Resumé

In this paper one of the classical questions in the study of politics will be
pursued, viz., whether foreign policy decisions are made in a less democratic
manner than policies pertaining to the domestic sphere of fully industrialized and
democratic states; and if so, how to account for such a condition. These two
queries will be examined in terms of three topics, each of which is explored in a
preliminary fashion: (a) the characteristic dimensions of the democratic process;
(b) the problems of policy comparability across sectorial boundaries; and (c)
possible explanatory hypotheses regarding sectorial variances in the democratic
process.

1. Introduction

It remains a curious fact of contemporary democratic politics that the conduct of foreign policy is still held to be the natural prerogative of the executive branch, or at least that this constitutes a sphere of politics in which the government of a country enjoys special privileges of access, secrecy, and exclusiveness. Various factors have been offered in explanation of this political anomaly, but the most common justification is probably the hallowed assumption that foreign policy pertains to a special domain, which in a peculiar and significant sense remains 'above' or 'beyond' the normal processes of domestic politics and partisan strife. In the words of John Locke's famous dictum, which is also something of a

---

* This paper has been written under the auspices of a project on 'Democracy and Bureaucracy in Domestic and Foreign Policy', financed by the Bank of Sweden Tercentenary Fund. For their comments and suggestions I wish to thank various individual colleagues, the other members of this ongoing project, as well as the participants at the seminars on this topic held at the universities of Uppsala and Stockholm.

locus classic-its in this context, 'the power of war and peace, leagues and alliances, and all the transactions with all persons and communities without the commonwealth' constitute the 'federative' power, which must 'necessarily be left to the prudence and wisdom of those whose hands it is in to be managed for the public good' (Wallace 1976, p. 1). But does this cornerstone of classical statecraft, so characteristic of the Westphalian system of diplomacy, continue to be valid with regard to the actual formulation of foreign policy or to the democratic means available to the public for affecting or controlling the pursuit of foreign relations? And is David Vital's contention - submitted some years ago - still true today, that not only is the above assumption of, or doctrine on, the role of government in foreign policy commonly accepted as fully legitimate, but also that for 'almost all practical purposes the Executive is unfettered in its exercise of this function' (Vital 1968, p. 49)? In short, does this claim - and others similar to it - accurately reflect the influence of the public in this sphere? And if so, what explanation(s) can be adduced to account for what to many must prima facie seem to be a democratically disturbing condition? These are obviously most pertinent queries, both with reference to normative democratic theory and the comparative analysis of contemporary policymaking in modern states.

In this research note I shall not, however, attempt to answer these questions, but merely suggest ways and means of going about doing so, and this in the following manner: first of all, the nature of the democratic process itself will be discussed in terms of three dimensions which, it will be claimed, crucially determine the nature of this process qua dependent variable. Thereafter I will propose a comparative framework for the empirical analysis of this process, which will involve us in a discussion of the notion of policy areas as distinct from sectorial, administrative and/or jurisdictional boundaries. In the final section a number of suggestions and hypotheses will be proffered for accounting for possible variances in the democratic process between the foreign policy area and other, more domestic sectors of government. It should also be stressed at this juncture that this paper constitutes part of a large and ongoing research project, and hence pretends to be nothing more than exploratory and - it is hoped - suggestive.

2. Dimensions of the Democratic Process

A traditional way of tackling our topic as a whole is to examine the
procedures which have been provided for the democratic control of a

state's foreign affairs. These may vary widely with reference to their democratic characteristics, depending on the institutions and constitutional arrangements of a country; and historically it is clear that these have e vol ved differently in different political systems, ranging from a very strict delimitation of the formal means available to the public for controlling the government's actions in this area, to provisions which put very clear limits on how freely a government can act without consulting in some way the public or its elected representatives. Let it be said immediately, however, that it is highly questionable if such a study will be able to tell us very much, since - amongst other things - the procedural aspects in this area very often either hide the actual processes of decision-making (which, e.g., may indeed be more democratic than these procedures might suggest), or give a picture which, from a formal democratic point of view, is far rosier than it in reality is. Institutions and procedural structures are never clear-cut with regard to their effects on the processes of a system; and because of their formal nature they also tend to obscure developments in actual practice, as well as influences on the concrete conduct of foreign policy which are not susceptible to procedural control.

The obvious alternative is an approach emphasizing the analysis of the political processes involved in the formulation and establishment of foreign policy. However, democracy qua political process is in itself a very broad conception, containing various and not always easily determinable or even compatible meanings. It therefore needs to be specified in a more analytic as well as empirical manner if it is to prove fruitful for our purposes.

(a) One such meaning - perhaps the most common one - refers to the participation of the populace in public affairs; or more specifically, given our present concerns, to the participation of the public in the establishment of foreign policy choices. An emphasis on this aspect also immediately propels us beyond a procedural definition of democracy, i.e., from merely analyzing formal access to power, to a consideration of the actual exercise of political power in a society. And clearly this is an aspect of the democratic process which enjoys a central position in the literature; indeed, often the democratic ideal is identified wholly with the participatory dimension. 'If democracy is interpreted as rule by the people', Sidney Verba and Norman H. Nie thus note in the opening lines of their widely quoted study of political democracy and social equality, 'then the question of who participates in political decisions becomes the question of the nature of democracy in a society. Where few take part in decisions there is

little democracy; the more participation there is in decisions, the more democracy there is.' As these two authors are the first to admit, this is, however, a crude measure of the democratic process, particularly if participation is defined broadly in terms of'those activities by private citizens that are more or less directly aimed at influencing the selection of governmental personnel and/or actions they take' (Verba and Nie 1972, pp. If.). For it is obvious that participation as specified by them involves various kinds of behaviour, on different levels of policymaking and influence, and with reference to disparate policy arenas. How these factors affect the democratic nature of participation is far from obvious, however, particularly in view of the variform implications of the different types of issues which are involved in the political system (see, e.g., Redford 1969).

Also, there is with regard to the democratic implications of participation an ongoing controversy between those democratic theorists who emphasize the sine qua non of leadership in the democratic process, and those who stress the pre-eminent and consummatory worth of'participatory'participat- democracy', both as a means and as a goal. This normative conflict has also been presented in terms both of an 'elitist' and a 'populist' version of the democratic ideal, and with reference to 'functional' and 'normativist' descriptions of the democratic model (see Pennock 1979, and Lewin 1970). The crucial point here is the ostensive imbalance which today exists in all Western societies between a substantial segment of uninterested nonparticipants in politics and a small group of full-time professional politicians, and the implications of this state of affairs for the democratic exercise of public power. Theorists like Robert A. Dahl have accepted the reality of this imbalance as an inescapable empirical fact, and have as a consequence revised classical democratic theory in favour of a 'functional' or 'operational' doctrine which reflects this reality, while a different (and generally younger) group of theorists have attacked such 'revisionist' attempts as normatively insidious, since in their view these doctrines tend to accept political facts which should, instead, be deplored and openly counteracted. Given the nature of this conflict it is not difficult to accept the appositeness of J. Roland Pennock's laconic comment, in his recent and massive treatise on Democratic Political Theory, that this discussion 'runs along ... lines, which have the characteristic of parallel lines in that they never meet' (Pennock 1979, p. 580). It is nevertheless an aspect which cannot be ignored, since this controversy pertains directly to participation as a criterion of democracy, and thus to the question how 'participation' should be used in the analysis of how 'democratic' foreign policy-making is in a given political system.

