Gouldner, Alvin W. 1960. "The Norm of Reciprocity: A PreliminaryStatement." American Sociological Review 25: 161-178.




Washington University at St. Louis

The manner in which the concept of reciprocity is implicated infunctional theory is explored, enabling a reanalysis of the conceptsof "survival" and "exploitation." The need to distinguish between theconcepts of complementarity and reciprocity is stressed. Distinctionsare also drawn between (1) reciprocity as a pattern of mutuallycontingent exchange of gratifications, (2) the existential or folkbelief in reciprocity, and (3) the generalized moral norm ofreciprocity. Reciprocity as a moral norm is analyzed; it ishypothesized that it is one of the universal "principal components"of moral codes. As Westermarck states, "To requite a benefit, or tobe grateful to him who bestows it, is probably everywhere, at leastunder certain circumstances, regarded as a duty. This is a subjectwhich in the present connection calls for special consideration."Ways in which the norm of reciprocity is implicated in themaintenance of stable social systems are examined.

"There is no duty more indispensable than that of returning akindness," says Cicero, adding that "all men distrust one forgetfulof a benefit."--Men have been insisting on the importance ofreciprocity for a long time. While many sociologists concur in thisjudgment, there are nonetheless few concepts in sociology whichremain more obscure and ambiguous. Howard Becker, for example, hasfound this concept so important that he has titled one of hisbooks Man in Reciprocity and has even spoken of man as Homoreciprocus, all without venturing to present a straightforwarddefinition of reciprocity. Instead Becker states, "I don't propose tofurnish any definition of reciprocity; if you produce some, they willbe your own achievements." (1)

Becker is not alone in failing to stipulate formally the meaningof reciprocity, while at the same time affirming its primeimportance. Indeed, he is in very good company, agreeing with L. T.Hobhouse, who held that "reciprocity . . . is the vital principle ofsociety," (2) and is a key intervening variable through which sharedsocial rules are enabled to yield social stability. Yet Hobhousepresents no systematic definition of reciprocity. While hardly anyclearer than Hobhouse, Richard Thurnwald is equally certain of thecentral importance of the "principle of reciprocity": this principleis almost a primordial imperative which "pervades every relation ofprimitive life" (3) and is the basis on which the entire social andethical life of primitive civilizations presumably rests. (4) GeorgSimmel's comments go a step further, emphasizing the importance ofreciprocity not only for primitive but for all societies. Simmelremarks that social equilibrium and cohesion could not exist without"the reciprocity of service and return service," and that 'allcontacts among men rest on the schema of giving and returning theequivalence." (5)

Were we confronted with only an obscure concept, which we had noreason to assume to be important, we might justifiably consign it tothe Valhalla of intellectual history, there to consort eternally withthe countless incunabula of sociological ingenuity. Howeverconvenient, such a disposition would be rash, for we can readily notethe importance attributed to the concept of reciprocity by suchscholars as George Homans, Claude Levi-Strauss, and Raymond Firth,(6) as well as by such earlier writers as Durkheim, Marx, Mauss,Malinowski, and von Wiese, to name only a few masters.

Accordingly, the aims of this paper are: (1) to indicate themanner in which the concept of reciprocity is tacitly involved in butformally neglected by modern functional theory; (2) to clarify theconcept and display some of its diverse intellectual contents, thusfacilitating its theoretical employment and research utility; and (3)to suggest concretely ways in which the clarified concept providesnew leverage for analysis of the central problems of sociologicaltheory, namely, accounting for stability and instability in socialsystems.


My concern with reciprocity developed initially from a criticalreexamination of current functional theory, especially the work ofRobert Merton and Talcott Parsons. The fullest ramifications of whatfollows can best be seen in this theoretical context. Merton'sfamiliar paradigm of functionalism stresses that analysis must beginwith the identification of some problematic pattern of humanbehavior, some institution, role, or shared pattern of belief. Mertonstipulates clearly the basic functionalist assumption, the way inwhich the problematic pattern is to be understood: he holds that the"central orientation of functionalism" is "expressed in the practiceof interpreting data by establishing their consequences for largerstructures in which they are implicated." (7) The functionalist'semphasis upon studying the existent consequences, the ongoingfunctions or dysfunctions, of a social pattern may be betterappreciated if it is remembered that this concern developed in apolemic against the earlier anthropological notion of a "survival."The survival, of course, was regarded as a custom held to beunexplainable in terms of its existent consequences or utility andwhich, therefore, had to be understood with reference to itsconsequences for social arrangements no longer present.

Merton's posture toward the notion of a social survival is bothpragmatic and sceptical. He asserts that the question of survivals islargely an empirical one; if the evidence demonstrates that a givensocial pattern is presently functionless then it simply has to beadmitted provisionally to be a survival. Contrariwise, if no suchevidence can be adduced "then the quarrel dwindles of its ownaccord." (8) It is in this sense that his position is pragmatic. Itis also a sceptical position in that he holds that "even when suchsurvivals are identified in contemporary literate societies they seemto add little to our understanding of human behavior or the dynamicsof social change. . . ." (9) We are told, finally, that "thesociologist of literate societies may neglect survivals with noapparent loss." (10)

This resolution of the problem of survivals does not seem entirelysatisfactory, for although vital empirical issues are involved thereare also important questions that can only be clarifiedtheoretically. Merton's discussion implies that certain patterns ofhuman behavior are already known to be, or may in the future be shownto be, social survivals. How, then. can these be explained in termsof functional theory? Can functional theory ignore them on thegrounds that they are not socially consequential? Consequential ornot, such social survivals would in themselves entail patterns ofbehavior or belief which are no less in need of explanation than anyother. More than that, their very existence, which Merton conceivespossible, would seem to contradict the "central orientation" offunctional theory.

Functionalism, to repeat, explains the persistence of socialpatterns in terms of their ongoing consequences for existent socialsystems. If social survivals, which by definition have no suchconsequences, are conceded to exist or to be possible, then it wouldseen that functionalism is by its own admission incapable ofexplaining them. To suggest that survivals do not help us tounderstand other patterns of social behavior is beside the mark. Thedecisive issue is whether existent versions of functional theory canexplain social survivals, not whether specific social survivals canexplain other social patterns.

It would seem that functionalists have but one of two choices:either they must dogmatically deny the existence or possibility offunctionless patterns (survivals), and assert that all socialbehavior is explainable parsimoniously on the basis of the samefundamental functionalist assumption, that is, in terms of itsconsequences for surrounding social structures; or, more reasonably,they must concede that some social patterns are or may be survivals,admitting that existent functional theory fails to account for suchinstances. In the latter case, functionalists must develop furthertheir basic assumptions on the generalized level required. I believethat one of the strategic ways in which such basic assumptions can bedeveloped is by recognizing the manner in which the concept ofreciprocity is tacitly involved in them and by explicating theconcept's implications for functional theory.

The tacit implication of the concept of reciprocity in functionaltheory can be illustrated in Merton's analysis of the latentfunctions of the political machine in the United States. Mertoninquires how political machines continue to operate, despite the factthat they frequently run counter to both the mores and the law. Thegeneral form of his explanation is to identify theconsequences of the machine for surrounding structures and todemonstrate that the machine performs positive functions which are atthe same time not adequately fulfilled by other existing patterns andstructures." (11) It seems evident, however, that simply to establishits consequences for other social structures provides no answer tothe question of the persistence of the political machine. (12) Theexplanation miscarries because no explicit analysis is made of thefeedback through which the social structures or groups, whose needsare satisfied by the political machine, in turn "reciprocate" andrepay the machine for the services received from it. In this case,the patterns of reciprocity, implied in the notion of the"corruption" of the machine, are well known and fully documented.

