For a usefulperspective on this process, imagine yourself all alone in somelarge openspace, perhaps lounging on a deserted beach, strolling across a parkmeadowor playing field, sitting in a empty subway car, or even driving alonganopen road, when another person or group of people appears on thescene.Although such intruders will at least initially "bear watching"(Anderson1978:129), their presence will not necessarily interfere withyour purposes,unless you are hoping to be alone. Their actions maypresent only a temporarydisturbance, until one party or the other leaves.In the meantime, "civilinattention" (see Goffman 1963:83-88) may sufficefor going about your ownbusiness.
However, the other party's behaviormight sometimes be extreme enough (anythingfrom shading your blanket onthe beach to driving slowly but erratically downthe center of the highway)that you find it highly effortful or even impossibleto continue your ownactivity in the face of disturbance. In PCT terms, yousuffer a prolongedperceptual error. What are your options for controllingyour perception ofthese other people's behavior? They too are self-guidedcontrol systems,so your problem comes down to making them readjust thereference valuesthey are using to control their own perceptions, in otherwords, to getthem to do something else.
Powers argues that at least some of a control system's referencesignals orintentions, even though they occur internally, can be reliablyascertained byoutside observers. The key to discovering a person'sintentions is to find aquantity in the environment which is beingstabilized by the person's actions,that is, something which couldchange but is not being permitted tochange. To test for this controlledquantity, Powers suggests applying adisturbance to a possible controlledquantity and watching to see if theperson's movements oppose thatdisturbance (Powers 1978, 1979c). Runkel haselaborated on the directionsfor "The Test" (1990a:117-118; 1990b), noting thatone should alsoascertain whether the person has perceptual access to thequantity oneassumes to be controlled and whether any blocking of that accessresults inloss of control.[8]
For practicalpurposes and under ordinary circumstances, elaborate tests ofintentionsare usually unnecessary (but see Jordan and Hershberger 1989). Wehave nodifficulty making reasonably accurate inferences about intentions atthelevel of everyday actions. There is no need, for instance, to pull achairout from under someone to see whether he had intended to sit or toplace anobstacle in someone's path to see if he intended to walk in thatparticulardirection. If someone does something repetitively or maintainsa line ofaction without change, our assumption that the action isintentional is usuallynot problematic. At this level, indeed, weordinarily name our own and otherpeople's actions by the goal or purposeof the action, or in other words, theapparent intention (see Vallacher andWegner 1987), rather than by the physicalmovements. "What are you doing?""Going to the store."
Still, making some inference, however crude,about the other person'sintentions is an essential first step in anysituation of potential conflictfor understanding what the other person isdoing and possibly getting him orher to do something else. Interpretinganother person's behavior usuallyinvolves reviewing one's own store ofmemories, which of course also serve asthe source of one's own referencevalues or intentions. Then, once thebehavior has been fit into aconceptual category, the label for the action willbecome a reference valuefor maintaining one's perception of what the otherperson is doing as onecontinues to observe the other's movements.
For example, when anintruder arrives on the scene, one's inner conversationmight go like this(though not necessarily in any verbal form): "What is thatperson doing?Maybe he's just passing through. He seems to want to talk tome. Is he amugger? Or just trying to be friendly? Maybe he thinks heknowsme. . . " As discrepancies arise between thereference valueone has adopted for the other person's action and themovements of the otherperson, one adjusts one's behavior to maintain thehypothesis one has formedabout the other's intentions, or else, if thediscrepancy continues andincreases, one is forced to recategorize theother's behavior by forming a newhypothesis about the other's intentions,which then becomes the new referencevalue to be maintained if possible inthe ensuing interaction (see Heise 1977).
A uses force on B when A acts withthe intent of creatinga disturbance for B which is serious enoughto cause B to losecontrol of one or more of the perceptualvariables B is currentlycontrolling.
My translation ofcommon-sense understandings into more technical languagesomewhat expandsthe definition of force by suggesting the parallel betweenphysical forceand other actions with the same intent as physical force andoften withsimilar effects. A disturbance need not be violent or even physicaltodisrupt perceptual control. For instance, an attempt to insult orembarrassanother person counts by this definition as a verbal act offorce, in that theintent of the action is to make the victim lose controlof his own perceptionof self-esteem. Likewise, a listener heckling aspeaker is using force if theobject is to disrupt the delivery of thespeech. Nonviolent demonstratorssitting in an office or blocking adoorway are by this definition also usingforce if their actions make itimpossible for others to carry on their work asusual. Moreover, by thisdefinition a use of force can be unsuccessful andstill constitute force:A's attempt to trip B is use of force,whether or notB actually falls down.
