Cognitive Process and Foreign Policy Decision-Making

Cognitive Process and Foreign Policy Decision-MakingAuthor(s): Michael J. Shapiro and G. Matthew BonhamSource: International Studies Quarterly, Vol. 17, No. 2 (Jun., 1973), pp. 147-174Published by: Blackwell Publishing on behalf of The International Studies AssociationStable URL: http://www.jstor.org/stable/2600226Accessed: 16/08/2010 19:20

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available athttp://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unlessyou have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and youmay 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 athttp://www.jstor.org/action/showPublisher?publisherCode=black.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

The International Studies Association and Blackwell Publishing are collaborating with JSTOR to digitize,preserve and extend access to International Studies Quarterly.

http://www.jstor.org

Cognitive Process and Foreign Policy Decision-Making

MICHAEL J. SHAPIRO Department of Political Science

University of Hawaii

G. MATTHEW BONHAM Department of Political Science

The American University

As the analysis of foreign policy decision-making has become a more theoretical enterprise, the production of historically oriented case studies has been supplemented by a growing number of investigations employing psychological or social- psychological perspectives. Early studies of foreign policy decision-making with a psychological orientation emphasized the influence of various psychological traits on those involved in foreign policy decisions (e.g., Levinson, 1957), but, more recently, the emphasis has been on perception, cognition, and information-processing. The cognitive process approaches, by contrast, have attempted, in varying degrees, to map out the belief structures of decision makers and explore the implica- tions of these structures for the way international events are understood and policy alternatives are considered. Studies under this rubric have focused upon the perceptions of particular foreign policy decision makers (Holsti, 1962), on the perceptions and choices of groups of persons simulating the roles of foreign policy decision makers (Driver, 1962; Hermann, 1969; Hermann and Hermann, 1967; Shapiro, forthcoming), and on the decision process of national decision groups focused upon particular policy problems (Steinbruner, forthcoming).

AUTHOR'S NOTE: This research was supported by the Institute of International Studies, University of California, Berkeley.

International Studies Quarterly, Vol. 17 No. 2, June 1973, ? 1973 I.S.A.

[1471

[148] INTERNATIONAL STUDIES QUARTERLY

The theory-building effort reported here relies on the cognitive process rather than the personality trait approach to foreign policy decision-making. The most immediate goal of this investigation is to understand the identification of and re- sponses to international events by foreign policy decision makers. Our approach has two dimensions which we will explicate throughout the rest of this paper. The first and most important is our explanation of the cognitive dynamics which occur when an individual receives information about an international event, processes it through his belief system (which contains concepts about actors and actions in the international system), and reaches a conclusion about what has happened and what should be done by his nation. The second dimension is the mode in which our explanation of foreign policy decision is to be expressed. Eventually, our theory will be in the form of a computer simulation model which will allow us to examine not only the decisional outcomes which would be expected, given our model of the decision process and the belief systems of our simulated decision makers, but also the likely decisional outcomes which would result by altering either belief systems, or rules for information processing, or both. After we have discussed some of the normative implications of our approach to the study of foreign policy decision-making, we will elaborate the background and structure of our theory.

Normative Implications

The choice of a cognitive process approach to foreign policy decision-making is based only partly on the expectation that it is a way to build a comprehensive theoretical framework which will allow us to explain and predict decision makers' responses to international events, including crisis decision-making and some situations that involve the dynamics of planning and anticipation. A variety of theoretical orientations, perhaps even a personality trait approach, might also yield predictive ac- curacy. The choice of a cognitive process approach is related to

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [149]

our long-range goal for research-namely, recommendations for policy planning and execution in international politics. Because our recommendations will take the form of ways of concep- tualizing decision situations and searching information, it is convenient if the explanatory model employed is somewhat congruent with the way in which decision makers justify their choices. Cognitive process models, particularly those which suggest relationships between cognitive components like values and beliefs, are similar in structure to justificatory arguments in which one makes a case for a particular choice on the basis of arguments about priorities (what values should be preeminent) and about beliefs (the evidence for believing in various connections between choices and objectives or values).

The ultimate goal of the investigation can thus be termed "policy-oriented." The implementation of such a goal begins not, as is often supposed, at the point where good predictions or, more generally, valid explanations have been obtained. Implementation is presupposed when one is selecting the concepts and thus the kind of explanation to be constructed. Clearly, an explanation must be sound on scientific grounds if it is to be useful for purposes of recommendation, but this is a necessary and not a sufficient condition for policy recommen- dation. The kind of explanation selected must be consistent with the kind of control over the situation that one envisions. Because we wish to be in a position to make recommendations about how information should be categorized and processed both in terms of adducing inductive support for beliefs and deductive support for explanations of prevailing situations and perceived policy alternatives, we are seeking an explanation that is consistent with the way persons use constructs and evidence to justify their choices.

Theoretical Perspectives

Our model of cognitive processing in foreign policy decision- making is a synthesis of three social-psychological perspectives.

[1501 INTERNATIONAL STUDIES QUARTERLY

The first theoretical perspective of relevance is concerned with the concept of cognitive complexity versus simplicity. For our purposes, the application of this concept is in the extent to which a decision maker's cognitive mapping of the international environment is conceptually simple or complex. In general, cognitive complexity has been shown to relate to the accuracy of an individual's predictions about people's behavior-e.g., the more complex their perceptual discriminations, the better their predictions (Bieri, 1955); and in the specific application of the complexity-simplicity conception to foreign policy decision- making (in a gaming situation), it was demonstrated that conceptual complexity is positively related to the range of behaviors-e.g., amount of trading, bargaining, and so forth with nations in other defined power blocs (Driver, 1962; Schroder et al., 1967). The cognitive complexity-simplicity concept, for our purposes, is reflected in the scope of the substantive explana- tion that a foreign policy decision maker maintains with respect to the international political environment. For example, to the extent that an individual is more complex or differentiated in the way he views the power configuration in the international system, he is apt to consider a broader range of approaches to conflict management. This was argued as well as demonstrated in a recent symposium. Two critiques of a speech by a Johnson Administration foreign policy spokesman both suggested that alternative policy postures would be entailed if foreign policy decision makers assumed that there now exists a significant Third World bloc rather than predicating policy alternatives on a more simple, bipolar, East versus West perspective (Washburn and Mitchell, 1967).

