2006 - The Distinction Between Positive and Negative Reinforcement Some Additional Considerations -...

download 2006 - The Distinction Between Positive and Negative Reinforcement Some Additional Considerations - Sidman_2.pdf

of 5

Transcript of 2006 - The Distinction Between Positive and Negative Reinforcement Some Additional Considerations -...

  • 8/13/2019 2006 - The Distinction Between Positive and Negative Reinforcement Some Additional Considerations - Sidman_2.pdf

    1/5

    The Distinction Between Positive andNegative Reinforcement:

    Some Additional Considerations

    Murray SidmanSarasota, Florida

    Because of the many constructivecontributions by Alan Baron andMark Galizio, I have learned to payclose attention to whatever they haveto say. In the present instance (Baron& Galizio, 2005), however, I am

    puzzledas I was when Jack Michaelpublished his original paper suggest-ing that we abandon the distinctionbetween positive and negative rein-forcement (Michael, 1975). It is notclear to me whether they are recom-mending simply a new terminologicalconvention or whether they are sug-gesting something more fundamen-tala change in basic principles, orat least a new way to conceptualizeour data.

    I am sure that Baron and Galizio(2005) and Michael (1975) are notrecommending that we ignore thedata from experiments on (or evenfrom daily observations of) escapeand avoidance behavior, the phe-nomena that give rise to the conceptof negative reinforcement. They areas committed as anyone to a behav-ioral science based on data, andwould never suggest that we throw

    data away simply because some ofthose data are difficult to classify. Iwondered, therefore, whether theywere simply looking for another termthat might encompass those datamore precisely. I certainly agree withtheir observation that the concept ofnegative reinforcement has causedconfusion. In my experience, thishas been true especially of readersand audiences who are being intro-

    duced to behavioral science. Forexample, readers of my book oncoercion (Sidman, 2000) who donot finish the book are most like-ly to have put it down perma-nently after they reached my pre-

    sentation of the distinction betweenpositive and negative reinforcement.(This remains true even after anextensive revision of that section.)I believe the difficulty here stemsfrom conventions of ordinary speech,in which the term negative usuallydenotes the opposite of somethingpositive; having accepted a defini-tion ofpositivereinforcement, and, inthe process, equating it at leastroughly with the everyday term re-ward, many then find it sensible toequate negative reinforcement withpunishment. This, of course, is notwhat we ask them to do, so we losethe ones who are not willing or whoare unprepared to work out theproblem.

    And so, I would welcome a sub-stitute for the term negative reinforce-ment, a new term that removes theconfusion of negative reinforcementwith punishment while retaining thenotion that escape and avoidance arekinds of reinforcement. I have notbeen able to come up with a satisfac-tory term, but I would not object inprinciple to such a suggestion. Orperhaps we could drop the character-ization of reinforcement as eitherpositive or negative and just talkabout different kinds of reinforce-ment: food, water, sex, escape, avoid-

    ance, and so on. Baron and Galizio(2005) however, seem to be unwillingto take the latter route, on thegrounds of the uncertainty in speci-

    Address editorial correspondence to MurraySidman, 1700 Ben Franklin Drive #9E, Sar-asota, Florida 34236 (e-mail: [email protected]).

    The Behavior Analyst 2006, 29, 135139 No. 1 (Spring)

    135

  • 8/13/2019 2006 - The Distinction Between Positive and Negative Reinforcement Some Additional Considerations - Sidman_2.pdf

    2/5

    fying whether production or escape isthe critical reinforcing event.

