in verbal behavior. A verbal response is
very different from the knee-jerk elicited by a tap on the patellar
tendon. W hat a speaker says is determ ined in part by the current lis
tener, in part by the recent verbal stim uli he has heard or seen, in
part by a nonverbal setting, and in large part o f course by his his
tory as listener and speaker. T hese variables can be sorted out by
id en tifyin g well-established behavioral processes.
T h ere was an excellent exam ple of the probabilistic control

exerted by a verbal stimulus at a recent symposium at Yale U n i
versity organized to discuss Beyond Freedom and Dignity. O n the
second evening, several students brought in a large banner reading
“ Rem em ber the A ir W ar,” w hich they hung from the balcony. It
could not be seen by m any in the audience, b u t it confronted the
five panelists on the platform throughout the evening. It had a
predictable effect: Everyone of us m entioned the w ar in V ietnam
at some p o in t in his discussion and the last speaker, Sir Denis
Brogan, p u t aside his manuscript and spoke only of the war.
T h a t was good behavioral engineering. W e should learn to
live w ith it.
i 8
Freedom, at Last,
From the Burden of Taxation
)
i
I
. \
N ew Ham pshire was first to have a lottery, perhaps because

it enjoys a unique opportunity to induce visitors to support its
government. B ut other states soon found that their ow n citizens
preferred voluntary support to taxation and lotteries quickly spread.
One can only applaud the zeal and ingenuity w ith w hich they have
been managed. M adison A ven ue has done its best. Lotteries are
advertised in airports and buses, in newspapers and magazines, on
television and the radio. Newspapers cooperate in pu blicizin g the
thrills of w inning. W hen it was discovered that some people could
not w ait for a deferred drawing, instant lotteries were invented. A ll
this is admirable, and we are grateful for the resulting reduction in
taxes, but I wish to point out that an im portant resource has been
overlooked— our schools.
People are not born gamblers. T h e y become gamblers when
exposed to certain sequences of lucky hits. W hy should our schools
not be used to expose everyone to such sequences? T h e necessary
behavioral technology is at hand. A ll that is needed is a system
of lotteries extending from kindergarten through high school in
which the odds are at first highly favorable but grow steadily worse
199
2 00 A MISCELLANY
until, upon graduation, the student w ill find the standard lottery
w ith its meager odds irresistible. I propose something like the
following.
In kindergarten the tickets w ill cost a penny and prizes w ill
be of the order of a dollar, w ith a grand prize now and then of
five dollars. T h e odds w ill be extrem ely favorable; at this stage the
state w ill lose money, but o f course the amounts involved w ill be
trivial. In the first three grades tickets w ill cost a nickel, prizes w ill
be in the five-dollar range, except for a grand prize of, say, fifty
dollars, and almost all the m oney collected w ill be returned in
prizes. T h e grand prizes w ill be awarded in ceremonies in the
several schools. In the next three grades tickets w ill cost a dime, the
prizes w ill range from ten to fifteen dollars w ith a grand prize of the
order of a hundred or two hundred dollars. T h e state w ill return
approxim ately 85% of the money collected, and the grand prizes
w ill be awarded in city-wide ceremonies. In ju n ior high school
tickets w ill cost a quarter, prizes w ill be on the order of twenty-five
dollars, w ith a grand prize of perhaps five hundred. T h e state w ill
return about 60% of the m oney it collects and winners w ill be
announced on local television. Finally, in high school, tickets w ill
cost fifty cents, prizes w ill be of the order o f fifty dollars, w ith a
grand prize of a thousand, and at this point the state w ill pay back
about 50% of what it takes. T h e grand prize w ill be awarded in a
ceremony on state-wide television w ith an adm ired figure partici
pating.
Since practically all the expenses of adm inistration w ill be
borne by the schools, the entire operation w ill be m uch more
profitable than the regular lottery. T h e result w ill be a yearly crop
of high school graduates who w ill continue to buy lottery tickets
for the rest of their lives, even though the lotteries continue to pay
back no m ore than 40% or 45% of the am ount wagered.
In other words our schools w ill be used to create vast numbers
of young people who come on the m arket each year as dedicated
(should we care if psychiatrists call them pathological?) gamblers.
T h e effect of one year’s crop may not be felt, but by the end of five
years I estimate that sales taxes can be abolished and that after
twenty-five years (and we must look ahead!) there w ill be absolutely
no need for state income taxes. A fter that the states w ill be able to
help cities reduce their taxes on real estate.
Freedom, at Last 201
W hen programs of this sort have been set up in all the states,
the full potential of our schools w ill be realized. T h e entire popula
tion above the age of six w ill know the joy and excitem ent of
weekly (or daily!) drawings. A huge national lottery w ill be inevi
table and Federal incom e taxes abolished. M y guess is that the
Pentagon w ill run its own lottery and thus escape forever from the
annoyance of those appeals to Congress. I do not think I am being
unduly sanguine in looking forw ard to the day when the support
of our governm ent— in city, state and nation— w ill be entirely
voluntary.
Economists w ill point out that m oney spent for lottery tickets
w ill not be spent for goods and services and that business w ill
suffer. But the loss w ill be more than offset by the absence of taxes
and by the money won. T h e only im portant economic change w ill be
a very considerable increase in the consum ption of lu xury goods
and services. T h e rich, released from the burden of taxation, w ill
be able to spend m uch more on luxuries and so w ill the b ig winners
— only one further proof of the virtue of voluntary action in the
support of a society of free and happy people.
Acknowledgments
T h e chapters of this book were presented and published as

