Some Fundamentals of B.

E Skinner's
Behaviorism
Dennis J. Delprato Eastern Michigan University
Bryan D. Midgley University of Kansas
Despite B. F. Skinner's prominence, his impressive written
corpus, and the many authoritative presentations by others
of his approach to psychology, the fundamentals of Skinner's
psychology have never been addressed in any comprehensive
manner. In this article, the authors take steps
to fill this gap by synopsizing Skinner's written corpus
into 12 fundamental points that seem to characterize his
behaviorism.
Behaviorism's impact on disciplines inside and outside
of psychology is exemplified by the prominence of its
leading advocate for much of this century, B. F. Skinner
(Gilgen, 1982; Heyduk & Fenigstein, 1984). Skinner's
version of behaviorism continues to exert a significant
influence on psychology and the culture at large. Reviewers
who have conducted quantitative (Wyatt, Hawkins,
& Davis, 1986) and qualitative assessments (Leahey, 1987)
agree that Skinner's psychology is alive and well. A random
sample of members of the American Psychological
Association ranked Skinner first in a survey of the most
important people in American psychology during the
post-World War II period (Gilgen, 1982). In another survey
of the most important events and influences in post-
World War II American psychology, a sample from the
same source ranked Skinner's contributions first, behavior
modification (largely associated with Skinner) second, and
the growth of behavioral psychology fourth (Gilgen, 1982).
Thus, it is an understatement to conclude that Skinner
has been, and is, influential and well-known.
Despite Skinner's influence, his impressive written
corpus, and the many authoritative and comprehensive
presentations by others of his approach to psychology
(e.g., Catania, 1980; Michael, 1985; Reese, 1986), no one
has forthrightly addressed the fundamental features, including
assumptions, of Skinner's approach to psychology
(but see Nye, 1979; Skinner, 1974; Verplanck, 1954, for
some preliminary attempts). Given Skinner's influence
and scholarship, this strikes us as an oversight. In an attempt
to fill this gap, we present what we consider to be
12 fundamental points of Skinner's behaviorism.
In this presentation, we adhere to a format that includes
a concise statement for each assumption, followed
by at least two quotations from which it was derived. This
is followed by a discussion of each. The quotations used
are those that, after a careful analysis of Skinner's published
works, seemed to best represent his position on
particular issues. Although we have attempted to minimize
interpretations and translations of these quotations,
we cannot be certain that a sufficient amount of context
was taken into account when drawing conclusions from
the written data. It is possible that variations in context
would have led to modifications of at least some of the
features on the basis that the quotations we used are not
representative samples. Arguing for their veracity is the
relatively high degree of internal consistency in Skinner's
overall system as we present it.
The features we identified are organized in a quasilogical
order such that those presented later build on those
presented earlier. This organizational scheme reflects our
own way of synthesizing Skinner's psychology into a coherent
whole. We do not critically assess the features either
singly or in toto. Our goal has been to synopsize the psychology
of the most eminent psychologist of the latter
part of the 20th century. The points we address pertain
to the purpose of science, methodology, determinism, locus
of behavioral control, consequential causality, materialism,
behavior as subject matter, reductionism, nonreductionism,
organism as the locus of biological change,
classification of behavior into respondent and operant,
stimulus control of operant behavior, and the generality
of behavioral principles.
Purpose of Science: The Primary Purpose of
Science is Prediction and Control
We undertake to predict and control the behavior of the individual
organism. (Skinner, 1953, p. 35)
The object [of my research] has been to discover the functional
relations which prevail between measurable aspects of behavior
and various conditions and events in the life of the organism.
The success of such a venture is gauged by the extent to which
behavior can, as a result of the relationships discovered, actually
be predicted and controlled. (Skinner, 1972, pp. 257-258)
If we have achieved a true scientific understanding of man, we
should be able to prove this in the actual prediction and control
of his behavior. (Skinner, 1972, p. 259)
The laboratory techniques . . . and their technological applications,
emphasize the prediction and control of behavior via
the manipulation of variables. Validation is found primarily in
We gratefully acknowledge H. S. Pennypacker and E. F. Malagodi for
their consummate assistance with the development of this article, W. A.
Balliet for his contributions to an earlier version of this article and
W. S. Verplanck and I. S. Schwartz for their comments on an earlier
version.
Correspondence concerning this article should be addressed to
Dennis J. Delprato, Department of Psychology, Eastern Michigan University,
Ypsilanti, MI 48197.
November 1992 • American Psychologist
Copyright 1992 by the American Psychological Association, Inc. OOO3-O66X/92/$2.OO
Vol. 47, No. 11, 1507-1520
1507
the success with which the subject matter can be controlled.
(Skinner, 1972, p. 41)
Skinner offered prediction and control as the primary
goals of science instead of hypothesis or theory testing.
He opposed deductive methods, which purported to postulate
a theory a priori and then test it against empirical
evidence. Skinner obtained empirical data first and then,
by induction, derived general principles or functional relations
between events. To ensure that the relations thus
described actually pertain to the events investigated, he
suggested that the scientist use them to make predictions
and to control subsequent events. Once the events are
successfully predicted and controlled, the relations discovered
are confirmed.
Skinner's emphasis on prediction and control over
theory and hypothesis testing directly relates to a much
misrepresented aspect of his systematic position. Although
he replaced theory testing with prediction and
control, he only abjured conventional psychological
theorizing. In his article "Are Theories of Learning Necessary?"
Skinner (1950) described the class of theory he
rejected as "any explanation of an observed fact which
appeals to events taking place somewhere else, at some
other level of observation, described in different terms,
and measured, if at all, in different dimensions" (p. 193).
However, Skinner vigorously stressed the importance of
a theory of behavior:
Behavior can only be satisfactorily understood by going beyond
the facts themselves. What is needed is a theory of behavior,
but the term "theory" is in such bad repute that I hasten to
explain [that not needed are theories of the conventional type].
(1947, pp. 27-28)
Whether particular experimental psychologists like it or not,
experimental psychology is properly and inevitably committed
to the construction of a theory of behavior. A theory is essential
to the scientific understanding of behavior as a subject matter.
(1947, pp. 28-29)
Beyond the collection of uniform relationships lies the need for
a formal representation of the data reduced to a minimum
number of terms. A theoretical construction may yield greater
generality than any assemblage of facts. (1950, pp. 215-216)
Skinner (1947) outlined three basic steps for constructing
a theory: (a) Decide on the basic data (the events we seek
to understand), (b) collect data (functional relations,
facts), and (c) inductively develop explanatory (theoretical)
concepts. The addition of facts permits the emergence
of collections of concepts (i.e., theory). The following
statement shows Skinner's (1947) affirmation of theorizing
and the central role of prediction and control in this
process:
We need to arrive at a theory of human behavior which is not
only plausible, not only sufficiently convincing to be "sold" to
the public at large, but a theory which has proved its worth in
scientific productivity. It must enable us, not only to talk about
the problems of the world, but to do something about them, to
achieve the sort of control which it is the business of a science
of behavior to investigate. The superiority of such a theory will
then be clear and we shall not need to worry about its acceptance
(P. 46)
Skinner's view that the essence of scientific behavior
is prediction and control comported with his position on
the epistemological question of the nature of scientific
knowledge. In the midst of discussions of operationism
in psychology, Skinner (1945a) argued against intersubjective
agreement as the major criterion for the acceptance
of scientific knowledge. He suggested that "whole-hearted
agreement on the definition of psychological terms . . .
makes for contentment but not for progress" (1945b, p.
293). As he put it:
The ultimate criterion for the goodness of a concept is not
whether two people are brought into agreement but whether
the scientist who uses the concept can operate successfully upon
his material—all by himself if need be. What matters to Robinson
Crusoe is not whether he is agreeing with himself but
whether he is getting anywhere with his control over nature.
