The end of ideology debate ranks as one of the major controversies in the social sciences during the postwar decades, and the debate was waged on truly a grand scale. It stretched over a period of more than fifteen years, from the mid-1950's into the 1970'5. To an unprecedented degree the debate engaged scholars on both sides of the Atlantic. Thousands of pages were written in several languages, and anthologies of contributions to the debate stand as monuments of the controversy 1968, Cox 1969, Allardt & Rokkan 1970, Rejai 1971).

In surveying the prolific writings of the debate, however, one is faced with the difficult question as to the importance of the debate with respect to theorizing about ideology. The purpose of this article is to deal with this question and examine several of the theoretical underpinnings of the end of ideology debate. This is done by focusing on the problem of defining ideology. This problem constitutes an appropriate focus because the influence of many of the definitional assumptions the debate still pervade thinking on ideology and much of the everyday usage of the term.

Resumé

This article deals with the definitional legacy of the end of ideology debate and the problem of defining The definitional legacy of the debate is described in terms of assumptions about the characteristics and functions of ideology, and several problematic features are identified. A major argument that certain theoretical difficulties characterizing the debate were the result of viewing various as definitional properties of ideology rather than hypothetical variables. In an effort to overcome the problematic features of the definitional legacy of the debate, the article recommends the construction of a minimal definition of ideology which converts characteristics and functions often viewed as a priori elements of ideology into an extensive battery of hypothetical variables.

---

*This article is a revised version of a paper presented at the workshop 'Ideology after the End of Ideology Recent Theorizing and Empirical Applications', the ECPR Joint Sessions of Workshops, 1983.1 would like to thank the workshop participants for their helpful criticisms and comments. In addition, Elias Berg, University of Stockholm, Barbara G. Haskel, McGill University, and Evert Vedung, University of Uppsala, have offered valuable suggestions.

A central question posed here is: what properties of ideology ought to be regarded definitional ones and which properties ought to be regarded as hypothetical The strategy employed in attempting to answer this question is to discuss the implications for political inquiry when a particular element is designated definitional, or alternatively a hypothetical, variable property of ideology. major argument of the article is that certain theoretical difficulties and dubious characterizing the end of ideology debate resulted from conceiving of certain properties as definitional components rather than variables.

Defining Ideology: Characteristics and Functions

A cursory inventory of the literature helps to pin-point the elements which have figured prominently in efforts to define ideology - and especially those which were important in the end of ideology debate. At this stage I shall merely enumerate properties frequently included in definitions of ideology, and below they will be discussed in more detail.

One main thrust of the efforts to define ideology has been to specify its distinctive usually in combination with a comparison of allied phenomena. the special attributes assigned to ideology is i_sharedness\ Ideology shared ideas, conceptions of reality and values. It consists of group beliefs (Parsons 1952, Lane 1962, Plamenatz 1970). A second property ascribed to ideology is stability. Ideology manifests permanence or an enduring quality as opposed to the transient nature of opinions (Minar 1961, Merelman 1969). -Thirdly, ideology is often defined as a system or set of ideas or beliefs, and stress is laid on the systematic element, coherence or consistency as a requirement for a body of ideas to be an ideology. Ideology is more than an isolated idea or attitude et al. 1960, Minar 1961, Converse 1964, Tingsten 1966a, b, Barnes 1966). A fourth distinguishing characteristic is intense belief or passionate commitment (Bell 1960, Sartori 1969). Fifthly, ideology is described as comprehensive all-embracing, providing a total view of society (Shils 1968). A sixth, much disputed attribute is the illusory nature or falseness of the ideas making an ideology. It is maintained that ideology is characterized by distortion, simplification and undue selectivity (Mannheim 1936, Johnson 1968). In this connection, ideology is often contrasted with theory and in some instances with philosophy (Bell 1960, Macridis 1980). A final characteristic which has also generated considerable controversy is its positional nature or its direction, that is, whether ideology is oriented toward change or the status quo (Christenson et al. 1972, Seliger 1976, Bryder 1983).

In addition to definitions focusing on distinctive characteristics, ideology has been defined by designating its functions. A wide variety of functions - psychological,social political - have been assigned to ideology in the literature, but the emphasis here is on the political functions of ideology. Many writers have pointed out that ideology has an action-orientation function. Ideology constitutes

a program, and it serves as a guide to political action (Friedrich 1963, Seliger 1976). A second function of ideology is legitimation. Ideology constitutes an argument, and it serves to justify various political objects, such as institutions and procedures, policies and leaders (Lane 1962, Easton 1965, Lewin 1967). Conversely,ideology possess a delegitimizing or unmasking function (Mannheim1936). or integration comprises an additional function (Parsons1952, 1963). Conversely, a divisive or partisan function of creating divisions and generating conflict is attributed to ideology (Mannheim 1936, Easton 1965). Lastly, ideology has a mobilization function (Bell 1960, Easton 1965, Barnes 1966, Sartori 1969).

An examination of the characteristics and functions of ideology forms the basis of the subsequent discussion. Firstly, I shall discuss how these various elements figured in the end of ideology debate. Secondly, I shall deal with the issue of whether these elements ought to be conceived of as definitional components of ideology or as hypothetical variables. These issues squarely address the definitional of the end of ideology debate.