indicated below. Permission to republish is gratefully acknowledged.
1. Presented: Am erican Psychological Association, W ashington, D.C.,

September, 1976. Published: Psychology Today, September, 1977.
2. Presented: W algreen Conference on Education for H um an U nder
standing, University of M ichigan, A p ril, 1973. Published: Im pact,
1973> 3 (0 » 5- ! 2-
3. Presented: A symposium on “ T h e Control of B ehavior: Legal,
Scientific, and M oral Dilem m as,” R eed College, M arch, 1975. P ub
lished: Crim inal Law B u lletin , 1975, 11, 623-636, and in T h e
Hum anist, January/February 1976.
4. Presented: H um anist Society, San Francisco, May, 1972. Published:
T h e Hum anist, Ju ly/A u gu st 1972.
5. A preface to a new p rin tin g of W alden Tw o (M acm illan, N ew York,
1976).
6. Presented as a H erbert Spencer Lecture, O xford U niversity, N ovem
ber, 1973. Published: In R . Harr6 (Ed.), Problem s of scientific
revolution: Progress and obstacles to progress in the sciences, (O x
ford: Clarendon Press, 1975), and American Psychologist, 1975, 30,
42-49.
7. Presented as a Phi Beta K appa O ration at H arvard University,
June, 1976. Published: H um an N ature, February, 1978.
8. Presented: Inter-Am erican Society of Psychology, M iam i, Florida,
December, 1975. Published: Behaviorism, 1977.
202
cknowledgments 203
9. Presented at a conference at the N ew Y o rk Academ y of Sciences on

