JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR 1982cl, 37, 149-155 NUMBER I (JANUARY)

DISTINGUISHING BETWEEN DISCRIMINATIVE AND

MOTIVATIONAL FUNCTIONS OF STIMULI
JACK MICHAEL
WESTERN MICHIGAN UNIVERSITY

A discriminative stimulus is a stimulus condition which, (1) given the momentary effective-
ness of some particular type of reinforcement (2) increases the frequency of a particular type
of response (3) because that stimulus condition has been correlated with an increase in the
frequency with which that type of response has been followed by that type of reinforcement.
Operations such as deprivation have two different effects on behavior. One is to increase the
effectiveness of some object or event as reinforcement, and the other is to evoke the behavior
that has in the past been followed by that object or event. "Establishing operation" is sug-
gested as a general term for operations having these two effects. A number of situations in-
volve what is generally assuined to be a discriminative stimulus relation, but with the third
defining characteristic of the discriminative stimulus absent. Here the stimulus change
functions more like an establishing operation than a discriminative stimulus, and the new
term, "establishing stimulus," is suggested. There are three other possible approaches to this
terminological problem, but none are entirely satisfactory.
Key words: stimulus control, establishing operation, establishing stimulus, deprivation, re-
inforcement

BACKGROUND CONCEPTS ing with this issue it is necessary to refine the

concept of the discriminative stimulus or SD
Technical Terminology and to review the way deprivation affects be-
Skinner (1957, pg. 430) described a three- havior.
stage process in the development of scientific
verbal behavior: first, more effective forms of The Discriminative Stimulus
verbal behavior are discovered; these are then It seems in keeping with current usage to
explicitly adopted and encouraged by the rel- describe the presentation of a discriminative
evant technical and scientific community; stimulus, or the change from SA to SD, in terms
finally, these verbal practices are themselves of three defining features. It is a stimulus
critically examined and altered to overcome change which, (1) given the momentary effec-
what seem to be inadequacies or limitations. tiveness of some particular type of reinforce-
This paper fits into the third state of this ment1 (2) increases the frequency of a particu-
process. Our way of talking about operant lar type of response (3) because that stimulus
stimulus control seems to include but fails to change has been correlated with an increase in
distinguish between two quite different forms the frequency with which that type of response
of control. We might improve our verbal prac- has been followed by that type of reinforce-
tices by adopting a new technical term for one ment. (Frequency of reinforcement is the most
of these forms of control or at least by explic- common variable used to develop the discrimi-
itly recognizing the problem. But before deal- native stimulus relation. However, this relation
1It would be more in keeping with current practice
The analysis in this paper developed as a result of to make frequent use of the term "reinforcer," rather
discussions with a number of colleagues and students than "type of reinforcement" or "form of reinforce-
but especially G. Dennehy, C. Cherpas, B. Fulton, B. ment," etc. However, I have argued in another paper
Hesse, M. Minervini, L. Parrott, M. Peterson, M. Sund- (Michael, 1975) that "reinforcer" implies a static event,
berg, P. Whitley, and K. Wright, who are also respon- which cannot have the behavioral implications that "re-
sible for many improvements in the present manu- inforcer" appears to have. "Reinforcement" properly
script. Reprints may be obtained from Jack Michael, implies stimulus change, and only change can function
Department of Psychology, Western Michigan Univer- as a consequence for behavior. Therefore throughout
sity, Kalamazoo, Michigan 49008. this paper "reinforcer" is deliberately avoided.
