The Analysis of Verbal Behavior 1988, 6, 3-9 Establishing Operations and the Mand Jack Michael Western Michigan University In Verbal Behavior Skinner identifies a small number of elementary verbal relations, one of which is the mand. Because its introduction is at first in terms of unlearned motivative variables, and because the mands relation to prior controlling events is quite complex, its general significance has probably been underestimated. An extensive treatment of establishing operations, including the warning and the blocked-response conditioned establishing operations is provided, followed by a description of the mand in terms of such operations. The importance of the mand for language training programs is suggested, as well as the reasons why it is typically neglected in such programs. In Verbal Behavior Skinner defines the type of pairing that causes stimuli to becomemand as ". . . a verbal operant in which the conditioned reinforcers or conditioned pun-response is reinforced by a characteristic con- ishers). Thus water deprivation (1) momen-sequence and is therefore under the func- tarily increases the reinforcing effectivenesstional control of relevant conditions of of any water that should be encountered;deprivation or aversive stimulation.... and and (2) momentarily increases the frequencyin contrast with other types of verbal of (evokes is a convenient synonym) the var-operants ... the response has no specified ious kinds of responses that have beenrelation to a prior stimulus" (1957, pp. 35-36). reinforced with water. Painful stimulation (1)In other words, with the mand what is said, momentarily increases (in this case actuallywritten, signed, etc. is primarily determined makes possible) the effectiveness of painby the motivative variable currently in effect. reduction as reinforcement, and (2) evokes the various kinds of responses that have ESTABLISHING OPERATIONS been reinforced by pain reduction. Keller and Schoenfeld (1950) introduced Variables such as deprivation and aversive the term establishing operation for a variablestimulation are identified as motivative that momentarily establishes the reinforcingbecause of two quite different, but related effectiveness of some other object or event.behavioral effects: (1) They momentarily alter Thus water deprivation momentarily estab-the reinforcing effectiveness of other events, lishes water as an effective form of reinforce-and (2) they momentarily alter (increase or ment. The term is useful in that it can referdecrease) the frequency of the kind of to any operation that has this effect (and theresponses that have been reinforced by those second effect seems always to accompanyother events. (Response frequency here refers the first), whether or not the operation seemsto number of response opportunities in to be a form of deprivation or aversivewhich a response occurs as well as number stimulation. Salt ingestion, for example, isof responses per unit of time, which makes not easily classified as either deprivation orit possible to avoid having to use the more aversive stimulation, but is clearly an estab-controversial probability or strength of lishing operation with effects very similar toresponse.) These changes only last as long those of water deprivation. Similarly,as the motivative variable is in effect and in changes in temperature above or below thethis sense are momentary, as contrasted with optimal level establish changes in the oppo-the changes produced by respondent or site direction as effective reinforcement, andoperant conditioning or extinction (or by the evoke behavior that has been reinforced by such changes.Reprints may be obtained from the author, Department The term establishing operation (EO) refersof Psychology, Western Michigan University, Kalamazoo, only to a variables capacity to produce theMI 49008. two defining effects described above. Many 3
4 JACK MICHAEL stimulus changes that have these two effects ful stimulus should be considered a UEO have other effects as well but those are more rather than an SD for the response that hasproperly identified with other terms. Thus, terminated it. This point can be most easilythe onset of painful stimulation is also an made by reference to a typical laboratoryunconditioned elicitor of smooth muscle and shock-escape procedure. The responsegland responses, and an unconditioned pun- which turns the shock off (often a leverisher for any behavior that immediately press) is clearly evoked by the shock onset,preceded the onset of the painful stimula- and the controlling relation is also clearlytion. These latter effects should not be con- operant rather than respondent since it wasfused with the motivational or establishing developed through the use of shock offset asoperation effects, even though they may a reinforcing consequence. These two factsoften occur along with them. (For a syste- might seem to qualify the shock as amatic organization