“ T h e Roots of Am erican Psychology,” A p ril, 1976. Published as
“ T h e Experim ental Analysis of O perant B ehavior” in R. W . R ieb er
and K. Salzinger (Eds.), T h e roots of Am erican psychology: H is
torical influences and implications for the future (Annals of the
New York Academy of Sciences, Vol. 291), N ew York: N ew Y o rk
Academ y of Sciences, 1977, pp. 374-385.
10. Published: In C. E. Thoresen (Ed.), Behavior modification in edu
cation (Chicago: N ational Society for the Study of Education, 1973,
pp. 446-456)-
11. Presented: New. Y o rk University, O ctober 19, 1972. Published: New
York University Education Quarterly, W in ter 1973, 4, 2-6.
12. Published: Daedalus, 1974, 103, 196-202.
13. Published in a ljnemorial volum e in honor of Jerzy Konorski: Acta
Neurobiologiae E x p e r im e n ta l, 1975, 35, 409-415, and Journal of
the Experim ental Analysis of Behavior, 1975, 24,117-120.
14. Published in a. festschrift for Sidney B ijou : B. C. Etzel, J. M.
LeBlanc, and D. M. Baer (Eds.), New developm ents in behavioral
psychology: Theory, m ethod, and application (Hillsdale, N ew
Jersey: Lawrence Erlbaum Associates, 1977, pp. 3-6) and in T h e
Hum anist, M ay/June 1977.
15. Published in a festschrift for I. A. Richards: R . Brower, H. Vendler,
and J. H ollander (Eds.), I. A. Richards: Essays in his honor (New
York: O xford U niversity Press, 1973, pp. 199-209).
16. Presented: T h oreau Society, Concord, Massachusetts, June, 1972.
Published: T h e Thoreau Society B u lletin , W in ter, 1973, 122, 1-3.
17. Published: New York Tim es, A ugust 11, 1972,p. 29.
18. Published: New York Tim es, July 26, 1977.
R obert Epstein has im proved the consistency of m any techni

cal expressions and in other ways given m uch appreciated help. I
also thank M. J. W illard for assistance in the preparation of the
manuscript.
Index
Abstraction, 98 Aversive control, 3, 7, 11, 22, 135, 143,

A b u lia , 38 152
A ccoun tability, 132 Awareness, 111
Acedia, 38
A d ap tation , 18
A d rian , E. D., 114
A dvice, 108 Bacon, Francis, 77
Affluence, 13 Bats, 164
Agencies, 9, 15 ' B eauty, 91
Aggression, 38 B ehavioral engineering, 57
A gricu lture, 84 B eh avioral sciences, Chapter 7
A id , 47 Behaviorism , Chapter 4, 72, 111
A lien ation , 12, 38, 90 B eh avior m odification, 10, 15, 40L, 45,
A llen , Laym an, 138 82
Alliteration , 182, 184 B eh avior therapy, 121
Am erican C iv il Liberties Union, 43 B ell, Sir Charles, 76
Anatom y, 69 Bellam y, E dward, 190
A nthropology, 59, 91 Bentham , Jerem y, 81
A p ath y, 38 B ernard, C lau d e, 79
A p p lie d analysis o f behavior, 41, 45 Bigness, 62
Aquinas, Saint Th o m as, 51 Books, 154
A rctic tern, 167 Boredom , 38
Aristotle, 77 Braverm an, 38
Assertion, 179 B ridgm an, P. W ., 117
Association, 97, 110 Brogan, Sir Denis, 198
Assonance, 182 B u dd h a, 66
A ttitu de, 48, 89 Burgess, A n th ony, 196
A ttneave, Fred, 104 Bussey Institute, 164
A u d iovisual devices, 130 B u tler, Sam uel, 190
205
2o6 Index
C alvin , John, 68, 82 Dostoevski, 195, 198

