The Behavior Analyst 2000, 23, 11-24 No.

1 (Spring)

Establishing Operations, Cognition, and Emotion

Michael J. Dougher and Lucianne Hackbert
University of New Mexico
In this paper we argue that behavior analysts have tended to neglect the study of important aspects
of complex human behavior, including cognition and emotion. This relative neglect has been costly
in terms of mainstream psychology's perception of the field of behavior analysis and in terms of
our ability to provide a more thorough account of human behavior. Observations and findings from
the clinical context are offered as examples of behavior that are not readily explained by the three-
term contingency, and we argue that an adequate account of these behaviors must include principles
derived from recent behavior-analytic work, in particular a better understanding of the short- and
long-term effects of establishing operations. The concept of the establishing operation and its im-
plications for understanding complex human behavior are discussed.
Key words: establishing operations, setting events, emotion, cognition, three-term contingency

THE RELATIVE NEGLECT OF clinical psychologists and other mental

COGNITION AND EMOTION BY health professionals. As an example, a
BEHAVIOR ANALYSTS psychologist appearing on a recent
It is common in psychology and in television talk show explained the re-
everyday discourse to separate the ac- ported alarming rise in the occurrence
tions of organisms into three catego- of "road rage" (violent confrontations
ries: behavior, cognition, and emotion. among drivers) as resulting from pent-
From this perspective, behavior typi- up anger and longstanding feelings of
cally refers to observable acts, and resentment and interpersonal alien-
these are usually defined by their form ation.
or topography (e.g., Gray, 1999, p. 3). In this dominant, tripartite division
Cognition refers to the activities of the of the human condition, behavior gen-
mind or unobservable mental processes erally takes a back seat to cognition
(e.g., Ellis & Hunt, 1993, p. 2), and and emotion. Inasmuch as it is often
they are usually seen as at least prox- assumed to be the result of more com-
imal causes of behavior. Emotions are plex and more interesting cognitive or
typically defined as bodily or affective emotional processes, behavior typically
states (e.g., Ellis, Ottoway, Varner, serves as the basis from which infer-
Becker, & Moore, 1997) and are often ences are made about those underlying
understood as the feelings (Gray, 1999, cognitive and emotional processes. In
p. 219) associated with our cognition line with the underlying mechanistic,
and behavior. Emotions are sometimes mentalistic, and structuralistic perspec-
given motivational and even explana- tive of mainstream psychology, most
tory status. As Skinner (e.g., 1974, psychological theories that attempt to
chap. 10) has pointed out, there is a explain cognition and emotion offer es-
common tendency to explain behavior sentially metaphorical descriptions of
in terms of emotions or other "inner inferred structures or processes. Once
causes," and this tendency extends to described, these processes are used to
derive predictions of behavior that are
This paper was first presented as part of a then subjected to experimental tests.
symposium at the annual meetings of the As- Quite often, the results of these exper-
sociation for Behavior Analysis in Chicago, iments are in accord with the predic-
May, 1997. tions derived from the hypothesized
Reprint requests should be addressed to Mi- structures and processes, and the model
chael J. Dougher, Department of Psychology,
Logan Hall, University of New Mexico, Albu- or theory that postulates their existence
querque, New Mexico 87131. and function is thus considered to be
11
12 MICHAEL J. DOUGHER & LUCIANNE HACKBERT

supported (see Donahoe & Palmer, service. As a field, behavior analysis

1994, pp. 8-10). already has a substantial public rela-
In line with the underlying contex- tions problem. Like it or not, from the
tualistic and functionalistic perspective outside, behavior analysis is sometimes
of behavior analysis, behavioral theo- seen as more of a cult than a scientific
ries have taken a different approach. community (e.g., Mahoney, 1989), and
Although it is commonly and errone- failing to fully address the phenomena
ously accused of ruling out private that many consider to be the most in-
events such as thinking and feeling on teresting characteristics of the human
scientific grounds, behavior analysis condition only adds to this perception.
instead views these as instances or as- As stated in most introductory texts
pects of behavior and then seeks to dis- and revealed even in casual conversa-
cover their determinants and reciprocal tions with nonbehavioral psycholo-
influence (e.g., Moore, 1980; Skinner, gists, the common view of behavior
1953, 1974). It could be argued, how- analysis is that it restricts its subject
ever, that although behavior analysis matter to relatively simple, publicly
certainly makes room for the study of observable behavior, explains behavior
cognition and emotion, these issues almost exclusively in terms of imme-
have not captivated the field. Although diate reinforcement contingencies, and
there does appear to be increasing in- is generally unconcerned with the rich-
terest among some behavior analysts in ness and complexity of the human con-
topics that fall under the general rubric dition. This is, of course, an inaccurate
of cognition (e.g., rule governance, cat- portrayal of the true behavior-analytic
egory learning, stimulus equivalence, position, and several writers have tried
relational responding, verbal behavior, to correct this misrepresentation. As
choice, and decision making), these early as 1950, Keller and Schoenfeld
topics can hardly be said to dominate presented a behavior-analytic view of
the field, and the work that has been emotional responding, and Skinner
done has had relatively little impact on (e.g., 1953, 1974) directly addressed
mainstream psychology. Furthermore, such topics as thinking, perceiving,
emotion and the relation between emo- knowing, emotion, motivation, the self,
tion and cognition are hardly studied at and personality. Still, it would be dif-
all. Some might argue that this is as it ficult to argue that the field has fol-
should be; Skinner, after all, catego- lowed up with the conceptual or ex-
rized emotions as epiphenomena perimental analyses that would allow
(Skinner, 1971, p. 12), and many be- an adequate understanding of these
havior analysts since have acted as if phenomena. Now, we are not arguing
emotions are unimportant. But even if that behavior analysts ought to adopt
emotions are not considered to be ad- the topics and methods of mainstream
equate explanations of behavior, that psychology. We agree, however, with
does not mean that it is unimportant to Anderson, Hawkins, Freeman, and
understand the determinants of emo- Scotti (2000) that more behavior ana-
tional responding or the role of emo- lysts should attend to a broader range
tional responding in complex behavior. of topics than they are looking at now,
Although we agree that the way that including those complex areas of hu-
cognition and emotions are typically man functioning, cognition and emo-
studied and understood in other fields tion, that have captivated the attention
of psychology is problematic, we dis- of mainstream psychology.
agree with the conclusion that these It is interesting that, except for the
phenomena are unimportant or should discussant, the contributors to this set
be ignored. When behavior analysts ig- of articles are all clinical behavior an-
nore or neglect what other serious be- alysts. That is, they work with verbally
havioral scientists see as important top- competent humans in clinical contexts.
ics of study, they do themselves a dis- One reason that clinical behavior ana-
 ESTABLISHING OPERATIONS 13