of the terms for respon- discriminative stimulus for the response thatdent and operant functional relations see terminates the shock, but it differs in twoMichael, 1983.) important ways from a discriminative stimu- Water deprivation, painful stimulation, lus.salt ingestion, and temperature changes pro- In the first place, a discriminative stimulus,bably have their reinforcer-establishing or SD, is a stimulus condition that has beeneffects without any learning. In other words, correlated with the availability of a type ofwe are probably prewired to be reinforceable consequence given a type of behavior. Pain-by water as a function of water deprivation, ful stimulation is not correlated with theor by pain reduction as a result of painful availability of the reinforcing event that itstimulation. Of course learning plays an establishes-in this case, the termination ofobvious role in the development of the par- painful stimulation. Being in pain is notticular repertoires that are evoked by water systematically correlated with being able todeprivation or by painful stimulation-the remove pain, sad but true, except in thesecond of the two defining effects of a sense that if there were no pain there wouldmotivative variable. It is convenient to refer be no pain to remove. In other words, the presence of pain is a necessary, but not a suf-to EOs with unlearned reinforcer-establish- ficient condition for its removal, just as fooding effects as unconditioned establishing or water deprivation is necessary but not suf-operations (UEOs). It is quite clear, however, ficient for food or water reinforcement. Thethat there are establishing operations whose stimulation associated with being foodreinforcer-establishing effects are learned, deprived-to the degree that there is anyand these can be referred to as conditioned stimulation-is not also differentiallyestablishing operations (CEOs). associated with the availability of food. Just The most familiar type of CEO is the because an organism is hungry doesntauditory or visual stimulus that is paired mean that food is likely to be available, aswith painful stimulation-usually electrical would be pointed out by many currentlyshock-in an avoidance procedure, the hungry organisms. Likewise with painfulso-called warning stimulus. In the typical stimulation: It is not differentially correlatedprocedure some arbitrary type of behavior- with the availability of some way to reducepressing a lever-terminates the warning the pain. An SD, on the other hand, is astimulus and also prevents the onset of the stimulus condition whose presence-absencepainful stimulus. As a result of the systematic must by definition be correlated with theand repeated relation of the warning stimu- availability of the relevant consequence givenlus to the painful stimulus, the offset of the the relevant response.warning stimulus acquires the capacity to Secondly, in the absence of shock, there isreinforce any response that precedes such no analog to the extinction responding thatoffset, and the onset and presence of the occurs in the absence of an SD. When shockwarning stimulus then comes to evoke such is absent, the failure of the lever press to ter-responses. minate the "nonpresent" shock is in no The warning stimulus should be con- sense extinction responding. In a typicalsidered a CEO rather than an SD for the laboratory discrimination procedure withresponse that has terminated it just as a pain- light-on as SD and light off as SA the
ESTABLISHING OPERATIONS AND THE MAND 5organisms responses in-the light-off condi- the behavior because of a relation with thetion fail to produce an event, food for exam- availability of the object or event, which mayple, which would clearly be reinforcing if it have been readily available before the stimu-were received. In shock-escape the absence lus change occurred, but because it causesof shock is much more like the absence of the object or event to assume reinforcingfood deprivation than like the absence of the value. In other words, it evokes the responselight. as an establishing operation rather than as a The argument for the CEO status of the discriminative stimulus.warning stimulus in an avoidance procedure For example, two people are walkingis identical in all respects to the argument for together and one sees something that mustconsidering painful stimulation to be a UEO be written down so that it will not be for-rather than an SD for the response that ter- gotten-a store name, an address, etc. Theminates it. The only difference between would-be writer, however, does not have athem is that the warning CEO acquires its pencil so requests one from the other person,reinforcer-establishing capacity whereas that who readily provides it. It might seemof the UEO is unlearned. reasonable to consider the stimulus respons- To state the relation in most general terms, ible for the request to have been an SD forany form of aversive