To state the issue generally: the demonstration that A isfunctional for B can help to account for A's persistence only if thefunctional theorist tacitly assumes some principle of reciprocity. Itis in this sense that some concept of reciprocity apparently has beensmuggled into the basic but unstated postulates of functionalanalysis. The demonstration that A is functional for B helps toaccount for A's own persistence and stability only on two relatedassumptions: (1) that B reciprocates A's services, and (2)that B's service to A is contingent upon A's performance ofpositive functions for B. The second assumption, indeed, is oneimplication of the definition of reciprocity as a transaction. UnlessB's services to A are contingent upon the services provided by A, itis pointless to examine the latter if one wishes to account for thepersistence of A.

It may be assumed, as a first approximation, that a social unit orgroup is more likely to contribute to another which provides it withbenefits than to one which does not; nonetheless, there are certaingeneral conditions under which one pattern may provide benefits forthe other despite a lack of reciprocity. An important case ofthis situation is where power arrangements constrain the continuanceof services. If B is considerably more powerful than A, B may force Ato benefit it with little or no reciprocity. This social arrangement,to be sure, is less stable than one in which B's reciprocitymotivates A to continue performing services for B, but it ishardly for this reason sociologically unimportant.

The problem can also be approached in terms of the functionalautonomy (13) of two units relative to each other. For example, B mayhave many alternative sources for supplying the services that itnormally receives from A. A, however, may be dependent upon B'sservices and have no, or comparatively few, alternatives.Consequently, the continued provision of benefits by one pattern,(14) A, for another, B, depends not only upon (1) the benefits whichA in turn receives from B, but also on (2) the power which Bpossesses relative to A, and (3) the alternative sources of servicesaccessible to each, beyond those provided by the other. In short, anexplanation of the stability of a pattern, or of the relationshipbetween A and B, requires investigation of mutually contingentbenefits rendered and of the manner in which this mutual contingencyis sustained. The latter, in turn, requires utilization of twodifferent theoretical traditions and general orientations, onestressing the significance of power differences and the otheremphasizing the degree of mutual dependence of the patterns orparties involved.

Functional theory, then, requires some assumption concerningreciprocity. It must however avoid the "Pollyanna Fallacy" whichoptimistically assumes that structures securing "satisfactions" fromothers will invariably be "grateful" and will always reciprocate.Therefore it cannot be merely hypostatized that reciprocity willoperate in every case; its occurrence must, instead, be documentedempirically. Although reciprocal relations stabilize patterns, itneed not follow that a lack of reciprocity is socially impossible orinvariably disruptive of the patterns involved. Relations with littleor no reciprocity may, for example, occur when power disparitiesallow one party to coerce the other. There may also be specialmechanisms which compensate for or control the tensions which arisein the event of a breakdown in reciprocity. Among such compensatorymechanisms there may be culturally shared prescriptions of one-sidedor unconditional generosity, such as the Christian notion of "turningthe other cheek" or "walking the second mile," the feudal notion of"noblesse oblige," or the Roman notion of "clemency." Theremay also be cultural prohibitions banning the examination of certaininterchanges from the standpoint of their concrete reciprocity, asexpressed by the cliche, "It's not the gift but the sentiment thatcounts." The major point here is that if empirical analysis fails todetect the existence of functional reciprocity, or finds that it hasbeen disrupted, it becomes necessary to search out and analyze thecompensatory arrangements that may provide means of controlling theresultant tensions, thereby enabling the problematic pattern toremain stable.



Thus far reciprocity has been discussed as a mutually contingentexchange of benefits between two or more units, as if it were an "allor none" matter. Once the problem is posed in this way, however, itis apparent that reciprocity is not merely present or absent but is,instead, quantitatively variable--or may be treated as such. Thebenefits exchanged, at one extreme, may be identical or equal. At theother logical extreme, one party may give nothing in return for thebenefits it has received. Both of these extremes are probably rare insocial relations and the intermediary case, in which one party givessomething more or less than that received, is probably more commonthan either of the limiting cases.

Having cast the problem of reciprocity in these quantitativeterms, there emerges an important implication for the question ofsocial survivals. The quantitative view of reciprocity. Thesefunctionalists made the cogent of the earlier notion of asurvival. It may now be seen that there a survival was tacitlytreated as one of the limiting cases of reciprocity, that is, one inwhich a pattern provides nothing in exchange for the benefitsgiven it.

The polemical opposition of the earlier functionalists to thisview of a survival rests implicitly on an unqualified principle ofreciprocity. These functionalists made the cogent assumption that asocial pattern which persists must be securing satisfaction of itsown needs from certain other patterns. What was further and moredubiously assumed, however, was that if this pattern continues to be"serviced" this could only be because it reciprocally provided somegratifications to its benefactors. In the course of the polemic, thequestion of the degree of such gratification-- the relation betweenits output and input--became obscured. To the early functionalists,the empirical problem became one of unearthing the hiddencontributions made by a seeming survival and, thereby, showing thatit is not in fact functionless. In effect, this enjoined thefunctionalist to exert his ingenuity to search out the hiddenreciprocities for it was assumed that there must be somereciprocities somewhere. This led, in certain cases, as AudreyRichards states, to "some far-fetched explanations. . . ." (15)

If, however, it had been better understood that compensatorymechanisms might have been substituted for reciprocity, or that powerdisparities might have maintained the "survival" despite its lack ofreciprocity, then many fruitful problems may well have emerged. Aboveall, the early functionalists neglected the fact that a survival isonly the limiting case of a larger class of social phenomena, namely,relations between parties or patterns in which functional reciprocityis not equal. While the survival, defined as theextreme case of a complete lack of reciprocity may be rare,the larger class of unequal exchanges, of which survivals area part, is frequent. The tacit conception of survivals as entailingno reciprocity led the early functionalists to neglect the largerclass of unequal exchanges. It is this problem which thefunctionalist polemic against survivals has obscured to the presentday.


It was, however, not only the functionalist polemic against theconcept of survivals that obscured the significance and inhibited thestudy of unequal exchanges. A similar result is also produced by thesuspicion with which many modern sociologists understandably regardthe concept of "exploitation." This concept of course is central tothe traditional socialist critique of modern capitalism. In the nownearly-forgotten language of political economy, "exploitation" refersto a relationship in which unearned income results from certain kindsof unequal exchange.

Starting perhaps with Sismondi's notion of "spoliation," andpossibly even earlier with the physiocrat's critique of exchange asintrinsically unproductive, the concept of exploitation can be tracedfrom the work of the Saint-Simonians to that of Marx and Proudhon.(16) It is also present in Veblen's notion of the Vested Interestwhich he characterizes as "the right to something for nothing" or, inother words, as institutionalized exploitation. Even after theemergence of sociology as a separate discipline the concept ofexploitation appears in the works of E. A. Ross, (17) von Wiese, andHoward Becker. (18) As it passed into sociology, however, the conceptwas generalized beyond its original economic application. Ross andBecker-von Wiese, for example speak of various types of exploitation:economic, to be sure, but also religious, "egotic " and sexual.However, just as the concept of exploitation was being generalizedand made available for social analysis, it almost disappeared fromsociological usage.