lysts might be particularly interested in yond the three-term contingency and

seeing behavior analysis extend its fo- rely to a large extent on concepts that
cus is that clinical work provides fre- have been relatively neglected. In par-
quent opportunities to observe very ticular, we argue that adequate ac-
complex and often highly emotional counts of these observations require in-
human behavior. In addition, these ob- clusion of behavior-analytic concepts
servations are often accompanied by that fall under the general heading of
clients' obviously subjective, but often cognition, emotion, and motivation.
very detailed, reports of their personal More specifically, they require inclu-
histories. With enough observations, sion of the concepts of stimulus equiv-
certain commonalities in response alence or relational responding and es-
styles and in the relations between re- tablishing operations. One article in
sponse styles and reported histories this set, by Wilson and Hayes (2000),
among clients with similar clinical focuses on the role of stimulus equiv-
problems become apparent. Many of alence and relational responding. Ac-
these observations, however, are not cordingly, except for a brief descrip-
readily explainable in terms of the tion of stimulus equivalence later in the
three-term contingency, and there is paper, we focus on the role of estab-
not much in the behavior-analytic lit- lishing operations. After discussing
erature that addresses or helps one to such operations at some length, we
understand the behavior observed in show how they may be involved in
these contexts. For example, almost some clinical phenomena.
anyone who has worked with de-
pressed clients will have observed that ESTABLISHING OPERATIONS
they differentially attend to (come un-
der the control of) aversive or poten- At the risk of being too rudimentary,
tially aversive current and past events, it may be helpful here to state the def-
and that this differential attention cor- inition of establishing operation and
relates with the clients' affective states. some of the history related to the use
The more depressed their affect, the of this and related terms. To our
greater the tendency to attend to and knowledge, Keller and Schoenfeld
report negative events and memories. (1950) first used the term to distinguish
Another observation with depressives motivational effects from reinforce-
is the very restricted range of events ment effects (Chase & Hyten, 1985).
that serve as effective reinforcers and Michael (1982) offered a more func-
the tendency for the size of this range tionally specific definition of the term
to covary with affective state. More in a paper that attempted to distinguish
generally, past experiences, even those between the discriminative and moti-
that are historically remote, often ap- vative functions of stimuli. He provid-
pear to have a sustained impact on cli- ed an expanded and more detailed dis-
ents' behavior, even when this behav- cussion of establishing operations in. a
ior results in aversive consequences. subsequent paper (Michael, 1993). Ac-
Another common observation is that cording to Michael, establishing oper-
simply talking about disturbing or trau- ations are environmental events, oper-
matic experiences such as assault, ations, or stimulus conditions that (a)
abuse, or the death of a spouse can momentarily establish or potentiate the
substantially alleviate the negative im- reinforcing effectiveness of other
pact of those experiences. events and (b) evoke behaviors rele-
We are not contending that there are vant to those reinforcing events. In the
no behavior-analytic explanations for 1993 paper, Michael elaborated upon
these observations. However, we con- the evocative effects of establishing
tend that an adequate behavioral expla- operations and specifically mentioned
nation requires an appeal to behavioral their effects on the evocative functions
principles and processes that go be- of relevant discriminative stimuli.
14 MICHAEL J. DOUGHER & LUCIANNE HACKBERT