(1945b, p. 293)
Thus, Crusoe's concern is prediction and control. Skinner
elaborated on this pragmatic theory of truth (cf. Zuriff,
1980) numerous times; for example:
Knowledge enables the individual to react successfully to the
world about him just because it is the very behavior with which
he does so. (1953, p. 409)
[Scientific knowledge] is a corpus of rules for effective action,
and there is a special sense in which it could be "true" if it
yields the most effective action possible. (1974, p. 235)
A proposition is "true" to the extent that with its help the listener
responds effectively to the situation it describes. (1974, p. 235)
Skinner's pragmatic epistemology carried over to
how he approached relations between science and technology
and to his concern for the culture at large. He
took the position that a technology most needs principles
for effective action and, in a like manner, that the very
survival of a culture depends on successful control over
conditions that threaten it (Skinner, 1971, 1978, 1987a,
1989). Skinner frequently argued that science based on
prediction and control was preferable to one founded on
theory testing when we seek to address applied (extralaboratory)
problems. Furthermore, Skinner's position on
the purpose of science and his pragmatic epistemology
directly related to his practice of using individual-organism
experimental design tactics. He suggested that "no
one goes to the circus to see the average dog jump through
a hoop significantly oftener than untrained dogs raised
under the same circumstances" (Skinner, 1956, p. 228).
Methodology: The Methodology Is Functional
Analysis, Which Relates Environmental
Independent Variables to Behavioral
Dependent Variables
[Experimentation means] we manipulate certain "independent
variables" and observe the effect upon a "dependent variable."
In psychology the dependent variable, to which we look for an
1508 November 1992 • American Psychologist
effect, is behavior. We acquire control over it through the independent
variables. The latter, the variables which we manipulate,
are found in the environment. (Skinner, 1947, p. 20)
We undertake to predict and control the behavior of the individual
organism. This is our "dependent variable"—the effect
for which we are to find the cause. Our "independent variables"—
the causes of behavior—are the external conditions of
which behavior is a function. Relations between the two . . .
are the laws of a science. (Skinner, 1953, p. 35)
Skinner's interest was psychology as an experimental science.
Experimentation allows the researcher to identify
reliable relations between one class of variable—the manipulated
environmental class—and behavioral ones.
Skinner labeled as functional relations those relations that
occur when a change in an independent variable results
in a change in a dependent variable. The process of experimentation
leading to the identification of functional
relations was labeled functional analysis. Functional
analysis yields functional relations that are the basic facts
of the science of behavior.
The key point of Skinner's methodology is the connection
among experimental functional analysis, functional
relations, and what he meant by controlling variables
(used synonymously with conditions under which
behavior occurs and with conditions, or variables, of
which behavior is a function). The independent variables
of functional relations are the controlling variables that
permit the scientist to predict and control behavior. Skinner
(1953, pp. 32-33) illustrated this process using the
example of predicting and controlling drinking a glass of
water. We can control drinking by manipulating variables
such as deprivation history, room temperature, exercise,
and amount of salt or urea in food ingested before the
experiment. To predict whether or not the subject will
drink, we must have information on each of these controlling
variables and on "extraneous" ones as well. Accordingly,
appeal to hypothetical states or conditions (e.g.,
motivation, drive, thirst, feelings) thought to be induced
by independent variables as causal variables is to propose
explanatory fictions that forestall scientific understanding
because they "allay curiosity and . . . bring inquiry to
an end" (Skinner, 1957, p. 6). That is, the search for controlling
variables that lie outside the organism is truncated
when, for example, we say the person perspired and stuttered
because of anxiety instead of searching for environmental
variables that control the excessive perspiration
and disfluent speech.
Classic applications of experimental methodology
are based on the assumption that it provides identification
of cause-and-effect relations; however, Skinner (1953) departed
from a strict adherence to this aspect of experimentation:
A "cause" becomes a "change in an independent variable" and
an "effect" a "change in a dependent variable." The old "causeand-
effect connection" becomes a "functional relation." The
new terms do not suggest how a cause causes its effect; they
merely assert that different events tend to occur together in a
certain order. This is important, but it is not crucial. There is
no particular danger in using "cause" and "effect" in an informal
discussion if we are always ready to substitute their more exact
counterparts, (p. 23)
A somewhat subtle aspect of Skinner's methodology
is that the independent variables of most interest are selective
contingencies (see Consequential Causality Section)
to which the organism was exposed before the occurrence
of the instance of the behavior to be explained.
This view of independent variables as temporally remote
in the past of the organism departs from conventional
applications of the experimental model in which "causal"
variables are required to be immediately antecedent to
effects, sometimes requiring hypothesized mental causes.
An experimental methodology with (causal) independent
variables whose effects are detected after a period of time
contrasts with the conventional behavioristic view of experimentation
in which independent and dependent variables
refer to temporally contiguous stimuli and responses,
respectively.
Determinism: Behavior Is Determined;
It Is Lawful
[Science] is more than the mere description of events as they
occur. It is an attempt to discover order, to show that certain
events stand in lawful relations to other events. . . . If we are
to use the methods of science in the field of human affairs, we
must assume that behavior is lawful and determined. (Skinner,
1953, p. 6)
To have a science of psychology at all, we must adopt the fundamental
postulate that human behavior is a lawful datum, that
it is undisturbed by the capricious acts of any free agent—in
other words, that it is completely determined. (Skinner, 1947,
P. 23)
I was working on a basic Assumption—that there was order in
behavior if I could only discover it—but such an assumption is
not to be confused with the hypotheses of deductive theory.
(Skinner, 1956, p. 227)
Man is not made into a machine by analyzing his behavior in
mechanical terms. Early theories of behavior . . . represented
man as a push-pull automaton, close to the nineteenth-century
notion of a machine, but progress has been made. Man is a
machine in the sense that he is a complex system behaving in
lawful ways, but the complexity is extraordinary. (Skinner, 1971,
p. 202)
[In reference to inconspicuous determining events and conditions
that are "easily overlooked"]: It is then easy to believe that
the will is free and that the person is free to choose. The issue
is determinism. The spontaneous generation of behavior has
reached the same stage as the spontaneous generation of maggots
and micro-organisms in Pasteur's day. (Skinner, 1974, pp. 53-54)
Skinner followed the commonly accepted position that
the scientific method begins with a deterministic assumption
rather than an indeterministic one. The scientist
assumes lawfulness, hence determinism, and proceeds to
look for lawful relations. Skinner was not different from
other early pioneers, such as Freud, who attempted to
bring human behavior into the realm of science by adopt-
November 1992 • American Psychologist 1509
ing the working assumption that it is orderly and that
regularities are able to be discovered by appropriate
methods.
In Skinner's approach, this determinism assumption
is fundamental for (a) making human behavior amenable
to scientific understanding and (b) what Skinner viewed
as the primary goals of science: prediction and control.
This assumption, however, does not imply any sort of
mechanistic determinism in which stimuli and responses
are contiguous and the former impel the latter (see Consequential
Causality). Indeed, in Skinner's (1935) early
work, he described a behavioral relation as the correlation
between a stimulus class and a response class, or what
might be described today as a definitely molar perspective.
To hold that behavior is determined (i.e., is not capricious)
is to hold that it is controlled whether we recognize
the lawfulness and sources of control or not. A
major point of Beyond Freedom and Dignity (Skinner,
1971) is that the most dangerous forms of control are
inconspicuous, thus permitting the controlled individual
to feel free. Skinner argued that individuals are better off
to shed the idea that they are beyond the bounds of controlling
factors. He advised individuals to identify how
they are controlled and thereby exercise maximum control
over their lives rather than leaving their fate in the
hands of others who may not have the individual's best
interests as a high priority. This point seems to have been
neglected uniformly by critics of Skinner's position on
freedom and control.