The Definitional Legacy of the End of Ideology Debate

Few of the end of ideology writers bothered to define ideology, and when they did offer definitions it was in their later contributions to the debate (e.g. Tingsten 1966b, Shils 1968). Several of them, however, either initially or eventually specified their observations pertained to particular ideologies. Among their specifications 'the nineteenth-century ideologies' (Bell 1960), 'total ideologies', 'ideologies of the extreme left and right', 'class conflict ideologies' (Lipset 1964), and the 'isms' (Tingsten 1966). Despite the introduction of this important qualification, often phrased their arguments in generic terms. An illustration of this tendency is provided by statements such as 'ideology which by its nature is an all-or-none affair' or 'ideology fuses these energies' (Bell 1960, 375, 371). The specific frequently shades into the general, and the reader comes away with the impression of generalizations about ideology and definitional assumptions concerning characteristics and functions of ideology.

Assumptions about the Characteristics of Ideology

Not all of the characteristics ascribed to ideology in the literature were major definitional of the debate. And as we shall see, not all of the writers put equal emphasis on the same properties. The attributes most frequently referred to in the contributions to the debate were comprehensiveness, passionate commitment, and the shared nature of ideology.

Comprehensiveness. Several of the writers stressed the comprehensive or totalisticnature

talisticnatureof ideology. Edward Shils was most emphatic in his emphasis on this property. In his first article he mentioned an eagerness for universal observation(Shils and later he described ideologies as aspiring to and claiming completeness. Ideologies were characterized by a high degree of explicitness of formulation over a very wide range of objects, and he referred to both their universalcomprehensiveness their unqualified insistence on complete realization of the ideal, through a total transformation of society. He further argued that ideologydiffered creeds and outlooks through its greater comprehensiveness (Shils 1968).

In contrast to Shils' categorical declaration of the totalistic nature of ideology, Seymour Lipset's writings were somewhat ambiguous on this point. In his second major contribution to the debate, Lipset specifically indicated that his remarks on the decline of ideology pertained to total ideologies, which would seem to imply the existence of ideologies which were not total in character. On the other hand, in the same article he contrasted ideological and pragmatic orientations. On the basis of his comments, ideological orientations appear to involve 'total weltanschauungen' 1964, esp. 281).

Passionate commitment. This was the characteristic most commonly associated ideology by the end of ideology writers. Most of them at least touched upon this element. Ideology was described as an emotionally charged set of beliefs (Lane 1962,15), as passionately believed (Lane 1966,660). The fervor in subscribing ideological principles and values in the past was noted by Herbert Tingsten who also spoke of the 'impassioned battles of ideas' (Tingsten 1955, 140-42, 148).

This characteristic was especially emphasized by Daniel Bell and Edward Shils. Bell wrote: 'What gives ideology its force is its passion' (Bell 1960, 371), and he underscored the emotive quality of ideology and its power to infuse passion into its followers. Shils went further and identified zealous espousal as an attribute set ideology apart from other belief systems about man and society:

Their [ideologies] acceptance and promulgation are accompanied by highly affective overtones. Complete individual subservience to the ideology is demanded of those who accept it, and it is regarded as essential and imperative that their conduct be completely permeated by it (1968,66).

A similar position was adopted by Giovanni Sartori, who can be classified as a late-comer to the debate inasmuch as he suggested that his typology of belief systems was suitable for testing the end of ideology thesis. Like Shils, he proposed that 'strong affect' or passion was a definitional component of ideology (Sartori 1969, 403-405) distinguishing it from other political belief systems. In fact, the emphasis on this property has prompted the rather dubious interpretation that the end of ideology debate can be reduced to either one of two propositions:'(l) a relative over the last two decades of the ultimacy with which ideological are stated, or (2) a relative attenuation of the emotive intensity with which ideological goals are pursued' (Rejai 1971, 17).¹

The systematic element. This quality was particularly important in the writings Tingsten and Shils. Tingsten clarified that his discussion dealt with systems of ideas - with the 'isms' - and he defined ideology as forming a systematic He further argued that ideologies had an elaborate structure, and that values alone did not constitute an ideology (Tingsten 1966a, b). In Shils' view, ideology was distinguished by consistency (1955), by internal integration or systematization (1968). He maintained: 'As compared with patterns of beliefs, ideologies are relatively systematized or integrated around one or a few pre - eminent values ...' (1968, 66).

The systematic aspect was also emphasized by Philip Converse (1964), who perhaps can be regarded as an end of ideology writer - at least in the sense that the findings reported in his influential article were grist for the mill of the end of idealogy Converse preferred the term belief system instead of ideology but he tended to equate the two (1964, 208). To the extent he used the two terms interchangeably more or less automatically incorporated the systematic component his conception of ideology. He defined a belief system as 'a configuration ideas and attitudes in which the elements are bound together by some form of constraint or functional interdependence' (1964, 207). The systematic nature of ideology in Converse's analysis was additionally underlined by labelling individuals beliefs displayed a high degree of constraint - or interrelatedness - as ideologues.2

Finally, Sartori also included systematization as a characteristic of ideology, and he specified that its systematic nature consisted of constraint rather than logical In his view, however, this property was not the discriminating element distinguished ideology from other patterns of political beliefs (Sartori 1969,400-401).

Stability. The enduring quality of ideology was cardinal in the writings of Sartori Shils, and it was referred to by Converse. According to Converse, those ideas or beliefs which were central to an ideology were the most impermeable to change. However, a change in these core beliefs would result in modifications in other idea-elements. Both Sartori and Shils described the stability of ideologies as rigidity, and they characterized ideologies as closed systems. Furthermore, they both regarded this property as one of the distinguishing attributes of ideology.