    The tendency to equate negativereinforcement with punishment isnot, however, a major source of

    Baron and Galizios (2005) concern.Their main discontent seems to arisefrom a perceived difficulty in distin-guishing between the production andthe removal (or prevention) of stimulias a basis for the classification ofreinforcing events:

    The argument is that positive and negativereinforcement are changes from one stimuluscondition to another, not the simple pre-sentation or removal of a stimulus. Withoutthis essential clarification, the claim thata reinforcer is exclusively positive or negativealways can be challenged by the assertion thatthe alternative form is the true basis for thereinforcing effect. (p. 87)

    One cannot, of course, claim that anyevent is exclusively a positive ornegative reinforcer in the sense thatit isalwaysjust one of these. Given anappropriate history and context, anyevent can function as either a positiveor a negative reinforcer. I mustassume, then, that Baron and Galizioare referring to an event that takesplace at a particular time and ina particular context, and are ques-tioning whether the reinforcement onthat particular occasion is positive ornegative. To illustrate this difficulty,they cite the classic Weiss and Laties(1961) experiment in which ratslearned to press a lever that turned

    on a heat lamp and increased thetemperature in their cold chamber. Isthe reinforcement here the produc-tion of heat (positive) or the re-duction of cold (negative)? Baronand Galizio then generalize this in-terpretive difficulty to all instances ofreinforcement. They argue that pro-duction of a stimulus (positive re-inforcement) always involves escapefrom a situation (negative reinforce-

    ment) in which the stimulus wasabsent; removal of a stimulus (nega-tive reinforcement) always involvesproduction of a situation (positive

    reinforcement) in which the stimulusis absent. I have two major difficul-ties with this argument.

    First, even though Baron andGalizio raise cogent objections to

    notions that reinforcement involveschanges in physiological or emotionalstates, they still find it necessary tobuttress their principal argument byappeal to such states. For example,the presentation of food is also saidto involve reduction in a state ofdeprivation (p. 87). Ingenious re-search designed to support exactlysuch an explanation for the effective-ness of food and water as reinforcers

    (e.g., Miller & Kessen, 1952; Miller,Sampliner, & Woodrow, 1957) failedto prove convincing, and the theorythat reinforcement must involve drivereduction is no longer in vogue.Similarly, Baron and Galizio ask,Is it better to speak of the conse-quence as increased attention or asrelief from loneliness? (p. 91). Theyask this question in support of theirthesis, in spite of their valid critique

    (p. 90) of the use of emotional pro-cesses as explanations.Second, and in my view, more

    important, is the too easy acceptanceby Baron and Galizio of the impos-sibility of determining whether pre-sentation or removal is the reinforcerin any particular case. In theirexemplar, the Weiss and Laties(1961) study, it is, of course, truethat both antecedent and consequenttemperatures must be taken intoaccount in specifying the reinforcer.But does the existence of instancesthat cannot be classified solely aspositive or negative reinforcementprovide sufficient reason to do awaywith the distinction? Could it not stillbe true that often it is specifiable andthat there is some value in doing so?Areas of definitional confusion existat the edges of many, if not most,concepts, yet this does not necessarily

    make the concepts unusable. Is some-thing large or is it small, fast or slow,high or low? In these and many othercases, one can only answer, That

    136 MURRAY SIDMAN

  • 8/13/2019 2006 - The Distinction Between Positive and Negative Reinforcement Some Additional Considerations - Sidman_2.pdf

    3/5

    depends. It is doubtless true that thetermination of a moneyless periodmay be critical in determining thereinforcing value of money, yet wesee billionaires working mightily to

    get more. The effectiveness of somepositive reinforcers depends less ontheir previous absence than on whatbehavior they make possible in thefuture. Such problems of definitionshould lead to research that will helpto clarify the existing definitions.Even punishment, which Baron andGalizio rely on as a replacement fornegative reinforcement, has its defi-nitional problems, about which I

    shall have more to say below.Very often, when an instance seems

    open to confusion about its status asa positive or negative reinforcer, waysexist to make the distinction clear,to determine whether a behavioralchange that results from productionor removal can be specified as some-thing more than just the absence ofan alternative. Responses that turnoff electric shocks also take the

    organism into an environment inwhich shocks are absent, but wouldthe animal work to produce (or tomaintain) that same environment in-dependently of any experience withshocks? That experiment can bedone, but the results are so predict-able that no one does so. Thetermination of shock will be reinfor-cing even if the environment pro-duced by the escape behavior keepschanging. When a child who acts insuch a way as to bring about hisremoval to a time-out room does soonly when he is being required toengage in a particular activity, onecan easily determine whether escapefrom that activity or whether thetime-out room per se provides thereinforcers for his disruptive behav-ior. This experiment is often worthdoing because production of thesocial interaction on the way to the