Thus, establishing operations act upon have been used to refer to these same
all three elements of the three-term factors. Goldiamond (1983), for ex-
contingency. Food deprivation is an ample, used the term potentiating var-
example of an establishing operation. iables, and Hawkins (1986) used the
As such, it establishes food and asso- term motivating variables, although he
ciated conditioned events as effective referred to their potentiating and de-
reinforcers, evokes behaviors that have potentiating effects on consequent
been previously reinforced with food, stimuli. Before that, Kantor (1959)
and increases the evocative effective- used the term setting factors, which he
ness of discriminative stimuli that have defined as those circumstances that in-
been associated with the availability of fluence which stimulus-response rela-
food reinforcers. tions would occur out of all those cur-
Parenthetically, the effects of estab- rently comprising a person's behavioral
lishing operations on discriminative repertoire built up through past per-
stimuli were implied in Michael's 1982 son-environment interactions. In their
paper, but were not fully articulated un- classic behavioral text on child devel-
til his 1993 paper. From our perspec- opment, Bijou and Baer (1961) used
tive, this is an important addition be- the term setting events, which they de-
cause, as we try to clarify below, it scribed as follows: "In contrast to
suggests a behavior-analytic explana- stimulus events, setting events are
tion of some interesting findings in the more complicated than the simple pres-
literature concerned with the relation ence, absence or change of a stimulus.
between emotion and cognition. To a ... Instead, a setting event is a stimu-
large extent, the cognitive phenomena lus-response interaction, which simply
referred to in this literature are essen- because it has occurred will affect oth-
tially examples of differential stimulus er stimulus-response relationships
control, and both the differential stim- which follow it" (p. 21). Leigland
ulus control and the emotional states (1984) reviewed the various terms that
described in these studies appear to be have been used in reference to the class
the effects of establishing operations. of variables that determine the effec-
In addition to establishing opera- tiveness of contingencies, and suggest-
tions, Michael also identifies and dis- ed that setting events and setting fac-
cusses what he calls abolishing opera- tors may be too general and function-
tions. Abolishing operations have the ally vague to suffice as technical terms.
opposite effects of establishing opera- He argues that, as Michael defines it,
tions. Abolishing operations reduce the establishing operation is more func-
probability of certain behaviors, depo- tionally specific and, therefore, prefer-
tentiate certain events as reinforcers, able as a technical term. We agree and,
and reduce the discriminative control for that reason, have opted to use the
exerted by certain stimuli. Where food term establishing operations rather
deprivation functions as an establishing than setting events or setting factors in
operation, food satiation functions as the present paper. Also, for purposes of
an abolishing operation. It reduces the convenience, we will use establishing
probability of behavior that has been operations in the present paper to col-
reinforced with food, depotentiates lectively refer to both establishing and
food as a reinforcer, and decreases the abolishing operations, unless specifi-
discriminative control exerted by stim- cally referring to the abolishing effects
uli correlated with the availability of of a particular operation.
food reinforcers. Behavior analysts who have written
Although we have focused on the on the topic have mentioned a number
term establishing operations to refer of different operations, procedures, and
collectively to function-altering factors events that might function as establish-
outside the three-term contingency, it ing operations. Kantor (1959), Bijou
should be pointed out that other terms (1976), and Bijou and Baer (1961)
 ESTABLISHING OPERATIONS 15
identified three classes of events con- win, Mayer, & Whitehurst, 1971; Rin-
sisting of (a) physiological conditions cover & Koegel, 1975; Wahler, 1980;
(e.g., conditions of deprivation or sa- Wahler & Fox, 1980) that had specifi-
tiation, illness or health, amount of cally manipulated setting events in an
rest, presence or absence of drugs), (b) attempt to modify behavior. Since then,
durational events (e.g., presence or ab- the number of studies that have specif-
sence of certain events, objects, or per- ically examined establishing operations
sons; instructions or verbal statements; remains comparatively small, despite
production requirements in a work set- reported successes and specific calls
ting; ambient noise or ambient temper- for this kind of research (Gardner, Kar-
ature), and (c) behavioral histories an, & Cole, 1984; Halle & Spradlin,
(e.g., family interactions before arriv- 1993; Homer, Vaughn, Day, & Ard,
ing at work or school, previous expe- 1996). This appears to be an important
rience with existing contingencies, the gap in the behavior-analytic literature.
development of relevant behavioral In their discussion of setting events,
repertoires). Hawkins (1986) lists five Wahler and Fox (1981) echo Kantor's
types of manipulations that could serve (1959) "friendly" criticism that the ex-
as establishing operations, and sug- perimental analysis of behavior has un-
gests that these might be helpful to duly emphasized the relatively simple
consider when developing programs to and temporally proximate conditions
change targeted behaviors. His five cat- that control behavior. They call for a
egories are chemical, health, emotion- conceptual and methodological expan-
al, precurrent stimuli, and learning. Mi- sion of behavior analysis that would
chael (1993) also identified a number include more complex and temporally
of potential establishing operations, in- distant environment-behavior relations
cluding deprivation, aversive stimula- and descriptive as well as experimental
tion, temperature changes, variables re- analyses. In that regard, they consider
lated to sexual reinforcement, a range Bijou and Baer's (1961) definition of
of operations that produce emotions, setting events to be especially impor-
and conditioned establishing opera- tant in that it emphasized the definition
tions. of setting events as stimulus-response
Taken together, these authors have interactions and acknowledged the im-
identified a very broad range of events, portance of factors that occur separate
operations, and conditions that might in space and time from the stimulus-
function as establishing operations. response relations they influence.
Yet, despite the early identification of
the importance of establishing opera- On the Duration of the Effects of
tions and the broad range of events that Establishing Operations
function as such, behavior analysts
have given relatively little conceptual The importance of considering tem-
or empirical attention to the role of es- porally distant factors raises the issue
tablishing operations in understanding of whether the effects of establishing
or modifying human behavior. As operations are momentary (Michael,
Hawkins (1986) and Michael (1982, 1982, 1993). Although the meaning of
1993) point out, the two establishing momentary is not specified, the term
operations that have received the most implies that the effects are brief and
attention are deprivation and aversive dissipate rather quickly. But why must
stimulation (e.g., Keller & Schoenfeld, this be so? What behavioral principles
1950; Skinner, 1938, chap. 9 and 10; mandate that the effects of establishing
1953, chap. 9 and 11; 1957, pp. 28-32, operations cannot be much longer, per-
212-214). As of 1981, Wahler and Fox haps weeks, months, or even years? If
could find only a handful of studies in it is possible for reinforcement to pro-
the applied behavior analysis literature duce changes in behavior that last for
(Fowler & Baer, 1981; Peterson, Mer- years, why is it not also possible for
 16 MICHAEL J. DOUGHER & LUCIANNE HACKBERT

