Power The of Positive Sanctions David A. Baldwin World Politics, Vol. 24, No. 1. (Oct., 1971), pp. 19-38.
Stable URL: http://links.jstor.org/sici?sici=0043-8871%28197110%2924%3A1%3C19%3ATPOPS%3E2.0.CO%3B2-X World Politics is currently published by The Johns Hopkins University Press.
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/jhup.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact [e-mail].
http://www.jstor.org Mon Jun 11 12:15:33 2007
T H E POWER OF POSITIVE SANCTIONS
By DAVID A. BALDWINX
OLITICAL science has made valuable contributions to the progressive clarification of the concept of power since World War 11. In view of the attention political scientists have traditionally lavished on the concept of power, it seems fitting that they should help clarify it. Thanks to the efforts of such men as Harold Lasswell and Robert Dahl, many political scientists today are keenly aware of the need to define power in relational terms, to distinguish power relations from power resources, to specify scope, weight, domain, and so on.' There is, however, one distinction that is rarely considered by political scientists-that between positive and negative sanctions. The purpose of this paper is to clarify this distinction and show how and why it matters. It is not that political scientists have said wrong things about the role of positive sanctions in power relations; it is just that they have said little.2 Most of their discussions of power have focused on severe negative sanctions. Can one influence more flies with honey than with vinegar? Can one influence more Vietnamese with economic aid than with napalm? The literature of political science not only gives few clues to the answers, it often implies that such questions are not even worth asking. Dahl recognizes but understates the problem: "The existence of both negative and positive coercion is sometimes a source of confusion in political analysis, since writers often either confound the two or ignore positive c~ercion."~ Although it is not the purpose of this paper to survey the literature,
" T h e author would like to thank the following colleagues who provided helpful comments on earlier drafts of this paper: Michael Banks, Robert Kleck, Michael Smith, and Richard Winters. Harold D. Lasswell and Abraham Kaplan, Power and Society (New Haven 1950); Robert A. Dahl, "The Concept of Power," Behavioral Science, 11 (July 1957), 201-15. For a bibliography of the more important contributions to the literature on power see Robert A. Dahl. "Power," International Encyclopedia of the Social Sciences, XI1 . (New York 1968), ' 4 ~ 4 - ~ 5.. 2 Although the primary focus of this discussion is political science, other social science disci~lineshave also underemphasized positive sanctions. The recent International ~ n ~ y c l o ~ e of i a Social ~ G e n c e s ,0; example, contains an index entry for d the "threat" but none for "promise" and an article on "punishment'' but none on "reward." See also James T. Tedeschi, "Threats and Promises," in Paul Swingle, ed., T h e Structure of Conflict (New York 1970)~155-91. 3Robert A. Dahl, Modern Political Analysis (Englewood Cliffs, N.J. 1963), 51.
P
20
WORLD POLITICS
a brief look at the various ways of handling positive sanctions is in order. Among the more common approaches are the following. I. Explicit rejection.-Most political scientists pay so little attention to the distinction between positive and negative sanctions that their exclusion of positive sanctions from the concept of power is implicit rather than explicit. A few, however, have explicitly rejected positive sanctions in defining power.4 Two sociologists who have influenced political science have also excluded positive sanctions from their concept of power. Both Talcott Parsons5 and Peter M. Blaue have carefully distinguished positive from negative sanctions and have explicitly defined power in terms of negative sanctions. Blau suggests that Max Weber also excluded positive sanctions from his concept of power.7 Although Weber's concept of power seems to include both positive and negative sanctions, his concept of political power emphasizes the actual or threatened use of force.Yuch emphasis tends to focus attention on negative rather than positive sanctions in power relations. 2. Conversion.-Positive sanctions can be conceptually converted into negative ones. Thus Dahl offers a definition of power in terms of severe penalties and then observes that large rewards "can be made to operate" rather like severe penal tie^.^ Similarly, the concept of B's opportunity costs of noncompliance with A's demands blurs the distinction between rewards and penalties.'' Regardless of whether A promises B a reward of $100 for compliance or threatens him with a penalty of $roo for failure to comply, the opportunity costs to B of noncompliance are the same. 3. Matatis matandis.-Perhaps the lack of attention to positive
This group includes David Easton, T h e Political System (New York 1953), 143-44; Andrew S. McFarland, Power and Leadership in Pluralist Systems (Stanford 1969), 13-14; and Gerald Garvey, "The Domain of Politics," Western Political Quarterly, XXIII (March 1g70), 120-37. In the second edition of Modern Political Analysis, Dahl joins this group: Robert A. Dahl, Modern Political Analysis, 2nd ed. (Englewood Cliffs, N.J. 1970), 32-33. Talcott Parsons, "On the Concept of Political Power," Proceedings of the American Philosophical Society, CVII (June 1963), 232-62. &PeterM. Blau, Exchange and Power i n Social Life (New York 1964). Ibid., I 15. 3Max Weber, T h e Theory of Social and Economic Organization, translated by A. M. Henderson and Talcott Parsons, and edited with an introduction by Talcott Parsons (New York 1947), pp. 152-57. Dahl (fn. 3), 50-51. loJohn C. Harsanyi, "Measurement of Social Power, Opportunity Costs, and the Theory of Two-Person Bargaining Games," Behavioral Science, VII (January 1962), 67-80. Throughout this paper, A refers to the actor exerting or attempting to exert power, and B refers to the actor over whom A is exerting or attempting to exert power.