    time-out room may be found to bethe reinforcer rather than escape perse. Would the production of anaspirin suffice to reinforce the open-

    ing of the bottle if the pill did notturn off the headache? I think weknow what would happen to aspirinconsumption if we filled aspirinbottles with falsely labeled placebo

    tablets. Might a child turn on a TVprogram even if it took her awayfrom a favorite game or a much-wanted dessert? Such an event iscertainly possible, and when it hap-pens, it would be difficult to identifythe reinforcement as escape fromboredom. Do spouses usually seekdivorces independently of their mar-ital experiences? Surely, we do notneed experimental data to classify

    many, if not most, divorces as escapebehavior. The possibility and utilityof identifying positive and negativereinforcers do not seem as discour-aging to me as they apparently do toBaron and Galizio.

    I suspect that one reason for Baronand Galizios (2005) unease with thepositivenegative reinforcement dis-tinction is their feeling that it reflectsdifferences in the reinforcement pro-

    cess itself. I am not sure where thatfeeling comes from. For anyone whodoes espouse such a view, theirconvincing discussion should settlethe issue. I am, however, uneasyabout their unquestioning substitu-tion of the reinforcementpunish-ment differentiation for the positivenegative reinforcement distinction.

    First of all, I have a problem withthe definition ofpunishment as a con-sequence of behavior that reduces thefuture probability of that behavior.When a teachers correction ensuresthat a students wrong answer willnever occur again, has the teacherpunished that wrong answer? Whena parent ends a childs eating betweenmeals by hiding the cookie jar, hasthe childs cookie eating been pun-ished? If a child stops misbehavingafter being asked to stop, has themisbehavior been punished? When I

    ask for something and the answer isno, does my failure to make therequest again signify that it has beenpunished? For such reasons, I do not

    IN RESPONSE 137

  • 8/13/2019 2006 - The Distinction Between Positive and Negative Reinforcement Some Additional Considerations - Sidman_2.pdf

    4/5

    believe that Baron and Galiziosdefinition of punishment is any morefree of ambiguity than are mostconcepts.

    The above definition of punish-

    ment was first proposed by Azrin andHolz (1966, p. 381), and it slippedinto the mainstream of the operantculture without the extensive discus-sion one might have expected fromsuch a radical change in B. F.Skinners definition. There seem tobe no more than a few today whoeven recognize that such a changetook place. I do not, of course,quarrel with Azrin and Holzs defi-

    nition simply because it differs fromSkinners, and perhaps the differencedoes not affect anything we might doin the laboratory or clinic, but itseems to me to possess a systematicstatus that would have been worthdiscussing before accepting it uncrit-ically.

    Skinner defined punishment aseither the response-contingent pre-sentation of a negative reinforcer or

    the removal of a positive reinforcer,a definition that was most clearlydetailed in Holland and Skinner(1961, p. 245). There are two typesof punishment simply because eachtype of reinforcement, positive andnegative, has a symmetric counter-part. Just as Skinner claimed nodifferences in principle between pos-itive and negative reinforcement, heclaimed no such differences betweenthe two types of punishment. As Ihave noted elsewhere, Skinners def-inition of punishment says nothingabout the effect of a punisher on theaction that produces it. It says neitherthat punishment is the opposite ofreinforcement nor that punishmentreduces the future likelihood of pun-ished actions (Sidman, 2000, p. 45).Unlike his definition of reinforce-ment, his definition of punishmentappeals to no behavioral outcome.

    The effects of any punishment are tobe empirically determined.

    To specify a stimulus or an eventas a reinforcer, one must observe

    a subsequent increase in the behavioron which it was contingent. Then, toexplain the increase in behavior, oneappeals to reinforcement. The circu-larity here is apparent, and although

    it can be obviated by a rather in-tricate chain of verbal reasoning, itremains an inelegant feature of re-inforcement theory. Azrin and Holzs(1966) definition of punishment im-parts the same inelegance to a secondfundamental concept. Skinners orig-inal definition avoids this awkward-ness because it requires no particularbehavioral outcome in order to iden-tify a punisher. Perhaps further dis-

    cussion of this matter would relievemy unease about appealing to thecurrently dominant definition of pun-ishment to solve any perceived prob-lems with the concept of negativereinforcement.