One further implication of the PCTdefinition is that any given instance offorce may be ambiguous. Was thefirst person deliberately trying to disruptthe perceptual control of thesecond, or was it just an accident? When apossible act of force has takenplace, we need information on the perpetrator'sintent to achieveperceptual control of what has happened. This is why customcalls for thepossible offender in such a situation to offer an account of hisactions(see Scott and Lyman 1968). By giving an apology or excuse, theoffendercan acknowledge responsibility for disruption of the victim's controlbutcan deny that the disturbance was intended to have that effect. Bygivinga justification, the offender can admit the action was intended butclaim itserved a higher-level goal, which presumably both victim andoffender share. The PCT perspectivealso exposes the intrinsic limits of using force as astrategy ofinterpersonal control. One can force a person to abandon hiscurrentreference level, but the choice of reference level to control nextisentirely up to him. The victim of force, frustrated in one initiative,selectsfrom his repertoire of reference signals a different line of actionto meet hisown higher-level goals, not the goals of the person using theforce. AsWartenburg puts it, force "can never make an agent doanything"(1990:93).[10] Moreover, force invites resistance and reprisal. Control systems arebuilt toresist disturbances with all the energy at their disposal.Disruption of aperson's control at the level of physical movements canthreaten his control athigher levels, and all such losses of control tendto provoke negativeemotions, as well. Thus, the conflict may rapidlyescalate, with the secondperson seeking to fight back by damaging in turnthe first person's ability tomaintain perceptual control. Physical forcemay have to be overwhelming tosucceed, and even when overwhelming mayrequire great expenditures of energy tocounter the resistance of thevictim, as any parent of a kicking and screamingtoddler can attest. Insum, force has limits which necessitate otherstrategies of interpersonalcontrol. A coerces B by acting to produce adisturbance which Aintends B to perceive as initiating asequence in which A'sactions will force B to lose perceptualcontrol of one or more ofB's currently operative goals. IfB then preemptively changes reference signals for thethreatenedsystem or systems in order to maintain perceptual control atsome higher level,the coercion has succeeded. In other words, coercionhappens when B'sperceptions of A's actions lead B toremember or imagine anincident of force, and B, with a zero ornear-zero reference value forperceiving himself as a victim of force,takes action to avoid this outcome. Coercion is a more efficientstrategy of interpersonal control than force,because it requires lessenergy to achieve the same goal of disrupting thevictim's perceptualcontrol, but coercion, like force, has many limitations.Again, a coercercan succeed in stopping an action but cannot determine whichaction willfollow (without resorting to influence, to be discussednext).Moreover, a coercer must depend on the victim to perceive the threatand toimagine that it will lead to some significant disturbance. Anintendedvictim who refuses to heed the threat cannot be coerced. Thus,would-becoercers must occasionally provide educational shows of force ifthey hope tointimidate others by their threats (Wartenburg 1990:102).Finally, with thevictim's perceptions an important part of theinteraction, coercion maysometimes occur when none was intended. Victimsmay conceive a threat andmodify their behavior in unsought ways simplybecause they fear the other partymay be likely to use force on them, say,in interactions of men and women orpolice officers and civilians.Coercion, then, may sometimes misfire orsometimes work too well,depending on the perceptions of the personbeing coerced. The use ofincentives, or so-called positive reinforcements, is ordinarily seenas anoncoercive method of interpersonal control, but from the PCTperspectivesuch manipulations are merely a variant of coercion (see Powers1992:64; Runkel1990a:162). Operant conditioning (Skinner 1974)works only when theexperimental animal is hungry. When satiated, anexperimental animal doesn'tbother jumping through the hoops set up for it,because the purpose of suchbehavior, from the animal's point of view, isto maintain its perception ofgetting enough to eat. Changingcontingencies of reinforcement result inchanging rates of response becausethe animal is trying to keep constant itslevel of food intake and theexperimenter's food trough is its only source.The less generous theexperimenter in providing the food, the more work in theform ofmaze-running or bar-pressing the animal must do to stave off itshungerpangs. Human workers drudging for a paycheck find themselves in asimilarlycoercive situation. They must submit to the interpersonalcontrol of theemployer or risk the loss of their wages. Under theseconditions, conflictbetween worker and employer may often develop, theincentives notwithstanding.Thus, incentive systems, like force andcoercion, are limited and oftenineffective strategies for controllinganother person's behavior. A influences B by acting to create a disturbancefrom which Bconstructs a reference level for perceptual control ofB'sown actions. In other words, influence occurs notonly when A tells B what todo and B complies, butalso when B observes A and resetshis own reference levels tothose he imagines A to be using. A personmay try to influenceanother by direct instruction or by setting an example.Both words anddeeds can have influence, but a second person must accept thesuggestion orimitate the action in order for influence to occur. Influence ispotentially a more effective strategy for interpersonal controlthan force,coercion, or incentives, because one can use it to prompt actionsas wellas to prevent them.