T H E POWER O F POSITIVE SANCTIONS
21
sanctions is partially explained by an implicit assumption that the distinction between positive and negative sanctions is not worth making. Even those who include both positive and negative sanctions in their concepts of power may believe that all or most generalizations about one are applicable to the other-if the proper assumptions are made." Since the relevance of generalizations about sticks to generalizations about carrots is so easy to see, why belabor the obvious? 4. Inconsistency.-One may explicitly include both positive and negative sanctions in his definition of power and then proceed to ignore the positive and accentuate the negative. References to power that are not applicable to positive sanctions are then likely to be passed off as mere "slips of the pen." When Dahl identifies several dimensions useful in measuring A's power, we hardly notice that he includes the degree of B's threatened deprivation and ignores the degree of B's promised reward.'' When Karl Deutsch defines power in terms of "expected capability to inflict sanction^,"^^ no harm is done. After all, one could talk of A inflicting rewards on B-although one never does. When Felix Oppenheim virtually ignores positive sanctions in his book-length study of power, it is all right because he includes a footnote which mentions "just in passing the promise of reward as another type of influence."14 When Bachrach and Baratz say that a necessary condition of a power relationship is B's perception of the threatened sanction as a deprivation,15or when they observe that the actual application of sanctions is an admission of defeat by A'" (a statement applicable only to negative sanctions), there is no reason for concern, because they have specifically included both rewards and penalties in their definition of sanctions.17 Any-but not all-of these examples can easily be dismissed as nothing more than unfortunate diction, a mere "slip of the pen." Scholarship pays little attention to "slips of the pen" on the assumption that they are random and tend to cancel out. When pens slip consistently in one direction, however, the effect is cumulative; the scholarly implications are different. In discussing the role of sanctions in power relations, the
l1 Kenneth E. Boulding appears to take this position in Conflict and Defense (New York 1962), 253-58. 12Dahl, "Power" (fn. I ) , 414. 13Karl W. Deutsch, The Nerves of Government (New York 1963), pp. 120-21; italics added. 14Felix E. Oppenheim, Dimensions of Freedom (New York 1961), 45. 15Peter Bachrach and Morton S. Baratz, "Decisions and Nondecisions: An Analytical Framework," American Political Science Review, LVII (September 1963), 634. laIbid., 636. l7Ibid., 634.
22
WORLD POLITICS
pens of political scientists often slip toward negative sanctions and almost never slip toward positive sanctions. No work published since 1945 has had a greater impact on the way political scientists think about power than Lasswell and Kaplan's Power and Society. Here too one finds a tendency to emphasize negative sanctions. Consider the following points: ( I ) a decision is defined as "a policy involving severe sanctions (deprivations)" (p. 74). (2) Power is defined as "participation in the making of decisions" (p. 75). (3) In differentiating power from influence in general, they focus on the "availability of sanctions when the intended effects are not forthcoming." "Power is a special case of the exercise of influence: it is the process of affecting policies of others with the help of (actual or threatened) severe deprivations for nonconformity with the policies intended" (p. 76). Since only negative sanctions are implemented (or threatened) for nonconformity, it appears that Lasswell and Kaplan are not referring to positive sanctions. (4) The word "deprivations" is often used as a synonym for "sanctions" (pp. 76, 84, 86). It is di6cult to believe that they intend to include both positive and negative sanctions in the term "deprivations," since they define a sanction as positive "when it enhances values for the actor to whom it is applied," and as negative "when it deprives him of values" (pp. 48-49, italics added). (5) Lasswell and Kaplan list control over B's "well-being," i.e., his physical health and safety, as a power base for A. Although this would seem to imply A's ability to add to B's well-being as well as to subtract from it, the forms of power based on control of well-being are decidedly negative-violence, terror, discipline, rape, brute force, brigandage, forced labor, and inquisition (p. 87). The discussion of power by Lasswell and Kaplan is clearly and consistently cast in terms of negative sanctions, that is, until one considers the last section of the chapter on power, the section labeled "Choice and Coercion" (pp. 97-102). This section explicitly incorporates both positive and negative sanctions into the concept of power. There seems to be no way to reconcile this section with the rest of the chapter on power. The value of this immensely useful book would be enhanced more by a frank recognition of this inconsistency than by pretending that it does not exist. Although D a h l l h t one time viewed Lasswell and Kaplan's concept of power as limited to negative sanctions, he later contended that a "close reading" of Power and Society "indicates that they include both negative and positive coercion
lsRobert A. Dahl and Charles E. Lindblom, Politics, Economics, a n d Welfare (New York 1g53), 106.