While it has been demonstrated that the amount of cognitive complexity of decision makers in general (including those involved in a simulated foreign policy decision-making situa- tion) has an effect on decisional outcomes, the implications of the degree of complexity has not been traced in the process of cognition which is triggered in a decision situation. In order to observe the way in which cognitive complexity articulates itself through a decisional process, we used a crisis game invented by

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [1511

Steinbruner ( 1970) and modified for our purposes. In brief, the crisis games we ran had three participants who read a scenario about a crisis situation and were asked to advise the President about what his response should be. The major event in the scenario was the placing of nuclear weapons in Syria by the Soviet Union.

Because each of the participants in our game differed in the degree of complexity in his image of the international political process, the crisis game provided an opportunity to scrutinize the relationship between cognitive complexity-simplicity and information-processing in foreign policy decision-making. To analyze the degree of cognitive differentiation in the mappings of the game participants, we examined their written forms and utterances, which we had taped. To illustrate our conclusions, we will discuss one of the decision-making groups that participated in our experiments. Participant 1, a specialist in international relations, tended to see the Middle Eastern situation in a broad context of an international power struggle featuring the United States, USSR, and various uncommitted nations. Participant 2, a specialist in comparative politics, focused more on relations within and between regimes in the Middle East. He related this to the international context primarily through speculations about Soviet policy in the Middle East as compared with its policy in other regions-e.g., Asia and Europe. Participant 3 is a Middle Eastern specialist and focused primarily upon the region itself. He saw the current situation almost exclusively as a logical consequence of histor- ical forces operating in the Middle East.

It should be evident that the participants arranged themselves in a convenient hierarchy on a differentiated-undifferentiated continuum from Participant 1, the most differentiated in his perceptual mapping of the Middle Eastern problem, to Partici- pant 3 who was the most undifferentiated in the number and breadth of categories which he applied to the Middle Eastern situation. The cognitive differences of the three participants revealed themselves in a variety of ways during the gaming. At the beginning of the gaming (before participants were informed

[152] INTERNATIONAL STUDIES QUARTERLY

of nuclear weapons in Syria), each participant was handed "Policy Planning Aid A" which asked, "What additional points of information would you like to have about the current situation in the Middle East as described in the documents for Excon III?"

In response to this question, Participant 1 (the international relations specialist) asked questions not only about the Arab positions (those of Lebanon and Syria) on the current situation but also about the policies and recent involvements in the area of the Soviet Union, the United States, England, and France. Participant 2 (the comparative politics specialist) asked about Soviet military strength and then inquired primarily about relations within and between Arab regimes and intragovern- mental conflicts in Israel. Participant 3 asked about Soviet intentions, the strength of some of the Arab countries, and the activities of the Palestinian commandos.

A transcript of the discussions during the gaming provides further insights into the impact of the different perceptual mappings of the three participants. The international relations specialist (Participant 1) saw many possible explanations for Soviet behavior-from a move to force Israel to surrender the occupied Arab territory to a move that was part of a global strategic interaction between the United States and the Soviet Union:

I still think that the basic question that faces the United States is an assessment of Soviet intentions. If, indeed, it is the Soviets' intention simply to force Israeli evacuation of occupied territories, that is one thing.... But we don't know whether that is the Soviet objective, and we don't know what other Soviet objectives there may be and how tenaciously they may be held.

The American response he urged was addressed to this image of the problem. In case the Soviets were only trying to force Israel to withdraw, he advocated the beginning of negotiations along the lines of the French proposal for a Middle East settlement that had been discussed in the first move period. Since Soviet intentions were not clear and might be more ambitious, he also

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [1531

recommended an assortment of public and private moves, the former including a major nuclear targeting and training exercise for American naval forces which would focus on Eastern European targets, an approach to India to give nuclear technical assistance to Israel (the United States is proscribed as a signer of the nonproliferation treaty), and the moving of the Sixth Fleet into a strategic position near Syria; the latter consisting of a targeting of Polaris missiles toward Syria and a threat to attack Cuba if the Soviets do not remove the weapons in Syria.

The comparative politics specialist (Participant 2) saw the Soviet move in somewhat less global-strategic terms than did the international relations specialist. He felt that Soviet behavior in Syria signaled an eagerness to start negotiations, supporting this interpretation with an argument based on the assumption that the Soviet Union liked to negotiate from a position of strength, and the evidence that the missiles were under Soviet control and had not been turned over to the Syrians. His recommendation for American policy did not involve an international signalling system of the breadth suggested by Participant 1, but involved instead a combination of protesting in the United Nations, demonstrations of military readiness in close proximity to the trouble spot, and attempts to begin negotiations.

The Middle East specialist (Participant 3), saw the Soviet move as a result not of the international strategic interaction system or of Soviet foreign policy alone but of developments in the Middle East. He saw the Arab-Israeli conflict as the major determinant of Soviet penetration into the Middle East. Arab regimes, he felt, looked to the Soviets because no other nation or international organization seemed sympathetic to their weakened position resulting from the 1967 War. His policy recommendations were oriented toward a more general, long- term settlement of the Arab-Israeli conflict. He advocated the immediate start of negotiations and use of international auspices to effect a settlement which would touch upon all the relevant sources of grievances in the area.

