Harnessing Conflict in Foreign Policy Making

xiaocisan · 发表于 2009-11-24 20:15:28

Harnessing Conflict in Foreign Policy Making：From Devil's to Multiple Advocacy

Presidential Studies Quarterly v. 32 no3 (Sept. 2002) p. 484-508

A large and alarming body of historical and laboratory evidence suggests that presidential decision making in foreign (and for that matter domestic) policy is plagued by a number of chronic impediments that may undermine the policy-making process and detract from its "quality" (e.g., De Rivera 1968; George 1980; Burke and Greenstein 1989; Vertzberger 1990; Haney 1997; Hess 2001; Parker and Stern 2002). Among these are insufficient diversity of expressed opinions and excessive conformity in advisory groups (Janis 1982; ’t Hart 1994; Stern and Sundelius 1997, 124-28; Schaefer and Crichlow 2002). Groups suffering from these problems deprive themselves of the kind of constructive debate that can "arouse discussion and stimulate creative thinking" (Amason and Thompson 1995, 2) in ways that "improve both decision making and acceptance of the decisions made" (Eisenhardt, Kahwajy, and Bourgeois, 1999, 171-72).

As an insightful student of the presidency has recently reminded us, a key challenge for presidents is "is to forge a team and get the most out of it, minimizing the tendency of subordinates to tell their boss what they sense he wants to hear" (Greenstein 2000, 195). In this article, we will examine two institutional compensations intended to make it more likely that adequate degrees of diversity and critical interaction will inform and energize the policy-making process. The first is the so-called devil’s advocate--a relatively modest measure designed to counteract pressures toward homogeneity and conformity in small groups and thus enhance deliberations. The second is the much more comprehensive and ambitious system of multiple advocacy, which is designed not only to cope with the problem of conformity but to moderate tendencies toward pathological conflict/competition among advisers such as those associated with more vicious forms of cabinet and bureaucratic politics (George 1972; Allison and Zelikow 1999; Stern and Verbeek 1998, 205-55; Garrison 1999).

Before examining each of these measures in turn, let us make an observation regarding the composition of the advisory network (Hult 1993, 114-20). As Irving Janis pointed out several decades ago, the knowledge and experience of individual group members are valuable resources for executives. Ceteris paribus, groups that are more diverse will be able to draw on a wider base of knowledge and experience in formulating and coping with policy problems. Longtime Washington hand and commentator David Gergen (2000) argues convincingly in his book Eyewitness to Power: The Essence of Leadership from Nixon to Clinton that a strong team should include "people who have known the president a long time to keep the flame alive; veterans of Washington to make sure the flame burns effectively. Youngsters to give the team energy; gray hairs to give it wisdom" (p. 184; see also Abshire 2001, viii-ix). Advisory groups that are overly homogeneous will thus be deprived of cognitive and social resources that might otherwise have informed their decisions. For example, Eisenhardt, Kahwajy, and Bourgeous (1999) argue strongly for diversity. Their advice to executives is "Assemble a heterogeneous team, including diverse ages, genders, functional backgrounds, and ... experience. If everyone in the executive meeting looks alike and sounds alike, then the odds are excellent that they think alike too" (p. 191). Therefore, choosing a team that is sufficiently diverse and composed of individuals with integrity is, of course, the first defense against excessive homogeneity and conformity of perspective (Janis 1982, 305; Russo and Schoemaker 1989, 156-57). Let us also note that diversity is potentially associated with costs as well as benefits (’t Hart 1997, 324-26; Kowert 2002), an issue to which we will devote more attention below. Thus, while diversity is clearly a good start in coping with the risk of excessive conformity in policy making, executives may find that additional measures are required and may thus wish to consider making use of devil’s advocacy.

DEVIL’S ADVOCACY: USES AND LIMITATIONS

Since the early sixties, much attention has been given to ways of ensuring that unpopular views are encouraged and given a proper hearing in small decision-making groups in which pressures for conformity often discourage expression of dissenting opinions. Among the organizational devices often recommended for this purpose is the time-honored institution of the "devil’s advocate." Following the Bay of Pigs fiasco, in which President Kennedy’s policy meetings had been marked by a seeming unanimity of opinion, his brother suggested that thereafter there always be a devil’s advocate to give an opposite opinion if none was pressed (Kennedy 1969, 90; De Rivera 1968, 61-64, 209-11; Janis 1972, 215-16). Indeed, JFK appears to have encouraged both his brother and Theodore Sorensen to take a more active "watch-dog" role in foreign-policy matters. Later, President Johnson is said to have referred to Undersecretary of State George Ball as his "devil’s advocate," thereby acknowledging Ball’s continued expression of his dissent over Vietnam policy. While something like a devil’s advocate role was perhaps played by these officials and no doubt by other advisers on other occasions, relatively little historical material describing activities of this kind is available to serve as a basis for evaluating the efficacy of this organizational device. Similarly, the experimental laboratory research on small groups that has been consulted in preparing this article remains inconclusive regarding the feasibility and utility of devil’s advocacy in real-world settings (Katzenstein 1996, 317-18, 329-30). However, many studies do strongly suggest that a group’s performance can be enhanced under certain conditions by leadership practices and the cultivation of a group culture that "protects" members who express minority views (see, e.g., Moscovici 1985; Nemeth and Staw 1989; Turner 1991; Schultz and Ketrow 1995; Kaarbo and Gruenfeld 1998).