T H E POWER OF POSITIVE SANCTIONS
23
in their definition of power, though the inclusion of positive coercion is not obvious."1QI shall leave it to the reader to decide for himself whether close reading of Power and Society reveals obscure inclusion of positive sanctions or outright contradiction. Either way it seems that the handling of positive sanctions in this influential book leaves much to be desired. The two most important tasks of this paper lie ahead. First, a clear distinction between positive and negative sanctions must be drawn; and, second, the relevance of this distinction to political analysis must be shown.
Positive sanctions are defined as actual or promised rewards to B; negative sanctions are defined as actual or threatened punishments to B. Although these definitions appear simple enough, there are both conceptual and empirical difficulties in distinguishing between positive and negative sanctions. Some things take the form of positive sanctions, but actually are not: e.g., giving a bonus of $roo to a man who expected a bonus of $200, or promising not to kill a man who never expected to be killed in the first place. Likewise, some things take the form of negative sanctions, but actually are not: e.g., a threat to cut by $roo the salary of a man who expected his salary to be cut by $200, a threat to punch in the nose, next week, a man who knows he will be hanged at sunrise, or the beating of a masochist. Is withholding a reward ever a punishment? Always a punishment? Is withholding a punishment ever a reward? Always a reward? The answers depend on B's perceptions of the sit~ation.~' In order to distinguish rewards from punishments one must establish B's baseline of expectations at the moment A's influence attempt begins." This baseline is defined in terms of B's expected future value position, i.e., his expectations about his future position relative to the things he values. Positive sanctions, then, are actual or promised improvements in B's value position relative to his baseline of expectations.
l9 Dahl (fn. 3), 51. In the second edition of Modern Politica2 Analysis (fn. 4), p. 32, Dahl abandons this interpretation and returns to his previous position. N o explanation is offered; nor is there any indication that Dahl recognizes the contradictory nature of Lasswell and Kaplan's discussion. The second edition of Modern Political Analysis gives even less attention to positive sanctions than the first edition-a seeming impossibility. Compare Modern Political Analysis (1963), pp. 50-51 with Modern Political Analysis (1970), pp. 32-33. 2 0 Cf., John R. P. French, Jr. and Bertram Raven, "The Bases of Social Power," in Dorwin Cartwright, ed., Studies in Social Power (Ann Arbor 1959), 158. 21The concept of the baseline is drawn from Blau (fn. 6), 116.
24
WORLD POLITICS
Negative sanctions are actual or threatened deprivations relative to the same baseline. Whereas conceptual establishment of B's baseline is vital but not difficult, empirical establishment of the baseline is both vital and difficult. Three pitfalls await those who would distinguish the concept of positive from that of negative sanctions. The pitfalls concern B's perceptions, time, and conditional influence attempts. As Bachrach and Baratz have reminded us, explanations of power relations should specify from whose point of view the situation is being ~ i e w e d . ~ Y n any given power relationship, A may perceive himself as employing carrots, while B may perceive A as using sticks. Although many Americans perceive their foreign aid program in terms of positive sanctions, many recipients perceive it differently. There is also a danger that the outside observer, i.e., the political scientist, will substitute his own baseline for that of B, e.g., "if someone gave me a million dollars, I would regard it as a reward." The second pitfall concerns time and is illustrated by Dahl's discussion of positive coercion. After defining power in terms of negative sanctions, he observes that substantial rewards can be made to operate in the same way: "For if . . . [B] is offered a very large reward for compliance, then once his expectations are adjusted to this large reward, he suffers a prospective loss if he does not comply."23 The italicized words indicate that time is not being held constant. Only after B's expectations are adjusted, does he perceive withholding the reward as coercive. What Dahl has done here is to use two different baselines. In referring to negative sanctions, he uses the baseline existing at the moment of A's influence attempt, while his references to positive sanctions use the new baseline after B has taken account of A's influence attempt. Since the purpose of A's influence attempt is to shift B's baseline, i.e., to cause B to change the expected values associated with doing X, Dahl's treatment tends to conceal the dynamics of the influence process. In distinguishing carrots from sticks one must be careful to specify not only B's baseline of expectations, but also the point in time at which that baseline was established. It is important, however, to recognize that the baseline changes over time. Today's reward may lay the groundwork for tomorrow's threat, and tomorrow's threat may lay the groundwork for a promise on the day after tomorrow. Thomas Schelling's" discussions of "comBachrach and Baratz (fn. 15), 640-41. Dahl (fn. 3), 50-51; italics added. 24 Thomas C. Schelling, The Strategy of Conflict (Cambridge, Mass. 1960), 195-99; and Thomas C. Schelling, A r m s and Influence (New Haven 1966), 6991.