Carr, A., 165, 168 D ualism , 49
Cassandra, 17 Dyer, H enry, 133
Castell, A., 57
Catholicism , 53
C ausality, 172 Econom ic incentives, 24
Chained responses, 21 Economics, 59
C hina, 31, 65, 92 Econom ist, 13
Classroom m anagem ent, 135 Education, 64, 129
Clockwork Orange, 196 E ducation al contingencies, 24
C lu b of Rom e, 29 E ducational philosophy, 152
Cognitive psychology, Chapter 8, 74 E ducational policy, 147
Cohen, H . J., 43, 62 E ducation for life, 159
Coincidence, Chapter 14 Edwards, Jonathan, 102
Colem an, P. J., 165, 168 E nlightenm ent, 66, 81
C ollins, James, 190 E nvironm ent, 85, 104
Com enius, 35 Epistem ology, 124
Com m unes, 193 E rvin Com m ittee R ep ort, 45L
Com m unication, 35
Escape, 6, 25
Com m unism , 65 Eskimos, 44
Com puter, 75, 106 E thical self-m anagem ent, 150
Concept, 99
Ethics, 46, 52, 158
Co nd ition al reinforcers, 22f. E thics o f h elp in g people, Chapter 3
C onditioned reflexes, 113 E thology, 82
Confidence, 86 E volu tion of cu ltu ral practices, 24, 78
C onfucius, 66 E xam inations, 137, 150
Conscience, 53 E xistentialism , 26, 48, 54, 150
Consciousness, 90, 111 E xperim ental analysis of behavior,
Consequences, 19 Chapter 9, 82, 134
Contingencies of survival, 167 E xteroception, 50
Contingency-shaped behavior, 12 E xtinction, 116
C ontrived reinforcers, 145
C ontrol, 14, 197
C o n tro llin g agencies, 24
C orporal punishm ent, 143 Face to face control, gf.
Countercontrol, 8, 27 F eelin g of freedom , 31
C overt behavior, 100 Feelings, 71, 85, 91, 101
Creative behavior, 150 Feigl, H erbert, 57
Credit, 78 Ferster, C. B., 122
Crim e, 62, 101 Filipczak, J., 43, 62
Crozier, W . J., 114 F in al causes, 18
C u ltu re, 9, 14; 34, 53, 125, 132, 158, 197 Forbes, A lexan der, 114
Cybernetics, 74 F ourth estate, 28
Fraser, R on ald, 168
Free and h ap p y student, Chapter n
D ’A lem bert Jean le R ond, 8if. Freedom, Chapter 17, 6, 25, 3 if., 63, 79,
D arw in, Charles, 18, 76
Davis, H allow ell, 114 i 43>191
Freedom to have a fu tu re, Chapter 2
D elinquency, 62 Free school, 26, 146
Dem ocracy, Chapter 1 Freud, Sigm und, 51, 59, 90, 187
D enm ark, 44 Froebel, F riedrich W ilh e im August,
Density o f m eaning, 180 140
Descartes, R ené, 51 F u ller, P au l, 121
D evelopm ent, 99 F uture, Chapter 2, 144L, 150
Dew ey, John, 144
Dews, Peter, 123
D iderot, Denis, 8if.
D ignity, Chapter 17, 79 G am bling, 199
D iogenes, 190 G en etic endow m ent, 32
D irections, 108 Goodm an, P au l, 130, 140
D iscovery, 13 1,15 0 Greeks, 104, 109
Doom sday prophecies, 17 G reenland, 44
Index 207
H am ilton College, 188 K eller, F. S., 136, 138, 146, 156