    Finally, something more needs tobe said about relating science toproblems of everyday life. Baronand Galizio (p. 91) object to thenotion that the positivenegative re-

    inforcement distinction might help toinform applied behavior analystsabout the undesirable aspects ofnegative reinforcement. In a sense,this is a straw man type of issue.First of all, undesirable here mustrefer to social rather than to scientificdesirability. Whether negative rein-forcement procedures are undesirableis not an empirical scientific questionat all, because it is not a question thatscience can answer. What science cando, however, and has done, is toprovide data that should be relevantto any such decisionsextensive da-ta, for example, on escape andavoidance behavior, on punishment,on conditioned suppression, on ex-tinction- and shock-induced aggres-sion, and so on. Our science shouldpresent these data to our society notto tell it that those procedures pro-duce undesirable effects but to tell it

    just what those procedures do. Ifa scientist wants to do more than justprovide information and also to helpmake societys judgment call, it must

    138 MURRAY SIDMAN

  • 8/13/2019 2006 - The Distinction Between Positive and Negative Reinforcement Some Additional Considerations - Sidman_2.pdf

    5/5

    be recognized that he or she is doingso not as a scientist but as a concernedcitizen.

    With respect to Michaels (1975)doubts about the wisdom of main-

    taining a distinction at the level ofbasic science because of its possiblesocial implications (p. 42), thosevery doubts may, in turn, be ques-tioned. If any science is to survive oris to be accepted by society assomething worthwhilethat is tosay, worth supportingthen it hadbetter relate its content to problemsof everyday life. Indeed, that is wheresciences start (Lee, 1988). People see

    things happening and wonder aboutthem, trying to understand and makeuse of them. Coercion, which I havedefined as punishment and negativereinforcement (Sidman, 2000), is atopic that has held peoples interestthrough the ages and up to now. Iwas interested in studying avoidanceand escape because even withoutany training in behavioral science Isaw those as the roots of what used

    to be called abnormal behavior. Ascience that fails to make a placefor topics that are considered impor-tant in the world at large may notonly fail to survive but may deservethat fate.

    And so, I find myself in disagree-ment with colleagues for whom Ihave unbounded respect. If those

    disagreements could be resolved, Iwould consider myself in one of thosesituations in which positive andnegative reinforcement (productionof agreement and escape from dis-

    agreement) are inextricably mixed.

    REFERENCES

    Azrin, N. H., & Holz, W. C. (1966). Punish-ment. In W. K. Honig (Ed.), Operantbehavior: Areas of research and application(pp. 380447). New York: Appleton-Centu-ry-Crofts.

    Baron, A., & Galizio, M. (2005). Positive andnegative reinforcement: Should the distinc-tion be preserved?The Behavior Analyst, 28,8598.

    Holland, J. G., & Skinner, B. F. (1961). Theanalysis of behavior. New York: McGraw-Hill.

    Lee, V. L. (1988). Beyond behaviorism. Hills-dale, NJ: Erlbaum.

    Michael, J. (1975). Positive and negativereinforcement: A distinction that is nolonger necessary; or a better way to talkabout bad things. Behaviorism, 3, 3344.

    Miller, N. E., & Kessen, M. L. (1952). Rewardeffects of food via stomach fistula comparedwith those of food via mouth. Journal ofComparative and Physiological Psychology,45, 550564.

    Miller, N. E., Sampliner, R. I., & Woodrow,P. (1957). Thirst-reducing effects of waterby stomach fistula vs. water by mouthmeasured by both a consummatory and aninstrumental response. Journal of Compara-tive and Physiological Psychology, 50, 15.

    Sidman, M. (2000). Coercion and its fallout(rev. ed.). Boston: Authors Cooperative.

    Weiss, B., & Laties, V. G. (1961). Behavioralthermoregulation. Science, 133, 13381344.

    IN RESPONSE 139