Distortion. A number of writers pointed to problems of distortion. Bell, for example, claimed that to be successful (to rouse people) ideologies must simplify ideas (1960, 372), and Sartori subscribed to a similar view. Shils acknowledged that true propositions could coexist alongside false ones in ideologies, but other characteristics he ascribed to ideology magnified the element of distortion. His distinguishing properties included closure or resistance to innovations. Similarly he saw the endeavor to achieve systematic integration and the totalistic demands of the ideological orientation as obstacles in the pursuit and attainment of truth (1968, esp. 73-74). Tingsten also commented on the intellectual deficiencies of ideologies, noting that ideologies consisted of a mixture of explicit and implicit or

perhaps purely unconscious desires, logical and factual errors, and generalizations founded on weak grounds. Robert Lane's thesis that the growth of scientific knowledge impinged upon the domain of ideology was based on the assumption of false beliefs. Ideology was sustained by uncertainty and ignorance, and new knowledge reduced the operative area of ideology (Lane 1966,660-661).

Sharedness. The end of ideology writers mentioned or referred to the shared nature of ideology. However, they tended to relate this property to diverse collectivities groupings in society: the ideological primary group (read: sect) (Shils 1968), social movements (Bell 1960, Lane 1966), classes (Lipset 1964) and parties (Tingsten 1966a, Lipset 1960). In the case of Bell and Shils, the shared nature of ideology was an important factor leading to distortion.

Before considering the implications of these assumptions about the characteristics ideology, I propose to examine the assumptions about the functions of ideology, and then comment on the problematic features of both sets of assumptions.

Assumptions about the Functions of Ideology

As implied by the rubric 'On the Exhaustion of Political Ideas in the Fifties' (Bell 1960), the end of ideology writers generally viewed ideology as a spent force, sapped of its capacity to fulfill major functions. In effect, the end of ideology was manifested in ideologies no longer performing their previous functions. Of the functions enumerated earlier, three were crucial to the arguments of these writers: action-orientation, mobilization and the divisive function.

Action-orientation. Both Bell and Tingsten called attention to the actionorientation of ideology, and Tingsten included it in his definition of ideology. spoke of ideology as a way of translating ideas into action, and he noted the Marxist philosophical tradition of claiming that truth was achieved through action. He also wrote:'... ideology, which once was a road to action, has come to be a dead end' (1960, 370), and he argued that few issues could be formulated any more, intellectually, in ideological terms.

Tingsten defined ideology as 'a collection of political ideas that is meant to form a more systematic whole and give general and specific directives for action' (1966b).³ An essential part of his end of ideology argument hinged on this definition. qualify as an ideology, the ideas must guide action in the form of policy. Tingsten argued that this was no longer the case. Ideological principles had been superseded by technical details in policy-making, and ideology had been reduced to phraseology and slogans 'to sanctify certain policy positions or to gather support a tradition' (1966a, 18).

Tingsten's assumptions about the function of ideology set the parameters of further discussion. The main tack of his critics was to discredit his thesis that ideology no longer influenced policy to any substantial degree. For example, in arguing the existence and vitality of ideologies, Leif Lewin (1967) utilized Tingsten'sdefinition

sten'sdefinitionof ideology and challenged this thesis by describing how ideologicalprinciples been and contined to be important in the area of economic policy-making. None of Tingsten's critics, however, questioned the narrowness of his definition or his assumption that a 'true' ideology must provide guidelines for action.

Mobilization. Several of the writers claimed that ideology no longer performed a mobilizing function at either the elite or mass level. As noted earlier, the proponents the end of ideology viewed passionate commitment as a basic ingredient of ideology. And this element was central to their discussion of the mobilization function and its decline.

More generally, Bell declared that ideology was the conversion of ideas into social levers, and he went on to suggest that the most important latent function ol ideology was to tap emotion and channel it into politics. In the 1950's ideologies failed to arouse passions, and they had lost their power to elicit commitment. Although noted that ideologies in the past had succeeded in infusing passion into their followers at the mass level, his specific comments dealt with the failure of ideologies at the elite level. Bell, and also Shils, emphasized the disenchantment intellectuals with ideologies (Bell 1960, 373, Shils 1955. Cf. Lipset 1960,439-50).

As for the mass level, Lipset (1964) pointed to déideologisation, a declining commitment to ideology among the mass public and various social classes, and especially the workers. According to Lipset, traditional leftist ideology had lost its appeal, and leftist sympathizers no longer subscribed to traditional ideological prescriptions. As evidence, Lipset cited the findings of various surveys tapping attitudes toward public ownership and class conflict. In essence, his argument was that the source of commitment to leftist parties was no longer ideological values principles but representation of interests (Lipset, 1964, 279-84). Elsewhere wrote that the 'democratic class struggle will continue, but it will be a fight without ideologies ...' (Lipset 1960,445).

The divisive Junction. The notion that ideologies contributed to divisions in society especially prominent in Lipset's and Tingsten's writings, although neither of them spoke explicitly of divisiveness as a function of ideology. They both focused on the decline of ideological differences and the observation thai ideologies no longer generated fundamental conflicts and violent controversy.

More specifically, Lipset regarded the far-reaching attenuation of 'serious intellectual among groups representing different values' as a fundamental change in the politics of Western democracies (Lipset 1960, 439). And in 'The End of Ideology?' his discussion referred primarily to the decline in ideological differences. That divisiveness may be interpreted as quintessential in Lipset's conception of ideology is also revealed by his use of the word 'ideologisation' which was synonymous with conflict.⁴

In a similar vein, Tingsten stressed the decline of ideological differences, and
he traced the growing agreement in a number of policy areas where sharp conflictsformerly

flictsformerlyprevailed (1955, 1966a). He commented that this development had been characterized as a 'deideologisation, an end of ideology, a coalescence of general beliefs'. He concurred with this description, stating that these characterizationswere because conservatism, liberalism, socialism were no longer systems of ideas presenting a theory of development and prescriptions (1966a, 18).