149
 19383711, 1982, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1901/jeab.1982.37-149 by Capes, Wiley Online Library on [14/04/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
150 JACK MICHAEL

can also be developed on the basis of rein- where they have been more frequently fol-
forcement quantity or quality, the delay to re- lowed by water reinforcement (the SD effect).
inforcement, the response requirement or effort, The distinction between these two ways to
and other variables.) The first feature is some- evoke operant behavior is the basis for the
times taken for granted but for the present suggestion of a new term which is the main
purposes it is better to be explicit: SDs do not point of the present paper.
generally alter response frequency when the
organism is satiated with respect to the type The Need for a More General Term:
of reinforcement relevant to that SD. The sec- the "Establishing Operation"
ond feature is important in distinguishing be- The term "deprivation" has been generally
havioral from cognitive accounts of stimulus used for relations such as the one discussed
control, where the stimulus supposedly "sig- above but does not adequately characterize
nals" the availability of reinforcement, without many of them. Salt ingestion, perspiration,
any direct implication for any particular type and blood loss have similar effects but cannot
of behavior. Whereas the first two features de- be accurately referred to as water deprivation.
scribe the controlling relation once it has been Aversive stimulation also establishes its ab-
developed, the third identifies the relevant his- sence as reinforcement and evokes the behav-
tory and thus makes possible a distinction be- ior that has in the past removed it. Likewise
tween the operant discriminative stimulus and temperature changes away from the organism's
the unconditioned and conditioned stimuli of normal thermal condition increase the effec-
the respondent relation. There are a number tiveness of changes in the opposite direction
of situations involving what is generally taken as reinforcement and also evoke behavior that
to be an SD because the relation seems so obvi- has resulted in such changes.
ously operant rather than respondent, but Skinner explicitly identifies deprivation-sati-
where the third defining feature is clearly ab- ation operations and aversive stimulation as
sent. An attempt will be made to show that motivational variables (1957, pp. 31-33 and
in some of these situations the stimulus change also 212), and with the term "predisposition"
is functioning more like a motivational oper- (1953, p. 162) includes the so-called emotional
ation such as deprivation or aversive stimu- operations in this collection. Again, the two
lation. different effects of such operations seem clear.
For example, to be angry is, in part, to have
The Behavioral Effects of Deprivation one's behavior susceptible to reinforcement by
How does deprivation, of water for example, signs of discomfort on the part of the person
affect behavior? It is necessary to distinguish one is "angry at," and also to be engaging in
two quite different effects which cannot be behavior that has produced such effects. Like-
easily derived from one another. One is an wise, "fear," from an operant point of view at
increase in the effectiveness of water as rein- least, seems to consist of an increased capacity
forcement for any new behavior which should for one's responses to be reinforced by the re-
happen to be followed by access to water. The moval of certain stimuli plus the high fre-
other is an increase in the frequency of all quency of behavior that has accomplished such
behavior that has been reinforced with water removal.
and in this respect is like the evocative effect A general term is needed for operations hav-
of an SD. Operant behavior can thus be in- ing these two effects on behavior. There is, of
creased in frequency (evoked) in two different course, the traditional "motive" and "drive,"
ways. Consider, for example, an organism that but these terms have a number of disadvan-
is at least somewhat water deprived and for tages, not the least of which is the strong im-
which some class of responses has a history of plication of a determining inner state. I have
water reinforcement. Assume further that the found "establishing operation" appropriate in
current stimulus conditions have been associ- its commitment to the environment, and by
ated with a low, but nonzero, frequency of abbreviating it to EO one may achieve the
water reinforcement for those responses. Such convenience of a small word without losing
responses can be made momentarily more fre- the implications of the longer term. An estab-
quent (1) by further depriving the organism lishing operation, then, is any change in the
of water, or (2) by changing to a situation environment which alters the effectiveness of
 19383711, 1982, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1901/jeab.1982.37-149 by Capes, Wiley Online Library on [14/04/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DISCRIMINATIVE AND MOTIVATIONAL FUNCTIONS OF STIMULI 151
some object or event as reinforcement and si- but in the absence of the SD it was food sati-
multaneously alters the momentary frequiency ated, and responding was not followed by
of the behavior that has been followed by that food. Such training would be considered of
reinforcement. (Sttudents of J. R. Kantor may dubious value in developing stimulus control
see some similarity to his term "setting factor," because during the absence of the SD, the sati-
but I believe that his term includes operations ated organism's failure to receive food after
that have a broader or less specific effect on the relevant response (if that response were
behavior than the establishing operation as to occtur) could not easily be considered "un-
defined above.) There is still a problem with reinforced" responding, since at that time food
this usage in that "establishing" implies only would not be functioning as reinforcement.