To bring some order to 'participation' as a democratically relevant and empirically manageable variable, Kjell Goldmann has recently suggested that the concept be differentiated in terms of three different scales or dimensions, defined with reference to - or in juxtaposition with - a bureaucratic ideal type involving only intra-administrative participation (Goldman 1981). All three of these dimensions have already been alluded to above, but Goldmann's elaboration gives them clearer contours at the same time as his discussion illustrates how decision-making in contemporary modern societies can vary significantly depending on - and these are his dimensions - bureaucratic, party-political, interest organizational and action group processes of decision-making, and especially on differences in the interplay between them (see Olsen 1978, Chaps. 1 and 2).

In terms of this perspective, democratic participation pertains to, or can be defined with reference to, the relative power and influence exercised in the decision-making process by, respectively, elected politicians, organized interest groups (and other private organizations), and more ad hoc and specially organized action groups. On the basis of this the following stipulations can be posited: when none of these three is decisively involved in the making of public policy, we have (i) a pure bureaucratic system (and hence no democratic participation to speak of); while if any one of these dominates the process to the virtual exclusion of the other two, then the system can be viewed as, respectively, (ii) a representative democracy, (iii) a corporativist state or (iv) a pure participatory democracy ¹ Usually, however, Western societies are characterized by a combination or mix of these ideal types, which not only varies from country to country but also over time in one and the same polity. Furthermore, and this is even more important in the present context, it is probably wise to assume that this mix also varies within societies at the same point in time, depending on the social sector or policy area which is involved. In other words, we here have a potential measure for comparing the participatory nature of the decision-making process in such sectors as foreign policy with other, more traditionally 'domestic' political arenas.

This does not mean, however, that we at the same time have been provided with a criterion for assessing which of these participatory forms comprises the more democratic process of the three: this remains a controversial issue in democratic theory, and therefore merits the kind of normative and philosophical discussion which cannot - and in fact need not for our purposes -be provided here. However, it is clear (and any such discussion will have to take cognizance of this claim) that while participation is central to the democratic process, it nevertheless constitutes only

one dimension of it, and thus needs to be supplanted by other criteria referring to aspects of this process which are not subsumed by this concept. This is particularly important to stress in view of the faet, emphasized for example by Verba and Nie, that participation is not an end in itself but only an instrumental activity, while democracy is usually thought to incorporate cerVd'mgoals of a more transcendental kind as well. One such hallowed notion is that democracy is to be considered not only as rule by the people, but also in an equally significant sense as rule for the people.

(b) This notion immediately leads us to a second major dimension of the democratic process, viz., to the prposition that government should represent the citizens of a state in one way or another, or, more specifically, that it should represent the interests or preferences of these citizens. This, too, is an übiquitous notion in the literature on democracy, despite the fact (as Hanna Pitkin has noted in what is perhaps the definitive contemporary study of this concept) that 'through much of their history both the concept and practice of representation have had little to do with democracy or liberty' (Pitkin 1967, p. 2). Indeed, as Pennock has suggested, political representation in the broadest meaning of the term is not necessarily democratic at all, since in 'a proper sense of the word, all legitimate governments are "representative" ', while not all legitimate governments are democratic (Pennock 1979, p. 309). It is for this reason that even the most absolute rulers - ancient monarchs and contemporary dictators alike - have attempted to legitimize their power by asserting that they represent their people, either by claiming to stand for their interests, ideals or aspirations or by being authorized to act on their behalf. Pitkin's penetrating analysis of the concept also provides an abundantly detailed illustration of the various conceptual and logical conundrums associated with the term - problems which are not alleviated by the fact that representation' is a cornerstone in the political philosophies of such central and at the same time mutually incompatible and disparate thinkers as Hobbes, Burke and Mill.

Due to these and other difficulties associated with the 'representative' aspect of the linkages between an elected body and the electorate, I suggest that we instead use the term responsiveness to denote this dimension of the democratic process. More specifically, I here have in mind 'policy responsiveness', defined as the degree to which political decisionmakers create policies which are congruent with manifest public demands. This aspect of the democratic process thus refers to the content of policies,

and to the relationship between it and the articulated preferences of the citizens of a given state. This link has also been described by Verba and Nie in somewhat more general terms as follows: 'The term responsiveness] refers to a relationship between citizen and government, one in which the citizen articulates certain preferences and/or applies pressure on the government and the government in turn - if it is responsive - attempts to satisfy these preferences. To measure responsiveness, therefore, one needs measures of citizen preferences and activities as well as information on the attitudes and activities of leaders' (Verba and Nie 1972, p. 300). It is important, furthermore, to stress the manifest, articulated nature of this dimension, and hence to exclude clearly from it references to latent needs or other nonexplicit preferences or desires. This does not mean, of course, that the quintessentially 'responsive' representative is merely an unquestioning instrument for the fulfilment of peoples' wishes; hence the above dimension does not implicate the notion which Heinz Eulau and Paul D. Karp have ascribed to this conception, viz., that this type of responsive representative 'is at best the executor of the group's will, indeed a human facsimile of Pavlov's dog' (Anckan, 1980, p. 33). As Dag Anchan has argued against this supposition, a responsive decisionmaker need not at all be a passive actor who does not, or cannot, attempt to mobilize and change the public's preferences and their notions of whatas- he sees it - is in their real or best interest. However, if the public does not wish to be mobilized in this way, responsiveness entails that - for better or worse - the decision-makers yield to the wishes of the populace. If this is not done, then these are no longer to be regarded as responsive: they then represent their people in a nonresponsive manner, in which case - if the process is to be regarded as democratic - the public must possess the electoral means for selecting new and presumably more responsive decision-makers to represent them and their interests. This also implies that the more palpable the availability of this instrument of public choice and control, the more democratic the process of government will be - at least with reference to responsiveness.

So far we have defined the democratic process in terms of two dimensions: participation, which refers to the process of decision-making in the broad sense of how decision-makers are chosen; and policy responsiveness, which pertains to the content of the policies made by these elected decision-makers, and particularly to the link between these policies and the manifest preferences of the public at large. The relationship between these two aspects of the democratic process is both obvious and close; or as Verba and Nie have put it, 'Responsiveness is what democracy is

supposed to be about and, more specifically, is what participation is supposed to increase' (Verba and Nie, 1972, p. 300). In short, responsiveness can be viewed in terms of leadership reaction to public participation, which in turn is a reflection upon this response. However, following a liberal tradition in classic democratic theory, a third dimension will also be introduced to characterize this process - a dimension which at first sight may seem less obvious than the two previous ones, and one which certainly has not enjoyed the same emphasis in the contemporary literature. It must, however, be regarded as an essential ingredient of any democratic process: indeed, as perhaps in a crucial sense more fundamental than the two dimensions already discussed.