At the same time, however, both Lipset (1964, 1966) and Tingsten maintained the emergent consensus constituted a new ideology. This conclusion hardly jibes with the meaning Lipset assigned to ideologisation. Nor does it square entirely with one of his major theses, that in European societies ideological orientations giving way to pragmatic ones.

Ambiguities and contradictions stand out more sharply in Tingsten's discussion. On the one hand, he characterized the attenuation of ideological differences as deideologisation and an end of ideology. On the other hand, he subsequently declared the resulting consensus did not entail the end of ideology but the emergence new ideology - a supraideology or an overarching ideology (1966a, 20-21). one cannot help wondering to what extent Tingsten viewed this 'new ideology' as meeting the two requirements set down by his own definition of ideology forming a systematic whole and providing guidelines for action. His own description of the new consensus as a result of incremental compromises and his emphasis on the replacement of values and principles by technical detail in policy-making do not suggest that he would have regarded the new ideology as meeting his definitional requirements. To some degree, this problem may have appeared less acute by introducing the concept of an overarching ideology. Nonetheless, it is difficult to avoid the impression that Tingsten was implicitly operating with two different conceptions of ideology.

Problematic Features of the Definitional Legacy

A number of problematic features of the definitional legacy of the end of ideology debate have suggested themselves in the previous discussion. One problematic feature was a rather narrow conception of ideology. Many of the authors attached special weight to certain elements and more or less made them definitional components ideology. Their argument, in part, consisted in noting the dilution or disappearance of these particular attributes or functions. In fact, I would suggest that it might have been the tendency to conceptualize ideology in this way that led them to describe their initial observations in the 1950's as 'the end of ideology?' rather than the decline of ideology or the decline of particular ideologies.

Many writers, for example, specified that ideology was characterized by passionate belief, and they pointed to an ebbing in the intensity of commitment to ideology. The erosion of one aspect of the shared nature of ideologies - their associationwith social classes - was a major theme of Lipset. Tingsten's argumentswere representative of this tendency. He defined ideology as providingguidelines

vidingguidelinesfor action, and evidence of deideologisation was its declining
importance in making policy. Similarly, the systematic nature of ideology was a
defining characteristic, and Tingsten noted the weakening of this element.

A second tendency was to describe the attributes of ideology in pejorative and, in some cases, absolute terms. One of the clearest illustrations of this tendency was Shils (1968). Belief in ideas was formulated as zealous espousal and complete subservience, stability as rigidity, and its scope as universal comprehensiveness. Concentrating on a smaller number of characteristics, Sartori assumed a stance which was basically the same as Shils'. Ideology was distinguished by adamant belief; it was closed and passionately felt. Its belief-elements were fixed, i.e. 'rigid, dogmatic, impermeable to argument and evidence' (Sartori, 1969, 404). Sartori's and Shils' positions represented the extreme cases. But most of the contributions the debate conveyed the idea that ideology was a negative phenomenon, to a large extent this impression was reinforced by the properties the authors attributed to ideology.

A third problematic tendency, which surfaced in the debate, was a reliance on dichotomies in conceptualizations of ideology. This feature was most pronounced in Sartori's discussion. He confronted ideology and pragmatism dichotomously, arguing that they ought to be viewed as polar types. Although Sartori conceded that pure types were seldom, if ever approximated in the real world, he justified the dichotomy by claiming that it was difficult to conceive of ideology and pragmatism blends of a same continuum and inquired: 'A continuum of what?' (Sartori 1969, 405). Likewise Lipset treated ideological and pragmatic orientations opposites (1964). Shils also relied on a dichotomous line of reasoning by contrasting ideology and civility - and ideological politics and civil politics (1958). Other contributors to the debate who were critics of the end of ideology writers also theorized about ideology in dichotomous terms (e.g. Himmelstrand 1970).

What is problematical about the use of dichotomies in theorizing about ideology? general, dichotomies and four-category typologies based on the intersection two dichotomies (cf. Rejai 1971, 14) pose at least two difficulties. First, although dichotomies and typologies based on dichotomies are often powerful devices and they may be especially useful in the early stages of theorizing, dichotomies entail gross simplifications by reducing complexities to two polar types. In many instances it is more appropriate to conceive of a property a variable with a range of values, rather than as a pure type. Second, the polar types may not be opposites except by definition. Dichotomies can block consideration of the possibility that the two properties, which are conceptually polar types, might coexist in the real world.

In the case of Sartori's dichotomous scheme, the latter weakness manifests itself.Conceptualizing open system and a closed system as polar types obviouslydoes involve any difficulties, but it is much less certain that rationalism ought to be viewed as the opposite of empiricism (cf. Diggins 1970, Mullins

1974) - and even less so that ideology and pragmatism ought to be conceived as polar types. Empirical findings caution against such a procedure. For example, Robert Putnam has documented the coexistence of pragmatic orientations (in the sense of an open cognitive structure) and ideology defined as 'a coherent, comprehensiveset beliefs'. On the basis of his empirical results, he warned that it was misleading to include dogmatism as a definitional property of ideology (Putnam1971).