stimulation, whether that request, but this is not correct. Theunlearned or learned, will momentarily immediate reinforcement for requests of thisincrease the reinforcing effectiveness of its type has clearly been receipt of the thingremoval or attenuation, and momentarily requested, in this case a pencil; but theincrease the frequency of any behavior that stimulus that evoked the request did not dohas preceded such removal or attenuation. so because it was an especially favorableLikewise, we must suppose that any form situation for obtaining pencils-the compan-of improvement will also momentarily ion would have provided the pencil when-increase the punishing effectiveness of its ever requested-but rather because itremoval or attenuation, and momentarily resulted in the increased reinforcing effec-decrease the frequency of any behavior that tiveness of pencils. That is, it did not evokehas preceded such removal or attenuation the request as an SD because of a correlation(although this situation has not been sub- with the availability of pencils, but rather asjected to laboratory investigation). a CEO because of a correlation with the rein- Another type of CEO, also likely to be mis- forcing effectiveness of pencils.taken for an SD, is related to the fact that the This type of CEO often seems to be aconditioned reinforcing effectiveness of stimulus event that functions as an SD for amany stimulus changes is itself correlated type of behavior which is in some sensewith the presence-absence of other stimulus blocked-cannot occur-until some otherconditions. This relation is sometimes object or event becomes available. The stim-referred to as a context effect, in that the ulus event then also functions as a CEO withfunction of stimulus events as conditioned respect to the behavior that has been rein-reinforcers is dependent on the context in forced by obtaining this other object or event.which they occur. Said another way, the rein- Perhaps this type of CEO could be usefullyforcing effectiveness of many objects or referred to as a blocked-response CEO. (This isevents is conditional upon the status of other the establishing stimulus, or SE, describedstimulus conditions. Now, when these other in Michael, 1982. In that paper EO referredconditions assume the proper value, the to unconditioned establishing operationsobject or event becomes an effective form of and SE to the blocked-response conditionedreinforcement, and behavior which has been establishing operation. It now seems betterfollowed by that object or event becomes to use UEO and CEO, with the latter refer-more frequent-is evoked. But in this case ring to any of the several kinds of CEO. EO can now refer to any type of motivativethe relevant stimulus change does not evoke variable, unleamed or learned.) In addition to the warning and the There is no good opposite for aversive stimulation. Rein- blocked-response CEOs there may be otherforcing stimulation refers too specifically (as well it should) kinds, but these two would seem to be theto the increase in future frequency of behavior that haspreceded the stimulation. Appetitive is still too closely main ones relevant to the mand, to which werelated to food. shall now retum.
6 JACK MICHAEL THE MAND stretch ordinary usage. The obvious solution With this more extended treatment of is to define the mand as a type of verbal motivative variables it becomes possible to operant where the response is determined supplement Skinners mand definition in by a prior EO, as contrasted with the other several directions. First, it might seem attrac- verbal operants where it is determined by ative to avoid dealing with establishing opera- prior SD.tions entirely and simply relate the mand to In Skinners statement that the responseits history of reinforcement: ". . . a verbal has no specified relation to a prior stimulus,operant in which the response is reinforced response is not as appropriate as response form,by a characteristic consequence" (1957, and stimulus should be replaced by discrimin-p. 35). But history of reinforcement only ative stimulus. He is saying that with theexplains the origin of the functional unit mand, unlike all of the other elementaryinvolving SD, EO, and R; not why the verbal relations, what it is that is actuallyresponse occurs on a particular occasion. In said, written, signed (as in the sign languageother words, if a child says water as a mand, of the deaf), etc. is not determined by a priorthe explanation of that instance of behavior discriminative stimulus. This does not meancannot consist simply in the statement that (1) that the frequency of occurrence of a mandsuch responses were reinforced with water is unrelated to prior discriminative stimuli,in the past, without adding that an EO nor does it mean (2) that the fonn of therelated to water (for example water depriva- response is not determined by prior stimulition) was in effect at the time the response functioning in some other way than as dis-occurred. The layman would say that the criminative stimuli-namely as establishingchild knew how to ask for water when thirsty operations.