(c) What I have in mind is the role in public decision-making of rationality in general and the availability of correct and unbiased information in particular. It is obvious that no democratic process worthy of the name is well served by either constituency or governmental ignorance and irrationality, however high participation may be and however responsive government is. Indeed, responsiveness to an ignorant but mobilized rabble is probably what mob rule is all about, while a misinformed but keen responsiveness on the part of the ruling elite is the surest way to executive irresponsibility. It is in this sense that correct and full information on political issues is a prerequisite for responsible and rational action on the part of both citizens and their governments.

The desideratum here is that the democratic process be characterized by (as Pennock has put it) 'responsibility in the sense of rationality, taking action on the basis of full information and careful deliberation after hearing arguments from all sides' (Pennock 1979, p. 302). And this clearly requires a political culture which is responsible both to public demands for full and unskewed information on any issue on the public agenda, as well as to the more specialized needs of elected decision-makers and their staffs. Three different aspects of rationality qua informed system of decision-making are thus involved here: an enlightened and critically reflective public, a corps of politicians sufficiently well-informed not to be the pawns of experts and professional bureaucrats, and a dynamic arena of public debate not beholden to any particular - private or public - interests. It is no secret that the ever increasing volume and complexity of the issues facing modern societies tend to have adverse and deeply problematic effects on all three of these aspects of public decision-making; hence, as both normative and empirical criteria for evaluating the rationality of the democratic process, their importance cannot be underestimated.

We can now once again return to our starting point, i.e., to the question whether political decision-making is less democratic in the area of foreign policy than it is in policy sectors of a more typical 'domestic' coloration; and if so, how to go about accounting for this. The first query can now be restated in a more differentiated and palpable manner as follows: is the process of decision-making in the area of foreign policy less democratic than in other areas in terms of, respectively, (i) participation, (ii) responsiveness and/or (iii) rationality? Our dependent variable has in other words now acquired three dimensions, at the same time as it has become more complicated, since it is not at all a priori obvious that the answer to our question will be the same with reference to all three of these. This is a consequence of viewing democracy as a multidimensional phenomenon rather than in terms of a single scale. With regard to our second query, it is clear that if significant democratic variances in the above sense indeed do exist, explaining their presence will perforce involve us in the examination of each dimension separately, since it is not at all evident that a single explanation (or hypothesis) can be adduced to account for the supposed democratic variances in toto. And even if we do find that the foreign policy sector is treated less democratically in terms of all three of our dimensions, it still does not follow that the same independent factors operate in each of these, or that they do so in a similar and comparable manner.

3. The Comparative Analysis of the Democratic Process

Before we can proceed further with these two queries, a different problem of analysis will first have to be resolved. The problem that I here have in mind is that of the feasibility of comparison in regard to the nature of the democratic process in various political sectors of a society. For it is obvious, first of all, that any and all questions of the kind raised above will have to take a comparative form if the answers are to be informed and fully meaningful. This is, of course, an entirely uncontroversial statement as such; but it is not equally clear and obvious what a significant comparison would comprise in our case. Thus, e.g., it would in my view not do simply to look at the foreign policy sector and, say, two or three 'domestic' sectors, and to compare them empirically with reference to the question if issues within the former are resolved less democratically than in the other sectors in terms of our three dimensions. This follows from the fact that it may very well be that issues typically belonging to, say, the domain of foreign policy are indeed handled less democratically in terms of these criteria, at the same time as this conclusion is entirely unproblematic from

a normative democratic point of view. For these issues may be so peripheral to the interests and concerns of the public at large that a low democratic score - in terms of either participation, responsiveness or rationality - may quite conceivably reflect not a lack of influence in any palpable sense but rather a lack of concern which is fully understandable and indeed defensible from the public's point of view. But as long as our comparisons are håndled solely in the above sectorial manner, we cannot really know this except on an intuitive and ad hoc basis; and hence our comparisons will not really be significant, since we cannot know if our cases are 'democratically' of a comparable kind.

From this it follows, secondly, that before we start comparing the democratic nature of the process of decision-making within various societal domains, we should first consider what types of policies are normally handled within them. I would thus maintain, e.g., that labour policies constitute something different from defence policies qua defence policies: they, so to speak, 'do' different things (and here I obviously do not have in mind the fact that they pertain to different functional areas). But in most comparative studies this essentially constitutive aspect is left aside, possibly on the assumption that this dimension is unimportant. But is it methodologically acceptable to pressuppose such constitutive equivalence? Let me illustrate this question as follows: does a decision to revise the Swedish constitution or to join NATO constitute qua policy a decision which is comparable to regulating the alcohol content of beer or the admission of new categories of students to institutions of higher learning? It would seem to me not: a policy is not a policy is not a policy. Now if this is indeed the case, if there are different policy levels, dimensions, or whatever we wish to call this constitutive aspect, then it must follow that this fact is not irrelevant to the comparative analysis of processes of decision-making: some policies, because of their constitutive nature, 'deserve' to be treated more seriously and hence more democratically than others. It must be noted, however, that this aspect differs from the point made in the previous paragraph; for we are here not dealing with how important or unimportant policies are perceived to be by the public (and hence with their response to it), but with the fundamental nature of policies themselves in relation to the political and social system as a whole.

A third aspect which has to be considered here is the by now commonplace assumption that the boundaries between respectively the foreign policy and the domestic policy areas are amorphous and often difficult to uphold in reality. Certainly I feel that it is almost impossible to

define it in terms of foreign policy' itself; that is to say, I would maintain that there is nothing in a foreign policy qua policy - however defined - which will make such a distinction tenable today. Instead, I suggest that we here simply have to accept institutional-bureaucratic criteria if we wish to maintain and work with this distinction, rather than analytic criteria of one kind or another. But, as adumbrated above, simply to compare decision-making processes within the institutionalized foreign policy sector with the decision-making process within given domestic sectors with regard to democratic aspects will not lead us very far: for we need - to repeat - comparable cases of policy-making if our conclusions are to have any clout. That is to say, in order to compare foreign policy decisionmaking with decision-making in other sectors, we need analytic framework which link the two without being defined in terms of either of them.

My suggestion here is that we posit as our starting point a framework based on the assumption - which has received increased acceptance during the past decade - that there exist different types of policy areas or arenas, and that these possess characteristics which not only cut across all sectorial boundaries, but which also decisively influence the nature and scope of the decision-making processes involved in these.