A fourth tendency was dualistic conceptions of ideology, which were not necessarily Although emphasis was on a narrow conception of ideology, some instances references were made to a broad conception (e.g. Lane 1966,66). The clearest case of a two-fold conception in the debate was a distinction between ideology as Weltanschauung and ideology as 'attitude structure'. Weltanschauung was defined in terms resembling those of Shils and Sartori. It referred to 'the traditional meaning of ideology' as a 'comprehensive, consistent, closed system of knowledge' involving 'commitment (both intellectual and emotional), orientation toward action, distortion and simplification of reality, hostility to critics and opponents, and goal-orientation (often of the millennial variety)' 1965,629). Upon closer inspection, however, it turns out that 'attitude structure' was defined as an opaque version of Weltanschauung. The distinction between the two was one of degree rather than kind. As put by the author, attitude structure fell short of the total belief system and tight logic of a 'world view'.

Dual conceptualizations of ideology can also be detected in ambiguities in the writings of the individual authors. Ideology was defined in one sense and later the term was used in an entirely different way. Tingsten's labelling the emergent consensus a new ideology is a case in point. Obviously ideology in this particular context meant something different from his definition of ideology.

Definitional ambiguities are also contained in Converse's celebrated article. On the one hand, he tended to equate ideology and belief system, thereby making constraint or interrelatedness a definitional component of ideology. On the other hand, he designated ideology as a source or type of constraint. 'The idea-elements go together... for more abstract and quasi-logical reasons developed from a coherent view as well. It is this type of constraint that is closest to the classic meaning of the term "ideology" ' (Converse, 1964, 211). In several contexts, ideology and ideological pertained to the liberal-conservative continuum or more generalized liberal and conservative views (esp. 219-23). Thus Converse used ideology to refer to (1) attitude structure or the patterning of beliefs and (2) specific orientations.

To conclude, the problematic features of the definitional legacy of the end of ideology debate can be summed up in the following points: (1) an emphasis on a narrow conception of ideology, so narrow that some authors found it necessary to refer to ideology in a broader sense or to use the term in a way which transcended their own definition or definitional assumptions, (2) a tendency to conceive of

certain properties and functions as definitional components whereas, as I shall subsequently argue, a more prudent course would have been to view them as hypotheticalproperties, a proclivity to describe the attributes of ideology in pejorativeand terms, and (4) a reliance on dichotomies in theorizing about ideology.

These problematic features have survived the ups and downs in the fashionability credibility of the various arguments in the debate. As political analysts and journalists point to a resurgence of 'ideological' politics in the Western countries, definitional assumptions of the debate have found new currency. How can political analysis divest itself of the difficulties of this tradition?

Beyond the Definitional Legacy

Intermittently I have suggested that the definitional assumptions of the end of ideology debate provide too narrow a focus for inquiry into the phenomenon of ideology. This suggestion is by no means novel. Already at the height of the debate, the objection was raised that the end of ideology writers did not mean 'any given set of values, beliefs, preferences, expectations and prescriptions regarding society' a particular set (La Palombara 1966,8). Among the more powerful criticisms at a particularistic conception of ideology was that its assumptions certain paths of empirical inquiry.

More recently, in a major contribution to the literature, Martin Seliger (1976) criticized what he called the restrictive conception of ideology, noting that several of the end of ideology writers were adherents of such a conception. He saw this as; a major stumbling block in efforts to develop a general theory of political ideology, he set as his task the formulation of an inclusive definition of ideology which would pertain to all categories of political belief systems. Although I find the desire to come to grips with restrictive conceptions of ideology commendable, this procedure has resulted in a lengthy and cumbersome definition of ideology - so lengthy that it cannot be quoted here (Seliger 1976, 119-20).

Instead I would like to suggest an alternative procedure which focuses on formulating minimal definition of ideology and in the process converts the characteristics functions which are often conceived as a priori elements of idealogy an extensive battery of hypothetical variables. A minimal definition does no more than attempt to identify the object of analysis, and 'all the properties or characteristics of an entity that are not indispensable for its identification are set forth as variable, hypothetical properties - not as definitional properties' and thus are left to empirical verification (Sartori 1976,61). As Sartori notes, minimal definitions have neither explanatory nor predictive power.

An initial step in formulating a minimal definition of ideology, then, is to consider elements ought to be definitional properties and which ought to be hypothetical variables. Such a consideration involves an examination of the pros and cons of conceiving of characteristics and functions as definitional properties.

Definitions focusing on the special attributes of ideology have not infrequently concentrated on undesirable qualities or have presented these attributes as undesirable. sort of emphasis has generally served to perpetuate the derogatory of the term of ideology, and it has precluded ideology from becoming a neutral concept of analysis. Moreover, the concentration on negative qualities has serious limiting effects. A striking example is Sartori's scheme based on two defining characteristics. Ideologies are defined as, firstly, closed systems and, secondly, felt. This definition greatly limits the potential use of the concept of ideology in political analysis. Ideology is to be reserved for the analysis extremist doctrines, political messianism and fanaticism (Sartori 1969, 411).

In the case of comprehensiveness and stability, one way to dispose of negative as well as unnecessarily restrictive connotations is to conceive of these properties as hypothetical variables - and to describe the attribute in neutral, general terms and to make the negative quality one facet of the variable. More precisely, characteristics odious associations - such as a totalistic nature, rigidity and dogmatism can be conceptualized as extremes of general hypothetical variables. Using Converse's term, we can speak of the range of ideology, and this property may vary from comprehensive to partial or narrow. Similarly, ideologies may vary in stability; they may exhibit differing degrees of rigidity or flexibility.