"Almost disappeared" because there remains one area inwhich unabashed, full-scale use of the concept is made bysociologists. This is in the study of sexual relations. As Kanin andHoward remark, "It has been the practice to speak of exploitationwhen males were found to have entered sexual liaisons with women ofcomparative lower status." (19) Kingsley Davis also uses the notionof exploitation implicitly in his discussion of the incest taboo,remarking that ". . . father- daughter incest would put the daughterin a position of subordination. While she was still immature thefather could use his power to take advantage of her." (20) What Davisis saying is that one function of the incest taboo is to preventsexual exploitation. He goes on to add that "legitimate sexualrelations ordinarily involve a certain amount of reciprocity. Sex isexchanged for something equally valuable." (21) This is aninteresting commentary, first, because Davis is quite clear abouttreating exploitation in the context of a discussion of reciprocity;and second, because he explicitly uses a notion of reciprocity in astrategic way even though it is not systematically explored elsewherein his volume, once again illustrating the tendency to use theconcept and to assume its analytic importance without giving itcareful conceptualization. (22)

The continued use of the concept of exploitation in sociologicalanalyses of sexual relations stems largely from the brilliant work ofWillard Waller on the dynamics of courtship. Waller's ambivalentcomments about the concept suggest why it has fallen intosociological disrepute. "The word exploitation is by no means adesirable one," explains Waller, "but we have not been able to findanother which will do as well. The dictionary definition ofexploitation as an 'unfair or unjust utilization of another' containsa value judgment, and this value judgment is really a part of theordinary sociological meaning of the term." (23) In short, theconcept of exploitation may have become disreputable because itsvalue implications conflict with modern sociology's effort to placeitself on a value-free basis, as well as because it is a conceptcommonly and correctly associated with the critique of modern societyemphasized by the political left. But the concept need not beused in such an ideological manner; it can be employed simply torefer to certain transactions involving an exchange of things ofunequal value. It is important to guarantee that the ordinary valueimplications of a term do not intrude upon its scientific use. It isalso important, however, to prevent our distaste for the ideologicalimplications of exploitation from inducing a compulsive and equallyideological neglect of its cognitive substance.

The unsavory implications of the concept of exploitation havenot excluded it from studies of sexual relations, althoughalmost all other specializations in sociology eschew it. Why this isso remains a tempting problem for the sociology of knowledge, butcannot be explored here. In the present context, the importantimplications are the following: If the possible sexual exploitationof daughters by fathers gives rise, as Davis suggests, to mechanismsthat serve to prevent this, then it would seem that other types ofexploitation may also be controlled by other kinds of mechanisms.These may be no less important and universal than the incest taboo.If the exploitation of women by men (or men by women) is worthy ofsociological attention, then also worth studying is the exploitationof students by teachers, of workers by management or union leaders,of patients by doctors, (24) and so on. If the notion ofexploitation, in a value-free sense, is useful for the analysis ofsexual relations then it can be of similar aid in analyzing manyother kinds of social relations.

Doubtless "exploitation" is by now so heavily charged withmisleading ideological resonance that the term itself can scarcely besalvaged for purely scientific purposes and will, quite properly, beresisted by most American sociologists. This is unimportant. Perhapsa less emotionally freighted--if infelicitous--term such as"reciprocity imbalance" will suffice to direct attention once againto the crucial question of unequal exchanges.

In any event, the present analysis of reciprocity opens uplong-neglected questions, yielding a new perspective on the relationbetween functional theory and the concepts of "survival" and'exploitation." In the latter case, moreover, intimations emerge ofsome of the ways in which two diverse theoretical traditions containsurprising convergences.

These two traditions are, first, that which is commonly ifquestionably (25) held to begin with Comte, was developed byDurkheim, and reaches its fullest current expression in the work ofParsons. The second tradition, while often ideologically distortednevertheless retains significant sociological substance derives fromMarx and Engels, was developed by Kautsky, and ended in Bukharin. Thelatent convergence between these two schools involves the implicitstress that each gives to reciprocity, albeit to polar ends of itscontinuum.

The "Comteian" tradition, of course, approached reciprocitythrough its emphasis on the division of labor, viewed as a majorsource of social cohesion. Characteristically focusing on the problemof social instability and change, rather than stability and cohesion,the "Marxian" tradition emphasized the opposite end of reciprocitynamely, exploitation. This, I suspect, is one of the major butoverlooked convergences in the history of sociological theory.

This latent convergence becomes most evident in Durkheim'slectures on "Professional Ethics and Civic Morals." (26) Durkheimcontends that the existence of social classes, characterized bysignificant inequalities, in principle makes it impossible for "just"contracts to be negotiated. This system of stratification, Durkheimargues, constrains to an unequal exchange of goods and services,thereby offending the moral expectations of people in industrialsocieties. The exploitation rendered possible by notable disparitiesof power among the contracting parties encourages a sense ofinjustice which has socially unstabilizing consequences. Thus bothDurkheim and Marx use a concept of "exploitation" for analyzingsocial instabilities. Durkheim, however, adds an important elementthat was systematically neglected by Marx, namely, that unequalexchanges of goods and services are socially disruptive because theyviolate certain pervasive values. But the specific nature of thisvalue element is never fully confronted and explored by Durkheim; wemust here take as problematic what Durkheim took as given.


First, however, the question of the meaning of the concept ofreciprocity should be reexamined. Consideration of some of the waysin which the reciprocity problem is treated by Parsons helps todistinguish reciprocity from other cognate concepts. "It is inherentin the nature of social interaction," writes Parsons, "that thegratification of ego's need-dispositions is contingent on alter'sreaction and vice versa." (27) Presumably, therefore, if thegratification of either party's needs is not contingent upon theother's reactions, the stability of their relation is undermined.This, in turn, implies that if a social system is to be stable theremust always be some "mutuality of gratification." (28) Social systemstability, then, presumably depends in part on the mutuallycontingent exchange of gratifications, that is, on reciprocity asexchange.

This, however, remains an insight the implications of which arenever systematically explored. For example, the implications ofdifferences in the degree of mutuality or in the symmetry ofreciprocity are neglected. Again, while the concept of "exploitation"assumes central importance in Parsons' commentary on thepatient-doctor relation, it is never precisely defined, examined, andlocated in his general theory.

One reason for Parsons' neglect of reciprocity is that he, likesome other sociologists, does not distinguish it from the concept ofcomplementarity. Parsons uses the two concepts as if they aresynonymous (29) and, for the most part, centers his analysis oncomplementarity to the systematic neglect of reciprocity rigorouslyconstrued. The term complementarity, however, is itself an ambiguousone and is not, in all of its meanings, synonymous with reciprocity.Complementarity has at least four distinct meanings: (30)

Complementarity-1 may mean that a right (x) of Ego against Alterimplies a duty (-x) of Alter to Ego. Given the often vague use of theterm "right," it is quite possible that this proposition, in oneaspect, is only an expansion of some definition of the concept"right." To that degree, of course, this is simply an analyticproposition. The interesting sociological questions, however, ariseonly when issues of empirical substance rather than logicalimplication are raised. For example, where a group shares a beliefthat some status occupant has a certain right, say the right of awife to receive support from her husband, does the group in fact alsoshare a belief that the husband has an obligation to support thewife? Furthermore, even though rights may logically or empiricallyimply duties, it need not follow that the reverse is true. In otherwords, it does not follow that rights and duties are alwaystransitive. This can be seen in a second meaning of complementarity.

Complementarity-2, may mean that what is a duty (-x) of Alter toEgo implies a right (x) of Ego against Alter. On the empirical level,while this is often true, of course, it is also sometimes false. Forexample, what may be regarded as a duty of charity or forebearance,say a duty to "turn the other cheek," need not be sociallydefined as the right of the recipient. While a man may beregarded as having an unconditional obligation to tell the truth toeveryone, even to a confirmed liar, people in his group might notclaim that the liar has a right to have the truth told him.

The other two meanings of complementarity differ substantially.Complementarity-3 may mean that a right (x) of Alter against Egoimplies a duty (-y) of Alter to Ego. Similarly, complementarity-4 maymean that a duty (-x) of Ego to Alter implies a right (y) of Egoagainst Alter.

In these four implications of complimentarity--sometimes calledreciprocal rights and obligations there are two distinctive types ofcases. Properly speaking, complementarity refers only to thefirst two meanings sketched above, where what is a right of Egoimplies an obligation of Alter, or where a duty of Alter to Egoimplies a right of Ego against Alter. Only the other two meanings,however, involve true instances of reciprocity, for only inthese does what one party receives from the other require somereturn, so that giving and receiving are mutually contingent.