The three participants thus utilized different substantive theories of the international environment in interpretating the

(1541 INTERNATIONAL STUDIES QUARTERLY

Soviet move in Syria, and, generally, the more differentiated or multifaceted the theory, the more complex the policy recom- mendations. What is particularly noteworthy is that the cognitive processing of the three participants paralleled, on the complexity-simplicity continuum, their substantive perspec- tives. The cognitive processing of the participants was deter- mined by an analysis of their responses to policy planning aids they were asked to fill out. T'wo of the policy planning aids were designed to construct a decision calculus for the re- spondent. Policy Planning Aid B asked the respondent to state the three major objectives that they saw for American policy in the Middle East, and Policy Planning Aid C asked them to indicate the extent to which a number of alternative policy proposals (including the one their policy-making group had selected) would achieve or block each of the objectives they had espoused (see Rosenberg, 1956). As in the case of the subsequent perceptual mappings, the cognitive processing of the participants was arranged from complex to simple with Partici- pant I the most complex, Participant 3 the least complex, and Participant 2 in between. Cognitive processing was analyzed by looking at the pattern of beliefs about which policy alternatives would achieve which objectives. A very simple cognitive processing tool is one in which the decision maker upholds a particular action principle as better in every way than any alternative type of action. A more complex cognitive style is one that views action alternatives on a gains and losses continuum with some actions better suited to some objectives and other types of actions more appropriate to alternative objectives. Thus, we can measure the complexity of cognitive processing from responses to Policy Planning Aid C by looking at the amount of variance in beliefs about the effects of alternative policies on objectives.

Participant 3 (the Middle Eastern specialist) was invariant in his beliefs about the effects of the various policy alternatives. He saw the alternative accepted by his group as having a moderately positive chance (a score of +3) of achieving each of his three objectives. He viewed the other three alternatives as

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [1551

blocking his objectives with one alternative given a -5 for all three objectives, another given a -4 for all three objectives, and the other given a -1 for all three objectives. The other two participants, by contrast, saw each policy alternative as having a differential effect on their espoused objectives. Participant I had the most variance in his beliefs about the effects of the different policy alternatives. He saw the policy alternative chosen by his group as being positively related to two of his objectives and negatively related to one of them. Generally, his beliefs varied widely, ranging, in one case, from a score of +5 to a score of -5 for one alternative as it affected his different objectives. Participant 2 also saw his group's alternative as positively affecting two of his objectives and negatively affect- ing one of them. His beliefs about the effects of the other alternatives also varied, but, generally, the variance was quite small compared with the beliefs of Participant 1.

The degree of cognitive differentiation of the participants affected the deductive component of their decision processes. The more undifferentiated the overall perspective on the Middle East of the participant, the more likely it was that a single policy commitment would be used to organize his calculus on the relationship between policy alternatives and objectives. We found, in addition, that the scope of a participant's cognitive mapping of the Middle Eastern situation affected the inductive component of his decision process. The transcripts and post- gaming interviews with participants indicated, among other things, that the range of historical events utilized in interpreting the event in the game varied with the participants' cognitive complexity. The actions proposed by Participant 1, for ex- ample, reflected his analysis of the Cuban missile crisis:

What I am doing is deliberately trying to avoid writing a scenario of reactions that simply replicates the Cuban missile crisis. That's what I am trying to avoid. I have a great many more phase-steps of escalation in there than were a part of the 1962 situation, and consequently, there is nothing in this, proposal which suggests a direct attack on the Soviet Union without a whole range of intermediate steps that could be taken (which I am not sure I am willing to take, incidentally).

( 156] INTERNATIONAL STUDIES QUARTERLY

Participant 2 also drew on analogies from situations in other geographical regions:

Quite frankly, I'm somewhat more concerned about the evidence of Soviet pilots in MIG-21's transgressing the Israeli frontier than I am about the weapons as such. I think we're moving here into a situation where you could have, potentially, a Viet Nam type of situation with the Russians actively participating in a conventional type of conflict, at least on a limited scale in the air.

Participant 3, however, referred in the game and the interviews only to past events that took place in the Middle East.

The second theoretical perspective forming the basis for our model of the foreign policy decision-making process was suggested in our discussion of cognitive complexity-that is, the deductive structure of the decision-making. The degree of cognitive complexity in a decision maker's perceptual system is represented in terms of the number of deductive paths that he may examine in explaining the significance of an event and selecting an appropriate response. In order to understand the logic of any deductive path, we must specify the nature of and relations between the cognitive components in a path.

A variety of social-psychological theories have been devel- oped to explicate the deductive process involved in both attitude formation and decision-making. One type, for example, is modeled after the rational choice theories of economists which assume that the decision calculus is one which maximizes expected gains minus expected losses as in the following:

n

E(U)1- ViP. i= 1

where E(U)1 is the expected utility of choice 1, n is the number of possible outcomes or values, Vi is the ith outcome affected by the choice, and Pi is the probability that choice 1 will result in the ith outcome. This general formulation is commensurate

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING (1571

with a number of cognitive models which predict attitude formation and choice on the basis of beliefs about objects and situations and evaluations as to the relative importance of the objects or situations. Fishbein's (1967) attitude theory, for example, is expressed algebraically as follows:

n

AO Biai i=l

where AO is the attitude toward object 0, Bi is the strength of belief i about 0 (i.e., the probability that 0 is related to some other concept Xi), ai is the evaluative aspect of Bi or the evaluation of Xi and n is the number of beliefs about 0.