Locus of Behavioral Control: The Causes of
Behavior Are Localized in the Environment
Initiating causes . . . lie in the environment and . . . remain
there. (Skinner, 1988e, p. 73)
The experimental analysis of behavior goes directly to the antecedent
causes in the environment. (Skinner, 1974, p. 30)
The environment made its first great contribution during the
evolution of the species, but it exerts a different kind of effect
during the lifetime of the individual, and the combination of
the two effects is the behavior we observe at any given time.
(Skinner, 1974, p. 17)
What we have learned from the experimental analysis of behavior
suggests that the environment performs the functions previously
assigned to feelings and introspectively observed inner states of
the organism. (Skinner, 1974, p. 248)
The major difference between Skinner and most psychologists
who object to his view of behavior as determined,
lawful, and controlled revolves around the locus
of determining or controlling variables. Skinner objected
to the idea that the critical variables for behavior are inside
the behaving organism. His view was that there are two
possible sources of behavioral control. The autonomous
individual approach leads us to search inside the organism
for mental structures and processes. The other choice is
to examine the environment of the organism. Skinner
(1947) found the former not conducive to a scientific approach
in contrast to work inspired by the environmental
perspective.
Skinner's insistence on nonmental, environmental
determinism was consistent with several other aspects of
his system. In terms of prediction and control, he maintained
that attempts to predict and control behavior by
means of organismic-centered causes had failed and that
only environmental variables enable the psychological
scientist to meet this fundamental goal. To locate controlling
variables in the environment is to locate formally
the controlling variables of functional relations.
The reliance on environment enables the psychologist
to steer clear of nonbehavioral (i.e., mentalistic) explanatory
attempts. Related to this is the definition of
behavior that includes "commerce with the outside
world" (Skinner, 1938, p. 6). In this way, the outside world,
or environment, is always an inherent component of behavior,
specification of which enables us to determine
what the organism is doing. Behavior cannot be separated
from the environmental context in which it occurs. Only
when one has identified the critical environmental factors
determining a bit of behavior has the behavior been defined.
Skinner objected to claims that private events can
be used to explain overt behavior: "Although speculation
about what goes on within the organism seems to show
a concern for completing a causal chain, in practice it
tends to have the opposite effect. Chains are left incomplete"
(Skinner, 1972, p. 268). The chains are incomplete
because the occurrence of the internal event has not been
explained. One must ultimately complete the chain by
going to the initiating causes in the environment. A behavioral
analysis is said to trace the causal chain of behavior
back no further than to:
The point at which effective action [prediction and control] can
be taken. That point is not to be found in the psyche, and the
explanatory force of mental life has steadily declined as the
promise of the environment has come to be more clearly understood.
(Skinner, 1974, p. 210)
In placing the causes of behavior in the environment,
Skinner maintained consistency with the assumption of
materialism (see later discussion) in that reliance is placed
on the physical world as opposed to a nonphysical world
such as one of mental structures and processes. In placing
controlling factors in the environment, the source of the
lawfulness in behavior is made consistent with the total
system.
Finally, Skinner's attempt to place the locus of behavioral
control in environmental events was not a denial
that independent variables may be isolated inside the organism,
as when injection of a pharmacological agent
places a controlling variable within the organism. He denied
only that invented internal structures, states, and
processes explain behavioral variability. Nor was his view
of the locus of behavioral control correctly taken as an
"environmentalism" that rules out genetic factors. For
Skinner, the opposite to environmental was not heredity
but mentality and the assumption of autonomous hu-
1510 November 1992 • American Psychologist
mans. Skinner's approach to causality, taken up next,
reflects how Skinner departed from earlier behavioristic
versions of environmentalism.
Consequential Causality: Selection by
Consequences Is the Primary Causal Mode
by Which Environment Determines
Outcomes in Living Systems
In certain respects operant reinforcement resembles the natural
selection of evolutionary theory. Just as genetic characteristics
which arise as mutations are selected or discarded by their consequences,
so novel forms of behavior are selected or discarded
through reinforcement. (Skinner, 1953, p. 430)
Selection by consequences is a causal mode found only in living
things, or in machines made by living things. (Skinner, 1981,
p. 501)
As a causal mode, selection by consequences was discovered
very late in the history of science.. . . The facts for which it is
responsible have been forced into the causal pattern of classical
mechanics. (Skinner, 1981, p. 502)
Science has largely subscribed to explanation involving
one or another version of mechanical causality. Skinner
noted that this influential causality of classical mechanics
requires initiating agents such as the stimuli of stimulusresponse
models. In the case of early behavioristic stimulus-
response models, the initiating agent was in the individual's
external environment. Not fundamentally different
from the stimulus-response models of behaviorism
was the Freudian and other internalistic applications
of mechanics that placed initiating agents inside the
organism.
Skinner departed from mechanical causality, derived
in part from study of nonliving things, to a type of causality
that was discovered in the study of living systems.
He argued that Darwinian natural (environmental) selection
represents the type of causality most applicable
to biology, psychology, and other life sciences. In contrast
to mechanical causality, which requires initiating causes,
Skinner argued for consequential causality in psychology.
According to the principle (Skinner would say "fact") of
selection by environmental consequences, behavior occurs
and concrete environmental conditions ensue. The effects
(changes in behavior) are usually delayed, sometimes
considerably, making it difficult for observers to detect
the selective process.
Skinner's emphasis on a new causal mode was based
on his work in operant conditioning. In the study and
application of operant conditioning in its most elementary
form, one makes certain consequences available after responses.
Over the years, Skinner and his colleagues built
on this procedure a sufficiently large body of data to justify
the principle most associated with him to the present:
Behavior is a function of its (past) environmental consequences.
Selection by consequences is a generalization
of this principle. In his earliest work, Skinner departed
from the antecedent causality of the environmental stimulus-
response model with the operant model that reverses
the terms: response->environment. Thus, it was
not difficult for him to raise to the level of a fundamental
assumption the principle of consequential causality.
Skinner's stance on causality also led him to a second
major departure from environmentalistic hypotheses. His
causal mode of selection by consequences applies to both
the phylogeny and ontogeny of behavior of species and
of individuals, respectively (Skinner, 1966b). Thus, he
left behind a bias frequently associated with environmentally
oriented theories (i.e., denial or rejection of genetic
factors in behavior). Phylogenic contingencies, or
contingencies of survival, are behavior-consequence relations
that select what appears to be inherited behavior.
Ontogenic contingencies, or contingencies of reinforcement,
are behavior-consequence relations that select behavior
originating during the life of the organism. Because
the contingencies responsible for inherited behavior occur
in the species' evolutionary history, they are more difficult
to confirm than are the reinforcement contingencies that
an experimenter can manipulate.
With the extension of selection by consequences
from the phylogeny of behavior to the ontogeny of behavior,
Skinner counteracted any assertion that his behaviorism
is a form of environmentalism that denies
heredity. There is clearly a role for genetic participation
in behavior through contingencies of survival. However,
it is overly simplistic to take genes as determiners of behavior.
Skinner's position was that behavior per se is not
inherited. This is clarified later in the assumption on the
organism as the locus of change (see Organism as the
Locus of Biological Change). Instead, susceptibility to
ontogenic contingencies is inherited; thus, genetics applies
to all behavior just as do those factors commonly referred
to as environmental.
Materialism: Dualism Is False; the Only
World Is a Physical World
Private and public events have the same kinds of physical dimensions.
(Skinner, 1963, p. 953)
The task of a scientific analysis is to explain how the behavior
of a person as a physical system is related to the conditions
under which the human species evolved and the conditions under
which the individual lives. (Skinner, 1971, p. 14)
[My] position can be stated as follows: what is felt or introspectively
observed is not some nonphysical world of consciousness,
mind, or mental life but the observer's own body. (Skinner, 1974,
p. 17)
A small part of the universe is contained within the skin of each
of us. There is no reason why it should have any special physical
status because it lies within this boundary. (Skinner, 1974, p. 21)
No special kind of mind stuff is assumed. A physical world
generates both physical action and the physical conditions within
the body to which a person responds when a verbal community
arranges the necessary contingencies. (Skinner, 1974, p. 220)
Behaviorists before Skinner attempted to shed dualism.