[11] Wheninfluenceis successful, the second party will have chosen a reference levelsuppliedor suggested by the first party, not just one that suits his ownpurposes.However, influence is not invariably successful. One can commandanintruder to leave, but it is up to him to comply. As a strategy ofcontrolinfluence has definite limitations, because its success depends socrucially onthe perceptions and actions of the person accepting theinfluence. In the first place, we have no simple way of transferringreference signalsfrom one person to another. According to PCT, eachperson develops a uniqueview of the world, and this makes the accuratecommunication of goalsinevitably problematic. Each person's set ofcontrol systems develops througha random process of reorganization, sothat every person's perceptual apparatusends up being somewhat differentfrom another's, and thus no two people everachieve precisely the samesense of "reality". Communication difficulties are not the only obstacle to successfulinfluence.The target person may understand precisely what the desiredreference level forhis action is, but still choose to substitute anentirely different referencelevel in order to eliminate perceptualdiscrepancies in some other higher-ordercontrol loop. I may know what youwant me to do but choose not to obey, if bydisobeying I can save mycountry or benefit my family or be true to myself orjust do what feelsgood. Compliance is always a matter of choice. Even when people chooseto obey, they remain in control of their ownperceptions. Verbal commands,for instance, can name a desired action and thusexplicitly specify thereference standard to be adopted, at least to the extentthat suchreference values are expressible in language. However, verbalcommands canrefer only to fairly high-order control loops in the humanperceptualcontrol system. Telling someone to "walk to the door" leavesentirelyimplicit the details of how this action will be accomplished, thatis,the particular sequences of arm andleg motions, musclecontractions,etc. Indeed, most of these lower-order reference values are hardtoverbalize, because we ordinarily do not exert any conscious controloverthem. Descriptions that get into the details of a desired motoroutput justcause confusion. Thinking too hard about an action often leadsto hesitationsand mistakes, and for most purposes it doesn't matterwhether one takes threesteps or four to reach the door. Thus, havingaccepted a higher-order goal, aperson still remains in perceptual controland must rely on his own lower-orderreference values in order to act. Insum, obstacles arising from the vagariesof communication, the independenceof human will, and the impossibility of everspecifying an action in fulldetail combine to make the use of influence tocontrol other people achancy business at best. Even this most effectivestrategy of controllingother people often falls short of the objective. Should we concludethen that social power is a myth, that people are simplyfree actorssuffering from the delusion that others hold them in thrall? By nomeans!Generations of sociologists have recoiled from this view. Toallappearances, structures of power--social arrangements in which oneperson haseffective control over the actions of another--abound incontemporary society.But how can PCT account for this ubiquity of socialpower, when the theorymakes interpersonal control seem so difficult? My argument will be not that power is an illusion, but that ourtendency to seepower as the property of the individual is mistaken. PCTcan help usunderstand the concept of power, but only when our analysisextends beyond theindividual and the dyad. Power, I will argue, inheresin social structures,not individuals, and to understand the workings ofpower we must consider theinteractions of at least three peoplesimultaneously. Social structures alwaysinvolve two or more people whoconsciously coordinate their actions by agreeingto a common goal; in PCTterms, they share the same reference values. Whensuch coordinationoccurs, social power appears. Just as a single stake driveninto theground cannot block the path of a walker, but a row of stakesproperlyaligned could fence him in, so a single individual may have littlecontrol overanother person's behavior while a group who coordinate theirefforts may havesubstantial power to limit the possibilities anotherperson experiences. Wartenburg (1990:150) has provided a convenient term fordescribing thiscoordination of action by means of equivalent referencevalues: "socialalignment." By looking at the ways reference values alignor fail to alignacross individuals, we can begin to conceptualize the waysin which theinteractions of humans-as-control-systems coalesce into themacro-levelpatterns of society. Whatever the processes by whichlarge-scale social alignments occur, a shift inour level of analysis froma focus on micro relations to an investigation ofmore macro patterns ofalignment leads to an important reconceptualization ofpower. The idea ofaligned reference standards provides the key to developinga PCT definitionof power in structural terms and distinguishing it from themere use offorce, coercion, incentives, or influence--micro-level interactionswhich,I would argue, are ways of achieving power but not power, per se.Forpractical purposes, alignment is social power. The socialstructures wethink of as powerful--armies, government bureaucracies,capitalistenterprises--derive their power from the willingness of largenumbers of peopleto align their perceptual reference standards and focuson shared goals. Whena group of individuals use the same language, sharecommon goals, think alikeabout events, and perceive their world in thesame way, this alignment placesgreat constraints on the options open toany individual trying to control hisown perceptions in a shared physicalenvironment. An individual who fails toconform to the prevailing socialalignment (or at least get out of the way) mayfind that virtually everyoneelse is seeking to use force, or coercion, orinfluence to control theirperceptions of his behavior. His freedom of actionmay be as limited as ifhe were locked in a jail cell. By equating social power with thealignment of reference standards, we can seehow the social situations mayvary in the sheer amounts of social power beinggenerated. At one extreme,if a given locality contains a scattering ofindividuals pursuing their owndisparate ends without any reference to oneanother, the power inherent inthe situation is at a minimum. At the otherextreme, when large numbers ofindividuals come together to seek closecoordination their perceptions andgoals, power is maximized. In general, thegreater the number ofindividuals aligned, the greater the power; the greaterthe number (andhigher the perceptual level) of the reference standardsaligned, thegreater the power; the more precise the alignment of(high-level)standards, the greater the power. As McPhail has pointedout in The Myth of the Madding Crowd (1991), mostpublic gatheringsof individuals are