In short, complementarity connotes that one's rights are another'sobligations, and vice versa. Reciprocity, however, connotesthat each party has rights and duties. This is more than an analyticdistinction: it is an empirical generalization concerning rolesystems the importance of which as a datum is so elemental that it iscommonly neglected and rarely made problematic. The Englishphilosopher MacBeath suggests that this empirical generalization maybe accounted for by the principle of reciprocity. (31) This wouldseem possible in several senses, one of which is that, were thereonly rights on the one side and duties on the other, there need be noexchange whatsoever. Stated differently, it would seem that there canbe stable patterns of reciprocity qua exchange only insofar aseach party has both rights and duties. In effect, then,reciprocity has its significance for role systems in that ittends to structure each role so as to include both rights andduties. It is now clear, at any rate, that reciprocity is by no meansidentical with complementarity and that the two are confused only attheoretic peril.


Renewing the effort to clarify the diverse meanings ofreciprocity, we turn to Malinowski's seminal contribution. This ismost fully elaborated in his Crime and Custom, (32) whichopens with the following question: Why is it that rules of conduct ina primitive society are obeyed. even though they are hard andirksome? Even under normal conditions, the savage's compliance withhis moral code is at best partial, conditional, and evasive. These,says Malinowski, are the elementary facts of ethnography, andconsequently we cannot assume that the savage's conformity is dueonly to his awe and reverence for traditional custom, or that heslavishly and spontaneously complies with its dictates.

Above all, Malinowski rejects the assumption that it is the sacredauthority of the moral code, or the "collective conscience," whichaccounts for the conformity given it. It is to this anti-Durkheimianpoint that he directs the brunt of his polemic. Conformity, saysMalinowski, is not sanctioned "by a mere psychological force, but bya definite social machinery. . . ." (33) Thus Malinowski expresslyrejects a psychological account of conformity and seeks instead adistinctively sociological explanation. (34) This he finds in the"principle of reciprocity."

One of Malinowski's central theses holds that people oweobligations to each other and that, therefore, conformity withnorms is something they give to each other. He notes, forexample, that almost every religious or ceremonial act is regarded asan obligation between groups and living individuals, and not only tothe immortal gods. For Malinowski, therefore, one meaning ofreciprocity refers to the interlocking status duties which people oweone another. Thus he speaks of reciprocity as taking place "within astanding partnership, or as associated with definite social ties orcoupled with mutuality in non-economic matters." (35)

Reciprocity also entails a "mutual dependence and [is] realized inthe equivalent arrangement of reciprocal services. . . ." (36) Herereciprocity is conceived as the complement to and fulfillment of thedivision of labor. It is the pattern of exchange through which themutual dependence of people, brought about by the division of labor,is realized. Reciprocity, therefore, is a mutually gratifying patternof exchanging goods and services.

As noted above, Malinowski speaks of reciprocity as involving anexchange of equivalent services; he further stresses this byinsisting that "most if not all economic acts are found to belong tosome chain of reciprocal gifts and counter-gifts, which in the longrun balance, benefiting both sides equally." (37) For Malinowski,then, the exchange of goods and services is not only mutuallygratifying but is equally so, "in the long run."

Speaking of the reciprocal exchange of vegetables and fish betweeninland communities and fishing villages, Malinowski remarks thatthere is a "system of mutual obligations which forces the fishermanto repay whenever he has received a gift from his inland partner, andvice versa. Neither partner can refuse, neither may stint, neithershould delay." (38) This is seen to be related to the group'sexistential beliefs about reciprocity. That is, men are not regardedas blindly involving themselves in reciprocal transactions; they areviewed as having some presentiment of the consequences of reciprocityand of its breakdown. In this vein, Malinowski writes: "Though nonative, however intelligent, can formulate this state of affairs in ageneral abstract manner, or present it as a sociological theory, yeteveryone is well aware of its existence and in each concrete case hecan foresee the consequences." (39) More specifically, it seems to beimplied that people believe that (a) in the long run the mutualexchange of goods and services will balance out; or (b) ifpeople do not aid those who helped them certain penalties will beimposed upon them; or © those whom they have helped canbe expected to help them; or (d) some or all of these.

It is clear that two basically different elements were caught upin Malinowski's "principle of reciprocity." One of these is a set ofsentiments or existential folk beliefs about reciprocity. The otheris a mutually contingent exchange of benefits or gratifications. (Thelatter conception converges, though it is not completely identical,with the ecological concept of symbiosis.) There is, however, a thirdanalytically distinct element which, if implicit in Malinowski,remained murky. This is a value element, the same value thatDurkheim, as mentioned earlier, invoked but did not clarify. LikeDurkheim, Malinowski never fully disentangles it from the otherelements.

In the exchanges between the fishing and the inland villages,cited above, we may suggest that each side lives up to itsobligations, not simply because of constraints imposed by thedivision of labor with its attendant mutual dependency, but alsobecause the partners share the higher level moral norm: "Youshould give benefits to those who give you benefits." Notethat this norm does not simply make it unconditionally imperative,say, for the fisherman to give the inland gardeners fish. I referhere not to the specific obligation to give fish but rather toa general obligation to repay benefits.

In sum, beyond reciprocity as a pattern of exchange and beyondfolk beliefs about reciprocity as a fact of life, there is anotherelement: a generalized moral norm of reciprocity which definescertain actions and obligations as repayments for benefitsreceived.

Malinowski frequently seems to confuse this general norm with theexistence of complementary and concrete status rights and duties. Itis theoretically necessary, however, to distinguish specific statusduties from the general norm. Specific and complementary duties areowed by role partners to one another by virtue of the sociallystandardized roles they play. These may require an almostunconditional compliance in the sense that they are incumbent on allthose in a given status simply by virtue of its occupancy. Incontrast, the generalized norm of reciprocity evokes obligationstoward others on the basis of their past behavior. In the first case,Ego's obligations to Alter depend upon Ego's status vis-a-vis Alter;in the second case, Ego's obligation toward Alter depend upon whatAlter has done for Ego. There are certain duties that people owe oneanother, not as human beings, or as fellow members of a group or evenas occupants of social statuses within the group but, rather, becauseof their prior actions. We owe others certain things because of whatthey have previously done for us, because of the history of previousinteraction we have had with them. It is this kind of obligationwhich is entailed by the generalized norm of reciprocity.


Contrary to some cultural relativists it can be hypothesized thata norm of reciprocity is universal. As Westermarck stated, "Torequite a benefit, or to be grateful to him who bestows it, isprobably everywhere, at least under certain circumstances, regardedas a duty." (40) A norm of reciprocity is, I suspect, no lessuniversal and important an element of culture than the incest taboo,although, similarly, its concrete formulations may vary with time andplace.

Specifically, I suggest that a norm of reciprocity, in itsuniversal form, makes two interrelated, minimal demands: (1) peopleshould help those who have helped them, and (2) people should notinjure those who have helped them. Generically, the norm ofreciprocity may be conceived of as a dimension to be found in allvalue systems and, in particular as one among a number of"Principal Components" universally present in moral codes. (The taskof the sociologist, in this regard parallels that of the physicistwho seeks to identify the basic particles of matter, the conditionsunder which they vary, and their relations to one another.)

To suggest that a norm of reciprocity is universal is not, ofcourse, to assert that it is unconditional. Unconditionality would,indeed, be at variance with the basic character of the reciprocitynorm which imposes obligations only contingently, that is, inresponse to the benefits conferred by others. Moreover, suchobligations of repayment are contingent upon the imputed valueof the benefit received. The value of the benefit and hence thedebt is in proportion to and varies with--among other things--theintensity of the recipient's need at the time the benefit wasbestowed ("a friend in need . . ."), the resources of the donor ("hegave although he could ill afford it"), the motives imputed to thedonor ("without thought of gain"), and the nature of the constraintswhich are perceived to exist or to be absent ("he gave of his ownfree will . . ."). Thus the obligations imposed by the norm ofreciprocity may vary with the status of the participantswithin a society.