One of the initial factors which gave impetus to this view - that the behaviour of groups, parties and other political participants is at least in part determined directly by the type of policy at stake - is the by now classic division within democratic theory between the so-called pluralist (or interest group, or group pressure) model of the political system, pre-eminently represented by Dahl, and the power elite school, with which C. Wright Mills' name is so often associated. The crucial point here is that although the views of these two schools are mutually incompatible, both of them have received persuasive empirical confirmation of their respective propositions in the literature. One common explanation of this disturbing fact within political science is to speak of 'ideological coloration,' i.e., of the a priori perspectives brought to the analysis by the scholars in question, and which are said to determine their research as a whole. Thus, e.g., William E. Connolly has argued that due to different conceptions or assumptions of how American society functions, and particularly of how power relations operate within it, 'strategic decisions [are made by Dahl and Mills] which reflect the perspective brought to the inquiry and which in cumulative effect push the outcomes of inquiry in the preferred direction', with the consequence that 'their resulting analyses are in large part self-fulfilling' (Connolly 1967, pp. 30, 49). The related

notion of paradigmatic' differences of one kind or another has also been
adduced to explain this and similar phenomena in the literature.

Another way of accounting for the above is to accept the terms of analysis of each school as they stand except for their claims to universality, and instead to relate them to different aspects or arenas of the political system. This is also the conclusion to which Theodore J. Lowi has come in a series of seminal articles over the past fifteen years. He thus contends that 'the American system is not all of a piece but is composed of several fundamental subsystems. Internally, these have developed around the most fundamental functions of the state. When public policy is facing the redistribution of resources, the system is elitist in very much the theoretical and empirical terms laid down by Mills and others. When public policy is facing the regulation of resources, the system is pluralist in all the specific issues fought out and decisions made... This is also perfectly consistent without present notions about differentiation in modern politics (Lowi 1967, pp. 297-98). Lowi has since this statement was made expanded considerably on this theme - that 'policies determine politics' - and on the general characteristics of the four types of policies (the distributive, constituent, regulative and redistributive) which, in his view, typify Western political systems today. Since this is a well-known discussion, I need not here go into a general description of this typology (see Lowi 1964, 1967, 1972).

However, there is a more particular aspect of this approach which needs to be explored somewhat more, since it is directly related to the questions which are of central concern to us in this research note. What I have in mind is the fact that Lowi's fourfold classification of politics and its attendant empirical hypotheses refer equally to the analysis of policies which are typically domestic in nature and political issues which traditionally belong to a state's foreign policy domain. In a real sense we thus here have a linkage notion; and this is certainly a fact speaking in favour of this type of analytic framework. Indeed, although Lowi has almost exclusively been concerned with the analysis of domestic politics, one of his very first articles explicating this approach contains an analysis of changes in America's foreign policy relations since 1945 (see Lowi 1967). His conclusion in that article is that apart from crisis situations in foreign policy, when decisions are usually made solely by a small elite of formally appointed officeholders (and this also holds true for policies involving no domestic resources, such as recognition of governments), American foreign policy during this period has tended to involve only two areas of power, viz., distribution and regulation; and the general trend has been

that the latter has increasingly become the dominant pattern. In other words, from involving issues which domestically could be disaggregated into nonconflictual, logrolling patterns, American foreign policy has increasingly evolved into policies - his examples here are trade tariffs, foreign aid and armament questions - which have involved conflict, bargaining, and compromise between different domestic interest groups. His explanation of this characteristic of present American foreign policymaking is that it reflects 'a political process so decentralized as to be almost completely susceptible to domestic political influences.' In short, except during the rare occasions involving crisis situations, US foreign policy has over these years become 'an extension of domestic processes, practices, and values' (Lowi, 1967, pp. 302-303). This is certainly a significant conclusion if valid; and as such it is a far cry from the classical conception of the division between a state's foreign policy and its domestic politics. It also - and this is more important here - points to the potential uses of his framework in the comparative analysis of both types of politics with reference to their democratic aspects.

There are, however, some significant problems associated with this approach which have to be addressed before any attempt is made to link it closer to our particular concerns. One of Lowi's most influential critics has been James Q. Wilson, who has argued that the Lowi typology is both logically and empirically deficient in various ways. From a logical point of view he has argued that it is hard to distinguish the various arenas of power from one another except in extreme cases; in his own words, 'there are broad arenas of power that seem to fit nowhere in the scheme; and there are important changes over time in the way groups behave.' In addition, there 'are a host of policies that could be classified under two or more categories'; thus, e.g., urban renewal programmes 'regulate the use of land, redistribute the housing supply, and distribute benefits to certain contractors and labour unions.' And with reference to the empirical implications posited by the Lowi framework, Wilson has argued, e.g., that although the pluralist view of politics often can be linked to the regulatory arena, 'there is a broad range of other regulatory issues in which group activity is modest, and the activity that does occur tends to be carried out by unsuccessful opponents.' Thus, to take one regulatory example, innovative consumer policies are, according to Wilson, often opposed rather than advocated by organized groups, and have been successful only as a result of 'the direct representation within key congressional committees of various points of view', a procedure aided by the activities of influential newspapermen and prominent consumer advocates such as

Ralph Nader (Wilson 1973, pp. 328-220).

There is undoubtedly some justification for these and similar criticisms. However, in my view they do not decisively refute the feasibility of using the general outline of the Lowi framework; and although I will not do so here, it is certainly possible to argue that at least some of the points made by Wilson - and particularly some of his examples - are not altogether persuasive.₂ Lowi's is after all an a priori and formalistic approach, and such approaches - like all others - possess both inherent strengths and weaknesses. Lowi himself is well aware of the many empirical exceptions to his formal categories, but is content to accept these as long as he is reasonably able to sustain the general tenor of his logic of analysis. In any case, Wilson himself admits that 'Lowi's fundamental insights - that the substance of a policy influences the role of organizations in its adoption - seems correct', while adding that two corrections to it are necessary: (i) a distinction should be made between the adoption of a new policy and the amendment of existing ones; and (ii) the incidence of costs and benefits should be taken into consideration when analyzing how policies affect politics (Wilson 1973, p. 330). What is thus called for is not the rejection in toto of this approach, but certain refinements and amendments of his categories.

Various such attempts at refining his typology have been made, but I shall here refer to only one of these, since in my view it is the most fruitful one for our purposes. Sverker Gustavsson, in a recent paper in which he discusses and suggests improvements of the Lowi approach, has taken his cue directly from Wilson's suggestion that the incidence of costs and benefits should be taken into consideration when analyzing how policies may affect politics. Without going into the details of Gustavsson's suggestive analysis, which requires more time and space than it presently available, I will very briefly try to summarize some of his most important points (Gustavsson 1980).