22
23
T H E POWER OF POSITIVE SANCTIONS
25
pellent threats" could be improved by recognition of this fact. The threat that compels, he says, often takes the form of administering the punishment until B acts, rather than if he acts." To call such a conditional commitment to withdraw punishment a "threat'' is counter to both common usage and the analysis presented above. Such situations could be more usefully described as ones in which A uses a negative sanction (the punishment) to lay the groundwork for the subsequent use of positive sanctions (the promise to withdraw the punishment if B complies). What A is doing in such situations is using the stick to shift B's baseline so as to make the subsequent promise of a carrot more attractive. A's offer to stop tipping the boat if B will row is unlikely to be perceived by B as a carrot unless A is actually tipping the boat at the time the offer is made. A tips the boat in order to shift B's expectation baseline, so that B will perceive the offer to stop tipping the boat as a reward. In his discussions of compellent threats Schelling blurs the distinction between positive and negative sanctions. Indeed, he turns the time sequence usually associated with threats around, so that a conditional commitment to punish and a conditional commitment to stop punishing are both called threats. Common usage, however, suggests a difference between offering to pull a thorn out of B's foot and a threat to stick a thorn in. The third pitfall is associated with conditional influence attempts, i.e., those in which A conditionally commits himself to reward or punish B for compliance or noncompliance." The problem is that it seems to be easier to distinguish rewards from punishments than to distinguish promises from threats. The possibility that withholding a reward may be regarded by B as a punishment tempts one to regard threats and promises as two sides of one coin. The argument runs as follows: "An unconditional commitment by A to reward (or punish) B regardless of whether he does X or not is not a promise (or threat). Thus, a promise to reward if B complies must imply a threat not to reward if B fails to comply. Likewise, a threat to punish B for noncompliance must imply a promise not to punish for compliance. Thus, all threats imply promises and all promises imply threats; they are simply different ways of describing the same conditional influence attempt." An implicit assumption along these lines may explain why so few political scientists bother to distinguish between threats and promises. An explicit example of such reasoning is found in Schelling's
Schelling (fn. 24), Strategy of Conflict, 196; and Arms and Influence, 70. 2eFollowing Harsanyi, we leave open the possibility that influence attempts may take other forms, including unconditional rewards or punishments (see fn. 10, 71).
25
26
WORLD POLITICS
Strategy of Conflict.27After considering several definitions and after admitting that the distinction between a threat and a promise is not obvious, he finally concludes that threats and promises are merely "names for different aspects of the same tactic of selective and conditional self-c~mrnitment."~~ The fallacy in this line of reasoning lies in the assumption that withholding a reward is always a punishment and withholding a punishment is always a reward. If rewards and punishment (and, correspondingly, promises and threats) are defined in terms of B's expectations at the moment A begins his influence attempt, it is clear that a conditional commitment not to reward if B fails to comply is not necessarily a threat. "If you do not do X, I shall not reward you" is a threat to punish if-and only if-B had a prior expectation of receiving the reward. "No bonus for you tomorrow unless you work hard today" means one thing on the day before Christmas bonuses are traditionally handed out and quite another on the day after such bonuses have been distributed. In order to distinguish threats from promises so that A may promise without necessarily threatening, one must identify three kinds of conditional commitments available to A: ( I ) conditional commitment to reward (promise), ( 2 ) conditional commitment to punish (threat), and (3) conditional commitment neither to reward nor punish (as~urance).~' The "assurance" would make it conceptually possible for A to promise a reward to B for compliance without simultaneously threatening to punish B for noncompliance. Instead of threatening B, A may "assure" him (explicitly or implicitly) that he will not be rewarded for noncompliance. Two advantages of conceptually isolating threats from promises should be mentioned. First, one may thereby distinguish between conditional commitments to deprive B of something he expects to have and conditional commitments to deprive him of something he does not expect to have. People respond differently to such situation^.^^ Second, it permits a more comprehensive description of the full range of policy options open to A in making his influence attempt. This, as Harsanyi has noted, is one of the main purposes for which social scientists use the concept of A's power over B.31 The following table
28 Ibid., 131-34. Schelling (fn. z4), 134. In Arms and Influence (p. 74) Schelling apparently uses the term "assurance" in this way. 30 Cf. Elton B. McNeil, "Psychology and Aggression," Journal of Conflict Resolution, 111 (September 1959), 203; and Kenneth E. Boulding, "Towards a Pure Theory of Threat Systems," American Economic Review, LIII (May 1963), 426. 31 Harsanyi (fn. IO), 69.