Happiness, 32, 61, 93 Koestler, A rth u r, 171
H arvard University, 188 K now ing, 48, 73, 104
H aw thorne, N ath an iel, 188 Know ledge, 105
H ealth, 64 Konorski, Jerzy, 119
Hedonism , 19 Kozol, Jonathan, 130, 140
H eilbroner, R ., 38
H elpin g, Chapter 3
H eron, W . J., 123 LaF arge Center, 53
H ip p y culture, 26 Lan ger, Suzanne, 101
H obbes, Th om as, 77 L a R och efou cauld , François, D u c de,
H olland, J. G., 121 70
H olt, John, 130, 140 L a w of Effect, 19, 35, 115
H om eric Greek, 74 Laws, 108
H om ing, 164 L earn in g, 130
H ubler, Edward, 177 ' Leisure, 63
H um an behavior and dem ocracy, Len in , 65
Chapter 1 Lew is, A rth u r, 89
H um anism , Chapter 4 License, 80
H um anistic psychology, 26, 54 L ife m agazine, 189
L ilien th a l, D . E., 87
L im its to Grow th, 29
Idea, 99 \ Lincoln, A., 3
Idols of the Schbol, 151 Lindsley, O gden R ., 121
Illich , Ivan, 131, 140 Linguistic com petence, n o
Images, 107 Linguists, 122
Im prin ting, 124 » Lo cke, J., 7, 51, 77
In d ividu al, 54, 80 \ Loeb , Jacques, 113
In d ivid u al pacing, 155 London Tim es, 88
Indochina, 47 Lotteries, Chapter 18
Industry, 38 L ucero, R . J., 42
Inform ation theory, 74 Luck, 175
Inside inform ation, 73
Institu tion al care, 40
In tellectu a l self-m anagem ent, 150, 173 M ach, Ernst, 117
Intelligence, 133 M achiavelli, N iccolò, 14
Intention, 48, 102, 179, 184 M acm illan Com pany, 57
Intern ation al Peace Research Associa M agnus, R u d o lp h , 114, 119
tion, 91 M althus, T . R ., 56
Interoception, 50, M an, 73
Introspection, 51, 72, 111 M ao Tse-tu n g, 32, 198
Iowa, 42 M arx, K arl, 12, 34, 39, 66, 81, 192
Im itation, 22 M ath em atical m odels, 74
Im m ediate gratification, 32 M axim s, 108
Incentives, 27 M cN eill, W ., 95
Interm ittent reinforcem ent, 21 M eaning, Chapter 1 5
Intervention, 15 M easurem ent, 149
I r a n ,87 M edicine, 77
M em ory, 76
M en tal apparatus, 110
Jakobson, Rom an, 176, 179, i82f., i85f. M en tal hospitals, 41
James, W illia m , 51 M ichotte, A ., 173
Jefferson, Th om as, 7 M idd le Ages, 100
Jensen, B., 44 M igration of birds, 166
Jesus, 66 M ill, J. S., 77, 196
Jones, 176, 179, 182L, 185L M iller, S., 119
Jouv£, Jean Pierre, 182 M ind, 100, 107
Julesz, Bela, 105 M innesota, 123
M olière, 110
M oney, 23, 46
K ant, Im m anuel, 51 M ontessori, 140, 143
2o8 Index
M orals, 52, 158 Psycholinguists, 122

M oynihan, D. P., 88 Psychology, 59
M yth, 174 Psychopharm acology, 123
Purpose, 19, 102
P u rsu it of happiness, 7
N atio n al Research Council, 58
N atio n al T ra in in g School for Boys, 43
N atu ra l Selection, i8f., 76 Ransom e, John Crow e, 179
N eill, A . S., 140 R apap ort, A n atole, 103
N ervous system, 123 R a te of eating, 1x5
N ew H am pshire, 199 R a te o f responding, 116
N ew York Tim es, 13 R e a l world, 131, 144, 152
N on con tin gen t reinforcers, i2f., 36 R eh ab ilitation , 43
N ozick, R obert, 47 Respondent conditioning, 21
Reston, James, 13
R evival of L earn in g, 66
O bjectives of teaching, 134, 146 Rew ards, 35
Onians, R . B., 73 R ichards, I. A ., i86f.
O perant behavior, 103, 120 R icks, Christopher, 196
O perant conditioning, 19, 42, 172 R ights, 4of.
O rder, 179 Robertson, J. M., 177
O rr, R . T ., 167 Rogers, C arl, 36, 125, 130, 144
Osterw ald, H ., 166 Rom ans, 77
Rousseau, Jean-Jacques, 131, 140, 191
R o y a l Society o f Arts, 86
Parts o f speech, 179 R u les, 12, 108
Pasteur, Louis, 79 Russell, Bertrand, 113, 117
Paul, E lliot, 190 Russia, 65, 83, 89
Pavlov, I. P., 74, 97, 1 13ft., 119, 198
Pavlovian conditioning, 10, 40
Pedagogy, 130 St. Paul, 36
Perfectionism , 192 Salmon, 167
Permissiveness, 6 Scherber, J., 42
Personalized system of instruction, 136 Schmidt, J., 168
Pestalozzi, 140 Scholastics, 77
Phenom enology, 26, 48, 54 Schum acher, E. F., 60
Philosophy, 113 Science of behavior, C hapter 6
Phylogenic behavior, Chapter 13 Scientific Am erican, 27
Physiology, 49, 69, 8 i, 111, 125 Scientific creativity, 110
P iaget, Jean, 109 Scientists, 29
Plato, 51, 73, 77, 190 Selection, 163
Poincar6, H enri, 117 Self-actualization, 54
P o litical action, 66 Self-control, 52
P o litical science, 59 Self-knowledge, 50, 52, 111
P ollu tio n, 61 Sense of self, 52
P opper, K arl, 98 Sense of tim e, 100
Positive reinforcem ent, 4f., n , 40, 145 Service, Elm an, 65
Possession o f facts, 105 Sexual identity, 99
Practice, 108 Shaping behavior, Chapter 13, 120, 163,
Press, 28 168
Prisoners, 43 Sherrington, C . S., 74, 114, 116; n 8 f.
P robab ility, 36 Silberm an, Charles, 130, 140
Program m ed contingencies, 13, 41 Social D arw inism 30
Program m ed instruction, 121, 135, 146, Social environm ent, 8, 83
156 Socialism, 65
Progress, 39, 53, 133 Social psychology, 91
Propositions, 104 Sociology, 59, 91
Proprioception , 50 Sonnet, 180
P rosthetic environm ent, 41 Spontaneous generation, 79
Protestant work-ethic, 61 State of m ind, 48
Proverbs, 108 Statistics, 174
PSI, 136 Stim ulus control, 169
Index 209
Stomach memory, 20 U N E SCO , 91