"increasing," but changes obviously occtur in The type of differential reinforcement rele-
both directions. "Deprivation" has the same vant to stimuilus control ordinarily implies that
limitation and is usually accompanied by its in the presence of the stimulus, the organism
opposite "satiation." It does not seem useful receives the reinforcement with greater fre-
at this time to introduce the term "abolishing" quency than it receives the reinforcement in
to serve a similar function, so perhaps in the the absence of the stimulus. If in the absence of
present context "establishing" should be taken the stimtulus, the critical event no longer func-
to be short for "establishing or abolishing." tions as reinforcement, then/ receiving it at a
The value of a general term is more than lower frequency is not behaviorally equivalent
just terminological convenience, however. The to a "lower frequency of reinforcement." This
absence of such a term may have been respon- supplement to the definition of the SD relation
sible for some tendency to disregard such ef- is not usually mentioned simply because the
fects or to subsume them under other head- SD-SA concepts were developed in a laboratory
ings. For example, it is common to describe setting with food and water as reinforcement
the basic operant procedure as a three-term and with the SD and SA alternating during the
relation involving stimulus, response, and con- same session. If food was reinforcing during
sequence. Yet it is clear that such a relation the presence of the SD, it would generally be
is not in effect unless the relevant establishing equally reinforcing during its absence. With
operation is at an appropriate level. A stim- establishing operations that affect behavior
ulus that is correlated with increased frequency more quickly, however, this aspect of the defi-
of water reinforcement for some class of re- nition becomes more critical. Aversive stimu-
sponses will not evoke those responses if water lation is just such an establishing operation.
is not currently effective as reinforcement. Consider a typical shock-escape procedure. The
Furthermore, aversive stimuli and other organism is in a situation where the shock can
events like temperature changes that quickly be administered until some response, say a
evoke behavior may appear to be discrimina- lever press, occurs. This escape response re-
tive stimuli but should not be so considered, moves the shock for a period, then the shock
although the argument is somewhat complex. comes on again, and so on. With a well-trained
As mentioned earlier, in order for a stimulus organism the shock onset evokes an immediate
to be considered a discriminative stimulus the lever pressing response, and since the relation
differential frequency of responding in its pres- is obviously an operant one it might seem rea-
ence as compared with its absence must be due sonable to refer to the shock as an SD for the
to a history of differential reinforcement in its lever press. For the shock to be an SD, it must
presence as compared with its absence. Al- have been a stimulus in the presence of which
though it is not usually mentioned, there is the animal received more frequent reinforce-
in this requirement the further implication ment-in this case shock termination-than it
that the event or object which is functioning received in its absence. But in the absence of
as reinforcement must have been equally effec- the shock, failure to receive shock termination
tive as reinforcement in the absence as in the for the lever press is not properly considered a
presence of the stimulus. It would not, for lower frequency of reinforcement. Unless the
example, be considered appropriate discrimi- shock is on, shock termination is not behav-
nation training if during the presence of the iorally functional as a form of reinforcement,
SD, the organism was food deprived and re- and the fact that the lever does not produce
ceived food as reinforcement for responding, this effect is irrelevant. The shock, in this sit-
 19383711, 1982, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1901/jeab.1982.37-149 by Capes, Wiley Online Library on [14/04/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
152 JACK MICHAEL
uation, is functioning more like an establish- like to suggest the term "establishing stimulus"
ing operation, such as food deprivation, than and SE for this relation.