27
29
T H E POWER OF POSITIVE SANCTIONS
27
indicates some of these possibilities. Note that the table allows for variations in A's behavior along the following dimensions: ( I ) type of sanction (positive or negative), (2) degree of probability that the sanction will be implemented, and (3) degree to which A is explicit in specifying his commitment. Although A must at least imply that he will behave differently if B complies than he would if B fails to comply, A may choose from a number of possible combinations. For example, A may say to B: "If you comply, I guarantee not to punish you and I may even reward you; but if you do not comply, I guarantee not to reward you and I may even punish you." The table makes sense, however, only if one distinguishes clearly between withholding a reward and punishing.
A's OPTIONS WHENMAKING CONDITIONAL ATTEMPT B INFLUENCE ON
If B complies
no indication (YOU guess) may reward probably reward guarantee reward may not punish probably not punish guarantee not to punish
If B does not comply
no indication (you guess) mav not reward probably not reward guarantee not to reward may punish probably punish guarantee to punish
A clear distinction between the concepts of positive and negative sanctions can be drawn in terms of B's expectations about his future value position. The possibility of making such a distinction, however, is not sufficient to justify doing so. As Cartwright points out, the study of power relations is already riddled with interminable theoretical distinction^.^^ The remainder of this paper will therefore be devoted to two propositions: ( I ) that positive and negative sanctions have different behavioral implications, and (2) that the difference is significant from the standpoint of political science.
The differences between positive and negative sanction situations are neither trivial nor obvious. Even though some of these not-so-obvious and not-so-trivial differences can be deduced from the logical opposi32Dorwin Cartwright, "Influence, Leadership, Control," in James March, ed., Handbook of Organizations (Chicago 1965), 4.
28
WORLD POLITICS
tion of positive and negative sanctions, this is rarely done. Deduction, however, can only clarify some of the differences. Although positive and negative sanctions are opposites in logic, they are not opposites in their behavioral consequence^.^^ Both A and B behave differently in positive sanction situations than they do in negative sanction situations. What follows is a survey of the more important hypothesized differences between such situations. I. A's burden of response.-When A's influence attempt is based on a promise, B's compliance obligates A to respond with a reward; whereas B's failure to comply calls for no further response from A.34 Thus, a nation using threats will invest time and effort in planning its response to B's noncompliance. Such a nation has little incentive to ponder what it will do if B complies, since compliance requires no response from A. In other words, threats provide an incentive for A to base its planning processes on the assumption that B will not cooperate. To the extent that nations (or other political groups) behave as other nations assume (or expect) they will behave, threats would seem to make compliance less likely. This is not a logical necessity, just a psychological probability. A nation using promises need not expect compliance, but it has an incentive to do so. The point is that A's responsibilities and planning processes are different when he uses promises rather than threats. 2. The role of costs.-One important consequence of the asymmetry between positive and negative sanctions is that promises tend to cost more when they succeed, while threats tend to cost more when they fail." Thus, threats and promises are related to the probability of success in different ways. The difference can be summarized as follows: The bigger the threat, the higher the probability of success; the higher the probability of success, the less the probability of having to implement the threat; the less the probability of having to implement the threat, the cheaper it is to make big threats. The bigger the promise, the higher the rob ability of success; the higher the probability of success, the higher the probability of having to implement the promise; the higher the probability of having to implement the promise, the more expensive it is to make big promise^.^' If A has doubts
33 Cf. Thomas W. Milburn, "What Constitutes Effective Deterrence?" Iournal of Conflict Resolution, 111 (June 1959), 139; and B. F. Skinner, Science and Human Behavior (New York 1953), 182-93. 3"The asymmetry between positive and negative sanctions has been noted by Boulding (fn. 30), 428; and Parsons (fn. 5), 239. 35 Schelling (fn. 24), Strategy of Conflict, 177. 36This assumes that other things remain equal. Two especially important assumptions are that cost varies directly with the risk of implementation and that the credi-
T H E POWER OF POSITIVE SANCTIONS
29