Storage, 106, 125 U nited States, 31, 47, 83
Stream of consciousness, 111 U tilitarianism , 35, 46, 154
Strength of behavior, 38
Structuralism , 'Chapter 15, 14, 20, 48, V ail, D . J., 42
54 V alues, 52, 92
Sum m erhill, 130, 141 V erb al behavior, 109, 139
Superstition, 20, 153, 172
Survival, 50, 53, 126
Susceptibility to reinforcem ent, 21 W addin gton, C. H ., 101
Sw inburne, A lgernon Charles, i83Íf. W alden, Chapter 16, 64
Systems analysis, 74 W alden Pond, 188
W alden Tw o, C hapter 5, C hapter 16,
29
Teachers, 28 W ar, 91
Television, 130 W arnings, 108
Tennessee V alley A u th ority, 87 W arren, R o b ert Penn, 57
T extb ook, 154 W atson, J. B., 113
T h e o ry of knowledge, 124 W egener, A., 168
T h erapy, 197 W elfare, 13, 64
T h ird w orld farmers, 84 W ill, 100, 102
T h oreau , H enry D avid, 64, i88ff. W illia m James Lectures, 122
T h orn d ik e, Edw ard L., 19, 35, n s f ., W olfson, A., i66ff.
119,130 1 W oodw orth, R . L., 116, 118
T h o u gh t, 100 W ord association, 98
T in b ergen , N., 101 W ordsw orth, W illiam , 183
Tokens, 4 if., 14s W ork, 39
T o lm a n , E. C., 117L W u n d t, W ilh elm M ax , 51
T o p ograp h y of •behavior, 169
T u rtles, 164
T y le r, A . F., 57 Zaharoif, Sir Basil, 47

Reflections On Behaviorism and Society PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Reflections On Behaviorism and Society PDF

Hochgeladen von

Copyright:

Verfügbare Formate

:REFLECTIONS

P r e n t ic e -H a ll , I n c ., Engletuood CliffSj N.J. 07632

Contingencies of Reinforcem ent: A T heoretical Analysis

© 1978 by Prentice-Hall, Inc., Englewood Cliffs, N.J. 07632

A ll rights reserved. No part of this book may be reproduced in any form or

Printed in the United States of America

THE SCIENCE OF BEHAVIOR

T h is is not a book to be read straight through. Most of the papers

1. B ehavior m odification is just the technology we need

environm ent, behaviorism makes it possible to achieve

tional practices can restore that essential function. H igher

18. “ State lotteries can become an ideal way in w hich

T h ere is more in each chapter than I have been able to pu t in

1 Human Behavior and Democracy

. . . T h a t this nation, under God, shall have a new

Abraham Lincoln? N o t exactly. L in coln said “governm ent of

Let us look at two problem s faced by all forms of governm ent

or avoid various kinds o f physical damage have had an obvious

T h ere has been a com parable miscarriage o f the process of

A social environm ent in w hich people thus behave as they

“ utopian ” in the pejorative sense of impossible. Nevertheless, we

It is probably im possible to keep these fields apart, and in its

better illustrate the failure to find appropriate means than the

M any people have had frightening visions of behavior m odifi­

being sequestered or m onopolized by individuals or organizations

T h e veiy substitution of positive reinforcem ent for aversive

produced. T h e separation of workers from the natural products of

do. N oncontingent reinforcers are characteristic of both affluence

These, then, are some of the principles to be observed in

T h e trouble is that any allusion to the control of hum an be­

solve their problem s and survive. It is an evolving culture, then,

T h ere w ill no doubt continue to be governm ental and eco­

It is often pointed out that I have specialized in the behavior

Doomsday prophecies are now a com m onplace of daily life. W e

W h at does it mean to say that a person “ takes the future into

with respect to the features w hich have played a part in selection.

pleasure and pain or satisfaction and annoyance w hich resulted

through a different process called respondent conditioning. T h e

shall see, but once established they usually continue to be supported

due prim arily to contrived governm ental contingencies, but it has

tion. T h e fact is that cultural practices have evolved in w hich con­

Few theologians of my acquaintance m ention hellfire, and most of

struction was once quite aversive. T h e Egyptians, the Greeks, and

urged to make penalties more severe, restoring the death penalty

W e cannot continue to leave the future to the occasionally

manage themselves effectively.” He has done this by constructing a

W e sometimes act for the good of others. W e feed the hungry,

who are too weak to reciprocate, or to protest effectively when we

T o begin w ith a very simple example, we may not really help

M ore serious un an ticipated effects of the good we do to others

vironm ents are constructed in w hich reinforcing consequences are

T h ose who are in a position to help others by givin g them

U nanticipated consequences w hich follow when we are said to J

which such things as food, shelter, and safety are guaranteed as

Organisms are at least as strongly disposed to take goods away

resentment that accompanies taking goods away from those who

A good exam ple of the neglect of relevant aspects of the en­

1 Heilbroner, R. Review of Braverman, Labor and Monopoly Capi­

A sim ilar correction needs to be m ade to offset the unwanted

behavior in helping others is determ ined to such a large extent by

T h e issue first arose w hen behavior m odification was used in

Troublem aking m ay be unintentionally reinforced, and if possible

“D eprivation is never to be used. N o patient is to be de­

M any people have had frightening visions of behavior m odifi

T h e trouble is that any allusion to the control of hum an be

T h ere w ill no doubt continue to be governm ental and eco

tion. T h e fact is that cultural practices have evolved in w hich con

A good exam ple of the neglect of relevant aspects of the en

1 Heilbroner, R. Review of Braverman, Labor and Monopoly Capi

“D eprivation is never to be used. N o patient is to be de

3 Behavioral Voice § 5 (C en ter for H um an D evelopm en t, D rake U n i

ever been achieved simply by supplying goods and services. G overn

6 “ In d ivid u al R ig h ts and the F ederal R o le in B eh av io r M od ifica

A critic contends that a recent book of m in e 1 does not con

havior? W hy have methods that have been so prodigiously success

I must confess to a predilection here for my own specialty, the ex

trists we may even try to establish an independent science of feel

to be of interest. T h e y are to be justified by their internal consis

have ever achieved before in art, music, literature, science, tech

m ained a m edieval society.1 C ertain rather inoffensive cultural prac