not very powerful in this sense ofalignment of goals orreference standards. McPhail argues that the notion of a"crowd mind"controlling the actions of individual participants is virtuallyunsupportedby the empirical evidence. In typical large-scale gatherings, heobserves,small groups of acquaintances or family members interact mainly witheachother, presumably in pursuit of their own private purposes. The levelsofcoordinated actions or emotions remain relatively low. McPhail andWohlstein'sexperiments (1986), however, show that with verbal instructionspreviouslyuncoordinated groups can achieve a powerful unity ofaction. Powerfulorganizations, such as armies, not only must bring together largenumbersof people, but they must impose close alignment of reference standardsonthese human control systems by drill and discipline. Military uniformsandmilitary jargon, the long hours of instruction and drill imposed onrecruits,and the segregation of military living quarters from the generalpublic are allintended to increase the power of the organization byperfecting the alignmentof the reference standards of the individualsinvolved. Thus, the power of thearmy lies in its coordination and unityof purpose. Some might object that anarmy's power depends even more onits command of sophisticated weaponry orother technological equipment,but, as Latour (1987) points out in connectionwith scientificlaboratories, the building and maintenance of any complextechnologicaldevice requires the efforts of large networks of people whosepurposes andefforts have been aligned. Hence, the importance of ritual and symbolism in group life. Myths,slogans,and classifications (see Lincoln 1989) can symbolize groups andhelp tointegrate them by making alignments more visible. In general,ritual occasionspresent tailor-made opportunities for people to bringtheir reference valuesinto alignment or to confirm their perceptions thatsuch alignment alreadyexists (see Collins 1988:192ff.). Morespecifically, displays of synchronizedspeech or movements, such as chantsand group prayers or parades by troopsmarching in formation, often figureprominently in the rituals that symbolizegroups because such displaysprovide immediate and demonstrable evidence of thepresumed alignment ofthe participants. Such publicly visible signs ofalignment cannot, ofcourse, guarantee that in their more complex or subtlesystems ofperceptual control the participants will be equally wellaligned.Worshippers, for instance, may be merely mouthing the prayerswhile theirimaginations wander elsewhere. But close adherence to theoutwardly visibleforms, especially if great attention and care arenecessary to produce thedisplay, is by default taken as sufficientevidence to indicate an inwardconformity to the shared standards of thegroup. Collins, drawing on the Durkheimian theory of rituals, arguesthat a group'srituals often affect the emotions of the participants(1992:30-47).Individuals can derive a "feeling of exultation and emotionalstrength" fromthe "emotion-producing machine" of a gathering for ritualperformances. ThePCT perspective on power as alignment of referencestandards gives some cluesabout why individuals might experience thisemotional charge from groupparticipation. We must start from thepremise that however powerful the group, however closethe alignment, theperceptual alignment of the participants cannot ever bequite perfect.Because individuals build up their control systems by a randomprocess ofreorganization, every individual's control systems will be uniqueand noindividual's reference standards can perfectly match another's.Thus,alignment can only be approximate at best. This inevitable mismatchbetweenthe standards of the individual and the collective standards of thegroup,together with the impossibility of knowing precisely what is inother people'sheads, leaves the individual always to some degree at oddswith the rest of thegroup or at least uncertain of the degree ofalignment. Relevant here are Scheff's descriptions (1990) of the"social emotions" ofpride and shame. Pride, says Scheff, reflectsfeelings of secure affiliationwith the group, while shame derives fromfears of threatened alienation fromthe group. Ritual performances provideopportunities for us to perceiveourselves as well attuned to a group'scollective standards, at least to theextent these standards are embodiedin the rituals. Thus, the increase in gainof one's own perceptions ofunity with the group will produce one positiveemotion--pride. When groupsaccomplish things that individuals cannot possiblydo on their own, thereduction of the individual's perceptual error is yetanother source ofpositive feelings, provided that individual's own referencelevels forperceiving the situation are close enough to the outcome whichiscollectively produced. If one's purposes are supported by the group,theexperience of participating in a powerful alignment can beexhilarating,because one gets good control with little effort. The definition of power as alignment makes clear that in manyinstances nosingle participating individual has any appreciable influencein determiningthe standards to which members of a group will alignthemselves. For instance,a language community fits this definition of apower structure. Speakers of acommon language must individually alignthemselves to a system of norms andconventions for speaking, writing, andunderstanding the common tongue. Noindividual can independently changethese rules and hope to be understood.Individuals are constantly makinginnovations in language, but the languagedoes not change until othermembers of the language community come to acceptthe new word or usage.Although individual innovators may thus achieve someinfluence in spite ofstrenuous efforts by grammarians to preserve the ancientforms, thestrength of a single individual to defy the language group's powerhasclear limits. As the example of a language community illustrates,individuals who participatein groups may often have the sense that powerof the group is entirely externalto them. Individuals who encounter awidespread alignment may feel themselvesto be in the grip of social forcesover which they effectively have no control.The power may thus beexperienced as coercive, with the individual seeming tohave no choice butto go along. Nevertheless, each time an individual adoptsthe group'sapparent reference standards, he adds to the overall power oftheconfiguration. Nevertheless, I would not want to overemphasize the