That a decision calculus such as Fishbein's assumes a deductive structure to decisions should be evident. If we process the calculus verbally, it says that an individual confronted with a policy alternative-e.g., sending the U.S. Sixth Fleet to a trouble spot in the Mediterranean-will consider the conse- quences such a move would produce by deciding the value of each consequence and the likelihood that it will be produced. He then will get a positive or negative score of some magnitude for the alternative, and his decision will then be (extending the logic of the calculus to a number of policy alternatives) to choose the alternative with the best (highest) score.

In the decision process we are positing as relevant to foreign policy decision-making, we are concerned not only with a decision calculus, but also with an explanatory calculus-i.e., the selection of one of a variety of plausible explanations of the event that has triggered the decision process. The choice among alternative explanations of a situation, like the choice among alternative possible actions, involves a calculus in which an evaluation is made of the amount of cognitive (deductive) support for various alternatives. For this reason, the social- psychological formulation of the economic, rational choice models is especially appropriate because it leaves to the decision

[ 1581 INTERNATIONAL STUDIES QUARTERLY

maker the specification of the number of beliefs about a given alternative, and thus allows for the combination of subjective, perceptual mappings of the subject matter (in this case foreign policy) and a structured decision calculus.

To illustrate the deductive structure of an explanatory calculus, consider the following hypothetical belief system of a foreign policy decision maker in the process of deciding the implications of Castro's Cuba for one of his major objectives, American security. Some cognitive support for an attitude might be adduced in a syllogistic form as follows (see Bem, 1970, for a similar presentation):

The Soviet Union's cold war activity constitutes a threat to American security. The Soviet Union and Cuba are cold war allies. Therefore Cuba is a threat to American security.

This derived belief might be further strengthened or lent more cognitive support as follows:

Hostile nations which are geographically close to the United States constitute a greater threat to American security than those at a distance Cuba is a hostile nation which is quite close to the United States. Therefore Cuba is a threat to American security.

And the same decision maker might reinforce this conclusion further:

The independence of Latin American nations from the Soviet sphere is essential to American security. Cuba threatens to bring more Latin American nations within the Soviet sphere of influence. Therefore Cuba is a threat to American security.

A partial belief system such as this is similar to what Abelson refers to, in his most recent reformulation of his cognitive theory, as an irnplication molecule, a self-consistent set of beliefs which is conjoined by one or two basic principles (Abelson, 1968).

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [1591

The deductive aspect of a decisional process as discussed above provides an understanding of the way in which informa- tion is combined in the foreign policy decision-making process. A comprehensive and effective approach to decision-making in such a setting must, in addition, address itself to the way in which information is selected by decision makers from among the welter of diverse cues which surround a decision situation. This third theoretical perspective contributing to our model of cognitive processing in foreign policy decision-making was also suggested in our discussion of cognitive complexity above. There we noted that, in varying degrees, the decision makers in our gaming situation used their perceptions of past situations as analogues to both explain and find a solution for their current decision situation. There is, in addition, considerable evidence that actual foreign policy decision makers have also attempted to resolve uncertainty in structuring and responding to decision situations by analogizing from the past. Former President Johnson (1971: 46), for example, reports:

When a President makes a decision he seeks all the information he can get. At the same time he cannot separate himself from his own experience and memory. This is especially true when his decisions involve the lives of men and the safety of the nation. It was natural, as I faced critical problems during those first months in office, that I should recall crises of the past and how we had met them or failed to meet them. No one who had served in the House or Senate during the momentous years of the nineteen-thirties, nineteen-forties, and nineteen-fifties, as I had, could fail to recall the many highs and lows of our performance as a nation, Like most men and women of my generation, I felt strongly that World War II might have been averted if the United States in the nineteen-thirties had not given such an uncertain signal of its likely response to aggression in Europe and Asia.

The theory of personality suggested by George A. Kelly provides a framework for building this part of the decisional process. An individual's cognitive system (for Kelly) is arranged in the form of an interrelated set of subordinate and superordi- nate constructs. This system of constructs, which Kelly saw as a

[160] INTERNATIONAL STUDIES QUARTERLY

series of dichotomous reference axes, provides the basis upon which the individual anticipates events and classifies his experience of them (Kelly, 1 955).

It is thus the categories decision makers utilize which determine the information that is selected and rejected in a decision situation. In order to prepare an informational base for the deductive decisional process such as that discussed above, we must know how the individual constructs of decision makers are utilized. Our model of the foreign policy decision-making process assumes that inductive as well as deductive support is frequently required for decisions to be reached. Some of the inductive support for decisions is "congealed" in the form of firm beliefs about aspects of international politics. These beliefs have been acquired on the basis of the individual's subjective understanding of past events and function as components in the deductive part of his decision process. Often, however, decision makers are uncertain about the significance of new events, and, in- the absence of firm beliefs about matters which directly bear upon possible decision alternatives, they scan past experience and attempt to analogize from what they regard, given the categories they utilize, as similar past experience. The selection of a past event as "similar" is a function of their constructs or conceptual model of the international system. Decision makers, thus, from our perspective, utilize history either implicitly, inasmuch as their firm operational beliefs are predicated on already digested and interpreted past events, or explicitly, as they verbalize or consciously employ analogies from the past. Thus, a major theoretical presupposition on which our model of the decisional process will be based is the impact of historical events on perceptions and images of foreign policy decision makers.

A Model of Foreign Policy Decision-Making

As a result of monitoring the decision process which took place during the gaming experiments, we drew conclusions not

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [161 ]

only about the structure and process of cognition in foreign policy decision-making but also about the sequence of phases within which the cognitive dynamics we discussed above are operating. Despite the differences in information-processing (primarily a function of degree of complexity) of each of the participants in our games, it is possible to ascribe to each decision-making group as a whole a single decision process. Although differences in substantive perspectives and informa- tion-processing seem to have had an impact on the information utilized and conclusions reached by each decision maker, individuals manifested the same kind of phasing with respect to the various components of the decision process. This similarity in phasing suggested a general model of foreign policy decision- making.