He accepts the general thrust of Lowi's framework on the structural assumption that society in actual fact is organized asymmetrically in terms of what Harry Eckstein has called 'patterns of authority', and that this macro-political condition is a function of the various ways in which the state attempts to govern society via the imposition of policies (Eckstein, 1973). However, he also accepts some of Wilson's major criticisms, and particularly his emphasis on the importance, when policies are analyzed, of how costs and benefits are concentrated and distributed. But here he wants to stress what he calls a marginalist perspective, i.e., an analysis of the marginal difference which a given policy will make to the equilibrium

holding in a particular policy situation. This is, in his view, not a structural question but one pertaining to the way particular policies areperceived, by the actors involved, materially to alter or sustain a given state of affairs. It is also a standpoint which explicitly contains a critique of Wilson's position, since in Gustavsson's view the latter refers only to policy changes, while in actual fact many policies which are pursued by the state are simply the continuation of past policies, which as a consequence do not fall within the purview of Wilson's framework. A marginalist analysis does not have this implication, since it views policies and their impact with reference to a concept of change defined in terms of a condition of static equilibrium (rather than, and here he appeals to Schumpeter's economic terminology, in terms of the 'dynamic' considerations which underlie Wilson's approach). As to Wilson's distinction between 'new' and 'amended' policies, Gustavsson maintains that at least in the Swedish context genuine examples of the former are rare today (Wilson defines 'new' very much in terms of societal issues and areas hitherto considered outside the legitimate realm of governmental intervention); and in any case, he adds, whether 'change is major or merely an amendment seems to be primarily a matter of historical and comparative perspective', and as such involves an analysis which has a different aim than the one pertinent here (Gustavsson 1980, p. 133).

The upshot of Gustavsson's analysis of both Lowi's and Wilson's typologies - and their implications - is that they pertain to two distinct types of phenomena and hence to two different types of questions, both of which are germane and legitimate with reference to determining how policies can be said to affect politics. More specifically, with regard to our present concerns, they lead us to raise the following two highly pertinent queries: (1) what difference does it make, with reference to variations within the democratic dimensions posited above, whether policies are, respectively, distributive, constituent, regulative or redistributive?; and (2) what are the effects, with reference to the same aspects, of how marginal costs and benefits are distributed? Gustavsson's tentative conclusion is that procedures and participation are affected the way Lowi has suggested, viz., with reference to the kind of policy which is at issue, but that the intensity of open conflict and public controversy - or what perhaps could be called politicization - in each case will be a function of how politically distinct or diffuse the marginal cost/benefit aspects are perceived to be by the actors involved. In this view, a redistributive policy, e.g., will structurally thus involve the public in the kind of conflictual relations hypothesized by Lowi; but the intensity with which the issue

in question will be fought out politically will depend on how it is seen to distribute costs and benefits between different sectors of society. This consideration of a redistributive policy may therefore lead to a basic and deeprooted conflict and controversy; but it may also be accepted without much fuss or public debate. In either case, however, it will involve a political process of the kind suggested by Lowi's framework, viz., the issues involved will be resolved as a result of interaction between the government, political parties, and peak organizations (all of which-if they so wish - can activate and command the support of large segments of the population, since redistributive issues are precisely those which affect whole 'classes' of society).

The result of Gustavsson's distinction between the structure of policies
and their marginal effects is the following matrix consisting of 16 cells:

It should be added that Gustavsson also provides us with clear stipulations as to the meanings of each of his classificatory terms; but, again for lack of time and space, I shall not discuss these here, but instead refer the reader to his stimulating analysis.

On the basis of this matrix and the categories on which it is based, we are now in a position to posit a genuinely comparative framework for analyzing such questions as the nature and scope of the democratic process in different public sectors, whatever their functional character or jurisdictional definition. We are also able to do this with reference to either one or more countries, and in diachronic or synchronic terms. Furthermore, if we so wish we can also control for either the effect of policy structure or the effect of marginal change by holding one of these constant; or we can

compare all the kinds of policies contained in the matrix with similar types of policies in other sectors. All of this is a consequence of the fact that comparability is obtained on the basis of a framework which is analytically independent of sectorial or institutional-structural characterizations or classifications of the political system. This does not mean, of course, that the above variables are the only feasible or germane ones, or necessarily those most fruitful, or that policy structure and the cost/benefit dimension cannot be defined with reference to other criteria or traits. But it is in my view imperative that a comparative framework of this analytical nature be utilized in a study of the kind envisaged here.

For practical and parsimonious reasons I suggest that we decide to hold one of the above dimensions constant and, furthermore, that we do this with Gustavsson's marginal perspective rather than with the structural dimension. This can be done in various ways, but this question should not, it seems to me, be settled until we have examined the type of empirical data, i.e., policies, which are available for analysis. The important point is to be able to find data regarding marginal change effects for all policy types and this within all of the sectors to be compared. We could obviously economize - and simplify matters - even more by also holding constant the policy type dimension in some suitable manner. One way of doing this would be, e.g., to compare only redistributive policies within each sector (these are possibly the most interesting from a democratic point of view). However, it would seem preferable not to curtail the analysis excessively in this way: 'comparatively' speaking, and given aceteris paribus clause, the more cases we have at our disposal, the merrier our subsequent analysis will tend to become.

4. Sectorial Variances in the Democratic Process: Suggestions and Selected Hypotheses

Once appropriate cases for comparison have been selected, we can once again face the twofold task previously discussed: determining if indeed processes of decision-making are less democratic in the foreign policy sector of government than in a typical domestic sector such as housing or educational policy; and if so, explaining the causes of, orreasons for, such variances in what are putatively fully democratic systems of government (it should be emphasized once more that we are here concerned only with such political systems).

With regard to the first question, a procedure involving two steps is

suggested. First ofall, it is in my view necessary as a starting point to give a general characterization of the policy configuration of a governmental sector as a whole. We need to know if a particular type of policy tends to dominate such a sector, and if so, to what extent this is the case. Such an evaluation should ideally be made on the basis of Gustavsson's framework, i.e., with reference to both the structural and the marginal dimensions. The point of this descriptive analysis is to ascertain if the sectors in which we are interested differ markedly from one another in terms of such configurations of their activities. For if these discrepancies are very apparent, we do not really have to go much further in our analysis, since we already here have a strong indication that a Lowian type of explanation may suffice for our purposes. That is to say, democratic differences between the two sectors could then be predicted to exist for the simple reason that in the one sector we have the dominance of policy issues which by their very nature are fundamentally more 'political' in their import (in the sense explicated by Lowi, but not discussed here) than would be the case with regard to the other example. And if our prediction turns out to be empirically valid, we are presented with a condition which, in my view, is not problematic either in normative or explanatory terms, despite clearly determinable democratic variances in the processes of decision-making in the respective sectors. Conversely, if the above prediction does not hold, and we for example find that the most 'political' sectors are those which are allocated the least amount of democratic resources, then it is obvious that we have a real problem at hand, both from a normative and an explanatory point of view.