permanence ofsocialalignments. PCT psychology implies that the reorganization of anyindividual'sperceptual systems is a constantly ongoing and essentiallyrandom process. Asexperiences accumulate, each individual continuallymodifies his or her view ofthe world in idiosyncratic ways. Thus, socialalignments of these individuallyheld reference standards always tendtoward disintegration, as peopleindependently reorganize theirperceptions. Something akin to an entropyprinciple guarantees that atleast a few of the individual participants in anyalignment will always bedrifting away from adherence to the socially acceptedstandards, so thatover time the alignment will start to crumble unlessperiodicallyreinforced by rituals or other uniformities of experience whichmight helpto repair the alignment. With any alignment, there is a tradeoffbetween the size of the group and theextent of conformity. The morepeople involved, the greater the number ofidiosyncratic ways of perceivingthe world that will have to be accommodated,and the cruder or moreabstract must be the character of the alignment, if itis to be achieved atall. A small primary group, with frequent interactionsover an extendedtime, can achieve a substantial convergence in theirperceptions of the world. Large groups cannot be so close-knit. In sum, theamount of power asocial structure can achieve is limited both by the tradeoffbetween groupsize and closeness of alignment and by the entropy which besetsalignmentsas individuals continually reorganize their perceptions. Theseemergent properties of social alignments, then, give structure tothesocial environment in which individuals attempt to control theirownperceptions. Further work applying PCT to sociology could take as apoint ofdeparture the insight that power-producing alignments are anessential propertyof all social structures and go on to explore theimplications of alignmentsfor such phenomena as social inequality,deviance, intergroup relations, andsocial change. Just as PCT provides an opportunity for sociologists to rethinktheirassumptions about power and control, it can shed some light onothersociological problems as well. For example, students of culturemight gain newinsights by defining culture as the sharing of closelysimilar perceptualcontrol systems, a necessary precondition for thecreation of social power byalignment of internal standards and goals.There are already many points ofcontact between PCT and various strands ofcurrent sociological thought. The fact that several of the prominentsociologists who have taken the lead inexploring control theory (e.g.,Burke, Heise, McPhail, Tucker) come from thesymbolic interactionisttradition is hardly an accident. In a classic article,Shibutani (1968)pointed out that George Herbert Mead's analysis of "the act"can be seen asan anticipation of modern negative feedback models (see alsoBuckley1967:94ff.). McPhail (1989; 1991:192-198) has elaborated on theconnectionbetween Mead's theories and PCT, and McPhail's work illustratesthepossibilities for convergence between the PCT model and thesymbolicinteractionist depiction of people as creativeactors. Ironically, in spite of immensestylistic differences in vocabulary andapproach, PCT also has someaffinity with insurgent movements in sociology likepostmodernism,feminism, and discourse analysis. The PCT constructivistepistemology (seevon Glasersfeld 1986) resonates strongly with the "epistemicsuspicion"(Seidman 1991:135; see also Brown 1987) postmodernists display abouttheepistemological realism embodied in conventional theories of sociology.ThePCT model of individuals as creators of unique perceptual worlds alsofitsunexpectedly well with attempts by discourse and rhetorical analysts(e. g.,Gilbert and Mulkay 1984; Latour 1987) to dismantle scientificclaims ofobjectivity, as well as with the insistence of feminist theorists(e.g., Smith1990) on the legitimacy of multiple points of view. Moreover,PCT, in spite ofits scientific models and jargon, joins with the morehumanistic branches ofsociology in reasserting the importance of theindividual and criticizingtraditional sociology's overemphasis onenvironmental constraint and themonolithic nature of the socialenvironment. By the same token, sociologists have much to contribute totheinterdisciplinary effort to extend and elaborate the PCT model. As itstandsalready, the PCT model of human psychology is very strongly social.Thehighest levels of perceptual control in the hierarchical PCT system arealmostentirely social in their origins and maintenance, depending as theydo onreference values expressed through the quintessentially social mediumoflanguage. The link between the social and psychological is obvious inthe formof the principles and system concepts which provide referencevalues at thesehighest levels of control. Even the lower levels ofperception, which may seemto be focused entirely on the physical world,regard a world which has beenalmost completely transformed by theoperations of other social actors. Nearlyevery blade of grass, every rockthat we see in the course of our daily liveshas been put into place andmust be maintained there by the efforts of someperson or group of peoplein the service of controlling their perceptions ofone or another socialstructure, be it a family, a business corporation, someother organization,or the state. Thus, the perceptual worlds we construct forourselves aresocial in their essence, and sociological analysis informed byPCT can helpshow how these worlds work. Acknowledgments: I am grateful toThomas Bourbon, Dag Forssell, DavidGoldstein, Robert Grey, ChristopherHunter, Lisa Kelly-Wilson, Richard Marken,Barry Markovsky, Clark McPhail,William T. Powers, Philip Runkel, DavidSchweingruber, Marci Sortor,Russell Tabbert, Charles Tucker, Greg Williams,and to the editors of thisjournal for their perceptive comments on earlierdrafts of thisarticle. Bates, Frederick L., and Clyde C. Harvey. 1975. TheStructure of SocialSystems. New York: Gardner Press. Berger,Peter, and Thomas Luckmann. 1966. The Social Construction ofReality.New York: Doubleday. Bourbon, W. Thomas. 1989. "A Control-TheoryAnalysis of Interference DuringSocial Tracking." Pp. 235-251 inVolitional Action: Conation andControl, edited by Wayne A.Hershberger. Amsterdam: North-Holland. Bourbon, W. Thomas. 1990."Invitation to the Dance: Explaining the VarianceWhen Control SystemsInteract." American Behavioral Scientist34(1):95-105. Bourbon,W. Thomas, Kimberly E. Copeland, Vick R. Dyer, Wade K. Harman, andBarbaraL. Mosley. 