However, the case for introducing a devil’s advocate into policy-making groups still rests in good measure on a priori grounds. While the case is a strong one in principle, the introduction and effective utilization of a devil’s advocate in real-world settings is by no means a simple matter (Katzenstein 1996, 329-30). Indeed, those who favor the idea of a devil’s advocate often have very different notions of what this would mean in practice. Some are content to suggest that the leader of a group should appoint one person on an ad hoc basis to serve as devil’s advocate if no one in the group will challenge the dominant view in a given situation. Others have identified a much more complex set of requirements and procedures for institutionalizing the devil’s advocate function, even going so far as to suggest that a subgroup rather than just one individual be assigned a continuing responsibility to make the opposition case even after a decision has been taken (De Rivera 1968).

Strictly speaking, the devil’s advocate performs a role; it is understood that the person performing this role will argue an unpopular position that should be considered but that no one else will speak up for and that the devil’s advocate does not really favor. The fact that he or she is performing an accepted role and is not a genuine dissenter is designed, of course, to protect that person from incurring sanctions for challenging the group’s opinion or its leader’s view. Thus defined, however, the limits as well as the potential utility of the role become manifest: for while the devil’s advocate introduces some diversity into the group’s deliberations or challenges some of the premises that enter into the leader’s judgment, he or she cannot persist in the challenge or, even more important, seek to develop a coalition within the group to oppose and, if possible, overcome the majority. Unlike a genuine policy dissident, the true devil’s advocate is not a political actor with policy commitments and organizational resources but is merely playing a role that, at best, facilitates a dialectical, multisided examination of the problem at hand. Nor should the role of devil’s advocate be confused with the much more comprehensive role of the "custodian-manager" of the policy-making process, which will be discussed below.(FN1).

Those who have observed with distress the repeated failure of policy makers to consider diverse views sometimes turn in desperation to the idea that installing a devil’s advocate would help. But the mere provision of a devil’s advocate in small decision-making groups is hardly a guarantee that the person will be able to perform the role well enough to contribute to improved policy making. However, the psychological literature does suggest that the perceived sincerity of a dissenter--as demonstrated through consistency, autonomy, and commitment--is an important factor in determining minority influence (Moscovici 1985, 355, 359-65; Kaarbo and Gruenfeld 1998, 230). We still know very little about what kinds of persons can perform effectively in this role and how it can be introduced and maintained so that it is not regarded as an awkward or time-wasting gimmick by members of the group.

Specialists on organizational decision making have formulated proposals for coping with some of these issues. For example, Russo and Shoemaker (1989, 159) argue that the function should be passed around on a rotating basis so the group will not come to dismiss or ignore the criticisms of a single permanent "dissenter." Despite such creative suggestions and a degree of knowledge accumulation over several decades, there is clearly much still to be learned about how best to structure and deploy devil’s advocacy. While it is far too soon to dismiss completely the potential of this innovation, one cannot be sanguine on the basis of the experience available to date. There is, first, some question whether the role can be performed with the integrity required to yield the desired impact. Second, there is sobering evidence that a devil’s advocate can be put to uses other than those for which the role is ostensibly intended. Let us examine both of these constraints on the utility of devil’s advocacy.

Accounts of Vietnam policy making suggest that the device of a devil’s advocate can be misused in an effort to "domesticate" advisers who genuinely oppose policy decisions being taken. George Ball, for example, repeatedly disagreed with the development of U.S. policy in Vietnam. From an early stage President Johnson took to calling Ball his "devil’s advocate"--a misnomer in this case since Ball was a genuine dissenter. Perhaps Johnson employed the euphemistic label of devil’s advocate to soften the import of Ball’s dissent and indicate that he would regard it as legitimate and acceptable only if Ball provided his views as a service to the group and kept them within the confines of the group.

What this suggests is that in contrast to the often-noted tendency in experimental laboratory groups for the majority to exert crude and extreme conformity pressures on dissident members, in real-world policy groups it is often unnecessary or undesirable to squelch or reject a dissident member. The possibility that dissidents are likely to increase malaise within the rest of the group is often accepted and legitimized in the expectation that they will strengthen on balance the ability of the group to cope with the problems of policy making. But in response, the dissident may moderate the style or manner in which he or she expresses dissent, if not also the full extent of the disagreement, by falling into what James C. Thomson (1968) calls "the effectiveness trap"--that is, the trap that keeps officials from resigning in protest and airing their discontent outside the government. It is possible to be overly cynical and uncharitable about such behavior. The reality of the dilemma, however, cannot be ignored. As Thomson puts it, "To preserve your effectiveness, you must decide where and when to fight the mainstream of policy by staying and not resigning one may be able to prevent a few bad things from happening" (p. 49). As for George Ball, who presumably acquiesced in his "domestication," Thomson is quick to concede that matters might have gotten worse faster if Ball had kept silent or left before his departure in the fall of 1966. Albert Hirschman (1970) has expressed concern over the extreme reluctance of Americans in public office to resign in protest against policies with which they strongly disagree. Hirschman’s general thesis is that "exit" has an essential role to play in restoring quality performance of government, as in any organization (pp. 114-19; cf. Meltsner 1991, 133-60).