I will assume, however, that no governmental sector in a highly developed and democratic state is this one-sidedly nonpolitical or apolitical - or purely functional - policy configuration. And even if such a sector did exist, it would most probably still contain some examples of policies which are politically salient; and we would most certainly wish to compare at least these with similar types of structural policies in other sectors with regard to their democratic characteristics. This leads us to our second step.

The suggested procedure here is to reintroduce our three democratic dimensions to the analysis, and to score comparable empirical cases of our policy types in the three sectors in terms of degrees of participation, responsiveness, and rationality. If as a consequence of this procedure we do find significant differences between different sectors with regard to the same policy types, then it follows that palpable variances do exist between these sectors with regard to how democratic their processes of decisionmaking

are - and then we most certainly have before us a subject matter crying out to be explained. In other words, we then have obtained an affirmative answer to the first question posed above, and can immediately proceed to the second and more significant one.

The assumption at this point is that the foreign policy sector will have been found to be less democratic than such domestic sectors as housing and education. We thus have before us a dependent variable in the form of three democratic dimensions characterized in each instance by lower scores in the former sector than in the latter. Our task is to explain these differences - to link them to variables in the form of one or more empirically persuasive explanations. On these final pages various suggestions and hypotheses will be proffered in a most skeletal and general form: they are to be regarded as mere starting points for the explanatory endeavor.

First of all two structural types of explanations will be mentioned which in my view are not germane to the task at hand. They are quite different from each other, and are consequently inappropriate for quite different reasons: at least so I shall argue.

(a) Policy-Structural Hypotheses: 'Policies Determine the Democratic Process'

It seems clear that if - as we have assumed - there do exist clear differences in how structurally similar policy types are handled in different sectors of a political system, then the Lowian approach will from this point on no longer give us any mileage. In other words, if it proves to be the case that 'policies do not determine politics' in certain cases (in the sense that the policies are of the same type but the politics differ), then we almost by definition have an exception to the Lowian dictum, with the consequence that we will have to appeal to other factors than the structural characteristics of policies in order to explain the above phenomenon. Anyhow, since we have used Lowi's framework to achieve comparability of cases, we should refrain from using it at the same time to account for variances between them. That is to say, as always we must resist the temptation of intertwining classificatory and explanatory variables.

(b) Systemic-Structural Hypotheses: 'Systemic Parameters Determine the Democratic Process'

A second type of structural hypothesis refers to the paramount importance of the boundary between national systems and their environments to explain differences - such as those pertaining to decision-making processes - between the domestic policy and foreign policy sectors. The assumption

here is that due to the existence of two fundamentally different systems, the one characterized by order and legitimate coercion and the other by anarchy and the primacy of power (usually military), political behaviour within the former will ipso facto be different from behavior involving the second system. In brief, it is usually deemed sufficient in these types of explanations to point to this fundamental structural difference in order to account for differences as to how decisions are made.

In my view the above approach must be rejected as inherently unacceptable for the simple reason that such an explanation is either true by definition or begs the very question which is posed here, viz., why foreign policies qua foreign policies are treated less democratically than domestic policies. It is analogous to explaining why adults but not children are mature by pointing out that the former but not the latter are adult. Furthermore, even when and if the contents and/or outcome of foreign policies pertain exclusively to conditions and goals lying 'outside' a given sovereign state, the process of foreign policy-making still remains thoroughly domestic (and thus 'sovereign') in terms of both the actors and the institutional-organizational framework involved. Hence, even though the structural parameters referred to above may have relevance to the outcome of a foreign policy decision - e.g., to the feasibility of its implementation or to its impact and effectiveness - this does not hold for the process leading to such a decision. Here, just as in all domestic policy areas, we are dealing with the interplay between a government and its various infrastructures, and between these and the public in whose service a government is supposedly to act. In this interplay the role of the boundary between domestic and international factors may or may not be important; but the extent to which it is the one or the other must be determined independently of how a given policy is classified in regard to this structural boundary.

The scepticism expressed above does not extent to the rest of our hypotheses, which are couched in terms of various cognitive-rational, bureaucratic-organizational, situational and contractual aspects of both domestic and foreign policy decision-making processes. It should be emphasized once again that these are merely examples of the types of hypotheses which may turn out to be fruitful in this area of inquiry.

(c) Rational Choice Hypotheses: 'Perception of Relative Pay-Off Determines the Democratic Process'

The notion that the relative pay-off of a given policy determines the
character of the political action involved in its pursuit can take various

forms and implicate various types of actors, both governmental and nongovernmental. Here I will use this notion with reference both to the behaviour of the policy-maker and the behaviour of the citizen at large (although, as we shall see below, with different consequences for the two). With regard to the former, the hypothesis here is that questions regarding foreign policy are treated less democratically than domestic policies because (i) foreign policy-makers in modern societies have been relatively more successful in retaining exclusive control over their jurisdictional sphere than other decision-makers; and that they have been able to do so (ii) as a result of a relatively sound strategic calculus regarding the expected behaviour of other political actors on the political arena. (It should perhaps be stressed that this argument refers to retaining decision-making power, not usurping or increasing it. For it is probably true that foreign policy decision-makers are far less independent of outside control or interference today than they were fifty years ago. The hypothesis claims, however, that in comparison with other political sectors, foreign policymakers have during this period lost less of their decision-making prerogatives than their counterparts in other areas.)

With regard to the public at large, the notion of implicit rationality can be used in a similar fashion, although here the calculus involved refers not to strategic considerations in the first hand, but to a balancing of the subcalculi of substance and preference. The former refers to an assessment of outcomes associated with available alternatives of political action, as well as to an assessment of the probabilities of these outcomes. The latter involves the ascription of given values to probable outcomes of political action, usually in terms of some order of preference (see Lundqvist 1980). The hypothesis here is that taking part more actively in the process of foreign policy decision-making is not deemed to be worth the candle in terms of the costs involved and the potential benefits - the pay-off- which may accrue as a consequence of such increased activity.

This general approach constitutes a potentially rich source for the generation of specific empirical hypotheses. Thus, e.g., James G. March and Johan P. Olsen have claimed that 'Individuals are seen to allocate available energy by attending to choice situations with the highest expected return' (March and Olsen 1979, pp. 14-15). And elsewhere, with specific regard to the concept of attention viewed as a form of rational action, they note the following: 'The routines of attention location tend to give priority to those things that are immediate, specific, operational, and doable; they tend to ignore things that are distant, general, and difficult to translate into action' (March and Olsen 1979, p. 50). It has also been

suggested that people do not get extensively involved in foreign policy issues because these are complex, vague and ambiguous: and that anyway, they have better and more rewarding things to do.The following is a somewhat more extensive syllogistic example of this type of argument: high rates of involvement in particular policy areas are a consequence of perceived efficacy of participation; such efficacy is low in areas characterized by high consensus (or as March and Olsen have noted, 'High participation rates among individuals who see themselves as efficacious are symptoms of a lack of consensus'); hence participation is low in the foreign policy area. And a final example, also from the pen of the above two scholars: 'there are competing arenas and competing objectives. An important reason for someone not being one place, is that he is somewhere else' (March and Olsen 1979, p. 50). And the point here is obviously that he/she is somewhere else by choice, as a consequence of having calculated that being in the latter place gives him or her a better trade-off than being in the former.