1990. "On the Accuracy and Reliability of PredictionsbyControl-System Theory." Perceptual and Motor Skills71:1331-1338. Brown, Richard Harvey. 1987. Society as Text: Essayson Rhetoric, Reason,and Reality. Chicago: University of ChicagoPress. Buckley, Walter. 1967. Sociology and Modern SystemsTheory. EnglewoodCliffs, NJ: Prentice Hall. Burke, Peter J.1991a. "Attitudes, Behavior, and the Self." Pp.189-208 in TheSelf-Society Dynamic, edited by Judith Howard and PeterCallero.Cambridge: Cambridge University Press. Burke, Peter J. 1991b."Identity Processes and Social Stress"American Sociological Review56(December): 836-849. Burke, Peter J., and Donald C. Reitzes.1991. "An Identity Theory Approach toCommitment." Social PsychologyQuarterly 54(3):239-251. Collins, Randall. 1988. TheoreticalSociology. San Diego: HarcourtBrace Jovanovich. Collins, Randall.1992. Sociological Insight: An Introduction toNon-Obvious Sociology.Second Edition. New York: Oxford UniversityPress. Cziko, GaryA. 1992. "Purposeful Behavior as the Control of Perception:Implicationsfor Educational Research." EducationalResearcher21(9):10-18,27. Fararo, Thomas J. 1989. TheMeaning of General Theoretical Sociology.New York: CambridgeUniversity Press. Fararo, Thomas J., and John Skvoretz. 1984."Institutions as ProductionSystems." Journal of MathematicalSociology 10:117-182. Frye, Marilyn. 1983. The Politics ofReality: Essays in Feminist Theory.Trumansburg, NY: Crossing. Gibbons, Hugh. 1984. "Justifying Law: An Explanation of the DeepStructureof American Law." Law and Philosophy 3:165-279. Gibbs,Jack P. 1989. Control: Sociology's Central Notion. Urbana,IL:University of Illinois Press. Gilbert, G. Nigel, and Michael Mulkay.1984. Opening Pandora's Box: ASociological Analysis of Scientists'Discourse. Cambridge: CambridgeUniversity Press. Goffman, Erving.1963. Behavior in Public Places: Notes on the SocialOrganization ofGatherings. New York: Free Press. Goffman, Erving. 1974.Frame Analysis. New York: Harper Colophon. Hagan, John, A. R.Gillis, and John Simpson. 1985. "The Class Structure ofGender andDelinquency: Toward a Power-Control Theory of Common DelinquentBehavior."American Journal of Sociology 90:1151-1178. Hagan, John, A. R.Gillis, and John Simpson. 1990. "Clarifying and ExtendingPower-ControlTheory." American Journal of Sociology 95(4):1024-1037. Hagan,John, John Simpson, and A. R. Gillis. 1987. "Class in the Household:APower-Control Theory of Gender and Delinquency." American JournalofSociology 92:788-816. Heise, David R. 1977. "Social Action asthe Control of Affect."Behavioral Science 22:163-177. Heise,David R. 1979. Understanding Events: Affect and the ConstructionofSocial Action. New York: Cambridge University Press. Hershberger,Wayne A. 1986. "An Approach Through the Looking-Glass."AnimalLearning & Behavior 14:443-451. Hochschild, Arlie Russell.1979. "Emotion Work, Feeling Rules, and SocialStructure." AmericanJournal of Sociology 85:551-575. Hochschild, Arlie Russell. 1983.The Managed Heart: Commercialization ofHuman Feeling. Berkeley,University of California Press. Jordan, J. Scott, and Wayne A.Hershberger. 1989. "The Behavioral Illusion:Misperception of VolitionalActs." Pp. 371-386 in Volitional Action:Conation and Control,edited by Wayne A. Hershberger. Amsterdam:North-Holland. Latour,Bruno. 1987. Science in Action: How to Follow ScientistsandEngineers Through Society. Cambridge: Harvard UniversityPress. Lincoln, Bruce. 1989. Discourse and the Construction ofSociety:Comparative Studies of Myth, Ritual, and Classification. NewYork: OxfordUniversity Press. Lord, Robert G., and Mary C. Kernan.1987. "Scripts as Determinants ofPurposeful Behavior in Organizations."Academy of Management Review 12:265-277. March, James G., andHerbert A. Simon. 1958. Organizations. NewYork:Wiley. Marken, Richard S. 1980. "The Cause of Control Movementsin a Tracking Task."Perceptual and Motor Skills 51:755-758.(Reprinted in Mind Readings:Experimental Studies of Purpose byRichard S. Marken. Gravel Switch, KY:Control Systems Group, 1992.) Marken, Richard S. 1982. "Intentional and Accidental Behavior: AControlTheory Analysis." Psychological Reports 50:647-650.(Reprinted inMind Readings: Experimental Studies of Purpose byRichard S. Marken.Gravel Switch, KY: Control Systems Group,1992.) Marken, Richard S. 1983. "'Mind Reading': A Look at ChangingIntentions."Psychological Reports 53:267-270. Marken, Richard S.1985. "Selection of Consequences: Adaptive Behavior fromRandomReinforcement." Psychological Reports 56:379-383. (ReprintedinMind Readings: Experimental Studies of Purpose by Richard S.Marken.Gravel Switch, KY: Control Systems Group, 1992.) Marken,Richard S. 1986. "Perceptual Organization of Behavior: AHierarchicalControl Model of Coordinated Action." Journal ofExperimentalPsychology--Human Perception and Performance 12:267-276.(Reprinted inMind Readings: Experimental Studies of Purpose byRichard S. Marken.Gravel Switch, KY: Control Systems Group,1992.) Marken, Richard S. 1988. "The Nature of Behavior: Control asFact andTheory." Behavioral Science 33:196-206. (Reprinted inMindReadings: Experimental Studies of Purpose by Richard S.Marken. GravelSwitch, KY: Control Systems Group, 1992.) Marken,Richard S. 1989. "Behavior in the First Degree." Pp. 299-314inVolitional Action: Conation and Control, edited by WayneA.Hershberger. Amsterdam: North-Holland. (Reprinted in MindReadings:Experimental Studies of Purpose by Richard S. Marken. GravelSwitch, KY:Control Systems Group, 1992.) Marken, Richard S. 1990."Spreadsheet Analysis of a Hierarchical ControlSystem Model of Behavior."Behavioral Research Methods, Instruments &Computers22(4):349-359. (Reprinted in Mind Readings: ExperimentalStudiesof Purpose by Richard S. Marken. Gravel Switch, KY: ControlSystemsGroup, 1992.) Marken, Richard S. 1991. "Degrees of Freedom inBehavior." PsychologicalScience 2(2):92-100. (Reprinted inMind Readings: Experimental Studiesof Purpose by Richard S.Marken. Gravel Switch, KY: Control SystemsGroup, 1992.) Marken,Richard S., and William T. Powers. 1989a. "Levels of IntentioninBehavior." Pp. 409-430 in Volitional Action: Conation andControl,edited by Wayne A. Hershberger. Amsterdam:North-Holland. Marken, Richard S., and William T. Powers. 1989b."Random-WalkChemotaxis: Trial and Error as a Control Process."Behavioral Neuroscience103:1348-1355. (Reprinted in MindReadings: Experimental Studies ofPurpose by Richard S. Marken.Gravel Switch, KY: Control Systems Group,1992.) McPhail, Clark.1989. "Meadian Versus Neo-Meadian Theories of Mind."SymbolicInteraction 12:43-51. McPhail, Clark. 