The disadvantages of making distortion or some related quality (simplification, falsity, etc.) a definitional property of ideology are numerous. First, perhaps more than any other attribute, this element has caused analysts to shun the use of ideology (e.g. Converse 1964,207,209). Second, inclusion of this element an a priori component also contributes to a continued politicization of the term and concept. Ideology is an epithet reserved for one's adversaries - political and intellectual. Third, there is the problem of the appropriateness of judging the evaluative aspects of ideology in terms of truth or falseness (Hamilton 1983,13). Fourth, if distortion or falsity is designated a definitional property, this creates enormous problems of empirical identification of phenomena in the real world qualifying as ideology (Vedung 1982, 89-90). Fifth, emphasis on distortion and falsehood easily leads to the pre-eminence of epistomological questions and the de-emphasis of cardinal issues in political analysis, such as the role of ideology in politics.

The inclusion of passionate commitment in a definition of ideology also presents The main problem lies in fashioning this attribute in such absolute and it is advisable to reformulate or reconceptualize it. Rather than passionate commitment, 'belief in' or acceptance ought to be conceived of as a definitional property of ideology and the intensity in belief or acceptance as a variable may range from vague sympathies to fanaticism.

The next question is whether sharedness should be regarded as a definitional
property of ideology. Here it needs to be stressed that there is an important differencebetween
the group nature of ideology, and extensiveness, the

wide dispersion of ideology. Several definitions have centered on extensiveness rather than sharedness. This is clearly illustrated by the claim: 'Ideology, to be ideology, must have a mass base' (Rejai, 1971, 9, italics original. Cf. Heeger in Vedung 1982,84 and Brzezinski 196 2). Sharedness reflects no such assumption and can occur with a minimum of two individuals (cf. Minar 1961,324-325). As a definitional property, sharedness is preferable to extensiveness because the latter'defines vital areas of political research. First, the fascinating question of social diffusion is dealt with by definition rather than by hypothesis. Second, the belief systems of decisionmakers might also be defined away to the extent they are not shared by the masses.

The systematic element is perhaps the most interesting and problematical of the attributes assigned to ideology. Many social scientists have argued that the systematic element is a definitional property of ideology. Internal consistency or constraint must exist in order for political attitudes or idea-elements to constitute an ideology. This assumption entails several difficulties which have not been adequately and it seems to pave the way for a periodic but perpetual 'end of ideology' debate.

A first difficulty is the criteria to be used in establishing the existence or nonexistence the systematic element. A variety of different requirements have been suggested: logical consistency, quasi-logical consistency, internal consistency functional interdependence or constraint. One problem shared by these requirements is that the researcher often imposes his or her sense of what is consistent with what. This danger might appear negligible or even to vanish in the case of functional interdependence evidenced in statistical associations and clustering. debate between Robert Lane and Philip Converse in the mid-1970's (Lane 1974, Converse 1975) indicates that problems of inference and interpretation A large part of the difficulty, according to Lane, stems from 'the analyst's in setting forth the idea-elements he thinks are important, developing the conceptual framework that the analyst regards as most likely to "govern" the more specific beliefs (exemplars), and thus providing a guided opportunity for measuring association and change' (Lane 1974, 99, italics original. Cf. Bennett 1975,6-18).

Utilizing constraint for functional interdependence as the criterion is also accompaniedby host of problems related to attitude measurement. A number of technical questions have been raised, for example, about Converse's coefficients measuring issue constraint (1964, 228-29), such problems as inferring a lack of issue constraint among individuals on the basis of aggregate data, the possible effectsof degree of heterogeneity of the mass public sample, and the construction of questions in tapping attitudes (e.g. Bennett 1975,15-24,181). In short an assessmentof degree of constraint and consistency is highly dependent upon the state of the art of attitude measurement. Despite considerable advances, it is still worth emphasizing that the techniques of survey research may not reveal the subtleties and nuances necessary in understanding how idea-elements are interrelatedor

relatedorthe 'logic' of respondents.

An additional difficulty involved in designating consistency a definitional component is determining the appropriate cut-off point between ideology and non-ideology. How much constraint is required of an ideology? And on what grounds? The advocates of consistency as an a priori element of ideology have seldom dealt with this issue. A further drawback is that a preoccupation with the degree of constraint among idea-elements could reduce empirical research to primarily in identification and labelling (ef. Sjöblom 1983).

Given these difficulties, what is a reasonable solution? One possibility is to specify the systematic element by negation: ideology is not a random, nonsensical of idea-elements (Heeger in Vedung 1977, 52-53). Less drastically but along similar lines, Willard Mullins has argued that ideology must 'make sense' and not result in logical absurdities within the confines of its basic conceptions and justifications (Mullins 1972,510). In any event, a prudent procedure to conceptualize this attribute in minimal terms. For example, a definition could stipulate that ideology is a body or collection of idea-elements, thus leaving the questions of interrelatedness and consistency to empirical investigation. This solution emphasizes the composite quality of ideology but makes its systematic nature a hypothetical variable. Such a procedure possesses the advantage of not settling the issue definitionally. Equally importantly, it does not make the identification ideology contingent upon consistency. It further minimizes another pitfall to designating systematization as a defining characteristic. Since ideologies be constructed by the analyst, there is an inherent danger that their systematic nature to some degree is the product of this construction process.

Finally, there is the matter of whether the functions of ideology ought to be definitional Functions have frequently figured in definitions, and emphasis functions has in part grown out of efforts to distinguish ideology from political (Friedrich 1963, Seliger 1976, Macridis 1980). In this context, functions have served as a basic discriminating element. Nonetheless, this asset may be outweighed by the drawbacks of including functions in a definition of ideology.