(had the relevant history of reinforcement) With regard to the first point, as Skinnerand did so at that time because of being points out (1957, p. 52) prior stimuli are notthirsty at that time (the relevant EO was in irrelevant to the actual occurrence of theeffect). Skinner expresses this general point mand response form. Consider the mandwhen he says that the mand is under the water where the response form is determinedfunctional control of relevant conditions of by water deprivation. Saying water in thedeprivation and aversive stimulation (1957, absence of an appropriate audience, orp. 35), which brings us to the next point. under circumstances where water has never It is clear that the terms deprivation and been available has typically undergoneaversive stimulation are not broad enough to extinction, and thus even under watercover all of the variables that control the deprivation the response will not ordinarilymand. Deprivation seems to be an operation occur until appropriate circumstances are inthat is primarily relevant to unconditioned effect. The audience or the circumstances areestablishing operations or UEOs, and clearly functioning as SDs, but not in thealthough aversive stimulation can include sense of determining the form of theany operation that can be considered a form response. The EO contributes to an increaseof worsening, the blocked-response CEO is in the momentary frequency of water as anot covered by either term, nor are UEOs response form, but SDS related to past rein-such as salt ingestion, temperature changes, forcement of such a response form also con-and some others. Of course one might try to tribute. In commonsense terms, the depriva-consider the blocked-response CEO to be a tion produces some tendency to ask forform of deprivation in the sense that some- water, but such asking will not occur underthing is absent, but again such absence is not circumstances where it has been systemat-the precipitating cause of the relevant behav- ically unsuccessful in the past. On the otherior. In the previous example it was not the hand, even in circumstances where theabsence of pencils that evoked the request, mand water has always been reinforced, theor there should have been requests for all the response form would not occur unless an EOother things that were absent. One might related to water reinforcement was in effect.also try to bring the blocked-response CEO This is what Skinner means by the statementinto the aversive stimulus category, in the that with the mand .... the response has nosense that it is aversive to need something specified relation to a prior stimulus" but itand not to have it, but this, too, seems to seems somewhat more precise to say that
ESTABLISHING OPERATIONS AND THE MAND 7with the mand the form of the response is not language is of little practical significance.determined by a prior discriminative stimulus. Normal children and adults do not need Now with respect to the second point, much professional support for their mands,prior stimuli functioning as EOs may well since the mand is the type of verbal behaviordetermine the form of the mand response. that directly benefits the speaker. If anything,UEOs such as painful stimulation and a more common concern is to induce thosetemperature changes are certainly prior who mand too much to be more consideratestimuli, and even more common, both warn- of the needs of others.ing CEOs and blocked-response CEOs are Of much more practical significance is theprior stimuli, and are clearly the determ- relative neglect of the mand in languageinants of response form in the mand relation. training programs for the developmentallyA revised description of the mand and its disabled. Such programs devote very littlecontrast with other verbal operants, then, is time to the mand, in favor of training the tactas follows: The mand is a type of verbal relation and what is referred to as receptiveoperant in which a particular response form language. There are several reasons for thisis reinforced by a characteristic consequence neglect, in addition to general ignorance ofand is therefore under the functional control Skinners analysis of verbal behavior. First,of the establishing operation relevant to that acquiring a verbal repertoire is seen by manyconsequence. In the mand the response in the speech and language area as learningform has no specified relation to a prior the meanings of words. It is assumed that whendiscriminative stimulus. The other elemen- such meanings have been acquired thetary verbal operants (tact, echoic, etc.) consist words can then be used in various ways withof response forms that are reinforced by gen- no further training. From this perspectiveeralized conditioned reinforcement (Skinner, receptive