Dualism holds that the world consists of two fundamental
November 1992 • American Psychologist 1511
realms: physical (body or matter) and nonphysical (mind
or spirit). Matter exists in space and time, whereas mind
exists nonspatiotemporally. Most contemporary dualists
claim that the physical realm can influence the nonphysical
realm, and vice versa (e.g., Eccles, 1989). Skinner
held that, because his analyses provided a complete account
of behavior without reference to such supposed
physical-nonphysical relations, the initial positing of the
dualist dichotomy was unnecessary and impedes effective
analysis. His alternative is materialism.
Materialism asserts that the world is composed of
physical or material things, varying in their states and
relations, and nothing else. This monism was clearly expressed
by Skinner (1945b): "What is lacking [in dualistic
analyses] is the bold and exciting behavioristic hypothesis
that what one observes and talks about is always the 'real'
or 'physical' world (or at least the 'one' world)" (p. 293).
Thus, private events refer to "real" events, and their ontological
status is identical to that of any other aspect of
the physical world.
Behavior as Subject Matter: The Subject
Matter of Psychological Science Is
Behavior and Behavior Only
If psychology is a science of mental life—of the mind, of conscious
experience—then it must develop and defend a special
methodology, which it has not yet done successfully. If it is, on
the other hand, a science of the behavior of organisms, human
or otherwise, then it is part of biology, a natural science for
which tested and highly successful methods are available. (Skinner,
1963, p. 951)
Whatever happened to psychology as the science of behavior?
(Skinner, 1987b, p. 780)
Psychology should confine itself to its accessible subject matter
[i.e., behavior]. (Skinner, 1987b, p. 785)
For Skinner the first step in developing a theory useful
for understanding in a science was to identify the nature
of fundamental events. Psychology had to begin with the
unequivocal understanding that its events were behavioral.
This placed Skinner in the behavioral movement that
began in the early 20th century. Taken together with the
pragmatic view that prediction and control are the purposes
of science, the value and success of any knowledge
claim are to be determined by the degree to which the
knowledge claimed is useful in predicting and controlling
behavior. The alternative to accepting behavior as the
subject matter is to opt for mind or mental life, which in
turn leads to mental explanations of behavior. Skinner
(1963) considered that giving primary emphasis, or any
at all, to mind was particularly undesirable because of
the failure of this strategy to contribute to the prediction
and control of behavior.
We sometimes hear that all of today's psychology is
at least partly behavioristic in the sense that measurement
always comes down in one way or another to the outcome
of observing individuals' behavior, if nothing else, their
verbal expressions, or marks on paper. Skinner (1974)
objected to this methodological version of behaviorism
when it treats behavior merely as a dependent variable
that is an indicator of something other than behavior (e.g.,
mind, cognition, brain activity), hence encouraging the
view that psychology is the science of behavior and mental
life. The result is that behavior remains subsidiary to
mentality or the nervous system. Skinner's (1938) often
reiterated position that "behavior may be treated as a
subject matter in its own right" (p. 440) does not question
the importance of the subject matter of the neurosciences
for behavior (Skinner, 1989, p. 130). He maintained that
"we can predict and control behavior without knowing
anything about what is happening inside [although] a
complete account will nevertheless require the joint action
of both sciences, each with its own instruments and
methods" (Skinner, 1989, p. 130).
There are two major subtopics to consider in discussing
Skinner's behavioral emphasis. The first concerns
his definition of behavior, which is rather abstract. For
example, to describe an episode of behavior, one cannot
be restricted to the behaving organism. Skinner held that
behavior is the action of the whole organism, not a piece
of it: "It is the organism as a whole that behaves" (Skinner,
1975, p. 44). "Behavior is what an organism is doing"
(Skinner, 1938, p. 6) that we can determine by observing
its relations with its environment, that is, "the action of
the organism upon the outside world" (Skinner, 1938, p.
6). Skinner's rejection of the activity of muscles and organs
per se as the essence of psychological behavior is elaborated
later in the section on Classification of Behavior
Into Respondent and Operant.
The second important issue that arises from Skinner's
thoroughgoing behavioral position concerns his apparent
rejection of everything mental. The history of the
behavioral movement involves confrontations between
mentalistic and behavioristic investigative and explanatory
attempts. Skinner viewed as misguided those behavioristic
attempts that (a) ignore the events that serve as
the controlling variables for talk of mind and consciousness,
(b) substitute self-reports for the events referred to
as mental, and (c) use behavior as an indicator of mental
activity that is assumed to be fundamentally different
from behavior. He distinguished public and private events,
although he did not consider the two to differ in the stuff
of which they are made (see Materialism section). Public
events are those accessible to other observers, whereas
private events are characterized by limited accessibility.
That is, private events (e.g., a toothache) are observable
only to the individual in whose body they are occurring.
Because others cannot directly observe certain psychological
events, observable behavioral (public) events are
separated from so-called mental (private) events, thereby
holding back a completely behavioral approach to the
subject matter of psychology. Skinner, of course, was not
embracing mental events per se but simply encouraging
an analysis of those behavioral events that are described
in mainstream psychology and in the vernacular as
"mentalistic." Skinner (1963) suggested that "the problem
of privacy may be approached in a fresh direction by
1512 November 1992 • American Psychologist
starting with behavior rather than with immediate experience"
(p. 953) and that
It is particularly important that a science of behavior face the
problem of privacy. It may do so without abandoning the basic
position of behaviorism. Science often talks about things it cannot
see or measure. When a man tosses a penny into the air, it
must be assumed that he tosses the earth beneath him downward.
It is quite out of the question to see or measure the effect on
the earth, but an effect must be assumed for the sake of a consistent
account. An adequate science of behavior must consider
events taking place within the skin of the organism, not as physiological
mediators of behavior but as part of behavior itself. It
can deal with these events without assuming that they have any
special nature or must be known in any special way. The skin
is not that important as a boundary. Private and public events
have the same kinds of physical dimensions, (p. 953)
Thus Skinner encouraged investigation of so-called
sensations, perceptions, images, thoughts, awareness, and
the like but did not treat these private events as fundamentally
different from any of the public organismic and
environmental behavioral events that served as the original
classes of events studied by behaviorism.
Reductionism and Nonreductionism
Skinner's position seems to be most ambiguous on the
issue of reductionism, so much so that we offer two conflicting
assumptions.
Reductionism: The Subject Matter of Psychology
Is Reducible (at Least to Biology)
Eventually, we may assume, the facts and principles of psychology
will be reducible not only to physiology but through
biochemistry and chemistry to physics and subatomic physics.
(Skinner, 1947, p. 31)
The behaving organism will eventually be described and explained
by the anatomist and physiologist. As far as behavior is
concerned, they will give us an account of the genetic endowment
of the species and tell how that endowment changes during the
lifetime of the individual and why, as a result, the individual
then responds in a given way on a given occasion. (Skinner,
1975, p. 42)
The physiologist of the future will tell us all that can be known
about what is happening inside the behaving organism. His account
will be an important advance over a behavioral analysis,
because the latter is necessarily "historical"—that is to say, it
is confined to functional relations showing temporal gaps.
Something is done today which affects the behavior of an organism
tomorrow. No matter how clearly that fact can be established,
a step is missing, and we must wait for the physiologist
to supply it. (Skinner, 1974, p. 215)
Nonreductionism: Behavior Cannot Be Completely
Explained in Terms of Biology or Any Other
"Lower-Level" Discipline
[This work] is not necessarily mechanistic in the sense of reducing
the phenomena of behavior ultimately to the movement
of particles, since no such reduction is made or considered essential.
(Skinner, 1938, p. 433)
Behavior is an acceptable subject matter in its own right, and
. . . it can be studied with acceptable methods and without an
eye to reductive explanation. (Skinner, 1961, p. 64)
We do not need an explicit account of the anatomy and physiology
of genetic endowment in order to describe . . . behavior.