The end of ideology debate is instructive concerning some of the difficulties. Of the end of ideology writers, Tingsten alone explicitly defined ideology in terms of functions, viz. action-orientation. However, the reasoning of Bell and Lipset suggeststhat viewed mobilization and the divisive function respectively as a definitionalcriterion ideology. A major part of these authors' argument consisted of speculations that political ideas no longer performed these functions and thus no longer qualified as ideologies. A first difficulty pointed to by the debate is that the empirical verification of these speculations would involve staggering research tasks. Hence, as in the case of several of the attributes discussed above, the incorporationof in a definition of ideology is afflicted with vast empirical problems of identifying the phenomena in the real world which could be designated as ideology in accordance with the definition. Even more importantly, the hypothesizedfunctions

pothesizedfunctionsof ideology ought to be a key area of empirical research. In fact, if the concept of ideology is worth saving, a major reason is precisely becausevarious of ideology have focused on the linkage between politicalideas political behavior.

Thus far this discussion has argued against retaining most attributes and functions definitional criteria of ideology. And at this point the reader may inquire: what is left of ideology? The answer lies in the content or subject matter of the ideas, beliefs and attitudes comprising ideologies, and I would argue that this aspect to be central in defining political ideology - and identifying the object of analysis. The literature on ideology offers a wide array of descriptions of subject They range from nebulous characterizations, such as ideology deals with the nature and purposes of man and society, to more specifically political statements emphasizing principles of government. The most essential points which emerge from the many descriptions, however, are the following: political ideology consists of idea-elements concerned with (1) explaining and evaluating the existing social order, (2) describing the nature of the good society and (3) designating means or strategy of attaining the good society (cf. e.g. Brzezinski 1962,4-5, Christenson et al. 1972,5-6, Sainsbury 1980,7-10, Aberbach et al. 1981,115-16).

Despite its importance, content has not always been included in definitional discussions. In some instances, it seems so central that it has been taken for granted. other cases, it has been neglected and even rejected. For example, Robert Putnam has maintained that it is 'not the what, but the how of political thought which makes it ideological' (Putnam i 971,65 7).⁵ However, to the extent that we are attempting to define ideology and not ideological, it is imperative to make subject matter a definiens of political ideology.

Much terminological diversity and confusion have developed because of an emphasis on 'ideological' rather than 'ideology'. The word 'ideological' has often had as its referent one of the characteristics or functions which were central in the definitional assumptions of the end of ideology writers. Accordingly, ideological' mean 'biased', 'extremist', 'impassioned', 'conflictual', 'polarized', 'structured or characterized by constraint', and so forth. The end result is multiple connotations of ideology and a pronounced case of concept stretching.

Concluding Comments

In an effort to overcome the difficulties of the definitional legacy of the end of ideology debate, the main recommendation here is to focus on formulating a minimaldefinition ideology. Only a few attributes ought to be retained as definitionalproperties ideology, and emphasis ought to be put on the content or subjectmatter idea-elements in defining political ideology. Simultaneously, many characteristics and functions attributed to ideology ought to be viewed as hypotheticalvariables than definitional properties. The conversion of these elementsinto

mentsintohypothetical variables does not entail a denial of their importance in
the study of ideologies. Quite the opposite is true. Crucial issues are not given a
definitional answer but are the focal point of empirical inquiry.

The approach to defining ideology outlined here has a number of assets. Among the requirements of an adequate definition is that it is neutral rather than pejorative. Although it is questionable whether neutral definitions and concepts actually be formulated, substantial headway ought to be possible in the case of ideology since it is still riddled with negative overtones. The suggested procedure offers a way of curtailing the pejorative meanings attached to ideology.

A related advantage is broad applicability. This is an asset because the development more neutral conceptualizations of ideology has been obstructed by overly narrow definitions where the meaning of ideology has been reserved for phenomena generally regarded as negative.

Besides limiting pejorative nuances, the approach here may help to reduce ambiguities dual conceptions of ideology of the type which sporadically to the fore in the end of ideology debate. Implicit in some of the conceptions the notion that a particular attribute could vary and was not necessarily absolute in nature.

An additional advantage derives from making the characteristics and functions of ideology hypothetical, variable properties rather than definitional components. Several definitions of ideology tend to settle major issues definitionally rather than making them the object of empirical investigation.

Finally, the emphasis on characteristics and functions as variable properties may also furnish a method of coming to grips with the schizophrenic treatment of ideology in some definitional discussions, especially found in textbooks. In attempting explain what ideology is, these discussions treat drastically different traits as definitional components of ideology rather than variables of ideologies.

---

NOTES

1. For additional interpretations of the major propositions of the debate, see Himmelstrand 1970 and Putnam 1971,655-656.

2. Among Converse's findings, which supported the arguments of the end of ideology writers, was the observation that the patterns of constraint registered among ideologues were not mirrored in the mass public. In large sections of the mass public he found fragmented and random beliefs, no coherence and a lack of overarching ideological orientations.

3. The initial formulation of this definition is found in a collection of essays from 1941. Similar formulations, not as an explicit definition, can also be found in Tingsten 1966a, 18, 49.

4. Lipset wrote: 'Intense ideologisation, sharp conflict, is characteristic of polities in which new emerging classes ... are struggling to achieve ... rights, but declines when these classes are admitted full citizenship' (1966, 17).

5. In a more recent work, Putnam and his co-authors have put considerably more emphasis on the what of political thought in their analysis of ideology (Aberbach et al. 115-169).

REFERENCES

Aberbach, J., Putnam, R. & Rockman, B. 1981. Bureaucrats and Politicians in Western Democracies.
Harvard University Press.

Allardt, E. & Rokkan, S. 1970. Mass Politics. New York: Free Press.

Barnes, S. 1966. 'Ideology and the Organization of Conflict: on the Relationship between Political
Thought and Behavior', Journal of Politics 28, 513-530.