(d) Environmental Dependence Hypotheses: 'Perception of Environmental Constraints Determines the Democratic Process¹

This also constitutes a cognitive-cum-rational type of approach, although it is more specific than the previous class and one which, in the view of many scholars, should be regarded in structural rather than perceptual terms. The factor which is stressed here is the constraining role which the international environment - broadly speaking - is said to play in domestic politics; or more specifically, the emphasis is placed on how the environment is perceived to affect the autonomy of a given state, and hence the efficacy of political action, both individual and organized. This notion in turn builds on the idea of a continuum in any modern political system between issues which are wholly resolved by - and are hence fully in the grasp of, so to speak - domestic actors and forces, and issues (or whole clusters of issues) which are more or less completely outside the effective domain of this decisional grasp.

I need not here go into the question if indeed the world today is significantly more interdependent than it was, say, half a century ago (see, e.g., Sundelius 1980). Rather, let us forthwith accept the proposition that whatever the merits of the case as a whole may be with regard to changes in the nature and volume of international transactions during this century, these changes have unquestionably had political implications which, for better or worse, have affected the perceptions of both politicians and citizens. And one of these perceptions is surmisably that the foreign policy

arena of any given state, both large and small, is to an unusual degree
intertwined with the international environment, and hence constitutes a
policy sector within which domestic political action is perceived to be less

- even much less - efficacious than in other areas. This assumption, conjoined with an assumption previously introduced in passing, viz., that the foreign policy sector tends to be characterized by a high level of consensus, provides us with the hypothesis that the perception, on the part of the general public, that foreign policy issues are highly constrained by factors, actors and/or conditions outside the control of a given state or its government, leads to low levels of democratic activity in this area.

(e) Organizational-Structural Hypotheses: 'The Type of Bureaucracy or Organization Involved Determines the Democratic Process'

This hypothesis has a well-established pedigree, reaching back to the time when foreign affairs was a royal prerogative rather than a governmental matter. However, even after this situation had changed, the unique position of the foreign affairs office has tended to perpetuate itself, and with it anachronistic practices and a secretiveness which would be considered wholly unacceptable within other governmental domains. What I have in mind here are organizational routines and standard operating procedures which in cumulative effect prevent decision-making in this sector from being conducted on the open political arena.

The most prominent analysis of the decision-making process in foreign affairs along these terms is of course that of Graham T. Allison, who in his early work (on the Cuba crisis) argued that policy output could inter alia be viewed as either the product of bureaucratic politics (in which bargaining plays the crucial role) or as the result of organizational processes (in which routines and standard operating procedures are central). Subsequently he has fused these two approaches into a framework involving the combined analysis of policy output in terms of both inter- and intraorganizational politics on various levels of a system. 'The organizational structures within the political system', W. I. Jenkins has characterized this method, 'and vested interests of political actors are considered to place constraints upon the decision process and hence on the form and nature of the output that can emerge' (Jenkins 1978, p. 33). The tendency here is perhaps to reify the organizational system as a closed entity which operates according to set patterns of behaviour; and yet the fact remains that organizational behaviour is possibly the factor above all others which has been adduced to explain the low levels of public participation, responsiveness, and debate on the foreign policy arena. Or as Jenkins also notes:

'If policy outputs are the products of political and organizational systems, variation in output may well be linked to variation in organizational performance' (Jenkins 1978, p. 35). This type of hypothesis merits very serious consideration and extensive empirical analysis, at the same time as one should probably resist the temptation of blaming bureaucrats and other organizational actors for all but the most innocuous social and political ills.

(f) Policy Style Hypotheses: 'The Particular Styles, Roles, etc., Characteristic of Organizations Determine the Democratic Process'

This type of explanation retains the organizational and/or bureaucratic framework, but specifies its propositions with particular reference to the psychological 'styles' or 'roles' which certain types of organizations tend to foster, and which in turn are presumed to affect both the procedures and content of decision-making within their political sectors. As one scholar has recently noted, 'policy styles may differ from one area to another and from one point in time to another... If each area of policy has a tendency to develop into a semi-watertight compartment inhabited by its own "policy elite", it is possible that quite different policy styles may develop within the same political system' (Richardson 1979, p. 341). In this research tradition there is, once again, an extensive contemporary literature, and hence I need not elaborate on this type of analysis here (see, e.g., Mellbourn, 1979).

(g) Choice Situation Hypotheses: 'The Conditions of Choice Determine the Democratic Process'

The basic notion here is that the definition of a choice situation on the part of either the public and/or the governing elite determines the nature of the democratic process, especially its participatory dimension. In this connection March and Olsen have written the following: 'In general, the longer a choice remains unresolved, the greater the potential range of issues that are defined as relevant. Similarly with respect to the activation of participants, everything else being equal, the longer time taken to reach a decision, the greater the potential number of participants who are activated. The longer the decision time, the broader the involvement in the decision.' To which they immediately add that the 'more time a decision takes, the more participants and issues; as a result, the more complicated the situation.' This is in fact a description of their famous 'garbage can' model, which builds on the assumption 'that each choice opportunity is an open receptacle into which any currently unresolved issues may be

dumped' (March and Olsen 1979, p. 86). The hypothesis here is that the domestic policy domain is far more littered with unresolved issues than the sphere of foreign policy, and hence that the latter possesses fewer garbage cans filled with questions and problems, which in turn explains why there is less participation in this area. Or to put it more bluntly: a state such as Sweden does not have much choice in its foreign affairs, and those decisions which it has had to make are usually clear-cut and lasting in effect. We thus here - in contrast to, e.g., the fields of education and housing - have an uncommonly 'well-decided' choice situation, with the consequence that there is little incentive to 'problematize' issues and thus open up the situation to the dynamics of involvement described above. Hence it is perhaps symbolically significant that the desks of foreign policy decisionmakers are commonly assumed to be well-polished and clean - and their wastebaskets empty.