like an SD. It evokes the escape response be-
cause it changes what functions as reinforce- General Conditions for the
ment rather than because it is correlated with Establishing Stimulus
a higher frequency of reinforcement. It would The circumstances for an establishing stim-
be quite possible, of course, to contrive a ulus (Figure 1) involve a stimulus change, Si,
proper discriminative stimulus in the escape which functions as a discriminative stimulus
situation. Let the escape response terminate for a response, R1, but under circumstances
shock only when a tone is sounding; the shock where that response cannot be executed or
remains on irrespective of the animal's behav- cannot be reinforced until another stimulus
ior when the tone is not sounding. Lever press- change, So, takes place. This second stimulus
ing is clearly unreinforced when the tone is change, then, becomes effective as conditioned
off, and the tone is thus clearly an SD for lever reinforcement, and the behavior that has in
pressing. the past achieved this second stimulus change,
In summary, we could improve our verbal R9, is evoked. S1, then, is an SD for R1, but
behavior about behavior if we could identify an SE for R2.
all environmental operations which alter the
effectiveness of events as reinforcement with A Human Example
the same term, and especially if this term has Suppose that an electrician is prying a face
also been explicitly linked to the evocative ef- plate off a piece of equipment which must be
fects of such operations. "Establishing opera- removed from the wall. The electrician's assis-
tion" might very well accomplish these pur- tant is nearby with the tool box. The removal
poses. of the face plate reveals that the equipment is
fastened to the wall with a slotted screw. We
THE ESTABLISHING STIMULUS can consider the slotted screw under the pres-
OR SE ent circumstances to be a discriminative stim-
ulus (Si) for the use of a screw driver in remov-
Establishing Conditioned Reinforcement ing the screw (R1). (The reinforcement of this
Most of the establishing operations discussed behavior is, of course, related to the electri-
so far have been the kind that alter the effec- cian's job. When the wall fixture is removed
tiveness of stimulus changes that can be classi- and a new one applied, payment for the job
fied as unconditioned reinforcement. Stimulus may become available, etc.) But removing the
changes identified as conditioned reinforce- screw is not possible without an appropriate
ment are also established as such by various
operations. The most obvious are the same
operations that establish the effectiveness of
BUT Rl CANNOT OCCUR
S"CCESSFULLY WITHOUT S2
the relevant unconditioned reinforcement. A EVOKES R1
light correlated with food becomes effective '
RI
conditioned reinforcement as a function of
food deprivation. Information about the loca-
(sD)
tion of a restaurant becomes reinforcing when
food becomes reinforcing. There is, however, a Slt
common situation in which a stimulus change ESTABLISHES AS Sr S2
establishes another stimulus change as con-
ditioned reinforcement without altering the
(sE)
effectiveness of the relevant unconditioned EVOKES
reinforcement. If the behavior which has >R2
previously obtained such conditioned rein- Fig. 1. General condition for an establishing stimu-
forcement now becomes strong we have an lus. S1, functioning as an SD evokes R, but this response
evocative relation like that produced by an es- cannot occur or cannot be reinforced without the pres-
ence of S. Thus S, also functions as an SE, establishing
tablishing operation but where the effect de- S2 as a form of conditioned reinforcement and at the
pends upon an organism's individual history same time evoking R& which has previously produced
rather than the history of the species. I would S2.
 19383711, 1982, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1901/jeab.1982.37-149 by Capes, Wiley Online Library on [14/04/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DISCRIMINATIVE AND MOTIVATIONAL FUNCTIONS OF STIMULI 153
screw driver (S2). So the electrician turns to not needed would result in the work area being
the assistant and says "screw driver" (R2). It is cluttered with unneeded tools (assume a small
the occurrence of this response which illus- or crowded work area, etc.). This second fea-
trates a new type of evocative effect. ture would seem usually to involve some form
It is reasonable to consider the offered screw of punishment for R2 that prevents it until Si
driver as the reinforcement of the request, and has made it necessary.