Similarly, this norm functions differently in some degree indifferent cultures. In the Philippines, for example, thecompadre system cuts across and pervades the political,economic, and other institutional spheres. Compadres are bound by anorm of reciprocity. If one man pays his compadre's doctor's bill intime of need, for example, the latter may be obligated to help theformer's son to get a government job. Here the tendency to govern allrelations by the norm of reciprocity, thereby underminingbureaucratic impersonality, is relatively legitimate, hence overt andpowerful. In the United States, however, such tendencies are weaker,in part because friendship relations are less institutionalized.Nonetheless, even in bureaucracies in this country such tendenciesare endemic, albeit less legitimate and overt. Except in friendship,kinship, and neighborly relations, a norm of reciprocity is notimposed on Americans by the "dominant cultural profile," although itis commonly found in the latent or "substitute" culture structure inall institutional sectors, even the most rationalized, in the UnitedStates.

In otherwise contrasting discussions of the norm of reciprocityone emphasis is notable. Some scholars, especially, Homans, Thurwald,Simmel, and Malinowski, assert or imply that the reciprocity normstipulates that the amount of the return to be made is "roughlyequivalent" to what had been received. The problem of equivalence isa difficult but important one. Whether in fact there is a reciprocitynorm specifically requiring that returns for benefits received beequivalent is an empirical question. So, too, is the problemof whether such a norm is part of or distinct from a more generalnorm which simply requires that one return some (unspecified)benefits to benefactors. Logically prior to such empirical problems,however, is the question of what the meaning of equivalence would bein the former norm of equivalent reciprocity.

Equivalence may have at least two forms, the sociological andpsychodynamic significance of which are apt to be quite distinct. Inthe first case, heteromorphic reciprocity, equivalence may mean thatthe things exchanged may be concretely different but should be equalin value, as defined by the actors in the situation. In thesecond case, homeomorphic reciprocity, equivalence may mean thatexchanges should be concretely alike, or identical in form, eitherwith respect to the things exchanged or to the circumstances underwhich they are exchanged. In the former, equivalence calls for "titfor tat"; in the latter, equivalence calls for "tat for tat."Historically, the most important expression of homeomorphicreciprocity is found in the negative norms of reciprocity,that is, in sentiments of retaliation where the emphasis is placednot on the return of benefits but on the return of injuries, and isbest exemplified by the lex talionis. (41)

Finally, it should be stressed that equivalence in the above casesrefers to a definition of the exchangeables made by actors in thesituation. This differs of course, from holding that the thingsexchanged by people, in the long run, will be objectively equal invalue as measured by economists or other social scientists. Here,again, the adequacy of these conceptual distinctions will bedetermined ultimately by empirical test. For example, can we findreciprocity norms which, in fact, require that returns be equivalentin value and are these empirically distinguishable from normsrequiring that returns be concretely alike? Are these uni-dimensionalor multi-dimensional? Similarly, only research can resolve thequestion whether a norm of retaliation exists in any given group, isthe polar side of the norm of reciprocity, or is a distinctive normwhich may vary independently of the reciprocity norm. Theseconceptual distinctions only suggest a set of research possibilitiesand have value primarily as guides to investigation. (42)


As mentioned above, sociologists have sometimes confused thenotion of complementarity with that of reciprocity and have recentlytended to focus on the former. Presumably, the reason for this isbecause of the importance of complementarity in maintaining thestability of social systems. Clearly, if what one party deems hisright is accepted by the other as his obligation, their relation willbe more stable than if the latter fails to so define it. But if thegroup stabilizing consequences of complementarity are the basis ofits theoretical significance, then the same consideration underwriteswith equal potency the significance of reciprocity. For reciprocityhas no less a role in maintaining the stability of social systems.

Note that there are at least two ways, not merely one, in whichcomplementarity as such can break down. In the one case, Alter canrefuse to acknowledge Ego's rights as his own duties. In the othercase, however, Ego may not regard as rights that which Alteracknowledges as duties. The former is commonly viewed as theempirically more frequent and as the theoretically more significantcase. That this often seems to be taken as a matter of coursesuggests the presence of certain tacit assumptions about basic humandispositions. It seems to assume, as Aristotle put it, that peopleare more ready to receive than to give benefits. In short, itpremises a common tendency toward what used to be called "egoism," asalient (but not exclusive) concern with the satisfaction of one'sown needs.

This or some cognate assumption appears to be eminently reasonableand empirically justified. There can be no adequate systematicsociological theory which boggles at the issue; indeed, it is one ofthe many virtues of Parsons' work that it confronts the egoismproblem. His solution seems to be side-tracked, however, because hisoverwhelming focus on the problem of complementarity leads to theneglect of reciprocity. If assumptions about egoistic dispositionsare valid, however, a complementarity of rights and obligationsshould be exposed to a persistent strain, in which each party issomewhat more actively concerned to defend or extend his own rightsthan those of others. There is nothing in complementarity as suchwhich would seem able to control egoism.

One way out may be obtained by premising that socializationinternalizes complementary rights and obligations in persons, beforethey fully assume responsible participation in a social system. Evenif socialization were to work perfectly and so internalize suchrights and obligations, there still remains the question as to whatmechanism can sustain and reinforce these during full participationin the social system. The concept of complementarity takes mutuallycompatible expectations as given; it does not and cannot explain howthey are maintained once established. For this we need to turn to thereciprocities processes because these, unlike pure complementarity,actually mobilize egoistic motivations and channel them into themaintenance of the social system. Benthamite utilitarianism has longunderstood that egoism can motivate one party to satisfy theexpectations of the other, since by doing so he induces the latter toreciprocate and to satisfy his own. As Max Gluckman might put it withhis penchant for Hegelian paradox, there is an altruism in egoism,made possible through reciprocity.

Furthermore, the existential belief in reciprocity says somethinglike this, "People will usually help those who help them." Similarly,the norm of reciprocity holds that people should help thosewho help them and, therefore, those whom you have helped have anobligation to help you. The conclusion is clear: if you want to behelped by others you must help them, hence it is not only proper butalso expedient to conform with the specific status rights of othersand with the general norm. Both the existential belief in and thenorm of reciprocity enlist egoistic motivations in the service ofsocial system stability. (43)

A full analysis of the ways in which the whole reciprocitiescomplex is involved in the maintenance of social systems wouldrequire consideration of the linkages between each of its variouselements, and their relation to other general properties of socialsystems. There is no space for such consideration here. Instead, Iexamine only one part of the complex, namely, the generalizednorm of reciprocity, and suggest some of the ways in which itcontributes to social system stability.

If, following Parsons, we suppose that social systems are stableto the extent that Ego and Alter conform with one another'sexpectations, we are confronted with the problem of why menreciprocate gratifications. Parsons holds that once a stablerelation of mutual gratification has been established the system isself-perpetuating; presumably, no special mechanisms are necessary tomaintain it. Insofar as this is not simply postulated in analogy withthe principle of inertia in physics, apparently reciprocity isaccounted for by Parsons, and also by Homans, as a result of thedevelopment of a beneficent cycle of mutual reinforcement. That is,Ego's conformity with Alter's expectations reinforces Alter'sconformity with Ego's expectations, and so on.

This explanation of reciprocity qua transaction isparticularly strange in Parsons' case since he often stresses, buthere neglects, the significance of shared values as a source ofstability in social systems. So far as the question here is notsimply the general one of why men conform with the expectations ofothers but rather, the more specific problem of why theyreciprocate benefits, part of the answer would seem to be thatthey have commonly internalized some general moral norm. Inshort, the suggestion is that the motivation for reciprocity stemsnot only from the sheer gratification which Alter receives from Egobut also from Alter's internalization of a specific norm ofreciprocity which morally obliges him to give benefits to those fromwhom he has received them. In this respect, the norm ofreciprocity is a concrete and special mechanism involved in themaintenance of any stable social system.