about his ability to calculate the probability of success he accurately, may want to hedge. Whereas, with promises, hedging is likely to take the form of scaling down the promise, with threats, hedging is likely to take the form of building up the threat: too big a threat is likely to be superfluous rather than ~ostly.~' 3. Propensity to use and prospects of success.-Since A's incentive to use promises instead of threats tends to increase as the probability of success decreases, one may hypothesize that A is more likely to use positive sanctions when he thinks his prospects of success are poor (ceteris paribus). This hypothesis suggests the need to re-examine some widely held opinions about the relationship between positive sanctions and political integration. Whereas the international political arena tends to be associated with anarchy and negative sanctions, the domestic political arena tends to be associated with integration and positive sanctions. Whereas the underworld tends to be associated with mutual distrust and negative sanctions, the overworld ( ? ) tends to be associated with trust, good faith, and positive sanctions. Advocates of increased use of positive sanctions in international politics are often pictured as unrealistic dreamers who want to use influence techniques that are obviously unsuitable in the international arena. The above hypothesis, however, suggests that positive sanctions are more relevant to international than to domestic politics, to the underworld than to the overworld. Since the prospects for success are worse in the international and underworld arenas, the incentive to use positive sanctions instead of negative ones should be greater (assuming, of course, that other things are equal). In a well-integrated domestic polity, however, negative sanctions are more useful. It is only because of the high probability that most people will obey the law that governments can afford to enforce laws with threats. It is much cheaper to punish the few who disobey than to reward the many who obey.38 Negative sanctions are often associated with situations in which A's influence attempt, if it succeeds, leaves A better off and B worse off than they would otherwise have been.39The above hypothesis suggests a re-examination. Other things being equal, A's prospects of success are worse when he tries to promote improbable outcomes than when
bility of a threat or promise is not affected by its size. Although either of these assumptions is questionable in other contexts, they do not seem to be directly related to differentiating between positive and negative sanctions. 37 Schelling (fn. z4), Strategy of Conflict, 177.
38 Cf. Karl W. Deutsch, T h e Analysis of International Relations (Englewood Cliffs,
N.J. 1968), 17-18. 39 See, for example, Boulding (fn. 3 ) 426. 0,
30
WORLD POLITICS
he tries to promote probable ones.40One may also assume that there is a low autonomous probability that B will do things that are bad for him and a high autonomous probability that B will do things that are good for him. Thus, one may formulate the following syllogism: "The lower the probability that X is bad for B, the higher the autonomous probability that B will do X; the higher the autonomous probability that B will do X, the higher the probability that A's attempt to get B to do X will succeed; the higher the probability of success, the more likely A is to use threats." Threats are most likely to be used in situations in which A is trying to make sure that B does something that B is quite likely to do anyway, e.g., refrain from murder, rape, larceny, or nuclear attack. It is not unlikely that B would agree that it is good for him to be deterred from such actions. In practice there may be a difference between the actual probability that A will succeed and the probability perceived by A. There is some evidence that A will tend to overestimate the probability that his influence attempt will succeed.41This would partially explain what appears to be a bias toward using negative sanctions in actual practice. 4. Indicators of success.-Empirical observation of influence attempts based on threats and promises is tricky, since the indicators of success for each differ. Whereas a successful threat requires no action by A, a successful promise obligates him to implement the sanction. In a well-integrated social system, where the probability that B will comply with A's wishes is relatively high, promises will be more visible than threats. In fact, it is precisely because threats are so successful in domestic politics that they are so difficult to detect.42One suspects that a lower rate of success makes threats more salient in international than in domestic politics. Let us assume that an equal number of threats and promises are made in the domestic and international political arenas (an unrealistic assumption). Let us further assume that the probability of success is lower in the international than the domestic realm (a realistic assumption). What will the empirical observer see? Whereas domestic life will appear harmonious and tranquil, international life will appear to be a war of all against all. The low rate of
40
Cf. Deutsch (fn. 38), 26-27.
41Cf. Deutsch (fn. 13), 212-13; Ward Edwards, "Utility, Subjective Probability,
Their Interaction and Variance Preferences," lournal of Conflict Resolution, VI (March 1962), 42-51: Robert Jervis, "Hypotheses on Misperception," World Politics, xx (April 1968), 454-79; and J. David Singer, "Inter-Nation Influence: A Formal Model," Amencan Polztrcal Science Review, LVII (June 1963), 426. 42This point has been noted by Talcott Parsons, "Some Reflections on the Place of Force in Social Process," in Harry Eckstein, ed., Internal W a r (New York 1964), 51-53.