establishing operations to have long- dures employed in such therapeutic in-

lasting effects on the events that func- terventions as systematic desensitiza-
tion as reinforcers (see Hawkins, tion and covert sensitization essentially
1986)? alter the positive and negative rein-
Wahler and Fox (1981) argue that forcing functions of specific stimuli
behavior analysts hold a "conceptual and can also be viewed as examples
bias" about time in their search for of establishing operations with long-
functional relations (p. 332). The em- term effects (Dougher, Crossen, &
phasis has been on brief temporal re- Garland, 1986). Rather than being
lations between stimuli and responses ephemeral, it may be that the effects
or the effects of immediate contingen- of some establishing operations persist
cies. They suggest that it seems rea- until they are supplanted or modified
sonable to put aside this bias in the by other events. Just as the establish-
study of setting events, and add that it ing effects of food deprivation contin-
is conceivable that setting events could ue until the organism eats, the estab-
be functional for hours. We certainly lishing effects of emotion-eliciting
agree, but go on to ask, why restrict it events may persist until they are mod-
to hours? ified by the occurrence of relevant
The assumption that establishing op- abolishing events. However, because
erations have momentary effects may of the paucity of research on estab-
be related to the relative emphasis lishing operations, all that can be said
placed on deprivation and aversive about the duration of their effects at
stimulation compared to the list of oth- this point is conjecture.
er events that might function as estab- In what follows, we attempt to show
lishing operations. The establishing ef- how the concept of the establishing op-
fects of mild or moderate food depri- eration might account for some behav-
vation or shock, for example, probably ioral phenomena that fall under the
are short lived. Once an organism has general heading of cognition and emo-
eaten or the shock has been terminated, tion. In particular, we focus on some
the relevant contingencies are no lon- clinically relevant phenomena, includ-
ger potentiated. But that does not mean ing the commonly reported symptoms
that the effects of all establishing op- of depression and the "cognitive bias-
erations or even severe levels of dep- es" associated with anxiety disorders,
rivation and aversive stimulation are some clinical observations of what ap-
momentary. If we examine the range pear to be temporally distant functional
of establishing operations suggested by relations, and the reported therapeutic
Kantor (1959), Bijou and Baer (1961), effects of verbalizing distressing ex-
Hawkins (1986), and Michael (1982), periences.
it is clear that some of them could con-
ceivably have very long-term effects. Establishing Operations and
For example, events that elicit strong Clinically Relevant Behavior
emotional reactions, such as the death
of loved one, rape, abuse, or the events In a previous paper (Dougher &
that lead us to "fall in love," are ex- Hackbert, 1994), we attempted to out-
amples of establishing operations with line a behavior-analytic account of
long-term effects. So are verbal state- clinical depression. In that paper we
ments or rules such as "Stay away identified three conditions that the de-
from snakes, they can be poisonous," pression literature indicates are com-
"Stay away from behaviorism, it re- mon antecedents of depression. These
jects the study of thoughts and feelings are (a) persistently insufficient levels
and believes people should be pro- of reinforcement (sometimes due to the
grammed like robots," or "Eating un- lack of an effective repertoire), (b) the
dercooked pork will make you sick." loss of a major source of reinforce-
The respondent conditioning proce- ment, and (c) persistent punishment or
 ESTABLISHING OPERATIONS 17

generally high levels of aversive stim- interact, relate, create, recreate, and de-
ulation. We offered a behavior-analytic rive pleasure from these activities and
account of how these antecedent con- their consequences.
ditions could account for the common Depressive contingencies, on the
symptoms of depression. In so doing other hand, are those that evoke and
we identified three conceptually dis- maintain depressed behavior. By de-
tinct potential behavioral effects of pressed behavior, we are referring to
these conditions. the verbal and nonverbal behavior
The first and most obvious effect is characteristic of depressed clients.
that all three may lead to low or de- These include excessive crying, com-
creased rates of appropriate behavior. plaining, blaming, worrying, self-dep-
For example, if repeated attempts to recation, inaccurate rule formulation
develop social relationships, succeed at about existing and future contingen-
work, or win parental approval go cies, inaccurate assessments of one's
largely unreinforced or are frequently ability to satisfy existing and future
met with aversive consequences, the contingencies, social avoidance, drug
frequency of these behaviors is likely or alcohol abuse, pessimistic predic-
to diminish. The second possible effect tions about the future, and sleep, eat-
of these antecedents is respondent elic- ing, and sexual disturbances. It is our
itation. Persistent low rates of rein- contention that the conditions that of-
forcement, extinction, or persistent ten precede depression function as es-
punishment not only reduce the fre- tablishing operations in that they estab-
quency of relevant operants but also lish depressive contingencies and abol-
can act as unconditioned or condi- ish nondepressive contingencies. We
tioned elicitors that produce very already mentioned the behaviors that
strong emotional reactions. Consider, are often evoked by these conditions,
for example, the emotional reactions to so we will turn our attention now to
the death of a loved one, the termina- their effects on behavioral consequenc-
tion of a valued relationship or job, the es and then to their effects on the dis-
inability to obtain or maintain desired criminative control exerted by relevant
friendships, or having to tolerate con- stimuli.
stant criticism or physical abuse. These The antecedent conditions of depres-
types of situations often elicit feelings sion may establish as effective rein-
of anger, frustration, despair, inadequa- forcers expressions of sympathy, com-
cy, and self-contempt. miseration, reassurance, offers of assis-
The third effect of these conditions, tance, or the removal of expectations,
and the one that we believe has been demands, or threats of punishment.
most often overlooked, is that they They may also differentially potentiate
may function as establishing (or abol- the reinforcing effects of food, sleep,
ishing) operations. In our previous pa- isolation, drugs, and alcohol. This may
per, we attempted to illustrate these ef- explain why depressives often overeat,
fects by heuristically dividing the oversleep, prefer to stay by themselves,
world into two very general types of and are prone to drug and alcohol
reinforcement contingencies: depres- abuse. Just as these conditions may es-
sive and nondepressive. We described tablish certain reinforcers, they may
nondepressive contingencies as those also abolish the reinforcing effects of
that are typically established when in- otherwise highly potent behavioral
dividuals are not depressed. They in- consequences. Such activities as social
clude the interpersonal, occupational, interactions, work, and hobbies may
recreational, and social contingencies not only lose their reinforcing effects
that characterize most people's daily but may even become aversive, func-
lives. When there is an adequate rep- tioning as negative reinforcers by their
ertoire and these contingencies are es- offset or as punishers by their onset.
tablished, individuals work, play, love, The loss of reinforcer effectiveness on
 18 MICHAEL J. DOUGHER & LUCIANNE HACKBERT
a wide scale is referred to clinically as that are intended to induce specific
anhedonia, and it may extend to such moods. Statements like "I'm a failure
primary reinforcers as food and sex. at most things I do" and "Even my
This anhedonia interferes with the parents have difficulty loving me" are
treatment of depressives, which often intended to induce a depressed mood,
includes increasing the density of re- whereas "I feel competent, happy, and
inforcement in their lives. But increas- secure" and "Things just go right for
ing the density of reinforcement usu- me" are statements intended to induce
ally requires clients to be motivated an elated mood. The effects of these
enough to emit the behaviors that pro- manipulations are typically measured
duce reinforcing consequences. That is, by mood inventories with acceptable
relevant consequences must be effec- psychometric properties. Even when
tive reinforcers. It is common, how- the demand characteristics of these
ever, to hear depressed clients com- studies are controlled, these procedures
plain that they just do not care about have been reported to induce some-
once potent reinforcers, or they just do times very significant mood changes in
not feel like doing what it takes to ob- otherwise normal subjects. To the ex-
tain them. This wide-scale loss of re- tent that verbally reported mood
inforcer effectiveness in itself entails a changes reflect respondent processes,
low density of reinforcement, which, verbal mood-induction procedures can
of course, is one of the previously be seen as conditioned elicitors. In the
mentioned antecedents of depression. same way, the negative self-statement
Thus, the abolishing function of the an- made by depressives can exacerbate
tecedents of depression may cause a their feelings of sadness and despon-
self-perpetuating cycle or, as it is more dence.
commonly called, a downward spiral Although the establishing functions
of depression. of mood-induction procedures have not
Another factor that may contribute yet been experimentally demonstrated,
to a self-perpetuating cycle in depres- it is reasonable to speculate that the
sion is the effect of the self-directed negative self-statements emitted by de-
verbalizations often emitted by de- pressives would not only affect their
pressed clients. As we mentioned pre- mood but would also have establishing
viously, these verbalizations are often and abolishing functions. In the same
very self-critical, self-deprecating, and way that a dinner companion's detailed
pessimistic. It is common for depressed description of the symptoms produced
clients to make statements like "I'm no by a recent bout with the flu might di-
good," "I'll never be success- minish one's appetite for dinner, a
ful," "Nobody likes to be around me," round of self-criticism about one's so-
"I am unlovable," and "Nothing will cial competence might also diminish
ever change." Although such state- the evocative and reinforcing effects of
ments may be reinforced with sympa- an upcoming social event.
thy and attention, they can also have A third factor that may maintain or
conditioned eliciting and establishing exacerbate depression is the reaction of
functions for the speaker's behavior. other people to the distressed behavior
The mood-induction literature demon- of depressives. Although it may ini-
strates that verbal statements can elicit tially be reinforced by expressions of
emotional reactions. In an attempt to sympathy and concern or the removal
study the relation between emotion and of demands, distressed behavior is per-
cognition, some investigators have em- ceived by others as aversive, and they
ployed verbal mood-induction tech- try to escape or avoid it (Coyne, 1976).
niques (see Bower, 1981, and Ellis & This, of course, is a form of extinction,
Ashbrook, 1989, for reviews of this lit- one of the previously mentioned ante-
erature). These techniques involve hav- cedents of depression. It is not difficult
ing subjects read a series of statements to see how this pattern of reinforce-
 ESTABLISHING OPERATIONS 19