Why should such a norm be necessary? Why is it that expedientconsiderations do not suffice to mobilize motivations to comply withother's expectations, thereby inducing them to provide reciprocalcompliances? One major line of analysis here would certainly indicatethe disruptive potentialities of power differences. Given significantpower differences egoistic motivations may seek to get benefitswithout returning them. (It is notable that Parsons fails to definethe power situation in his basic model of Ego-Alter equilibrium.) Thesituation is then ripe for the breakdown of reciprocity and for thedevelopment of system-disrupting exploitation. The norm ofreciprocity, however, engenders motives for returning benefits evenwhen power differences might invite exploitation. The norm thussafeguards powerful people against the temptations of their ownstatus: it motivates and regulates reciprocity as an exchangepattern, serving to inhibit the emergence of exploitative relationswhich would undermine the social system and the very powerarrangements which had made exploitation possible. (44)

As we have seen, Parsons stresses that the stability of socialsystems largely derives from the conformity of role partnersto each other's expectations, particularly when they do their duty toone another. This formulation induces a focus on conformity anddeviance, and the degrees and types of each. Presumably, the morethat people pay their social debts the more stable the social system.But much more than conformity and deviance are involved here.

The idea of the reciprocities complex leads us to the historicalor genetic dimension of social interaction. For example, Malinowski,in his discussion of the Kula Ring, carefully notes that the giftsgiven are not immediately returned and repayment may take as long asa year. What is the significance of this intervening time period? Itis a period governed by the norm of reciprocity in a double sense.First, the actor is accumulating, mobilizing, liquidating, orearmarking resources so that he can make a suitable repayment.Second, it is a period governed by the rule that you should not doharm to those who have done you a benefit. This is a time, then, whenmen are morally constrained to manifest their gratitude toward, or atleast to maintain peace with, their benefactors.

Insofar as men live under such a rule of reciprocity, when oneparty benefits another, an obligation is generated. The recipient isnow indebted to the donor, and he remains so until he repays.Once interaction is seen as taking place over time, we may note thatthe norm of reciprocity so structures social relations that, betweenthe time of Ego's provision of a gratification and the time ofAlter's repayment, falls the shadow of indebtedness. An adequateanalysis of the dynamics of social interaction is thus required to gobeyond the question of deviance from or conformity with the partiesobligations to one another. A second basic dimension needs to beexamined systematically, namely, the time period when there is anobligation still to be performed, when commitments which have beenmade are yet to be fulfilled.

These outstanding obligations, no less than those already givencompliance, contribute substantially to the stability of socialsystems. It is obviously inexpedient for creditors to break offrelationships with those who have outstanding obligations to them. Itmay also be inexpedient for debtors to do so because theircreditors may not again allow them to run up a bill of socialindebtedness. In addition, it is morally improper under thenorm of reciprocity to break off relations or to launch hostilitiesagainst those to whom you are still indebted.

If this conclusion is correct, then we should not only look formechanisms which constrain or motivate men to do their duty and topay off their debts. We should also expect to find mechanisms whichinduce people to remain socially indebted to each other andwhich inhibit their complete repayment. This suggests anotherfunction performed by the requirement of only roughequivalence of repayment that may be involved in one of the norms ofreciprocity. For it induces a certain amount of ambiguity as towhether indebtedness has been repaid and, over time, generatesuncertainty about who is in whose debt. (45) This all hinges,however, on a shared conception of the moral propriety of repaymentengendered by the norm of reciprocity.

Still another way in which the general norm of reciprocity isimplicated in the maintenance of social system stability is relatedto an important attribute of the norm, namely, its comparativeindeterminancy. Unlike specific status duties and like other generalnorms this norm does not require highly specific and uniformperformances from people whose behavior it regulates. For example,unlike the status duties of American wives, it does not call uponthem to cook and to take care of the children. Instead, the concretedemands it makes change substantially from situation to situation andvary with the benefits which one party receives from another.

This indeterminancy enables the norm of reciprocity to performsome of its most important system-stabilizing functions. Beingindeterminate, the norm can be applied to countless ad hoctransactions, thus providing a flexible moral sanction fortransactions which might not otherwise be regulated by specificstatus obligations. The norm, in this respect, is a kind of plasticfiller, capable of being poured into the shifting crevices of socialstructures, and serving as a kind of all-purpose moral cement.

Not only does the norm of reciprocity play a stabilizing role inhuman relations in the absence of a well developed system ofspecific status duties, but it contributes to social stability evenwhen these are present and well established. Status dutiesshape behavior because the status occupant believes them binding intheir own right; they possess a kind of prima facie legitimacyfor properly socialized group members. The general norm ofreciprocity, however, is a second-order defense of stability; itprovides a further source of motivation and an additional moralsanction for conforming with specific status obligations. Forexample, the employer may pay his workers not merely because he hascontracted to do so; he may also feel that the workman has earned hiswages. The housewife may take pains with her husband's meals notmerely because cooking may be incumbent on her as a wife; she mayalso have a particularly considerate husband. In each case, thespecific status duties are complied with not only because they areinherent in the status and are believed to be right in themselves,but also because each is further defined as a "repayment." Insum, the norm of reciprocity requires that if others have beenfulfilling their status duties to you, you in turn have an additionalor second-order obligation (repayment) to fulfill your status dutiesto them. In this manner, the sentiment of gratitude joins forces withthe sentiment of rectitude and adds a safety-margin in the motivationto conformity.

The matter can be put differently from the standpoint of potentialdeviance or non-conformity. All status obligations are vulnerable tochallenge and, at times, may have to be justified. If, for anyreason, people refuse to do their duty, those demanding compliancemay be required to justify their claims. Obviously, there are manystandardized ways in which this might be done. Involving the generalnorm of reciprocity is one way of justifying the more concretedemands of status obligations. Forced to the wall, the man demandinghis "rights," may say, in effect, "Very well, if you won't do thissimply because it is your duty, then remember all that I have donefor you in the past and do it to repay your debt to me." The norm ofreciprocity thus provides a second-order defense of the stability ofsocial systems in that it can be used to overcome incipient devianceand to mobilize auxiliary motivations for conformity with existentstatus demands. (46)


Two distinct points have been made about the social functions ofthe norm of reciprocity. One is that this norm serves a groupstabilizing function and thus is quite familiar in functionaltheory. The second point, however, is the view that the norm is notonly in some sense a defense or stabilizing mechanism but is alsowhat may be called a "starting mechanism." That is, it helps toinitiate social interaction and is functional in the early phases ofcertain groups before they have developed a differentiated andcustomary set of status duties.

In speaking of the norm of reciprocity as a "starting mechanism,"indeed in conceiving of starting mechanisms, we find ourselvesoutside the usual perspective of functional theory. Functional theorycommonly focuses on already-established, on-going systems, and on themechanisms by means of which an established social system is enabledto maintain itself. Although functional theory is concerned with theproblems of how individual actors are prepared by socialization toplay a role in social systems, its general theoretical models rarely,if ever, include systematic treatment of the beginnings of a socialsystem as such and, consequently, do not formally raise the questionof the nature of the mechanisms needed to start such a system. (47)

Every social system of course has a history, which means that ithas had its beginnings even if these are shrouded in antiquity.Granted that the question of origins can readily bog down in ametaphysical morass, the fact is that many concrete social systems dohave determinate beginnings. Marriages are not made in heaven, andwhether they end in divorce or continue in bliss, they have someidentifiable origins. Similarly, corporations, political parties, andall manner of groups have their beginnings. (Recent studies offriendship and other interpersonal relations in housing projects havebegun to explore this problem.)

People are continually brought together in new juxtapositions andcombinations, bringing with them the possibilities of new socialsystems. How are these possibilities realized? Is such realizationentirely a random matter? These are the kinds of questions that werefamiliar to the earlier students of "collective behavior," who, infocusing on crowds, riots, and rumors, were often primarily concernedwith investigating the development of groups in statunascendi. (48) Although this perspective may at first seemsomewhat alien to the functionalist, once it is put to him, he maysuspect that certain kinds of mechanisms, conducive to thecrystallization of social systems out of ephemeral contacts, will insome measure be institutionalized or otherwise patterned in anysociety. At this point he would be considering "starting mechanisms."In this way, I suggest, the norm of reciprocity provides one amongmany starting mechanisms.