Bell, D. 1960. The End of Ideology. Glencoe: The Free Press.

Bennett, W. L. 1975. The Political Mind and the Political Environment. Lexington: Lexington Books.

Bryder, T. 1983. 'Ideology, Self-esteem and Political Mobilization', Paper, ECPR Workshop on
Ideology after the End of Ideology Debate, Freiburg.

Brzezinski, Z. 1962. Ideology and Power in Soviet Politics. New York: Praeger.

Campbell, A., Converse, P. E., Miller, W. E. & Stokes, D. E. 1960. The American Voter. New
York: Wiley.

Christenson, R., Engel, A., Jacobs, D., Rejai, M. & Waltzer, H. 1971 (1972). Ideologies and
Modern Politics. London: Nelson.

Christoph, J. B. 1965. 'Consensus and Cleavage in British Political Ideology', American Political
Science Review 59, 629-642.

Converse, P. E. 1964. 'The Nature of Belief Systems in Mass Publics', in Apter, D., ed., Ideology
and Discontent. New York: Free Press.

Converse, P. E. 1975. 'Public Opinion and Voting Behavior', in Greenstein, F. & Polsby, N., eds.,
Handbook of Political Science, Vol. 4. Reading: Addison- Wesley.

Cox, R. H., ed., 1969. Ideology, Politics and Political Theory. Belmont: Wadsworth.

Diggins, J.P.I 970. 'Ideology and Pragmatism: Philosophy or Passion?', American Political Science
Review 64, 899-906.

Easton, D. 1965. A Systems Analysis of Political Life. New York: Wiley.

Hamilton, M. B. 1983. 'The Elements of the Concept of Ideology', Paper, ECPR Workshop on
Ideology after the End of Ideology Debate, Freiburg.

Himmelstrand, U. 1970. 'Depoliticization and Political Involvement', in Allardt, E. & Rokkan, S.,
eds.., Mass Politics. New York: Free Press.

Johnson, H. M. 1968. 'Ideology and the Social System', in Sills, D., ed., International Encyclopedia
of the Social Sciences, Vol. 7. New York: Macmillan & the Free Press.

Lane, R. E. 1962. Political Ideology. New York: Free Press.

Lane, R. E. 1966. 'The Decline of Politics and Ideology in a Knowledgeable Society', American Sociological
31,649-662.

Lane, R. E. 1974. 'Patterns of Political Belief, in Knutson, J., ed., The Handbook of Political Psychology.
Francisco: Jossey-Bass.

LaPalombara, J. 1966. 'Decline of Ideology: a Dissent and an Interpretation', American Political
Science Review 60, 5-16.

Lewin. L. 1967. Planhushållningsdebatten. Stockholm: Almqvist & Wiksell.

Lipset, S. M. 1964. 'The Changing Class Structure and Contemporary European Polities',Daedalus
93,271-303.

Lipset, S.M.I 966. 'Some Further Comments on "The End of Ideology"', American Political Science
60, 17-18.

Macridis, R. 1980. Contemporary Political Ideologies. Cambridge: Winthrop.
Mannheim, K. 1936. Ideology and Utopia. London: Routledge Kegan & Paul.

Merelrnan, R. M. 1969. 'The Development of Political Ideology: a Framework for the Analysis of
Political Socialization', American Political Science Review 63, 750-767.

Minär. D. 1961. 'Ideology and Political Behavior', Midwest Journal of Political Science 5, 317-331.

Mullins, W. A. 1972. 'On the Concept of Ideology in Political Science', American Political Science
Review 66, 498-510.

Mullins, W. A. 1974. 'Sartori's Concept of Ideology: a Dissent and an Alternative', in Wilcox, A.,
ed., Public Opinion and Political Attitudes. New York: Wiley.

Parsons, T. 1952. The Social System. London: Tavistock.
Plamenatz, J. 1970. Ideology. London: Pall Mall.

Putnam, R. D. 1971. 'Studying Elite Political Culture: the Case of "Ideology"', American Political
Science Review 65,651 -681.

Rejai, M., ed., 1971. Decline of Ideology? Chicago: Aldine. Atherton.

Sainsbury, D. 1980. Swedish Social Democratic Ideology and Electoral Politics 1944-1948.
Stockholm: Almqvist & Wiksell International.

Sartori, G. 1969. 'Politics, Ideology and Belief Systems', American Political Science Review 63,
398-411.

Sartori, G. 1976. Parties and Party Systems. Cambridge: Cambridge University Press.
Seliger, M. 1976. Ideology and Politics. London: Allen & Unwin.

Shils, E. 1955. 'The End of Ideology?', Encounter 5, 52-58.

Shils, E. 1958. 'Ideology and Civility: on the Politics of the Intellectual', Sewanee Review 66,450-480.

Sjöblom, G. 1983. 'Notes on the Study of Political Belief Systems', Paper, ECPR Workshop on
Ideology after the End of Ideology Debate, Freiburg.

Tingsten, H. 195 5. 'Stability and Vitality in Swedish Democracy', Political Quarterly 26,140-151.
Tingsten, H. 1966a. Från Idéer till idyll. Stockholm: Norstedt.

Tingsten, H. 1966b. Strid kring idyllen. Stockholm: Norstedt.

Vedung, E. 1977. Det rationella politiska samtalet. Stockholm: Aldus/Bonnier.
Vedung, E. 1982. Political Reasoning. Beverly Hills: Sage.

Waxman, C. 1., ed., 1968. The End of Ideology Debate. New York: Funk & Wagnalls.