T H E POWER OF POSITIVE SANCTIONS
31
international compliance will activate negative sanctions and obviate positive ones, while the high rate of domestic compliance will activate positive sanctions and obviate negative ones. Thus, one may formulate the following hypothesis: "Empirical observers will tend to overestimate the role of threats relative to promises in international politics and to underestimate the role of threats relative to promises in domestic 5. Deterrence.-The asymmetrical way in which threats and promises are related to success helps to explain the tendency to associate negative sanctions with policies of deterrence. Logically, of course, one could just as well speak of deterring B (that is, reducing the probability that B will do X) with promises as with threats; for example, "I will give you a reward if you will refrain from doing X." Time and again, however, writers imply that deterrence is a matter of threats alone, or perhaps threats combined with promises, but never a matter of promises alone.44The connection between threats and deterrence is rarely explained, merely implied as a matter of trutheither intuitively obvious or definitional. How does one explain this propensity to imply a special relationship between deterrence policies and negative sanctions? Consider three possible explanations: ( I ) Dictionary definitions of "deterrence" usually depict it as discouragement through fear. Such definitions make the concept of a deterrent promise a contradiction in terms. It is difficult to believe, however, that semantics can account for the United States government's reliance on threats rather than promises in order to lower the probability of Russian attack. ( 2 ) Talcott Parsons has
43 There is probably a tendency to overestimate the importance of threats relative to promises in underworld politics also. 44For an example of explicit association of negative sanctions and deterrence, see Talcott Parsons (fn. 5), 239-40 Less explicit examples are Blau (fn. 6), 116-17; Glenn H . Snyder, "Deterrence and Power," Journal of Conflict Resolution, N (June 1960), 163-78; Richard A. Brody, "Deterrence," International Encyclopedia of the Social Sciences, IV (New York 1968), 130-32; Schelling (fn. 24). Schelling implies the association between threats and deterrence by the scant attention he gives to promises relative to threats. The association is also implied in observations such as the following: "It is a paradox of deterrence that in threatening to hurt somebody if he misbehaves, it need not make a critical difference how much it would hurt you tooi you can make him believe the threat." ( A r m s and Influence, p. 36.) This is not f true of attempts to deter that are based on positive sanctions. It is true of all influence attempts based on conditional use of negative sanctions-regardless of whether such attempts seek to deter or compel. Thus the paradox should be called a paradox of negative sanctions, not a paradox of deterrence. A few writers have attempted to link positive sanctions with deterrence policies: See Milburn (fn. 33), 138-45; E. James Lieberman, "Threat and Assurance in the Conduct of Conflict," in Roger Fisher, ed., International Conflict and Behavioral Science (New York 1964), 110-22; and Jerome D. Frank, Sanity and Survival (New York 1967), 162-63.
32
WORLD POLITICS
suggested that negative sanctions have more intrinsic effectiveness Parsons produces no than positive sanctions in deterrence ~ i t u a t i o n s . ~ ~ evidence and very little argument to support this contention, however. (3) The asymmetrical way in which threats and promises are related to success offers another explanation. As Deutsch has pointed out, the autonomous probability of B's doing X when X is defined as anything but Y is usually high relative to the autonomous probability of B's doIn ing X when X is defined as a narrowly specific act.4B other words, deterrence is usually a matter of promoting a highly probable outcome. As noted above, there is a special relationship between threats and situations in which A's influence attempt is likely to succeed. Thus, the association of threats with policies of deterrence can be partially explained in terms of the broader special connection between threats and attempts by A to promote probable outcomes. A is likely to use threats rather than promises in attempting to deter B from doing X because the probability of success tends to be relatively high. 6. B's immediate response.-B's immediate reaction to sticks usually differs from his immediate reaction to carrots. Whereas fear, anxiety, and resistance are typical responses to threats, the typical responses to promises are hope, reassurance, and attraction. Three important examples of the difference this can make are as follows: ( I ) threats cause B to feel stress, which is likely to affect (that is, enhance or impair) B's problem-solving capacity, i.e., his rationality." ( 2 ) Threats tend to generate resistance by B. Cartwright, and French and Raven, have suggested that it is useful to distinguish between opposition to an influence attempt and resistance generated by the influence attempt itself." (3) Whereas positive sanctions tend to convey an impression of sympathy and concern for B's needs, negative ones tend to convey an impression of indifference or actual hostility toward B. Cartwright argues that such impressions have a profound effect upon the outcome of any particular influence attempt.49 7. After-effects and side-effects.-Positive sanctions differ from negative ones not only in their immediate effect on B, but in their afterand side-effects on him as well. An important side-effect is the "spillover effect" on B's relations with A with respect to other issues. While positive sanctions tend to enhance B's willjngness to cooperate with
4 6 Deutsch (fn. 38), 26-27. Parsons (fn. 5), 239-40, Singer (fn. 41), 429. 4 8 Cartwright (fn. 32), 33-34; French and Raven (fn. zo), pp. 156-57. 49 Cartwright (fn. 3z), 15. On the importance of B's perceptions of A's motives see also Tedeschi (fn. z), and H. H. Kelley, "Attribution Theory in Social Psychology," in D. Levine, ed., Nebraska Symposium on Motivation (Lincoln 1967), 192-238.