From the standpoint of a purely economic or utilitarian model,(49) there are certain difficulties in accounting for the manner inwhich social interaction begins. Let us suppose two people or groups,Ego and Alter, each possesses valuables sought by the other. Supposefurther that each feels that the only motive the other has to conductan exchange is the anticipated gratification it will bring. Each maythen feel that it would be advantageous to lay hold of the other'svaluables without relinquishing his own. Furthermore, suppose thateach party suspects the other of precisely such an intention, perhapsbecause of the operation of projective or empathic mechanisms. Atleast since Hobbes, it has been recognized that under suchcircumstances, each is likely to regard the impending exchange asdangerous and to view the other with some suspicion. (50) Each maythen hesitate to part with his valuables before the other has firstturned his over. Like participants in a disarmament conference, eachmay say to other, "You first!" Thus the exchange may be delayed oraltogether flounder and the relationship may be prevented fromdeveloping.

The norm of reciprocity may serve as a starting mechanism in suchcircumstances by preventing or enabling the parties to break out ofthis impasse. When internalized in both parties, the normobliges the one who has first received a benefit to repay itat some time; it thus provides some realistic grounds for confidence,in the one who first parts with his valuables, that he will berepaid. Consequently, there may be less hesitancy in being the firstand a greater facility with which the exchange and the socialrelation can get underway.


I have limited this discussion of the norm of reciprocity to itsfunctions and its contribution to the stability of social systems,omitting examination of its dysfunctions and of the manner in whichit induces tensions and changes in social systems. That the normcommonly imposes obligations of reciprocity only "when the individualis able" to reciprocate does not guarantee agreement concerning theindividual's "ability." Furthermore there may be occasions whenquestions as to whether the individual's return is appropriate orsufficient (apart from whether it is equivalent) that arise by virtueof the absence of common yardsticks in terms of which giving andreturning may be compared. Moreover, the norm may lead individuals toestablish relations only or primarily with those who can reciprocate,thus inducing neglect of the needs of those unable to do so. Clearly,the norm of reciprocity cannot apply with full force in relationswith children, old people or with those who are mentally orphysically handicapped, and it is theoretically inferable that other,fundamentally different kinds of normative orientations will developin moral codes. I hope to explore these and related problems insubsequent discussions.



* Sections of this paper were read at the annual meeting of theAmerican Sociological Association, September, 1959. The author isindebted to Robert K. Merton, Howard S. Becker, John W. Bennett,Louis Schneider, and Gregory Stone for reading an earlier draft butknows of no adequate "reciprocity" for their many valuablesuggestions.

1. Howard Becker, Man in Reciprocity, New York: Prager,1956, p. 1.

2. L. T. Hobhouse, Morals in Evolution: A Study in ComparativeEthics, London: Chapman & Hall, 1956, First edition, 1906, p.12.

3. Richard Thurnwald, Economics in Primitive Communities,London: Oxford University Press, 1932, p. 106.

4. Ibid., p. 137. See also, Richard Thurnwald, "BanaroSociety: Social Organization and Kinship System of a Tribe in theInterior of New Guinea," Memoirs of the American AnthropologicalAssociation, 8, 1916; among other matters of relevance to theanalysis of reciprocity, Thurnwald's discussion here (p. 275) opensthe issue of the "exchange of women," which Levi-Strauss laterdeveloped.

5. Georg Simmel, The Sociology of Georg Simmel, translatedand edited by Kurt H. Wolff, Glencoe. Ill.: Free Press, 1950, p. 387.

6. See, respectively, George Homans, "Social Behavior asExchange," American Journal of Sociology," 68 (May, 1958) pp.597-606; C. Levi-Strauss, Les Structures elementaires de laparente, Paris: Presses Universitaires, 1949; and Raymond Firth,Primitive Polynesian Economy. New York: Humanities Press,1950.

7. R. K. Merton, Social Theory and Social Structure,Glencoe, Ill.: Free Press, 1957, pp 46-47.

8. Ibid., p. 33.

9. Ibid., p. 34.

10. Ibid.

11. Ibid, p. 73. Among the functions of the politicalmachine to which Merton refers are: the organization andcentralization of power so that it can be mobilized to satisfy theneeds of different groups, provision of personalized forms ofassistance for lower-class groups, giving political privileges andaid to business groups, and granting protection for illicit rackets.

12. An initial statement of this point is to be found in A. W.Gouldner, "Reciprocity and Autonomy in Functional Theory," in L.Gross, editor Symposium on Sociological Theory, Evanston,Ill.: Row, Peterson, 1959, pp. 241-270.

13. For fuller discussion of this concept, see Gouldner,ibid.

14. Use of terms such as "pattern" or "unit" is intended toindicate that the present discussion deliberately collapsesdistinctions between institutional, interpersonal, group, or rolereciprocities, treating them here under a single rubric for reasonsof space.

15. Raymond Firth, editor, Man and Culture: An Evaluation ofthe Work of Bronislaw Malinowski, New York: The Humanities Press,1957, p. 19.

16. The views of these and other analysts of exploitation are ablysummarized in C. Gide and C. Rist, A History of EconomicDoctrines, translated by R. Richards, Boston: Heath, revisededition, 1918.

17. See, e.g., E. A. Ross, New-Age Sociology, New York:Appleton-Century, 1940, esp. Chapter 9.

18. Note von Wiese and Becker's comment: "The Marxians trace thesocial process of exploitation to the 'capitalistic' economic order;their thesis is that capitalism creates exploitation. We, on theother hand, do not deny the existence of capitalistic exploitation,but it is for us only one of the forms which are found among thephenomena of exploitation. The destruction of capitalism will notsignalize the end of exploitation, but will merely prevent theappearance of some of its forms and will open up new opportunitiesfor others." L. von Wiese and Howard Becker, SystematicSociology, New York: Wiley, 1932, p. 700. It would seem that 20thcentury history amply confirms this view.

19. E. Kanin and D. H. Howard, "Postmarital Consequences ofPremarital Sex Adjustments," American Sociological Review, 23(October, 1958), p. 558. (My italics.)

20. Kingsley Davis, Human Society, New York: Macmillan,1949, p. 403.

21. Ibid., p. 404.

22. Note Davis's tendency to assume that legitimate sexualrelations entail an exchange of equal values even though hisprevious sentence indicates that there may be no more than "a certainamount of reciprocity" involved. The latter is a way of talking aboutunequal exchanges and thus implies that these occur ininstitutionalized and not only in illicit relations. This is animportant problem that cannot be developed here.

23. Willard Waller, The Family: A Dynamic Interpretation,revised by Reuben Hill, New York: Dryden, 1951, p. 163.

24. The point is not to stress, as Parsons does, the uniqueexploitability of the patient or the peculiar power of the physician,but to see this relationship as but one dramatic case of a largerclass of phenomena of basic theoretic significance which should beexplicitly dealt with in systematic theory rather than given onlyad hoc treatment in specific empirical contexts. See TalcottParsons, The Social System. Glencoe, Ill.: Free Press, 1951,p. 445.

25. The thesis that this is more mythological than real isdeveloped in my introduction Emile Durkheim, Socialism andSaint-Simon, translated by C. Sattler and edited by A. W.Gouldner, Yellow Springs: Antioch Press. 1958, esp. p. ix.

26. Emile Durkheim, Professional Ethics and Civic Morals,translated by C. Brookfield, Glencoe, Ill.: Free Press, 1958; seeesp. pp 209-214.

27. Parsons, op. cit., p. 21.

28. Talcott Parsons and Edward A. Shils, editors, Toward aGeneral Theory of Action, Cambridge: Harvard University Press,1951, p. 107.