the history of such reinforcement to be the Now, for the animal analogue consider a
basis for the occurrence of this request under food-deprived monkey in a chamber with a
the present circumstances. It might also seem chain hanging from the ceiling and a retract-
reasonable to consider the sight of the slotted able lever. Pulling the chain moves the lever
screw as an SD for this response, but here we into the chamber. Pressing the lever has no
should be cautious. If it is proper to restrict effect unless a light on the wall is on, in which
the definition of the SD to a stimulus in the case a lever press dispenses a food pellet. To
presence of which the relevant behavior has prevent the chain pull from functioning as a
been more frequently reinforced, then the standard preparatory component of the behav-
present example does not qualify. A slotted ioral sequence, we could require that the chain
screw is not a stimulus that is correlated with be held in a pulled condition, or we could ar-
an increased frequency of obtaining screw range that each chain pull makes the lever
drivers. Electricians' assistants generally pro- available for only a limited period, say five
vide requested tools irrespective of the use to seconds. In either case we would expect a well-
which they will be put. The presence and at- trained monkey ultimately to display the fol-
tention of the assistant are SDS correlated with lowing repertoire: while the wall light is off
successful asking, but not the slotted screw. (before the' electrician has seen the slotted
However, it is an SD for unscrewing responses screw), the chain pull does not occur (a screw
-a stimulus in the presence of which unscrew- driver is not requested), even though it would
ing responses (with the proper tool) are corre- produce the lever (even though the assistant
lated with more frequent screw removal-but would provide one). When the light comes
not for asking for screw drivers. The evocative on (when the slotted screw is observed), the
effect of the slotted screw on asking behavior monkey pulls the chain (the electrician re-
is more like the evocative effect of an establish- quests the screw driver) and then presses the
ing operation than that of an SD, except for its lever (and then unscrews the screw) and eats
dependence on the organism's individual his- the food pellet that is delivered (and removes
tory. The slotted screw is better considered an the piece of equipment, finishes the job, etc.).
establishing stimulus for asking, not a discrim- Returning to Figure 1, S, is the onset of the
inative stimulus. wall light, which evokes lever pressing (R1).
Lever pressing, however, cannot occur without
An Animal Analogue the lever (S2), and thus the availability of the
It is not difficult to describe an SE situation lever becomes an effective form of reinforce-
in the context of an animal experiment, but ment once the wall light is on. The chain pull
first it is necessary to describe two secondary is R9, evoked by the light functioning as an
features of the human situation giving rise to SE rather than as an SD because the light is not
this concept. First, it must be possible to pro- correlated with more frequent lever availabil-
duce the second stimulus change (S2) at any ity (the reinforcement of the chain pull) but
time and not just after the first stimulus change rather with greater effectiveness of the lever as
(S1). Otherwise we are dealing with simple a form of reinforcement.
chaining. Furthermore, the second stimulus
change should not be one which once achieved ALTERNATIVE SOLUTIONS
remains effective indefinitely with no further
effect or cost to the organism, or it will become Larger Units of Behavior
a standard form of preparatory behavior. In It may seem reasonable to consider the re-
the electrician's situation, if there were some sponse evoked by the SE to be simply an ele-
tool that was used on a high proportion of ment in a chain of responses evoked by an SD.