45
47
T H E POWER O F POSITIVE SANCTIONS
33
A on other issues, negative sanctions tend to impede such cooperation. For example, America's use of negative sanctions to fight communism increases the difficulty of simultaneous cooperation with Communist nations on cultural exchanges, international trade, and air pollution. Whereas the "spill-over effect" refers to the effect on A's concurrent relations with B, the "scar effect" refers to the effect on A's future relations with B. If A uses positive sanctions today, B will tend to be more willing to cooperate with A in the future, but if A uses negative sanctions today, B will tend to be less willing to cooperate with A in the future." Today's choice by A between positive and negative sanctions affects not only today's response by B, but tomorrow's as well. Just as the memory of America's past interventions colors today's relations between Latin America and the United States, so the memory of Russia's intervention will color relations between Czechoslovakia and the USSR for years to come. Likewise, the memory of American influence attempts in Vietnam will play an important part in international politics long after the shooting has stopped. 8. Legitimation.-It is usually easier to legitimize demands based on positive sanctions than demands based on negative ones.51 An example is provided by one of the most important social institutions in the world-the institution of private property. Most societies have one set of rules specifying the conditions under which a man may be deprived of his property and quite a different set of rules specifying conditions under which a man's property may be augmented. In general, these rules make it much easier for A to add to B's property than to subtract from it. The man who threatens to deprive B forcibly of $10 if he fails to do X is likely to wind up in jail. The man who offers B $10 to do X may not succeed, but he is less likely to be jailed. The international political arena may be an exception to the general rule on legitimation. It is true, of course, that aggressive warfare is illegal while foreign aid is not. But it is also true that many people would view deterring nuclear attack by a threat to retaliate in kind as more moral than deterring nuclear attack by offering a bribe to the potential attacker. Likewise, the contemporary mores of statecraft seem to make it more moral to drop napalm on a man to prevent him from becoming a Communist than to bribe him to do the same thing.
5 0 On the "scar effect" see Cartwright (fn. 32), 35-36; French and Raven (fn. 20), 156-58; Morton Deutsch, "Trust and Suspicion," ]ournal of Conflict Resolution, 11 (December 1958), 265-79; John R. Raser, "Learning and Affect in International Politics," lournal of Peace Research, No. 3 (1965), 216-26; and Bertram H. Raven and Arie H. Kruglanski, "Conflict and Power," in Paul Swingle, ed., T h e Structure of Conflict (New York 1970), 69-109. French and Raven (fn. 20), 15660; Dahl and Lindblom (fn. 18), 107-108.
34
WORLD POLITICS
A useful research project could focus on comparing the relative ease of legitimating demands based on positive sanctions in the domestic arena with the legitimation of such demands in the international arena. g. Symbolic importance.-Negative sanctions have become psychologically linked with such characteristics as courage, honor, and masculinity. Soldiers, not diplomats, symbolize masculine virtues. In international politics these psychological links are especially pronounced. The statesman who would use positive sanctions risks being perceived by both foreigners and his domestic public as soft, weak, or lacking in toughness. When the North Koreans seized the Pueblo, it was "unthinkable" that President Johnson would offer to buy it back. National honor was at stake. It is "honorable" to fight but "dishonorable" to try to buy one's way out of a fight. My purpose here is neither to condemn nor to condone the symbolic functions of positive and negative sanctions; it is merely to note that there is a difference between their psychological functions in politics. 10. Human nature.-The propensity to use either positive or negative sanctions is partially determined by A's view of human nature. It has been hypothesized that those who view man as basically lazy, wishing to avoid responsibility, and desirous of security above all else are more likely to emphasize negative sanctions than are those with a more optimistic view of human nature.52 Regardless of the viability of this particular hypothesis, it seems worthwhile to explore the correlation between views of human nature and propensities to use positive or negative sanctions. I I. Eficacy.-The relative effectiveness of positive and negative sanctions in getting B to do X has been the focus of much research in psy~hology.~~ Although the precise conditions under which one type of sanction is more effective than the other have yet to be spelled out, it is quite unlikely that positive and negative sanctions will be equally effective in any given influence attempt. The tendency of poCartwright (fn. 32), 13-15. NO attempt to cite this vast literature will be made here. For interesting starting points and further references see Blau (fn. 6), 224-27; Lewis A. Froman, Jr. and Michael D. Cohen, "Threats and Bargaining Ef...