Representation of Beliefs

The major theoretical presupposition of our model is that beliefs of foreign policy decision makers are central to the study of decision outputs and probably account for more of the variance than any other single factor. Beliefs represent both the congealed experiences of the decision maker and his expecta- tions about the decision environment. In the decision-making process, beliefs act like templates for channeling information and for relating possible policy options to perceptions about the intentions and behavior of other nations, and also to the policy objectives of the decision maker.

Beliefs of decision makers are represented in the simulation as a map of causal linkages among three types of concepts. "Affective" concepts refer to the policy objectives or interests of the actors in the international system; "cognitive" concepts denote beliefs about actions that occur in the international system; and "conative" concepts indicate possible alternatives from which the decision maker selects policy recommendations. The linkages between concepts are represented as arrows which carry either positive or negative signs in order to distinguish the

0~~~~~~~~~~~~~~~~~~~~~~0C

0

0 ~~~~~~~~~~~cc

(0 00.~L 4 4 4 < 0 0(0 ~~~~~~0

4.' V~~~~~~4 o 0~ 0 ~ ~ ~ ~ ~ ~ ~~~~ (~~)0

0)

[162 1~~~~~~~~~~~~~~~~~~~~~~~~~C

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [163]

direction of the causal relationship perceived by the decision maker. Taken together, the concepts and the causal linkages between them form a "cognitive map" of the decision maker's belief system. It is this cognitive map which allows a decision maker to relate an event or a series of events to policy alternatives and policy objectives.

This representation of decision makers' beliefs reflects the proposition that decision makers tend to believe that interna- tional events are causally related and thus try to infer causal relationships underlying events and the actions of other nations, even when there is little or no evidence of a causal nature. Jervis (1970: 29) has observed this tendency in his study of the use of signals and indices by foreign policy decision makers. This proposition is also supported by research on attribution phenomena in social psychology, which holds that a person's motivation to exercise control over his environment is related to his attribution of causal relationships to the behavior of others (see Abelson and Reich, 1969; Kelly, 1971; Kanouse, 1971).

Amplification of Beliefs and Search for Explanations

Four processes are invoked in the simulation model when a decision maker is confronted with a new international situation that requires a response from his government: the amplification of beliefs, the search for an explanation, the search for policy options, and the choice of a policy (see Figure 1).

During the amplification of beliefs process, the decision maker attempts to put the new international situation into the context of his experiences. This is a process of bringing together various components of the situation with his existing beliefs about the nations and actions involved so that the decision maker can define the situation. It is similar to the cognitive problem that Abelson (1968: 136-139) has simulated with his ideology machine, which makes novel events understandable by referring them to a structure that has interpreted them in the past: a set of beliefs about past events which are stored at different levels of abstraction.

[164] INTERNATIONAL STUDIES QUARTERLY

At the amplification of beliefs stage of the simulation, concise statements describing new international developments are fed into the model to activate concepts in the decision maker's belief system. When an event exemplifying a concept is received as an input, that particular concept is "highlighted," and this information is stored in the model's memory. After all concepts directly pertaining to a new international situation have been highlighted, the simulation searches for additional concepts that would be highliglhted by implication (i.e., those that are reachable from the initially highlighted concepts). Once causal paths have been followed from initially highlighted concepts, the decision maker has an amplified set of concepts which is a subset of his conceptual overviews of the interna- tional environment. This subset is then utilized in the subse- quent phases of the decision process.

The second decision-making process, search for explanations, is somewhat more involved than the first process. As noted in the flow chart (see Figure 1), a simulated decision maker must determine whether he possesses an adequate explanation of what has occurred. An explanation in our model consists of the arrows or "paths" connecting a set of two or more highlighted concepts. The initial concept in an explanatory path is most often an actor's intention or motivation. The path is a sequence from intention concept to the consequence or set of conse- quences that were initially input in the amplication of beliefs phase of the decision process. Once new information has activated concepts in a decision maker's cognitive system, the model searches for arrows between highlighted concepts and isolates them for further processing. For example, a participant in one of the political games we have used to study decision- making processes made the following statement about a crisis that had developed in the Middle East:

It would appear to me that the continuous occupation of Arab territories by Israel is precisely this causal factor that makes the Arab regimes open to proffers of Soviet military aid and brings more and more Soviet penetration into the Middle East, therefore bringing

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [165]

us all to a sort of brink in the world situation. It tends to polarize more the situation, and it tends to make the Arab states more dependent on the Soviet Union.

This statement was coded as part of his cognitive map and is shown in Figure 2 as a structure consisting of six concepts linked together by five positive arrows. When this person's cognitive map was used later in a simulation of a real crisis, the 1970 war in Jordan, the concepts shown in Figure 2 were all activated by the model and then isolated from the rest of his cognitive structure as one possible explanation for the crisis in Jordan.

The first two processes, amplification and search, are a representation in the model of the proposition that decision makers tend to fit incoming information into their existing theories and images (this is noted by Jervis, 1968: 455). The crucial role of images in international relations has been discussed by Boulding (1959, 1956). More recently, Jervis (forthcoming; 1968) has analyzed effects of images on informa- tion-processing by policy makers and has illustrated this proposition with numerous historical examples. In addition, Holsti (1962) and Finlay et al. (1967) have tested this proposition in a study of former Secretary of State John Foster Dulles (see also Axelrod, forthcoming; 1972).