1991. The Myth of theMadding Crowd. New York:Aldine de Gruyter. McPhail, Clark, andRonald T. Wohlstein. 1986. "Collective Locomotion asCollectiveBehavior." American Sociological Review 51: 447-463. McPhail,Clark, and Charles W. Tucker. 1990. "Purposive CollectiveAction."American Behavioral Scientist 34(1):81-94. Mehan, Hugh,and Houston Wood. 1975. The Reality of Ethnomethodology.New York:Wiley. Miller, James Grier. 1978. Living Systems. New York:McGraw-Hill. Pavloski, Raymond. 1989a. "A Control SystemApproach to CardiovascularReactivity: Behavioral Models That Behave."Psychophysiology26(4):468-481. Pavloski, Raymond. 1989b."The Physiological Stress of ThwartedIntentions." Pp. 215-232 inVolitional Action: Conation and Control,edited by Wayne A.Hershberger. Amsterdam: North-Holland. Pavloski, Raymond P., Gerard T.Barron, and Mark A. Hogue. 1990."Reorganization: Learning and Attentionin a Hierarchy of Control Systems."American Behavioral Scientist34(1):32-54. Petrie, Hugh G. 1981. The Dilemma of Enquiry andLearning. Chicago:University of Chicago Press. Plooij, Frans X.1984. The Behavioral Development of Free-LivingChimpanzee Babies andInfants. Norwood, NJ: Ablex. Powers, William T. 1973.Behavior: The Control of Perception.New York: Aldine deGruyter. Powers, William T. 1978. "Quantitative Analysis of PurposiveSystems: SomeSpadework at the Foundations of Scientific Psychology."PsychologicalReview 85:417-435. Reprinted in Powers, William T.1989a.Living Control Systems: Selected Papers of William T.Powers. GravelSwitch, KY: Control Systems Group. Powers, WilliamT. 1979a. "A Cybernetic Model for Research in HumanDevelopment."Pp. 11-66 in A Cybernetic Approach to the Assessment ofChildren: Towarda More Humane Use of Human Beings, edited by Mark N.Ozer. Boulder,CO: Westview. Reprinted in Powers, William T. 1989a.LivingControl Systems: Selected Papers of William T. Powers. GravelSwitch,KY: Control Systems Group. Powers, William T. 1979b."Cybernetics and the Assessment ofChildren." ASC Cybernetics Forum9:57-65. Powers, William T. 1979c. "The Nature of Robots: Part1. DefiningBehavior." Byte 4(June):132-144. Powers, William T.1979d. "The Nature of Robots: Part 2. SimulatedControl System."Byte 4(July):134-152. Powers, William T. 1979e. "TheNature of Robots: Part 3. A CloserLook at Human Behavior." Byte4(August):94-116. Powers, William T. 1986. "Interaction: The Controlof Perception.Commentary on a Paper by J. M. White" CSR Working PaperNo. 86-4.Edmonton: Center for Systems Research, University ofAlberta. Powers, William T. 1989a. Living Control Systems:Selected Papersof William T. Powers. Gravel Switch, KY: ControlSystems Group. Powers, William T. 1989b. "QuantitativeMeasurement of Volition: APilot Study." Pp. 315-332 in VolitionalAction: Conation and Control,edited by Wayne A. Hershberger.Amsterdam: North-Holland. Powers, William T. 1992. Living ControlSystems II: Selected Papers ofWilliam T. Powers. Gravel Switch, KY:Control Systems Group. Richardson, George P. 1991. Feedback Thoughtin Social Science and SystemsTheory. Philadelphia: University ofPennsylvania Press. Robertson, Richard J., and William T. Powers(Editors). 1989. Introductionto Modern Psychology: The Control-Theory View. Gravel Switch, KY:Control Systems Group. Runkel,Philip J. 1990a. Casting Nets and Testing Specimens.NewYork: Praeger. Runkel, Philip J. 1990b. "Research Method forControl Theory."American Behavioral Scientist34(1):14-23. Scheff, Thomas J. 1990. Microsociology:Discourse, Emotion, and SocialStructure. Chicago: University ofChicago Press. Scott, Marvin B., and Stanford M. Lyman. 1968."Accounts." AmericanSociological Review 33:46-62. Seidman,Steven. 1991. "The End of Sociological Theory: The PostmodernHope."Sociological Theory 9(2):131-146. Shibutani, Tamotsu. 1968. "ACybernetic Approach to Motivation." Pp. 330-336in Modern SystemsResearch for the Behavioral Scientist: A Sourcebook,edited by WBuckley. Chicago: Aldine. Simon, Herbert A. 1973. "The Organizationof Complex Systems." Pp. 3-27 inHierarchy Theory: The Challenge ofComplex Systems, edited by Howard H.Pattee. New York: GeorgeBraziller. Skinner, B. F. 1974. About Behaviorism. New York:Knopf. Smith, Dorothy E. 1990. The Conceptual Practices of Power:A FeministSociology of Knowledge. Boston: Northeastern UniversityPress. Stryker, Sheldon 1987. "The Vitalization of SymbolicInteractionism."Social Psychology Quarterly 50:83-94. Thoits,Peggy A. 1985. "Self-Labeling Processes in Mental Illness: The RoleofEmotional Deviance." American Journal of Sociology91:221-249. Vallacher, Robin R., and Daniel M. Wegner. 1987. "WhatDo People ThinkThey're Doing? Action Identification and Human-Behavior."PsychologicalReview 94(1):3-15. von Bertalanffy, Ludwig. 1967.Robots, Men, and Minds; Psychology in theModern World. New York:G. Braziller. von Glasersfeld, Ernst. 1986. "Steps in the Constructionof `Others' and`Reality": A Study in Self-Regulation." Pp. 107-116 inPower, Autonomy,Utopia: New Approaches toward Complex Systems,edited by Robert Trappl.New York: Plenum. Wartenburg, Thomas E. 1990.The Forms of Power: From Domination toTransformation.Philadelphia: Temple University Press. White, James M. 1986."Interaction: The Control of Perception. Reply to aComment by W. T.Powers." CSR Working Paper No. 86-4. Edmonton:Center for SystemsResearch, University of Alberta. Wiggins, Beverly, and David R. Heise.1987. "Expectations, Intentions, andBehavior: Some Tests of AffectControl Theory." Journal of MathematicalSociology 13:153-169.(Reprinted in Analyzing Social Interaction:Advances in Affect ControlTheory, edited by Lynn Smith-Lovin and David R.Heise. New York:Gordon and Breach, 1988.)APCT Definition ofCoercion
Coercion refers to the threat of force. One persondissuades another fromdoing something by somehow reminding him of thepossibility that force may beused. For example, one might warn off anintruder by assuming a fightingposture or displaying a weapon. Putting itin PCT terms,APCT Definition ofInfluence
All of three of these strategies are more effectivewhen combined withinfluence. We say one person has influenced anotherwhen the first has beenable to tell the second what to think or do. Wealso describe it as influencewhen the second person imitates the behaviorof the first. A single PCTdefinition for influence covers both thoseevents.