ment for depressed behavior followed ceive, attend to, interpret, and remem-
by extinction can lead to deeper levels ber anxiety-relevant stimuli compared
of depression. to normal ones. An example of cogni-
Having discussed the types of be- tive bias is the difference in reaction
havior evoked by the antecedents of times between anxious and normal
depression and their potential effects subjects on emotional Stroop tasks. On
on relevant behavioral consequences, these tasks, anxiety-related and non-
we turn our attention now to their ef- anxiety-related words are presented in
fects on relevant discriminative stimu- various colors. Subjects are then asked
li. Among other so-called "cognitive to report the color of the words as
deficits" (Beck, 1967; Beck, Rush, quickly as they can. Anxious subjects
Shaw, & Emory, 1979; Ellis, 1990; El- take longer to name the colors of the
lis et al., 1997), a core symptom of de- anxiety-related words than they do to
pression is the tendency to selectively name the colors of the non-anxiety-re-
attend to and remember negative lated words. Normal subjects show no
events. When asked to recall a list of reaction-time differences to the two
words varying in affective content, or sets of words. The assumption is that
the words used in a previously read the anxious subjects unconsciously at-
paragraph, or just to freely recall early tend to the anxiety-related words, and
memories, depressives tend to differ- that interferes with their reaction times
entially remember words and experi- in naming the colors. Memory bias is
ences associated with sad and unpleas- demonstrated by the differences in per-
ant affect (Ellis, 1990). When giving formance between anxious subjects
self-descriptions, depressives overem- and normal subjects on a memory task.
phasize their negative characteristics When given a list of words to study
and deemphasize their positive attri- and later recall, anxious subjects tend
butes. In social situations, they selec- to remember more anxiety-related
tively attend to the negative reactions words than non-anxiety-related words.
of others and seem oblivious to posi- Normal subjects show no differential
tive reactions. Moreover, this selective recall.
attention tends to covary with the in- Cognitive psychologists have postu-
tensity of depressed affect; the more lated a variety of mental processes
depressed they are, the more they focus (e.g., Ellis, 1990; Ellis et al., 1993) to
on the negative (Beck, 1967; Beck et explain these data, but, in behavioral
al., 1979). This tendency to differen- terms, it appears that they may be ex-
tially attend to negative stimuli is also plained by appeal to establishing op-
found in normal subjects who have erations. Most of the cognitive tasks
been exposed to the kind of depressed employed in these studies assess the
mood-induction procedures described stimulus control exerted by the stimuli
above. In fact, mood-induction proce- included in the tasks. In the Stroop test,
dures have been reported to impair for example, there are two dimensions
subjects' performance on a range of of the words that compete for stimulus
cognitive tasks including memory, at- control: the content of the words and
tention, and problem solving (Ellis, their color. Similarly, the word recall
Varner, & Becker, 1993). tests used in these studies assess the
Although the discussion thus far has differential stimulus control exerted by
focused on depression, it is not the the words on the lists the subjects were
only emotion or affective state related asked to recall. Evidently, relevant af-
to cognitive performance. McNally fect-related words exert more stimulus
(1996) reviewed a series of studies that control than non-affect-related words
demonstrate a relation between anxiety for clinical subjects but not for normal
and "cognitive bias." Cognitive bias subjects. It seems reasonable to suggest
refers to the tendency of anxiety-dis- that the mood-induction procedures
ordered clients to differentially per- and other factors that led to the devel-
20 MICHAEL J. DOUGHER & LUCIANNE HACKBERT