. . . Nor do we need to consider anatomy and physiology in
order to see how the behavior of the individual is changed by
his exposure to contingencies of reinforcement during his lifetime
and how as a result he behaves in a given way on a given
occasion. (Skinner, 1975, pp. 42-43)
A science of behavior will be needed for both theoretical and
practical purposes even when the behaving organism is fully
understood at another level. (Skinner, 1975, p. 43)
Discussion
The reductionism assumption states a reductionistic position
whereby events on one level can be explained in
terms of another supposedly more simple or basic level.
However, statements such as those following the nonreductionism
assumption argue against such a resolution
of psychological events to biology or any other discipline
and assert that behavior can and should be treated as a
subject matter in its own right without appeal to another
level of explanation.
If Skinner's position were unequivocally reductionistic,
then it would be in agreement with conventional
scientific materialism, according to which science deals
with a material realm, as contrasted with a spiritual or
idealistic one, and it is possible and desirable to attempt
to explain events in terms of materialistic concepts that
are at a lower level. The first quotation under Reductionism
exhibits an extreme form of materialistic reductionism
when subatomic physics is considered to be the eventual
basis of explanation of psychological events.
Skinner exhibited a nonreductionistic side to his behaviorism,
as the quotations under Nonreductionism
show. At one point in About Behaviorism, he explicitly
disavowed reductionism. He argued that his approach in
no way advocated reducing any aspect of humans, stressing
that behaviorism does not "reduce feelings to bodily
states . . . reduce thought processes to behavior . . . or
reduce morality to certain features of the social environment"
(Skinner, 1974, p. 241).
Ambiguity is introduced into Skinner's seemingly
antireductionistic position when, in referring to the physiologist
of the future, he stated that "his account will be
an important advance over a behavioral analysis, because
the latter is necessarily 'historical'—that is to say, it is
confined to functional relations showing temporal gaps"
(Skinner, 1974, p. 215). Does this imply that someday
psychology will be reduced to physiology? Evidence for
such an implication seems to exist:
In general I reject any appeal to physiology in explaining behavior
on the grounds that physiology is still [as of the present]
far less advanced than the analysis of behavior and has yet to
[but may someday?] deal with the processes responsible for the
behavior attributed to contingencies of reinforcement. (Skinner,
1982, p. 190)
November 1992 • American Psychologist 1513
It seems that the key to understanding Skinner's position
on reductionism is that he endorsed a physiological
analysis only insofar as it may prove useful in filling the
temporal gaps in the functional analysis of the relations
between the individual's exposure to causal environmental
contingencies and ensuing behavior. Insofar as the
purpose of behavior analysis is the prediction and control
of behavior (see Purpose of Science), a functional analysis
definitely has the upper hand over a physiological analysis.
That the latter is a possibility, however, is seemingly not
ruled out.
Organism as the Locus of Biological Change:
The Organism Changes Through Evolutional
and Environmental Histories, and the
Changes are Biological
Evolutionary and environmental histories change an organism.
(Skinner, 1971, pp. 195-196)
[The physiologist of the future] will be able to show how an
organism is changed when exposed to contingencies of reinforcement
and why the changed organism then behaves in a
different way, possibly at a much later date. What he discovers
cannot invalidate the laws of a science of behavior, but it will
make the picture of human action more nearly complete. (Skinner,
1974, p. 215)
Both kinds of contingencies [phylogenic and ontogenic] change
the organism so that it adjusts to its environment in the sense
of behaving in it more effectively. (Skinner, 1966b, pp. 1211-
1212)
People are changed by the contingencies of reinforcement, they
do not store information about them. (Skinner, 1988e, p. 53)
Contingencies of reinforcement change the individual; as a result
the individual now behaves in a different way. (Skinner, 1988d,
p. 409)
The relation between environmental history and contemporary
behavior is so important that we address this
issue with a separate assumption. This assumption, concerning
the locus and form of changes induced by selective
consequences, clarifies why Skinner addressed biologicalbehavioral
relations in the first place. It does not clarify
where Skinner stood on reductionism-nonreductionism
per se.
Above and beyond behavior, what changes when an
organism is exposed to effective contingencies? Skinner's
answer was that the organism changes but not in any
conventionally psychological way (e.g., psychically, mentally,
cognitively). In the case of natural selection, the
organism has been changed by endowing it with a physiology
that makes conditioning possible: "What has been
selected appears to be a susceptibility to ontogenic contingencies"
(Skinner, 1966b, p. 1208). In the case of operant
conditioning, reinforcement contingencies change
organisms biologically during their individual life spans.
Skinner's emphasis on selective contingencies
changing organisms biologically counteracted explanations
in terms of traditional psychological concepts of
acquisition and storage devices. To restrict our attention
to ontogenic contingencies, the explanatory problem is
the temporal gap between exposure to reinforcement
contingencies and subsequent behavioral change. Skinner
viewed a biologically empty organism approach as unsatisfactory
for handling the mediation of reinforcement
effects over time, but, consistent with the fundamental
assumption of materialism (see previous discussion), he
found nonphysical (i.e., mentalistic) accounts equally
lacking. He rejected the notion that the organism cognitively
internalizes contingencies of reinforcement in
such forms as information, knowledge, or expectations
that require unknown storage mechanisms so they can
be activated in the future on occasions when behavior
occurs. The alternative to modified mental structures and
processes with exposure to reinforcement contingencies
is a changed organism, one changed biologically. To deny
alterations of mental conditions is not to deny that reinforcement
contingencies modify "what is felt as feelings
or introspectively observed as states of mind . . . [for
these] . . . are the products of certain contingencies of
reinforcement" (Skinner, 1988b, p. 175).
Classification of Behavior Into Respondent
and Operant: There Are Two Major Classes
of Behavior or, More Completely, Functional
Relations: Respondent and Operant
The kind of behavior that is correlated with specific eliciting
stimuli may be called respondent behavior and a given correlation
a respondent. The term is intended to carry the sense of a relation
to a prior event. Such behavior as is not under this kind of
control I shall call operant and any specific example an operant.
(Skinner, 1938, p. 20)
[Two processes evolved] through which individual organisms
acquired behavior appropriate to novel environments. Through
respondent (Pavlovian) conditioning, responses prepared in advance
by natural selection could come under control of new
stimuli. Through operant conditioning, new responses could be
strengthened ("reinforced") by events which immediately followed
them. (Skinner, 1981, p. 501)
This assumption extends the definition of behavior
brought up under Behavior as Subject Matter. The classification
of behavior has its origins in the distinction
between involuntary and voluntary behavior. According
to Skinner, the term involuntary, when used properly, refers
to elicited behavior or reflexes of the sort most associated
with Pavlov's work. Mechanical causality applies
to behavior of this class; the responses of stimulus-response
functional relations are true responses insofar as
they depend upon immediately prior events (Skinner,
1953, p. 64). The term respondent applies to this class of
stimulus-response functional relations, and they are
modified only in the sense that the eliciting stimuli of
responses can be changed. Respondent conditioning refers
to this process of changing eliciting stimuli. Thus, Skinner
included work inspired by Pavlov in his behaviorism but
found that class of behavior known in the vernacular as
1514 November 1992 • American Psychologist
"voluntary" behavior of much more interest and relevance
for psychology.
Perhaps most of Skinner's novel contributions to
behavioral science revolve around the operant class of
behavior to which the new causal mode of selection by
consequences (see Consequential Causality section) applies.
Operant behavior is denned by functional relations
between classes of responses (not specific instances of responses)
and environmental consequences. The term operant
"emphasizes the fact that the behavior operates
upon the environment to generate consequences" (Skinner,
1953, p. 65).
A more complete definition of operant behavior requires
a distinction between response instances and response
classes (e.g., Skinner, 1953, p. 65, 1969, p. 131).