At the operational level, amplification and search processes are accomplished by treating the cognitive maps of decision makers as directed graphs or "digraphs" and using the rules of digraph theory for making calculations. Digraph theory, a formal system with elaborate rules for moving about in a network of interrelated elements, provides an inference struc- ture that is convenient for both representing the elements in a cognitive structure and calculating cognitive processes. Calcula- tion is greatly facilitated because the inventors of digraph theory have worked out the relationships between digraphs and matrix algebra, so it is possible to manipulate relationships between and among elements. The set of rules (axioms, primitives, theorems, and so on) which constitute the theory of

0 00

/~~~

Ul*

QL.@

0 >~~~ Cl)~~~~~~~,

Oo Cl) CC~~~~~ < 0 0

X ,0 O

.N z~~~~ 4-~~~~~~~~~~.

0~~~

0 0

+ +* Z

0 cc~~~~~~~. 0 LL~~~~~~~~

COm 4)

0.4~~~~~~~~~~~~~~~~~~0

LL

[166]

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [167]

directed graphs is far too elaborate to treat here (see Harary et al., 1965; Harary, 1961). We shall simply note, as we proceed, those rules which directly pertain to the explication of our model of foreign policy decision-making. For example, we use the idea of "reachability" in the amplification and search calculations. In digraph theory, reachability suggests that concepts connected by paths to initially highlighted concepts will also be highlighted. Thus, if concept V1 in digraph D (see below) is initially highlighted, then the other three concepts, V2' V3, and V4, which are reachable from V1, would be highlighted by implications.

D=V1 V3

V4

The search for additional concepts is accomplished in the computer by constructing a reachability matrix as a basic input for the simulation and then checking relationships in the matrix.

When a decision maker possesses only one explanation for a policy situation, the simulation goes on to the next stage, the search for alternative courses of action. However, if there is more than one explanation in his cognitive system, the model selects the one with the most "deductive support." For our purposes, deductive support is defined as the number of logically independent reasons that reinforce an explanation, and it is calculated by counting the number of separate paths comprising a given explanation. At the operational level, this is accomplished by constructing and manipulating an adjacency matrix.'

This subroutine of the simulation model follows directly from the proposition that decision makers tend to rely on

1. "An adjacency matrix A of a digraph D is a square matrix with one row and one column for each point of D, in which the Entry Aij 1 if the line vivj is in D and Aij = 0 if vivj is not in D" (Harary et al., 1965: 408). The adjacency matrix A(D) for digraph D above is as follows:

[168] INTERNATIONAL STUDIES QUARTERLY

explanations supported by several logically independent rea- sons. This is stated in this form by Jervis (forthcoming), who notes that it is an application of consistency theories and is compatible with theories dealing with avoidance of psycho- logical conflicts (see Kelman and Baron, 1968; Abelson and Rosenberg, 1958; Abelson, 1968). In addition, research on cognitive complexity provides some evidence for this proposi- tion (Bieri, 1955; Driver, 1962; Schroder et al., 1967).

A decision maker's belief system is usually dense enough for him to find an explanation for a policy situation, but there are occasions when he lacks an adequate explanation for an event or series of events. If there is not any explanation in a decision maker's belief system with sufficient deductive support, then the model begins searching for inductive explanatory support, as indicated in the flow chart. The simulated decision maker scans his memory and attempts to find similar past events which might provide an explanation for the current situation. During the search, the relevance of historical events is determined for the decision maker by the categories or constructs through which he views the international system.

The logic of the inductive search process (in both the explanatory and choice of an alternative phases) is similar to the precedent search in Alker and Greenberg's (forthcoming) model

V1 V2 V3 V4

V1 [ 1 0 1

V2 I 0 1 0

V3 [ 0 0 1

V4 0 0 0 ? J

Each nonzero element in an adjacency matrix (reading by rows) indicates the number of paths (of length 1) from the row element to all other elemtns. If we raise the matrix to the power of 2, the nonzero elements in the matrix will indicate the number of paths of length 2. Generally, by raising the adjacency matrix to the nth power, we can discover the number of paths of length n from each element to every other element. Thus by raising the matrix to the power equal to the longest path in the reachability matrix, we can count, by adding across matrices, the number of paths from each highlighted concept to every other highlighted concept. In terms of our theory of decision-making, this process is one of counting the deductive support for various explanations of the international situation.

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [169]

of UN decision-making. The search of past events scans the stored events and initially chooses the event with the most concepts in common with the current situation. If more than one event has the same number of shared concepts, the most recent event is selected. The search then proceeds lexicographically. If the explanatory or choice of alternative rule in the first historical event leads to a selection from among two or more possible paths, the search is ended. If the first event selected does not successfully continue or terminate the decision process, a second one is selected, using the same decision rules, and so on, until a past event is found that overcomes the uncertainty that led to the inductive search process.

The inductive search process of this part of the model is responsive to the proposition that decision makers, in the absence of firm beliefs about new events, tend to rely on previous experience. Writers on foreign policy have often made reference to the significance of historical events in shaping images and decisions. This position has been stated succinctly in Holsti et al. ( 1968: 125):

Essentially, then, it is by projecting past experience into the future that human beings make decisions; and statesmen, in this respect, are not exceptions. Foreign policy decisions, like other human decisions, imply not only an abstraction from history, but also the making of "predictions"-the assessment of probable outcomes.

Effects of historical events have been studied by Deutsch and Merritt (1965), and Jervis (1968: 470-472) has analyzed how decision makers analogize from historical experience. Further- more, the proposition is a basic assumption of the CASCON (Beattie and Bloomfield, 1969) and CACIS (Tanter, 1971) models.

Search for Options and Policy Choice

The next decision-making process is the search for an acceptable course of action. Once various alternative explana-

[170] INTERNATIONAL STUDIES QUARTERLY

tions of the situation have been sorted out, the decision maker searches for alternative policies with which to respond. To accomplish this, he follows the explanatory paths from the policies he might select (conative concepts) to the objectives he feels are at stake in the situation (the affective concepts) and calculates which alternative or combination of alternatives will result in the maximum net gain in objectives.