opment of the subjects' depression or one serious romantic relationship that

anxiety disorders served as establish- terminated 5 years before when she
ing operations that differentially en- discovered her boyfriend was involved
hanced the stimulus control exerted by in a sexual relationship with another
the affect-related stimuli. Just as stim- woman. Since then, she dated a few
uli that have been correlated with the men, but these interactions failed to de-
availability of food gain control over velop into lasting relationships. She
the behavior of a hungry individual, had heard from others that the men she
events correlated with relevant poten- dated found her aloof and withholding.
tiated consequences might gain stimu- She added that her girlfriends also
lus control over the behavior of de- complained that she was emotionally
pressed and anxious individuals. This distant and reserved. She went on to
is, of course, a post hoc explanation of say that she was afraid of revealing
these results, and direct experimental more of herself to other people for fear
analyses are required before this claim of criticism, rejection, and humiliation.
can be verified. Nevertheless, it seems She also reported that she was becom-
to be a reasonable assertion, and it ing increasingly lonely, depressed, and
points to the importance of establishing pessimistic about ever finding a satis-
operations in understanding interesting fying long-term relationship. When she
findings in the clinical and cognitive was most despondent, she would often
literatures. binge on substantial amounts of junk
food, with a particular preference for
The Potential Long-Term Effects of potato chips.
Establishing Operations The client had a difficult history.
She was given up for adoption by her
There are clinical observations that biological mother and lived in foster
suggest that some experiences can care until she was adopted at the age
have what appear to be long-lasting es- of 4 years by a couple who was told
tablishing and abolishing effects. by their family physician that they
These include traumatic events like would be unable to conceive a child. It
rape, combat, physical assault or abuse, was a surprise to everyone when, 2
the death of a child's parent, or the years later, her adoptive mother be-
death of a parent's child. Rape victims, came pregnant and gave birth to a girl.
for example, frequently suffer from Two years after that, she became preg-
nightmares, anxiety reactions, depres- nant again, and this time had a boy.
sion, generalized fear and distrust of Reportedly, things changed dramatical-
men, and a loss of interest in any kind ly after the birth of the first child, and
of sexual behavior that can last for deteriorated even more after the birth
weeks or even years. But events need of the second. The client's adoptive
not be so extreme or traumatic to have mother grew harsh and demanding, of-
what appear to be long-term establish- ten ridiculing the client and even re-
ing and abolishing effects. A clinical sorting to physical punishment for mi-
example may illustrate this point. nor infractions. The client was made to
The client was a 27-year-old woman feel unwanted. She was given a dispro-
seen by the first author at a university- portionate number of household chores
based outpatient psychology clinic. and received just the bare necessities
She had a pleasant appearance and de- in terms of clothing, toys, and personal
meanor, was slightly overweight, and items. When her adoptive mother dis-
worked as a health professional at a lo- ciplined her, she invariably reminded
cal hospital. She came to treatment the client that her biological mother
seeking help in controlling her weight had given her up for adoption because
and in feeling more comfortable in in- the mother did not want her. She often
terpersonal relationships. She had a ridiculed the client as being dumb, fat,
few casual friendships and reported difficult, and an unwanted burden.
 ESTABLISHING OPERATIONS 21

Obviously, there are many past and or writing about distressing, emotional,
immediate factors that must be consid- and traumatic experiences can reduce
ered in trying to account for the client's
their negative impact and even produce
current behavioral patterns. In addition,
positive outcomes. The emotional ex-
because many of the events reported in periences in these studies included the
therapy occurred in the past, an exper-death of a spouse, physical assault, the
imental analysis of their effects is not
loss of a job, divorce, and physical ill-
possible. It is possible, however, to in-
ness. The positive benefits included
terpret the client's behavior in line with
fewer visits to therapists and physi-
known behavioral principles. In that re-
cians, improved school performance,
gard, it is at least plausible and con-increased immune function, shorter pe-
ceptually justifiable to interpret the riods of time to find reemployment af-
mother's harsh treatment of the client ter a job loss, enhanced social relation-
as, among other things, a set of estab-ships, and a reduction in self-reported
lishing operations with long-term ef- physical and psychological symptoms
fects on the contingencies surrounding of distress. Interestingly, these benefits
the client's interpersonal interactions.
By her report, the client had always were observed even when the subjects
perceived interpersonal situations as received no feedback from therapists
frightening and dangerous. Rather than about their disclosures.
opportunities to share and interact, she We suggest that an adequate behav-
viewed them as situations in which ior-analytic account of these data re-
others might discover her intensely quires an appeal to the concepts of
embarrassing history and assumed per- both stimulus equivalence and the es-
sonal flaws. The client believed that, tablishing operation. By participating
given the opportunity, others would in a stimulus equivalence class (Sid-
come to see her as dumb, incompetent, man, 1994) or relational frame of co-
awkward, and boring, just as her adop- ordination (Hayes, 1991) verbal stimuli
tive mother had. Her history had effec-can acquire the functions of the events
tively determined the behavioral func- to which they refer or stand for (for a
tion of social interactions for her. They
more extended discussion see Dougher
had the potential to become punishing, & Markham, 1996). In this way, words
and events that were correlated with can bring past (or future) events into
the development of intimacy func- the psychological present. Thus, ver-
tioned effectively as discriminative balizing past traumatic experiences
stimuli that evoked avoidance behav- brings them into the present, where
ior. The client interpreted her previous
they can occur in a safe context. This
boyfriend's sexual infidelity as proof can be seen as a verbally based process
that the fears she had about herself and
that can alter both the respondent and
the consequences of social intimacy operant functions of the verbally de-
were well founded. The client's avoid- scribed events. As such, it can be un-
ance of social interactions and intima-derstood as an establishing or abolish-
cy also served as an establishing op- ing operation. As an example, verbal-
eration. It caused her feelings of lone-
izing previous experiences of child
liness, evoked her binge eating, and abuse may extinguish the emotional
eventually led her to therapy.
elicitation functions of those experi-
Verbalizing Distressing Events ences as well as alter the reinforcing
Reduces Their Impact and evoking functions of relevant con-
tingencies. In this way, the reinforcing
Pennebaker (1997) reviewed a series effects of hurting others may be abol-
of studies (see also Pennebaker, 1995, ished (or at least diminished), as would
and Pennebaker, Colder, & Sharp, the evocative effects of other people as
1990) that suggest that simply talking discriminative stimuli for aggression.
22 MICHAEL J. DOUGHER & LUCIANNE HACKBERT