The former is specifiable in terms of topography or structure
and refers to a particular, specific occurrence of a
response. For example, "The pigeon pecked a key at 1:
30 p.m." and "Edwardo drank a glass of water between
4:00 and 4:02 p.m. today." On the other hand, key pecking,
for example, regardless of when specific instances
occur, defines a response class. It is a set of acts defined
by a measurable impact on the environment (such as
activation of a relay) that transcends particular instances
and forms of response. Given this distinction between
response instances and response classes, Skinner defined
the operant as follows:
The term emphasizes the fact that the behavior operates upon
the environment to generate consequences. The consequences
define the properties with respect to which responses are called
similar. The term will be used both as an adjective (operant
behavior) and as a noun to designate the behavior defined by a
given consequence. (Skinner, 1953, p. 65)
According to this definition, an operant is identified with
a response class that can be strengthened by events (reinforcers)
that immediately follow it, but whether or not
the operant is strengthened has no bearing on definition
of the operant.
In theoretical, research, and practical applications
of the operant construct, Skinner stressed contingencies
between operants and consequential events as crucial for
predicting and controlling operant behavior. In fact,
Skinner (1969) also tied the definition of operant behavior
to experimental demonstrations in which a response class
is modified as a function of its consequences. For example,
[In discussing the act of "flipping on a light switch":] The topography
of the response is described accurately enough as
"flipping the switch." If the appearance of light is reinforcing
[italics added; "flipping the switch" becomes more likely]. . .
the topography and the consequences define an operant. (p. 128)
It is always a response [instance] upon which a given reinforcement
[consequence] is contingent, but it is contingent upon
properties which define membership in an operant. Thus a set
of contingencies defines an operant. (p. 131)
Allowing water to pass over one's hands can perhaps be adequately
described as topography, but "washing one's hands" is
an "operant" defined by the fact that, when one has behaved
this way in the past, one's hands have become clean—a condition
which has become reinforcing [italics added] because, say, it has
minimized a threat of criticism or contagion. Behavior of precisely
the same topography would be part of another operant if
the reinforcement had consisted of simple stimulation (e.g.,
"tickling") of the hands, (p. 130)
Catania (1973) referred to the latter use of the term operant
as a functional one ordinarily used in theoretical
discussion. He noted that in the method section of experimental
reports the term is used descriptively, that is,
without requiring modifiability as a function of consequences.
Rate of Responding as Fundamental Datum
Skinner offered probability of response as the primary
conceptual measure of his science. However, probability
is not directly measurable, and measures such as latency
and magnitude are not appropriate for operant behavior
because such behavior is not elicited by antecedent stimuli.
Skinner suggested that "in operant conditioning we
'strengthen' an operant in the sense of making a response
more probable or, in actual fact, more frequent" (Skinner,
1953, p. 65). Thus, Skinner's solution to the unique character
of the operant was to measure probability primarily
by way of frequency of response, or, more precisely, "the
length of time elapsing between a response [instance] and
the response [instance] immediately preceding it or, in
other words, the rate of responding" (Skinner, 1938, p. 58).
Purpose
The operant concept, in conjunction with several of the
assumptions we have presented up to this point, is crucial
for understanding Skinner's handling of the purpose,
meaning, and intention of behavior. Consider an episode
in which an individual "washed his or her hands." This
response instance is part of history and is not an operant.
However, what the individual was doing here may be
properly regarded as operant behavior on the basis that
when the person has behaved similarly in the past their
hands were made cleaner, a condition that has become
reinforcing because it has been socially praised, has reduced
the likelihood of criticism, or has made contagion
less likely. In this case, we may speak of the operant class
of "hand washing." It is possible, as some of those who
work with certain classes of individuals with disabilities
will note, that the same topography participates in defining
another operant such as "stimulating body" on the
basis that the critical consequences of past instances have
not been those that define hand washing but rather mere
tactual stimulation. Intention and purpose are best understood
in terms of controlling variables (see Methodology
section):
Purpose is not a property of behavior itself; it is a way of referring
to controlling variables. (Skinner, 1953, p. 88)
When someone says that he can see the meaning of a response,
he means that he can infer some of the [potentially manipulable]
variables of which the response is usually a function. (Skinner,
1957, p. 14)
November 1992 • American Psychologist 1515
Clean hands and body stimulation are very different consequences
of performing the same topographies; the future
orientations (purposes) of the two operants distinguish
them.
The replacement of purpose with selection, in combination
with consequential causality (see previous discussion),
led Skinner to a proposal concerning the evolution
of cultural practices. He fully discussed the implications
of this (e.g., 1953, 1969, 1971, 1972, 1981, 1987a),
beginning with the Utopian novel Walden Two (1948). In
our previous discussion of consequential causality, we
pointed out that Skinner applied selection by consequences
to phylogeny (biological natural selection) and
to ontogeny (the behavior of individuals). At the level of
the evolution of practices characteristic of a group of people,
Skinner suggested that the selecting consequences
are those that contribute to the survival of the group.
Thus, cultural evolution is a third kind of natural selection.
He argued that the delayed consequences of cultural
practices (e.g., the controlled use of fire) are "too remote
to reinforce the behavior of any member of the group"
(Skinner, 1989, p. 117).
Stimulus Control of Operant Behavior:
Operant Behavior Can Be Brought Under the
Control of Antecedent Stimuli, and
Description of Operant Behavior Usually
Requires Three Elementary Terms and Their
Functional Interrelations
The occasion upon which behavior occurs, the behavior itself,
and its consequences are interrelated in the contingencies of
reinforcement.. . . As a result of its place in these contingencies,
a stimulus present when a response is reinforced acquires some
control over the response. It does not then elicit the response
as in a reflex; it simply makes it more probable that it will occur
again. (Skinner, 1974, pp. 73-74)
An adequate formulation of the interaction between an organism
and its environment must always specify three things: (1) the
occasion upon which a response occurs, (2) the response itself,
and (3) the reinforcing consequences. The interrelationships
among them are the "contingencies of reinforcement." (Skinner,
1969, p. 7)
Under certain conditions, experimenters can reliably
"turn on" and "turn off" operant behavior by presenting
and removing stimuli. Such stimulus-response relations
may even have the appearance of respondent relations;
however, more detailed examination will reveal that these
effects are systematic outcomes of the organism's previous
exposure to response-consequence contingencies conditional
on the presence or absence of particular stimuli.
In the simplest case, occurrences of instances of a response
class are followed by reinforcement only in the presence
of a stimulus. This stimulus is said to acquire discriminative
stimulus control over the operant. After discriminative
stimulus control has developed, the experimenter
can manipulate the stimulus as an independent variable
to control the operant behavior. In this sense, the stimulus
definitely is a controlling variable of the behavior in question.
However, its status as a controlling variable is conditional
inasmuch as functional relations between discriminative
stimuli and behavior are dependent on a history
of selective behavior-consequence relations (see
Consequential Causality and Classification of Behavior
Into Respondent and Operant sections) in the presence
of the stimulus. In contrast to the eliciting function of
stimuli in respondent relations, the discriminative stimuli
of operant relations are said to "set the occasion" for
responding. Even when discriminative stimuli have the
surface characteristics of elicitors, because the history of
their functional effects is traceable to operant behaviorconsequence
relations, it is not proper to fit them into
the stimulus-response reflex causal mode.
To take into account the development of discriminative
stimulus control of operants, description of behavior
requires not only response-reinforcer functional
relations but discriminative stimulus-response relations
as well. The result is the "three-term" contingency—discriminative
stimulus: response-reinforcing consequence.
The three-term contingency forms the fundamental unit
of analysis in the study of operant behavior.