language training is clearly one of1957, pp. 53-54) but in the presence of charac- the easiest ways to teach such meanings, andteristic discriminative stimuli, and are there- tact training is probably next. Based onfore under the functional control of those experience with normal children and adults,discriminative stimuli. Ideally these other once a person has learned what an object isverbal operants have no specified relation to called (by learning to point to it when givenany establishing operation (but see Chapter its name, or to say the name when the object6 of Verbal Behavior). is shown) it is reasonable to assume that when the object becomes important the PRACTICAL IMPLICATIONS learner will be able to ask for it without Considering only mands controlled by further training.UEOs, one could easily underestimate the It is clear that this does not happen withubiquity of the mand. When mands related low functioning individuals, many of whomto the warning CEOs and to the blocked- have had a good deal of receptive languageresponse CEOs are added in, however, it is and tact training but are said to lack a func-reasonable to assume that about half of the tional language repertoire, which is thenadults ordinary daily verbal interaction con- explained in terms of their general intel-sists of mands. In addition to the mands for lectual deficit. They can often point to severalobjects and actions there are the mands for kinds of objects when the name is spoken,SDs and CEOs-that is, for information- and they can sometimes even say the namethat constitute such a large share of what we when the object is shown, but have nosay to others. Much of the verbal behavior tendency to request the object when it iscontrolled by other ongoing verbal behavior clear from other evidence that it would be anin the same speaker, autoclitic verbal effective form of reinforcement for them.behavior is also a type of mand, but due to This point was dramatically made in a studythe complexity of this relation it cannot be by Hall and Sundberg (1987). Two subjectsdealt with in a paper of the present scope (for were taught to perform a sequence of activi-more on the autoclitic mand see Skinner, ties (without responding verbally in any1957, pp. 311-367, but especially, pp. 321-330; way) which culminated in the production ofalso Peterson, 1978, pp. 177-180). something that was known to be effective as Underestimation of the mands impor- reinforcement. For example, both subjectstance in our speculative analysis of everyday were taught to make a cup of soup using a
8 JACK MICHAELpackage of instant soup, a bowl, hot water, Sundberg study mentioned above, andand a spoon, with the last step being con- could be a major part of any language train-suming the soup. In a different setting the ing program.same subjects were taught to tact all of the Mand training may also be neglectedobjects used in the other setting, but not in because it is not well appreciated that it is thethe process of producing and consuming the only type of verbal behavior which directlyreinforcer. Then when again attempting to benefits the learner. Receptive and tact reper-produce the reinforcer, but the sequence toires permit the learner to follow directionscould not be completed because a critical given by others, and to provide informationobject was missing (the hot water in the case to others. Of course such directions and suchof making soup), the subjects had no ten- information may well be to the long rangedency to mand the missing object, although advantage of the learner, but long rangethey tacted the object in the other setting. advantage is seldom effective as reinforce-When the same subjects were next taught to ment. There is some evidence (Carroll &mand the missing objects by either duplic or Hesse, 1987; Stafford, Sundberg, & Braam,tact prompting, they readily learned to do so, 1988) that mand training makes other aspectsand fairly soon acquired a more general of language training more effective. The EOtendency to mand other things when they and specific consequences can be used inbecame effective as reinforcement even combination with other variables (nonverbalthough they had only been formally taught stimuli, verbal prompts, etc.) to help evoketo tact those things. This suggests that a little a verbal response, and then faded out oncebit of mand training might have dramatic the response is strong. A student who caneffects with respect to the development of successfully mand for, and receive, specificfunctional language. reinforcement is often much more willing to Another reason for the neglect of mand participate in training sessions. Receptive,training, even by those who might well tact, and intraverbal trials can also beappreciate its significance, is that the trainer interspersed with mand trials, and languagemust contrive appropriate EOs, or take training becomes much more like typical ver-advantage of those that develop naturally. bal interactions, rather than the standardContriving a variety of effective EOs for the situation where the trainer does all the man-learner seems at first glance much more diffi- ding (e.g., "Whats that?" "Touch red.") artdcult than providing a variety of objects the student simply complies.