CONCLUSIONS and an examination of more temporally

distant environment-behavior relations
In this paper we attempted to show than is typical in behavior-analytic re-
the importance of the concept of the search. Toward this end, they suggest
establishing operation in providing an the use of correlational or descriptive
adequate behavior-analytic account of research methods when examining
some examples of behavior that fall events that are not directly manipula-
under the general heading of cognition ble. When possible, these would need
and emotion. In so doing, we focused to be followed by experimental analy-
on clinical data and observations be- ses, but at least they would begin to
cause they provide examples of the identify potentially important estab-
kinds of behavioral phenomena that we lishing operations and lead to hypoth-
think have been neglected by behavior eses that could be explored through ex-
analysts but are important to address. perimental analysis.
In our discussion of establishing oper- If we think that establishing opera-
ations, we noted the discrepancy be- tions are important in understanding
tween the considerable range of events, complex human behavior, and we ar-
situations, and processes that have gue that they are, then it is clear that
been mentioned as potential establish- we need to develop methods to study
ing operations and the dearth of con- them in both laboratory and real-world
ceptual and empirical papers that have settings. Clearly this is a challenge, but
addressed the role of establishing op- the potential benefit seems well worth
erations in the analysis and modifica- the effort. It is just this kind of research
tion of human behavior. We also ques- that could demonstrate to nonbehavior-
tioned the assertion that the effects of al psychologists the relevance and even
establishing operations are momentary, the advantage of behavior-analytic ap-
and suggested that their effects could proaches to important human behavior.
instead be quite long lasting. In that re-
gard, we offered a conceptual account REFERENCES
of how some potential establishing op-
erations might have long-term effects Anderson, C. M., Hawkins, R. P., Freeman, K.
on relevant contingencies of reinforce- A., & Scotti, J. R. (2000). Private events: Do
ment. they belong in a science of human behavior?
Obviously, this conceptual analysis The Behavior Analyst, 23, 1-10.
Beck, A. T. (1967). Depression: Causes and
is much easier to do than actually con- treatment. Philadelphia: University of Penn-
ducting the necessary empirical analy- sylvania Press.
ses to demonstrate the functional rela- Beck, A. T., Rush, A. J., Shaw, B. F, & Emory,
tions that have been suggested here. G. (1979). Cognitive therapy of depression.
New York: Guilford.
From a research perspective, it is very Bijou, S. W. (1976). Child development: Vol. 3.
difficult to separate the effects of se- Basic stage of early childhood. Englewood
lected historical events and operations Cliffs, NJ: Prentice Hall.
from the effects of intervening and im- Bijou, S. W., & Baer, D. M. (1961). Child de-
mediate variables. This is especially velopment: Vol. 1. A systematic and empirical
theory. Englewood Cliffs, NJ: Prentice Hall.
true in research with verbally compe- Bower, G. H. (1981). Mood and memory.
tent, normal humans in which the op- American Psychologist, 36, 129-149.
portunity to conduct experimental Chase, P, N., & Hyten, C. (1985). A historical
analyses in real-world contexts is quite and pedagogic note on establishing opera-
tions. The Behavior Analyst, 8, 121-122.
restricted. This may very well be why Coyne, J. C. (1976). Depression and the re-
there is so little research on establish- sponse of others. Journal of Abnormal Psy-
ing operations. As Wahler and Fox chology, 85, 186-193.
(1981) suggest, the conceptual and Donahoe, J. W, & Palmer, D. C. (1994). Learni-
functional analyses of setting events ing antd complex behavior. Boston: Allyn &
Bacon.
(establishing operations) may require Dougher, M. J., Crossen, J. R., & Garland, R.
more molar or global units of analysis (1986). An experimental test of Cautela's op-
 ESTABLISHING OPERATIONS 23
erant explanation of covert sensitization. Be- Conceptual foundations of behavioral assess-
havioral Psychotherapy, 14, 226-248. ment (pp. 331-385). New York: Guilford.
Dougher, M. J., & Hackbert, L. (1994). A be- Hayes, S. C. (1991). A relational control theory
havior analytic account of depression and a of stimulus equivalence. In L. J. Hayes & P
case report using acceptance-based proce- N. Chase (Eds.), Dialogues on verbal behav-
dures. The Behavior Analyst, 17, 321-334. ior (pp. 19-40). Reno, NV: Context Press.
Dougher, M. J., & Markham, M. R. (1996). Homer, R. H., Vaughn, B. J., Day, H. M., & Ard,
Stimulus classes and the untrained acquisition W. R. (1996). The relationship between set-
of stimulus function. In T. R. Zentall & P. M. ting events and problem behavior. In L. K.
Smeets (Eds.), Stimulus class formation in hu- Koegel, R. L. Koegel, & G. Dunlap (Eds.),
mans and animals (pp. 71-90). Amsterdam: Community, school, family, and social inclu-