Skinner's conception of the three-term contingency
as the fundamental unit of stimulus control was a radical
departure from the reflexological-derived stimulus-response
model, according to which the organism could
only respond (in the conventional sense of the term) to
prior physical or mental stimuli. In terms of the operant
framework, at the most elementary level, the crucially
important class of behavior called voluntary is a matter
of the selective altering of the probability of behavior by
discriminative stimuli in the presence of which the behavior
was selected by consequences in the history of the
organism.
On the Generality of Behavioral Principles:
The Full Complexities of Human Activity—
Including Language, Thinking, Consciousness,
and Science—Are Behaviors to Which All
These Features Apply
This final assumption of Skinner's psychology reveals
perhaps the most revolutionary portion of Skinner's careerlong
attempt to permit no aspect of the human experience
to remain untouched by scientific understanding.
The breadth of his applications that this section reveals
is further amplified by articles on the technology of education,
psychotic behavior, artistic creativity, the genesis
of a poem, literary products, and telepathy experiments
(Skinner, 1972).
On Language
Our first responsibility is simple description: what is the topography
of this subdivision of human behavior? Once that question
has been answered . . . we may advance to the stage called
explanation: what conditions are related to the occurrence of
the behavior—what are the variables of which it is a function?
Once these have been identified, we can account for the dynamic
1516 November 1992 • American Psychologist
characteristics of verbal behavior within a framework appropriate
to human behavior as a whole. (Skinner, 1957, p. 10)
Language has the character of a thing, something a person acquires
and possesses. Psychologists speak of the "acquisition of
language" in the child. The words and sentences of which a
language is composed are said to be tools used to express meanings,
thoughts, ideas . . . and many other things in or on the
speaker's mind. A much more productive view is that verbal
behavior is behavior. It has a special character only because it
is reinforced by its effects on people—at first other people, but
eventually the speaker himself. As a result, it is free of the spatial,
temporal, and mechanical relations which prevail between operant
behavior and nonsocial consequences. (Skinner, 1974, pp.
88-89)
Skinner (1945a, 1945b) addressed the four behavioral
complexities included in this final assumption that affirm
a thoroughgoing behavioral psychology. We obtain the
first glimpse of an approach to language that is neither
mentalistic nor in terms of stimulus-response behaviorism
and one that Skinner subsequently used as a keystone
to handle thought, what it means to be conscious, and
the nature of science. Skinner (1957) rejected conventional
views of language as the "use of words," the "communication
of ideas," the "sharing of meaning," the
"expression of thoughts," and so on. Consistent with his
position on the subject matter of psychology (see Behavior
as Subject Matter section), Skinner took language as behavior
that is understandable in its own right. Given this,
he found that (a) the term verbal behavior is preferable,
(b) the behavior is of the voluntary (operant) class (see
Classification of Behavior Into Respondent and Operant
section), (c) the behavior is selected (see Consequential
Causality) by environmental (see Locus of Behavioral
Control) consequences, (d) the behavior is subject to
functional analysis (see Methodology), and (e) the starting
point for description is the three-term contingency (see
Stimulus Control of Operant Behavior).
Although bearing a certain superficial similarity to
nonhuman behavior under conditions of food-reinforced
lever pressing or key pecking, verbal behavior operates in
a fundamentally different way. We can describe in physical
terms the nonhuman pressing a key or the human in a
nonsocial situation who walks toward an object and picks
it up. In contrast, humans frequently act "only indirectly
upon the environment from which the ultimate consequences
of [their] behavior emerge.. . . Instead of going
to a drinking fountain, a thirsty man may simply 'ask for
a glass of water' " (Skinner, 1957, p. 1). The special aspect
of verbal behavior is that other persons are crucially involved,
at least indirectly, in mediating the consequences
of the speaker's behavior. As such, verbal behavior has
special controlling variables (i.e., social ones). The role
of others in controlling verbal behavior is favored over
reference or correspondence theories of language. Skinner
objected to the conventional view that words or collections
of words stand for or refer to objects. Consistent with the
discussion of purpose, intention, and meaning under
Classification of Behavior Into Respondent and Operant,
Skinner treated the meaning of words in terms of the
variables determining their occurrence in any given instance.
According to Skinner's treatment, we account for
a remark and understand what it means by identifying
the variables that control it. For example, we say that we
see different meanings of "fire" when given as a command
to a firing squad, when made in the presence of a burning
building, and when preceded by "wind and rain."
Skinner's analysis of the role of verbal behavior led
him to recognize a complexity in the nature of the relations
controlling operants that took his behaviorism ever
more fully into what psychologists typically regard as
cognitive processes. Specifically, he proposed that operant
behavior can be classified into two categories: contingency
shaped and rule governed.
The response which satisfies a complex set of contingencies, and
thus solves the problem, may come about as the result of direct
shaping by the contingencies . . . or it may be evoked by contingency-
related stimuli constructed either by the problem solver
himself or by others. The difference between rule-following and
contingency-shaped behavior is obvious when instances are
pretty clearly only one or the other. (Skinner, 1966a, p. 241)
A person who is following directions, taking advice, heeding
warnings, or obeying rules or laws does not behave precisely as
one who has been directly exposed to the contingencies, because
a description of the contingencies is never complete or exact
. . . and because the supporting contingencies are seldom fully
maintained. (Skinner, 1974, p. 125)
To this point, we have discussed one of the two classes
of operant behavior. Contingency-shaped behavior is the
class that most obviously (a) follows the principle of consequential
selection and (b) comes under stimulus control
and is described using the three-term contingency. Skinner
(1966a) greatly extended the scope of his approach
in the realm of human behavior with the addition of the
rule-governed class of operant behavior. Rule-governed
behavior occurs when the individual is behaving in accordance
with explicit rules, advice, instructions, modeling
performances, plans, maxims, and the like. Rules
are contingency-specifying stimuli. Either directly or by
implication from past experience, the rule specifies an
environmental consequence of behaving in a certain way
(e.g., "A free gift will be given to the first 100 individuals
who enter the store" and "The student is required to attend
all class meetings to pass this course").
If we advise someone how to "make friends and influence
people," his or her behavior may change to a point
at which he or she acts more in accordance with our
rules. However, the ultimate effect of the advice rests on
the environmental contingencies enjoined by our advisee's
rule-following behavior. Rules function as discriminative
stimuli (see Stimulus Control of Operant Behavior section),
and a person will follow rules in the first place to
the extent that "previous behavior in response to similar
verbal stimuli [i.e., rules, advice] has been reinforced"
(Skinner, 1966a, p. 244). Thus, consequential selection
is central in the development and maintenance of rulegoverned
behavior albeit less directly so than with many
November 1992 • American Psychologist 1517
cases of straightforward contingency-controlled behavior.
The pervasiveness of social factors in human behavior
makes most cases of human behavior the product of both
contingencies and rules.
On Thinking
The simplest and most satisfactory view is that thought is simply
behavior—verbal or nonverbal, covert or overt. It is not some
mysterious process responsible for behavior but the very behavior
itself in all the complexity of its controlling relations, with respect
to both man the behaver and the environment in which he lives.
(Skinner, 1957, p. 449)
Mental life and the world in which it is lived are inventions.
They have been invented on the analogy of external behavior
occurring under external contingencies. Thinking is behaving.
The mistake is in allocating the behavior to the mind. (Skinner,
1974, p. 104)
Implicit in Skinner's (1957) interpretation of verbal behavior
is a listener who responds to the speaker's verbal
stimuli. Those cases in which speakers also become listeners
(e.g., talk to themselves), particularly when others
cannot observe their behavior, describe (but not fully, see
later discussion) the "special human achievement called
'thinking' " (Skinner, 1957, p. 433). This is not to suggest
that Skinner took the early behavioristic position that
identified thinking with subaudible talking. His view was
considerably more sophisticated. The main characteristic
of thinking is that persons behave with respect to themselves,
which is to say that controlling relations do not
involve other persons. Thus, the person's own behavior
has an overt or covert self-stimulatory effect, once again
illustrating that Skinner's framework for behavior was
not formulated on the S-R, wind-weather vane model.