(usually pictures of objects) to be named or A final point concerning the importancepointed at. And relying on naturally occur- of mand training with the low functioningring EOs in a language training setting will developmentally disabled client is that itnot usually result in sufficient variety, will often lead to a considerable reduction inalthough the variety can be increased by the frequency of various kinds of inappro-providing language training under other priate behavior-crying, aggressive behavior,circumstances not instituted for that pur- loud unintelligible vocal responses, etc. Aspose. The procedure called incidental teaching Sundberg (1987) points out, much of this(Hart & Risley, 1975) makes some use of this behavior is actually under the control oflatter approach, in that verbal prompts for some strong EO, but the behavior is eithermands are provided whenever the learner insufficiently specific for the trainer or care-needs help in obtaining some kind of rein- giver to comply, or compliance does occurforcement during any training or care-giving which functions as continued reinforcementactivities. Although it might appear dif- for the inappropriate behavior. An appro-ficult to contrive EOs in the artificial setting priate mand response, if generally successfulof a language training program an under- will displace the inappropriate behavior, andstanding of the blocked-response CEO the client can then function more normally.should make it easier. In general, where The neglect of mand training is naturallysome known form of effective reinforcement paralleled by a neglect of the mand in lan-cannot be obtained without some additional guage assessment procedures. This wouldobject or action, that object or action becomes be expected to result often in assessmentsthe basis for a reinforceable mand. This that credit the client with better languagestrategy is well illustrated in the Hall and skills than are actually available. The relative
ESTABLISHING OPERATIONS AND THE MAND 9lack of functional language would then be REFERENCESthe basis for underestimating the clients Carroll, R. J., & Hesse, B. E. (1987). The effects of alter-actual potential for language performance. A nating mand and tact training on the acquisition ofmore detailed treatment of language assess- tacts. The Analysis of Verbal Behavior, 5, 55-65.ment from the perspective of Skinners Hall, G., & Sundberg, M. L. (1987). Teaching mands by manipulating conditioned establishing operations.analysis of verbal behavior is available in The Analysis of Verbal Behavior, 5, 41-53.Sundberg (1983). Hart, B. M., & Risley, T. R. (1975). Incidental teaching of language in the preschool. Journal ofApplied Behavior CONCLUSION Analysis, 8, 411-420. Keller, F. S., & Schoenfeld, W. N. (1950). Principles of Skinners analysis of language is a major psychology. New York: Appleton-Century-Crofts.behavioral breakthrough, with many Michael, J. (1982). Distinguishing between discrim- inative and motivational functions of stimuli. Journaltheoretical and practical implications. Its of the Experimental Analysis of Behavior, 37, 149-155.advantage over traditional language theory Michael, J. (1983). Evocative and repertoire alteringis especially clear in its identification of effects of an environmental event. TheAnalysis of Verbalelementary verbal relations, and its implica- Behavior, 2, 21-23.tions for teaching and learning these rela- Peterson, N. (1978). An introduction to verbal behavior. Grand Rapids: Behavior Associates.tions. Because of the mands introduction in Skinner, B. F. (1953). Science and human behavior. Newterms of unlearned motivative variables, and York: Macmillan.because its relation to prior controlling events Skinner, B. F. (1957). Verbal behavior. New York:is quite complex, its general significance has Appleton-Century-Crofts.probably been considerably underestimated. Stafford, M. W., Sundberg, M. L., & Braam, S. J. (1988). A preliminary investigation of the consequences thatA more extensive treatment of establishing define the mand and the tact. The Analysis of Verbaloperations and a description of the mand Behavior, 6, 61-71.relation in such terms will, I hope, prevent Sundberg, M. L. (1983). Language. In J. L. Matson &such a mistake, as well as lead to a more S. E. Breuning (Eds.) Assessing the mentally retarded (pp. 285-310). New York: Grune & Stratton.appropriate emphasis on the mand in Sundberg, M. L. (1987). Teaching language to thelanguage training programs for subjects who developmentally disabled: A course manual. Princedo not develop normal language. George, B.C.: College of New Caledonia Press.