North Holland. sion through positive behavioral support (pp.
Ellis, H. C. (1990). Depressive deficits in mem- 381-402). Baltimore: Paul Brookes.
ory: Processing initiative and resource allo- Kantor, J. R. (1959). Interbehavioral psychol-
cation. Journal of Experimental Psychology: ogy. Granville, OH: Principia Press.
General, 119, 392-395. Keller, F S., & Schoenfeld, W. N. (1950). Prin-
Ellis, H. C., & Ashbrook, P. W (1989). The ciples of psychology. New York: Appleton-
"state" of mood and memory research: A se- Century-Crofts.
lective review. Journal of Social Behavior and Leigland, S. (1984). On "setting events" and
Personality, 4, 1-21. related concepts. The Behavior Analyst, 7, 41-
Ellis, H. C., & Hunt, R. R. (1993). Fundamen- 46.
tals of cognitive psychology (5th ed.). Madi- Mahoney, M. J. (1989). Scientific psychology
son, WI: Brown & Benchmark. and radical behaviorism. American Psycholo-
Ellis, H. C., Ottoway, S. A., Vamer, L. J., Beck- gist, 44, 1372-1377.
er, A. S., & Moore, B. A. (1997). Emotion, McNally, R. M. (1996). Cognitive bias in the
motivation, and text comprehension: The de- anxiety disorders. Nebraska Symposium on
tection of contradictions in passages. Journal Motivation, 43, 211-250.
of Experimental Psychology: General, 126, Michael, J. (1982). Distinguishing between dis-
131-146. criminative and motivative functions of stim-
Ellis, H. C., Varner, L. J., & Becker, A. S. uli. Journal of the Experimental Analysis of
(1993). Cognition and emotion: Theories, im- Behavior, 37, 149-155.
plications, and educational applications. In L. Michael, J. (1993). Establishing operations. The
A. Penner, G. M. Batsche, H. M. Knoff, & D. Behavior Analyst, 16, 191-206.
L. Nelson (Eds.), The challenge in mathemat- Moore, J. (1980). On behaviorism and private
ics and science education: Psychology's re- events. The Psychological Record, 30, 459-
sponse (pp. 83-111). Washington, DC: Amer- 475.
ican Psychological Association. Pennebaker, J. (1995). Emotion, disclosure and
Fowler, S. A., & Baer, D. M. (1981). "Do I health. Washington, DC: American Psycho-
have to be good all day?" The timing of de- logical Association.
layed reinforcement as a factor in generaliza- Pennebaker, J. (1997). Writing about emotional
tion. Journal of Applied Behavior Analysis, experiences as a therapeutic process. Psycho-
14, 13-24. logical Science, 8, 162-166.
Gardner, W. I., Karan, 0. C., & Cole, C. L. Pennebaker, J., Colder, M., & Sharp, L. K.
(1984). Assessment of setting events influ- (1990). Accelerating the coping process.
encing functional capacities of mentally re- Journal of Personality and Social Psychology,
tarded adults with behavioral difficulties. In
A. S. Halpem & M. J. Fuher (Eds.), Func- 58, 528-537.
tional assessment in rehabilitation (pp. 171- Peterson, R. F., Merwin, M. R., Mayer, T. J., &
185). Baltimore: Paul H. Brookes. Whitehurst, G. J. (1971). Generalized imita-
Goldiamond, I. (Chair). (1983, May). The in- tion: The effects of experimenter absence, dif-
terdependence of fornal, basic, and applied ferential reinforcement, and stimulus com-
behavior analysis: Can the bell toll for just plexity. Journal of Experimental Child Psy-
one? Symposium conducted at the annual chology, 12, 114-128.
meeting of the Association for Behavior Anal- Rincover, A., & Koegel, R. L. (1975). Setting
ysis, Milwaukee, WI. generality and stimulus control in autistic chil-
Gray, P (1999). Psychology (3rd ed.). New dren. Journal ofApplied Behavior Analysis, 8,
York: Worth Publishers. 235-247.
Halle, J. W, & Spradlin, J. E. (1993). Identi- Sidman, M. (1994). Equivalence relations and
fying stimulus control of challenging behav- behavior: A research story. Boston: Authors
ior: Extending the analysis. In J. Reichle & D. Cooperative.
Wacker (Eds.), Communicative alternatives to Skinner, B. F (1938). The behavior of organ-
challenging behavior: Integrating functional isms. New York: Appleton-Century-Crofts.
assessment and intervention strategies (pp. Skinner, B. F (1953). Science and human be-
83-109). Baltimore: Paul Brookes. havior. New York: The Free Press/Macmillan.
Hawkins, R. P. (1986). Selection of target be- Skinner, B. F (1957). Verbal behavior. New
haviors. In R. 0. Nelson & S. C. Hayes (Eds.), York: Appleton-Century-Crofts.
24 MICHAEL J. DOUGHER & LUCIANNE HACKBERT
Skinner, B. F (1971). Beyond freedom and dig- Wahler, R. G., & Fox, J. J. (1981). Setting
nity. New York: Knopf. events in applied behavior analysis: Toward a
Skinner, B. F (1974). About behaviorism. New conceptual and methodological expansion.
York: Knopf. Journal of Applied Behavior Analysis, 14,
Wahler, R. G. (1980). The insular mother: Her 327-338.
problems in parent-child treatment. Journal of Wilson, K. G., & Hayes, S. C. (2000). Why it
Applied Behavior Analysis, 13, 207-219. is crucial to understand thinking and feeling:
Wahler, R. G., & Fox, J. J. (1980). Solitary toy An analysis and application to drug abuse.
play and time-out: A family treatment package The Behavior Analyst, 23, 25-43.
for aggressive and oppositional children. Jour-
nal of Applied Behavior Analysis, 13, 23-29.