Although thinking frequently is both covert (not readily
observable by others) and verbal, it might be overt (if
someone were in the presence of the behaver, he or she
could readily observe it), and it is not restricted to verbal
behavior. The critical mark of thinking is not tied to distinctions
between verbal and nonverbal, overt and covert,
and private and public nor to weak and strong behavior
(Skinner, 1957, 1989). According to Skinner, thinking is
behaving, either verbal or nonverbal, overt or covert,
weakly or strongly.
In our view, one of the most subtle and pregnant
features of Skinner's behaviorism concerns the assumption
that thinking can be covert or overt nonverbal behavior.
Recall that the definition of verbal behavior requires
a special character: the participation of social consequences
that may include speakers themselves.
Nonverbal behavior entails no participating social consequences;
the consequences are components of the behaver's
nonsocial environment. What this comes down
to is that
To think is to do something that makes other behavior possible.
Solving a problem is an example. A problem is a situation that
does not evoke an effective response; we solve it by changing
the situation until a response occurs. Telephoning a friend is a
problem if we do not know the number, and we solve it by
looking up the number. (Skinner, 1989, p. 20)
To us, it seems that Skinner took the radical position that
all thinking is operant behavior and all operant behavior
is thinking. This appears to be the message "behind"
Skinner's (1957) assertion that "so far as a science of
behavior is concerned, Man Thinking is simply Man Behaving"
(p. 452) and in remarks such as "in the broadest
possible sense, the thought of Julius Caesar was simply
the sum total of his responses to the complex world in
which he lived" (Skinner, 1957, pp. 451-452) and "cognitive
processes are behavioral processes; they are things
people do" (Skinner, 1989, p. 23). If we are correct, operant
behavior—Skinner's foremost subject throughout
his career—completely encompasses one of the main
strongholds of the mentalistic psychology that he aspired
to replace with his behaviorism.
On Consciousness
Being conscious, as a form of reacting to one's own behavior,
is a social product.. . . The individual becomes aware of what
he is doing only after society has reinforced verbal responses
with respect to his behavior as the source of discriminative
stimuli. The behavior to be described (the behavior of which
one is to be aware) may later recede to the covert level, and (to
add to a crowning difficulty) so may the verbal response. It is
an ironic twist, considering the history of the behavioristic revolution,
that as we develop a more effective vocabulary for the
analysis of behavior we also enlarge the possibilities of awareness,
so denned. (Skinner, 1945a, p. 277)
I believe that all nonhuman species are conscious in the sense
[that they see, hear, feel, and so on], as were all humans prior
to the acquisition of verbal behavior.. . . But they do not observe
that they are doing so. . . . A verbal community asks the individual
such questions as, "What are you doing?," "Do you
see that?," "What are you going to do?," and so on and thus
supplies the contingencies for the self-descriptive behavior that
is at the heart of a different kind of awareness or consciousness.
(Skinner, 1988a, p. 306)
The previous discussion of Skinner's position on behavior
as the subject matter of psychology (see Behavior as Subject
Matter) indicates that he did not avoid private events.
His handling of thinking is an example of how he forthrightly
confronted the problem of privacy. Consciousness
or awareness is another aspect of human experience that
is often taken as private. Skinner interpreted consciousness
such that it is not restricted to humans but yet he
retained the uniqueness of human consciousness. Far
from treating humans as but complicated rats or pigeons,
Skinner argued that there is something special about human
behavior.
Skinner (1974) distinguished between two classes of
consciousness. Consciousness i (our designation) pertains
to organisms' "awareness" of their environment, as when
we say they are "conscious of their surroundings." Persons
who have been rendered unconscious are no longer under
the stimulus control of events either inside the body or
outside it, and they may talk without being "conscious
of their effects on listeners." Humans and nonhumans
alike are conscious in this respect.
1518 November 1992 • American Psychologist
They feel pain in the sense of responding to painful stimuli, as
they see a light or hear a sound in the sense of responding appropriately,
but no verbal contingencies make them conscious
[consciousness2] of pain in the sense of feeling that they are
feeling, or of light or sound in the sense of seeing that they are
seeing or hearing that they are hearing. (Skinner, 1974, p. 220)
Thus, for Skinner, consciousness2 was (probably) restricted
to humans because of its social-verbal nature.
Other persons arrange verbal contingencies for behavior
descriptive of our behavior, permitting us to state rules
about it and its relation to controlling variables. In tying
together the uniquely human consciousness2 with verbal
contingencies that are always social, Skinner's behaviorism
has a "radically" social cast that does not seem to
be much appreciated in conventional social psychology.
On Science
The behavior of the logician, mathematician, and scientist is
the most difficult part of the field of human behavior and possibly
the most subtle and complex phenomenon ever submitted to a
logical, mathematical, or scientific analysis, but because it has
not yet been well analyzed, we should not conclude that it is a
different kind of field, to be approached only with a different
kind of analysis. (Skinner, 1974, p. 235)
Scientific knowledge is verbal behavior, though not necessarily
linguistic. It is a corpus of rules for effective action, and there
is a special sense in which it could be "true" if it yields the most
effective action possible. (Skinner, 1974, p. 235)
The scientist first interacts with the world, like everyone else,
in contingency-shaped behavior. He becomes a scientist when
he begins to describe the contingencies and to design experiments
which make them clearer. The ultimate product, the "laws" of
science, governs scientific behavior as a corpus of rules to be
followed. The behavior of the scientist in following them is reinforced
by the same consequences as the original contingencyshaped
behavior, but the controlling stimuli are different. (Skinner,
1988c, p. 197)
In the article "The Operational Analysis of Psychological
Terms," Skinner (1945a) took the standpoint of functional
analysis (see Methodology section) in his approach
to the question of terms, concepts, and constructs. Basically,
his position was that because the knowing profession
of science offered as products verbal behavior, the
preferred approach to epistemology was to describe the
behavior of the scientist. Thus, a scientific term or knowledge
claim has the same status as any other piece of human
behavior. He later elaborated on the implications of
an empirical epistemology (e.g., Skinner, 1957, 1963,
1974). According to this way of approaching knowledge,
terms are not important. What is important are the conditions
under which the scientist uses the term (i.e., the
controlling variables). In other words, the meaning of scientific
terms is handled as is the meaning of any other
behavioral episode (see Classification of Behavior Into
Respondent and Operant). Terms and knowledge claims
do not mean anything more than their controlling variables.
They do not convey otherwise unknowable information
about things.
Skinner's radically naturalistic approach to epistemology
has received little attention to date, and we doubt
that few but the most devoted students of behaviorism
are familiar with it. However, some writers have begun
their own extensions of Skinner's epistemology to issues
in the philosophy of science (Hineline, 198O;Lamal, 1983;
Malagodi, 1986; Malagodi & Jackson, 1989; Schnaitter,
1978, 1984; Williams, 1986; ZurifF, 1980, 1985). As
workers continue to take an operant standpoint to delve
into behavioral complexities (e.g., those associated with
the distinction between contingency-shaped and rulegoverned
behavior), it seems likely that we will hear more
of this behavioral-psychological interpretation of what it
means to know.
Eventually we shall be able to include, and perhaps to understand,
our own verbal behavior as scientists. If it turns out that
our final view of verbal behavior invalidates our scientific structure
from the point of view of logic and truth-value, then so
much the worse for logic, which will also have been embraced
by our analysis. (Skinner, 1945a, p. 277)
Conclusion
The purpose of the present research was to examine the
assumptive base of Skinner's behaviorism. This revealed
a coherent set of fundamental assumptions. For those
interested in deepening their understanding of Skinner's
complex and influential approach to psychology, we recommend
that they keep these assumptions available while
they examine authoritative presentations such as those
of Catania (1980), Day (1975, 1983), Michael (1985),
and Reese (1986) and, above all, Skinner's writings themselves
(e.g., 1953, 1957, 1974).