29. Parsons' tendency to equate complementarity and reciprocitymay be illustrated by his comment that "Role expectations organize .. . the reciprocities, expectations, and responses to theseexpectations in the specific interaction systems of ego and one ormore alters. This reciprocal aspect must be borne in mind since theexpectations of an ego always imply the expectations of one or morealters. It is in this reciprocity or complementarity thatsanctions enter. . . ." Ibid., pp. 190-191 (my italics); seealso p. 105. The burden of Parsons' analysis attends to theconditions and consequences of complementarity, by which he meansthat a role player requires of himself what his role partner requiresof him. It is precisely for this reason that Parsons emphasizes thatvalues must be held in common by the actors, if their expectationsare to be compatible. The equation of reciprocity withcomplementarity is not peculiar to Parsons. It is evident in the workof other sociologists who sometimes speak of the rights andobligations in a pair of roles as "reciprocal" and other times as"complementary." And, like Parsons, others state that rights andduties, or role expectations, are always complementary.

30. The analysis here closely follows W. D. Ross, The Right andthe Good, Oxford: Clarendon Press, 1950.

31. Alexander MacBeath. Experiments in Living, London:Macmillan, 1952; see esp. Pp. 127 ff.

32. Bronislaw Malinowski, Crime and Custom in SavageSociety, London: Paul, Trench, Trubner, 1932.

33. Ibid., p. 55.

34. This, by the way, is why I cannot concur in Parsons' judgmentthat Malinowski never disentangled a social system level of analysisfrom an encyclopedic concept of culture. See Talcott Parsons"Malinowski and the Theory of Social Systems," in Man andCulture . . ., op. cit., pp. 53-70. Malinowski's Crimeand Custom transcends a clinical case analysis of specificprimitive societies and presents a generalized and basic contributionto the theory of social systems when it addresses itself to theproblem of reciprocity. Parsons, however, does not mention thesignificance of reciprocity in Malinowski's work and is able tosupport his claim that it ignores social system analysis only by thisnoteworthy omission. Parsons' neglect of the principle of reciprocityin Malinowski's work, it would seem, is consistent with his ownneglect of the distinction between reciprocity and complementarity.

35. Malinowski, op. cit., p. 39.

36. Ibid., p 55.

37. Ibid., p. 39.

38. Ibid., p. 22.

39. Ibid., p. 40. This is not to cay, however, thatMalinowski regards reciprocity qua transaction asalways intended by all the actors or as something of whichthey are always aware. In brief-- and I agree--there are both latentand manifest reciprocities.

40. Edward Westermarck. The Origin and Development of the MoralIdeas, London: MacMillan, 1908, Vol. 2, p. 154.

41. It is further indicative of our terminological difficulties inthis area that this is often what Piaget spoke of as "reciprocity."For example, ". . . reciprocity stands so high in the eyes of thechild that he will apply it even where to us it seems to border oncrude vengeance." J. Piaget, The Moral Judgment of the Child,New York: Harcourt, Brace, 1932, p. 216.

42. A further point that fuller discussion should develop concernsthe terms "roughly" equivalent. Use of the term "roughly," in onepart, indicates that a certain range of concrete behavior will beviewed by the actors as compliance with this reciprocity norm andthat more than one specific return will be acceptable and defined asequivalent. The norm of reciprocity qua equivalence is thuslike most other norms which also tolerate a range of variability. Thedemand for exact equality would place an impossible burden even onactors highly motivated to comply with the reciprocity norm and wouldyield endemic tensions. Conversely, a notion of "rough" equivalenceheld by the actors allows for easier compliance with the norm and canbe regarded as one of the mechanisms sustaining it. Recognition thatthe requirement is for "rough" equivalence, however, should not beallowed to obscure the fact that there may be a specific reciprocitynorm which does in fact call for equivalence. This would be adistinguishing feature of the hypothesized norm and should no more beconcealed by reference to a "rough" equivalent than should thedistinctive content of any other norm be obscured by the fact that avariable range of behaviors will be acceptable to those holding it.

43. I suppose that one can take two different attitudes towardthis transmutation of the base metal of egoism. One can deplore thesituation and say with Eliot:

"The last temptation is the greatest treason;
To do the right thing for the wrong reason."


Or one can adopt the older and perhaps sociologically wiser viewthat here, once more, "private vices make public benefits," andprovide an indispensable basis for the spontaneousself-regulation of social systems.

44. This line of analysis is further strengthened if we considerthe possibility that Ego's continued conformity with Alter'sexpectations may eventually lead Alter to take Ego's conformity for"granted" and thus lead Alter to reciprocate less for later acts ofconformity by Ego. In short, the value of Ego's conformity mayundergo an inflationary spiral in which his later conforming actionsare worth less than earlier ones, in terms of the reciprocities theyyield. As reciprocities tend to decline, the social system mayexperience mounting strain, either collapsing in apathy or beingdisrupted by conflict. In this connection, the general norm ofreciprocity may serve as a brake, slowing the rate at whichreciprocities decline or preventing them from declining beyond acertain (unknown) level, and thus contributing to the stability ofthe system. This is more fully developed in A. W. Gouldner,"Organizational Analysis," in R. K. Merton et al., editors.Sociology Today, New York: Basic Books, 1959, esp. pp. 423 ff.

45. An interesting case of a mechanism serving to create andmaintain outstanding obligations is part of the Vartan Bhanji, a formof ritual gift exchange in Pakistan and other parts of India. Eglar'sstudy of this pattern makes it clear that a fundamental rule ofVartan Bhanji is reciprocity, that a gift should be returned for agift, and a favor for a favor. It is also notable that the systempainstakingly prevents the total elimination of outstandingobligations. Thus, on the occasion of a marriage, departing guestsare given gifts of sweets. In weighing them out, the hostess may say,"These five are yours," meaning "these are a repayment for what youformerly gave me," and she then adds an extra measure, saying, "Theseare mine." On the next occasion, she will receive these back with anadditional measure which she later returns, and so on. See Z. EEglar, Vartan Bhanji: Institutionalized Reciprocity in a ChangingPunjab Village, Ph.D. thesis, Columbia University, 1958.

Other mechanisms for maintaining outstanding obligations may befound in cultural prescriptions which require men not to be overlyeager to repay their social obligations. It still seems to beunderstood that there is a certain impropriety in this, even if we donot go as far as Seneca in holding that "a person who wants to repaya gift too quickly with a gift in return is an unwilling debtor andan ungrateful person."

46. A cogent illustration of this is provided by William F. Whyte:"When life in the group runs smoothly, the obligations bindingmembers are not explicitly recognized. . . . It is only when therelationship breaks down that the underlying obligations are broughtto light. While Alec and Frank were friends I never heard either oneof them discuss the services he was performing for the other, butwhen they had a falling out . . . each man complained to Doc that theother was not acting as he should in view of the services which hadbeen done for him." Street Corner Society, Chicago: Universityof Chicago Press, 1945, p. 256.

47. Modern functionalism emerged in a world in which Newtonianmechanics was the overshadowing scientific achievement and a basicmodel for the development of social science. The Newtonian standpointwas not, of course, a cosmology concerned with the question ofplanetary origins but took the existent relations among planets asgiven. Today, however, two developments of global significanceencourage and perhaps require a shift in social perspectives. In one,rocket engineering, the question is raised as to how new, man-made,planets may be "shot" into stable orbits. Secondly, internationalpolitics require us to help "underdeveloped" countries to begin abeneficent cycle of capital accumulation which will beself-sustaining. In both instances, practical "engineering" problemsforcefully direct attention to the question of "starting mechanisms"and would seem likely to heighten dissatisfaction with generalsociological models that largely confine themselves to alreadyestablished systems.

48. I am indebted to Howard S. Becker for this and many otherinsights into what seemed to be the guiding impulses of the "ChicagoSchool" of collective behavior.

49. Some indications of the utilitarian approach to this problemmay be derived from the stimulating paper by T. C. Schelling, "AnEssay on Bargaining, American Economic Review, 46 (June,1956), pp. 281-306

50. Cf. M. Deutsch, "A Study of Conditions Affecting Cooperation,"New York: Research Center for Human Relations, 1955, p. 25, dittoed.


Back tothe Syllabus