The signs of the causal linkages (as well as the probability weightings) in the decision maker's cognitive map are crucial at this point in the process, because the model must calculate how each policy alternative will affect every policy objective to which it is connected by an explanatory path. At the operational level, we used a signed digraph, which takes into account whether relationships between concepts are positive or negative. The signed digraph is converted into an adjacency matrix which preserves the signed relationships. This kind of matrix, which is called a valency matrix, makes it possible to determine whether there is a positive or negative path from a specific conative concept to one or more affective concepts.2

The search for policy options in our model is restricted to those that are embedded in a decision maker's explanation of a situation. The logic of the search is simply that a policy maker looks for options that he thinks will give him some control over events in the international system. He thus selects a policy that he believes will set off a series of events that will have an impact on his policy objectives. This proposition is similar to the notion of an "isomorphism of experience and action," which was first introduced by Asch (1952) and has since been discussed extensively by Campbell (1963). They hold that a person's views of the world and his tendency to respond to the world are essentially equivalent.

As in the case of the explanatory phase of the decision process, when a decision maker finds that he is ambivalent about possible alternatives or none seems adequate, he reflects on relevant historical experiences in an attempt to adduce additional concepts which will lead to an acceptable alternative.

2. The procedure parallels the explanatory calculus described above. The valency matrix is raised to powers to capture policy implication paths and their signs.

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [171]

The choice of a policy option from among a number of possible alternatives is the final decision-making process in our simulation. There are, of course, a variety of decision models one can employ to deal with the trade-offs involved in this calculus. For our simulation, we employ a lexicographic decision calculus which assumes that the decision maker first uses his most important policy objective to see if the alterna- tives affect it differently. If this objective does not distinguish between alternatives, he then moves to his second objective, and so on, until he gets to an objective that distinguishes one alternative as better than the others.

In the simulation, the choice process is operationalized as follows: A valency matrix is used to trace the connections between policy options and policy objectives, and then the model invokes a lexicographic decision calculus. To decide whether a particular policy objective distinguishes among two or -more policy alternatives, the model takes into account not only the sign of the relationship but also the number of paths involved. For example, there may be three positive paths from a particular policy alternative to a given policy objective, whereas other alternatives have only one or two positive paths to the same objective. After a policy option is selected, the model recycles by adjusting the decision maker's cognitive system to take into account any new or strengthened causal linkages and stores information about the decision situation for future reference.

Conclusion

In another paper (Bonham and Shapiro, forthcoming) we reported that, on the basis of a post-Jordanian War interview with the Middle East specialist whose cognitive mapping we have used here for purposes of illustration, we successfully predicted what his response would have been as an advisor to the President. In this paper, however, we are more concerned with the theory-building and policy recommendation implica-

[172] INTERNATIONAL STUDIES QUARTERLY

tions of our model. At a theoretical level, the model is coherent with various social-psychological perspectives that have been successful predicates for experiments on attitude formation and decision-making. We have integrated these perspectives in a cognitive process model of foreign policy decision-making, while preserving both the conceptual bases and operational stipulations implied in each perspective.

The policy implications of the model are particularly noteworthy. One implication which suggests itself would be examining the kinds of cognitive adjustments, both in terms of the structure of a cognitive mapping of the international political system and in terms of the stored perceptual histories of past events, that would result in policy choices different from those that have been actually pursued. This is one of the major implications we intend to pursue once we have enough data to represent the cognitive mappings of the various foreign policy groups that have actually been responsible for foreign policy choices.

REFERENCES

ABELSON, R. P. (1968) "Psychological implication," pp. 112-139 in R. Abelson et al. (eds.) Theories of Cognitive Consistency: A Source Book. Chicago: Rand McNally.

---and C. M. REICH (1969) "Implicational molecules: a metlhod for extracting meaning from input sentences," in D. E. Walker and L. M. Norton (eds.) Proceedings of the International Joint Conference on Artificial Intelligence, (May).

ABELSON, R. P. and M. J. ROSENBERG (1958) "Symbolic psychologic: a model of attitudinal cognition." Behavioral Sci. 3: 1-13.

ALKER, H. R., Jr. and W. J. GREENBERG (forthcoming) "The U.N. charter: alternative pasts and alternative futures," in J. D. Ben-Dak (ed.) The Simulation of Inter-Societal Relations. New York: Gordon & Breach.

ASCH, S. E. (1952) Social Psychology. Englewood Cliffs, N.J.: Prentice-Hall. AXELROD, R. (forthcoming) "Schema theory: an information processing model of

perception and cognition." Amer. Pol. Sci. Rev. 67 (December). --- (1972) "Psycho-algebra: a mathematical theory of cognition and choice with an

application to the British Eastern Committee in 1918." Papers of the Peace Research Society (International) 18: 113-131.

BEATTIE, R. R. and L. P. BLOOMFIELD (1969) "CASCON: Computer-Aided System for Handling Information on Local Conflicts." Center for International Studies ACDA/SEC-141, Cambridge, Massachusetts, December.

Shapiro, Bonham / FOREIGN POLICY DECISION-MAKING [ 1731

BEM, D. J. (1970) Beliefs, Attitudes, and Human Affairs. Belmont, Calif.: Brooks/Cole.

BIERI, J. (1955) "Cognitive simplicity-complexity and predictive behavior." J. of Abnormal and Social Psychology 51: 263-268.

BONHAM, G. M. and M. J. SHAPIRO (forthcoming) "Simulating foreign policy decision-making: an application to the Middle East," in J. D. Ben-Dak (ed.) The Simulation of Inter-Societal Relations. New York: Gordon & Breach.