activities it would be kept always available. Thus, with the electrician, asking for the screw-
On the other hand, to ask for a tool when it is driver might be interpreted as a part of a
 19383711, 1982, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1901/jeab.1982.37-149 by Capes, Wiley Online Library on [14/04/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
154 JACK MICHAEL

larger unit of behavior evoked by the slotted change can only be produced when the light
screw and reinforced by successful removal of is on. This involves adding a static component
the screw. From this point of view, the SE con- to our reinforcing stimulus change, which may
cept might seem unnecessary. That is, the be theoretically sound, but is certainly not the
slotted screw could be considered a discrimi- usual way we talk about reinforcement. In the
native stimulus for the larger unit, which is human example this would mean that the
more likely to achieve the removal of the screw screw driver does not reinforce the request, but
in the presence of a slotted screw than in the rather the change from looking at a slotted
presence of a wing nut, a hex nut, etc. This screw without a screw driver in hand to look-
analysis is not too plausible, however, since it ing at one with a screw driver in hand. As with
would imply that requests are not sensitive to conditional conditioned reinforcement this is
their own immediate consequences but only to not an issue regarding the facts of behavior,
the more remote events related to the use of but rather our verbal behavior concerning
the requested item. But even if this were true, these facts. Will we be more effective in intel-
we would still have to account for the forma- lectual and practical ways by introducing a
tion or acquisition of the large unit of behav- new stimulus function and retaining a simple
ior. This has usually involved reference to re- form of verbal behavior about reinforcement,
peated occurrence of chains of smaller units or will we be better off retaining the SD inter-
(for example, Skinner, 1938, 52ff and 102ff; pretation for both types of evocation but com-
Keller 8c Schoenfeld, 1950, Chapter 7; Millen- plicating our interpretation of reinforcement
son, 1967, Chapter 12), and the initial element for one of them? I clearly favor the former.
of this particular chain would require the SE
concept. That is, we might be able to do with-
out the SE in the analysis of the current func- SUMMARY
tion of a large unit but would need it to ac- In everyday language we can and often do
count for the first element of the chain of distinguish between changing people's behav-
smaller units out of which the large unit was ior by changing what they want and changing
formed. their behavior by changing their chances of
getting something that they already want. Our
Conditional Conditioned Reinforcement technical terminology also makes such a dis-
The notion that a form of conditioned rein- tinction, but only in the case of establishing
forcement may be conditional upon the pres- operations such as deprivation and those kinds
ence of another stimulus condition is quite of reinforcing events called "unconditioned."
reasonable and requires no new terminology. Much more common are those stimulus
This could be referred to as conditional con- changes which alter the reinforcing effective-
ditioned reinforcement, and the SE of the pre- ness of events ordinarily referred to as condi-
vious sections can be seen to be the type of tioned reinforcement, and which evoke the
stimulus upon which such conditioned rein- behavior that has previously produced this re-
forcement is conditional. This general ap- inforcement. We do not have a convenient
proachi, however, fails to implicate the evoca- way of referring to such stimulus changes, and
tive effect which is the main topic of the because of this they may be subsumed under
present paper and thus seems less satisfactory the heading of discriminative stimuli. I have
than the new terminology. suggested the term "establishing stimulus" for
such events, thus linking them with establish-
Retaining the SD by Complicating ing operations such as deprivation, and I hope,
the Reinforcement suggesting the relation to the individual's his-
If we consider the chain pull to be rein- tory by the replacement of "operation" with
forced, not by the lever insertion into the "stimulus."
chamber but by the more complex stimulus
change from light on with lever out to light on
with lever in, we may be able to retain the no- REFERENCES
tion of the light -onset as an SD for the chain Keller, F. S., & Schoenfeld, W. N. Principles of psychol-
pull, because the more complex stimulus ogy. New York: Appleton-Century-Crofts, 1950.
 19383711, 1982, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1901/jeab.1982.37-149 by Capes, Wiley Online Library on [14/04/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DISCRIMINATIVE AND MOTIVATIONAL FUNCTIONS OF STIMULI 155
Michael, J. L. Positive and negative reinforcement, a Skinner, B. F. Science and human behavior. New
distinction that is no longer necessary; or a better York: Macmillan, 1953.
way to talk about bad things. Behaviorism, 1975, 3, Skinner, B. F. Verbal behavior. New York: Appleton-
33-44. Century-Crofts, 1957.
Millenson, J. R. Principles of behavioral analysis. New
York: Macmillan, 1967.
Skinner, B. F. The behavior of organisms. New York: Received September 22, 1980
Appleton-Century-Crofts, 1938. Final acceptance August 31, 1981