(h) Situational Source Demand Hypotheses: 'Organizational Slack Determines the Democratic Process'

This, too is a type of hypothesis emanating from organizational theory, and particularly from the notion of 'slack', defined as the difference between existing resources and activated demands. 'High' slack thus indicates a situation involving a surfeit of resources relative to demands, while 'low' slack indicates the converse condition. March and Olsen have written the following in this connection: 'Thus, the greater the slack, the fewer the number of people activated in a decision. When slack is reduced, more people will be activated... When there is a great amount of slack, decisions are made with minimal participant involvement; they are made by the relatively full-time "management", acting as though they were solving problems ... When slack is further reduced... more part-time participants are activated and it becomes obvious that there is no way all the demands can be met. As the participants confront each other with an overt conflict of interest, the "managerial" style of leadership is replaced with a "political style" ...' (March and Olsen 1979, p. 88). A typical example of this mechanism is the activation of the American public on the question of the Vietnam War (or the comparable Korean situation); and more recently we have experienced a similar phenomenon in connection with the Iranian hostage issue. Sweden, on the other hand, has not in its recent history had similar instances of reduced slack in regard to its foreign affairs, mainly because it has not committed its resources in a manner comparable to great powers such as the United States. However, it can be hypothesized that the more problematic the economy becomes, and the

clearer our dependence on the outside world is made to be seen, the more active will be demands to curtail foreign aid, for example, or the free entrance of foreign goods into this country. Thus the implication of this assumption is that crises in foreign affairs are beneficial to the democratic process, while good times foster passivity and satisfaction. The question then becomes: shall we choose the former or the latter? Unfortunately, however, it is not for us to make this choice, since the thrust of this type of hypothesis is, after all, that the situation makes the decision for us.

(i) Contractual Hypotheses: 'Political Precommitments Determine the Democratic Process'

A final hypothesis rests on a contractual notion with some Rousseauian overtones. Although this may immediately suggest a mystifying rather than enlightening account, it nonetheless refers to an idea which is less farfetched than perhaps first supposed. Jon Elster has dealt extensively with it in his highly suggestive elaboration on the theme of 'political precommitment', which he illustrates metaphorically in terms of Ulysses's s decree to his fellow mariners, on setting out towards the Sirens, that '... you must bind me hard and fast, so that I cannot stir from the spot where you will stand me... and if I beg you to release me, you must tighten and add to my bonds' (The Odyssey). This form of imperfect rationality, of purposely limiting one's future options in certain respects in order to forestall incontinent, vacillating or zigzag actions and policies, is said to be particularly pertinent to the task of keeping democratic systems within the limits required for efficiency and stability.

A number of institutions within modern democracies can, in Elster's view, be interpreted as devices for societal precommitment in this sense. Thus, for example, the exchange rate was under classic liberalist policy regarded as a crucial parameter which under no circumstances was to be tampered with by politicians or governments. The same holds for the autonomy of such public institutions as the British Broadcasting Corporation (but not its French equivalent, the ORTF), as well as - which is obviously more significant here - the foreign ministries of many countries. 'For these institutions', he adds in clarification, 'it is possible to identify, with varying degrees of precision, the act of abdication whereby politicians have decided that certain values are too important... to be subject to the current control of politicians' (Elster 1979, p. 90). It should be emphasized, once again, that we are here speaking of intentional measures, in the sense that the 'binding' of oneself in this manner is seen by all to constitute an explicit political act of will in the name of certain transcendental

values - e.g., efficiency and democratic stability - rather than being the consequence merely of passivity or other ordinary sins of omission on the part of the body public. (Hence no theories regarding the benign effect of electoral passivity are implied here.) For the same reason this hypothesis differs fundamentally from the central notions associated with the Lockean tradition referred to in the beginning, despite superficial similarities between the two. For while there the stress is placed on the special essence or nature of foreign policy per se, which 'necessarily' must exclude the general public from the foreign policy sphere, no such intrinsic attributes are implicated here. Rather, the core of this hypothesis refers to an act of political precommitment in the name of certain values transcending the arenas of both domestic an d foreign policy. It is in this respect also the most normative of our hypotheses, in that it refers to the pre-eminence and acceptance of certain rules of the political game - rules which are deemed to be so important that the system would not be able to sustain itself if these are not embraced by one and all in the name of the common good. The question then becomes - and this is perhaps also a proper note on which to end this discussion as a whole - whether democracy requires such a precommitment with regard to the pursuit of foreign policy, or if this is merely a remnant from an undemocratic past for which there is no real justification today.

---

NOTES

1 These are my own and perhaps not wholly felicitous terms.

2 Thus, e.g., he should distinguish between the intent of a given policy viewed in terms of the policy actor's perception of it, and the outcome of a given policy, which may be quite different from that which the policy intended to accomplish. Lowi's framework applies quite explicitly only to the former, i.e., to the categorization of issues with reference to perceived intention.

REFERENCES

Anckar, D. 1980. 'Responsivitet - Om begreppet, företeelsen och forskningsområdet',
Poliikka 1, 28-38.

Connolly, W. E. 1967. Political Science and Ideology. New York: Atherton.

Eckstein, H. 1973. 'Authority Patterns: A Structural Basis of Political Inquiry', American
Political Science Review LXVII, 1142-61.

Goldmann, K. 1981.' "I utrikes- och försvarsfrågor fungerar inte demokratin": Precisering
av en hypotes', Statsvetenskaplig Tidskrift, forthcoming.

Gustavsson, S. 1980. 'Types of Policy and Types of Polities', Scandinavian Political Studies
3 (New Series), 123-142.

Jenkins, W. I. 1978. Policy Analysis: A Political and Organizational Perspective. New York:
St. Martin's Press.

Lewin, L. 1970. Folket och eliterna. Stockholm: Almquist & Wiksell.

Lowi, T. S. 1964. 'American Business, Public Policy, Case Studies and Political Theory',
World Politics 16, 677-715.

Lowi, T. S. 1967. 'Making Democracy Safe for the World: National Politics and Foreign

Lowi, T. S. 1972. 'Four Systems of Policy, Politics, and Choice', Public Administrative
Quarterly 32, 298-310.

Lundqvist, L. J. 1980. The Hare and the Tortoise: Clean Air Policies in the United States and
Sweden. Ann Arbor: University of Michigan Press.

Mellboum, A. 1979. Byråkratins ansikten: Rolluppfattningar hos svenska högre statstjänstemän.
Stockholm: Liber.

Oisen, J. P. 1978. Politisk organisering. Bergen: Universitetsforlaget.

Olsen, J. P. and March, J. G. 1979. Ambiguity and Choice in Organizations. Bergen:
Universitetsforlaget.

Pennock, J. R. 1979. Democratic Political Theory. Princeton: Princeton University Press.
Pitkin,H. F. 1967.The Concept ofßepresentation. Berkeley: University of California Press.

Redford, E. S. 1969. Democracy in the Administrative State. New York: Oxford University
Press.

Richardson, J. J. 1979. 'Policy Making and Rationality in Sweden: The Case of Transport',
British Journal of Political Science 9, 341-353.

Sundelius, B. 1980. National Security in an Interdependent Environment. Stockholm: FOA
Report C 10155-M3.

Verba, S. and Nie, N. H. 1972. Participation in America: Political Democracy and Social
Equality. New York: Harper & Row.

Vital, D. 1968. The Making of British Foreign Policy. London: George Allen & Unwin.
Wallace, W. 1976. The Foreign Policy Process in Britain. London: George Allen & Unwin.