This examinationof strategies interpersonal control from a PCT standpointsupports theclaim that our attempts to exert power over other people aresometimes"paradoxical" and that power itself may often be "an illusion"(Collins1992:60-85). Whatever the strategy employed, it seems that control ofaperson's perceptions always remains with the person, and thus that nopersoncan achieve direct control of another's behavior. While force caninterferewith perceptual control, thus preventing action, and force takento the extremecan kill a person, this analysis suggests that force is theleast effectivemethod of getting someone to do anything positive. All ofthe other strategieswe have examined--coercion, incentives, andinfluence--rely on the complianceof the person supposedly under controland are ineffective if target person sochooses. AlignmentasSocial Power
The concept of social alignment provides leveragefor understanding theprocesses by which power is achieved, as well as,more generally, forunderstanding the ways that human control systems canaggregate into socialstructures of all kinds. We may define the conceptof alignment for PCTpurposes by saying that whenever two or more peopleperceive themselves toshare the same reference values, an alignment hasformed. McPhail and Tucker(1990; McPhail 1991) have outlined the ways inwhich alignments can develop atthe micro level as co-present individualsadopt similar reference signals byorienting themselves to the samephysical disturbances or by talking to eachother or by listening to athird party. Presumably, macro-scale alignments canarise as theaggregation of such micro events. Groupsand Alignment
Thedefinition of social power as alignment makes it clear that power isagroup, not an individual, phenomenon, and that groups differ in the powertheyexhibit. The power achieved by a collection of individuals who aligntheirreference standards belongs to the whole group, not to anyindividual. Everygroup inherently possesses some degree of power, becauseof and to the extentof the group's success in aligning the referencestandards of the participants.Obviously, groups riven by disagreement areless powerful than those operatingaround a strong consensus. A collectionof individuals who fit a socialcategory but have not aligned their ownreference standards do not evenconstitute a group in this sense. Ingeneral, the larger the number ofindividuals participating in the groupalignment, the more powerful the group;and the more thoroughly theindividuals of a group have achieved alignment (interms of numbers andperceptual levels of standards), the more powerful thegroup. Power,Perception, andEmotion
The equation of social power with alignment of referencestandards raises thepossibility of misalignment or incomplete alignment,an important practicalproblem in group life. Power is often hard to gaugein interactions becausealignments are not directly observable. We have noway of looking directlyinto other people's heads to see if they trulyshare our reference standards.At best, we must resort to inferences basedon their overt behavior. But PCTimplies that such inferences are oftenfaulty; in particular, the basicprinciple of PCT is that people controltheir perceptions in the face ofunpredictable disturbances by varyingtheir output. In other words, theperceptual input remains constant, whilethe behavioral output varies. Thus,the "same" overt behavior can atdifferent times be serving differentstandards, while "different" behaviorsmay serve the same standard. Thisunavoidable ambiguity presents aproblem, not only for social scientificobservers, but also for the personengaged in practical action. One can hardlymake a accurate assessment ofthe power supporting one's line of action in asocial situation withoutknowing who among those present does or does not shareone's goals.
A reconceptualization ofsocial power as alignment of reference standardsreveals the deficienciesin many of our common-sense images of power,especially those which dependon images of power as the interpersonal controlof other people's actions.We often think of the leader of a group (the personwith the mostinfluence) as possessing the group's power, but this idea is "anillusion"(Collins 1992:85). Should it happen that a group falls out ofalignmentwith each other or with the leader, the group's power hastherebydispersed. Leaders can lead only as long as followers follow.Leaders, ofcourse, do well to foster the illusion that they personallyhold the group'spower; such mystification serves the purposes of anyleader who wishes to exertmaximal influence, and leaders may seek toimpose rituals which convey thefalse perception that they personally are"in control." The leader, however,who personally succumbs to the illusionruns the risk of being disillusionedwhen events change or new leadersappear.FurtherImplications of DefiningPower as Alignment
Finally, this definition of power as alignmentaccounts for the often-observedconservatism of power structures and alsoreveals some inherent limitations onpower. Once perceptual alignment of alarge number of individuals has beenachieved, making changes in thatalignment may be difficult. As I arguedabove, no single individual islikely to have much influence on a widespreadalignment. Even people inprominent leadership positions in a group cannot besure of getting theattention of everyone participating in the alignment whenthey issuecommands or suggestions, nor can they guarantee that everyone willchooseto go along with a proposed change in reference levels. This accountsforthe frustration frequently felt by leaders such as politiciansinsupposedly powerful positions. In order to bring the members of a largegroupinto a new alignment, a leader may have to package the proposedchange in thecrudest, most oversimplified terms--in the manner oftelevision ads forcandidates. When the powerful alignment is notconnected to any organizedgroup but is spread more widely across thepopulation, as for instance arepatterns of institutionalized racism andsexism, the difficulties of initiatingsubstantial changes in the alignmentare even greater. In this sense, the morewidespread the alignment, themore conservative and resistant to change it islikely to be.
This attempt to use perceptual controltheory as a way of exploring andredefining the control of other people'sbehavior has shown that issues ofcontrol are more complex and paradoxicalthan most sociologists have heretoforeassumed. The key insight ofPCT--that people behave by controlling their ownperceptions--demolishesthe easy assumption that one person can control thebehavior of another,although from a more macro perspective we have seen howstructures ofsocial power do arise from the interactions of humanperceptual-controllersand how individuals might well perceive that othersexert control overthem.
Anderson,Elijah. 1978. A Place on the Corner. Chicago: University ofChicagoPress.
REFERENCES