BOULDING, K. E. (1959) "National images and international systems." J. of Conflict Resolution 3 (June): 120-131.

(1956) The Image. Ann Arbor: Univ. of Michigan Press. CAMPBELL, D. T. (1963) "Social attitudes and other acquired behavioral disposi-

tions," pp. 94-172 in volume 6 of S. Koch (ed.) Psychology: A Study of a Science. New York: McGraw-Hill.

DEUTSCH, K. W. and R. C. MERRITT (1965) "Effects of events on national and international images," pp. 130-187 in H. C. Kelman (ed.) International Behavior. New York: Holt, Rinehart & Winston.

DRIVER, M. J. (1962) "Conceptual structure and group processes in an inter-nation simulation." Ph.D. dissertation. Princeton University.

FINLAY, D. J., 0. R. HOLSTI, and R. FAGEN (1967) Enemies in Politics. Chicago: Rand McNally.

FISHBEIN, M. (1967) "A behavioral theory approach to the relations between beliefs about object and the attitude toward the object," in M. Fishbein (ed.) Readings in Attitude Theory and Measurement. New York: John Wiley.

HARARY, F. (1961) "A structural analysis of the situation in the Middle East in 1956." J. of Conflict Resolution 5 (June): 167-178.

---R. Z. NORMAN, and D. CARTWRIGHT (1965) Structural Models: An Introduction to the Theory of Directed Graphs. New York: John Wiley.

HERMANN, C. F. (1969) Crises in Foreign Policy. Indianapolis: Bobbs-Merrill. --- and M. G. HERMAN (1967) "An attempt to simulate the outbreak of World

War I." Amer. Pol. Sci. Rev. 61 (June): 400-416. HOLSTI, 0. R. (1962) "The belief system and national images." J. of Conflict

Resolution 6 (September): 24-52. ---R. C. NORTH, and R. A. BRODY (1968) "Perception and action in the 1914

crisis," pp. 123-158 in J. D. Singer (ed.) Quantitative International Politics: Insights and Evidence. New York: Free Press.

JERVIS, R. (forthcoming) Perception and International Relations. --- (1970) The Logic of Images in International Relations. Princeton, N.J.:

Princeton Univ. Press. ---(1968) "Hypotheses on misperception." World Politics 20 (April): 454479. JOHNSON, L. B. (1971) The Vantage Point. New York: Holt, Rinehart & Winston. KANOUSE, D. E. (1971) Language, Labeling, and Attribution. New York: General

Learning Press. KELLEY, H. H. (1971) Attribution in Social Interaction. New York: General

Learning Press. KELLY, G. A. (1955) The Psychology of Personal Cpnstructs. New York: W. W.

Norton. KELMAN, H. C. and R. BARON (1968) "Inconsistency as a psychological signal,"

pp. 331-336 in R. Abelson et al. (eds.) Theories of Cognitive Consistency. Chicago: Rand McNally.

[1741 INTERNATIONAL STUDIES QUARTERLY

LEVINSON, D. J. (1957) "Authoritarian personality and foreign policy," J. of Conflict Resolution 1 (March): 37-47.

ROSENBERG, M. J. (1956) "Cognitive structure and attitudinal affect." J. of Abnormal and Social Psychology 53: 367-372.

SCHRODER, H. M., M. J. DRIVER, and M. STREUFFERT (1967) Human Information Processing. New York: Holt, Rinehart & Winston.

SHAPIRO, M. J. (forthcoming) "Cognitive rigidity and moral judgments in an inter-nation simulation," in H. Guetzkow (ed.) Simulation in International Relations: Findings from Research.

STEINBRUNER, J. D. (forthcoming) Decisions Under Complexity. Princeton: Princeton Univ. Press.

---(1970) Excon III (A Middle Eastern Crisis Game), Instruction Manual. Cambridge, Mass.: MIT Center for International Studies.

TANTER, R. (1971) "A computer aided conflict information system (CACIS)." University of Michigan, Department of Political Science, December. (mimeo)

WASHBURN, A. M. and W. H. MITCHELL (1967) "W. Rostow, Vietnam and the future of American foreign policy." Princeton, N.J.: Princeton University Center of International Studies Policy Memorandum 33.

SCANDINAVIAN POLITICAL STUDIES A Yearbook published by the Political Science Associations in Denmark,

Finland, Norway and Sweden

With Volume 8 (1973), Sage Publications is pleased to announce it has assumed co-publishing responsibility for this annual series, which focuses attention on the politics of the Scandinavian societies, and offers challenging insights into the political implications of experiments with soci welfare programs, as well as such topics as taxation policy, prejudice, political parties, rapid industrialization and labor radicalism in the Scandinavian nations whose societal innovations have long been of interest to other post-industrial societies.

VOLUME 8 (1973), scheduled for publication November/December 1973, and edited by Stein Rokkan, Helen Aareskjold, and Helga Hernes, will contain articles on the theory of negotiation recent Danish elections, referenda held in Norway and Denmark concerning entry into the Corm mon Market, and reviews of other recent political developments in Scandinavia. Code #9032

Approx. 320 pp Bibliography & Index $15.00 (clothbound

ASK YOUR LIBRARY TO ENTER A STANDING ORDER FOR THIS ENTIRE SERIES Special offer-to those entering a standing order with Volume 8, the entire 7-year back volume set may be purchased for only $84 (regular price $105). To take advantage of this limited-time offer, send orders by October 1, 1973.

SAGE PUBLICATIONS, INC. SAGE PUBLICATIONS LTD 275 S. Beverly Dr. / Beverly Hills, CA. 90212 44 Hatton Garden, London E C 1