The concept of social movement Mario Diani Abstract Recent deveiopments in social movement research have evidenced a greater underlying consen.sus in the field than one might have assumed. Efforts have been made to bridge different perspectives and merge them into a new synthesis. Yet, comparative discussion of the concept of 'scKial movement' has been largely neglected so far. Tliis article reviews and contrasts systematieally the definitions of 'social movement' formulated by some of the most influential authors in the field. A substantial convergence may be detected between otherwise very different approaches on three points at least. Sociai movements are defined as networks of informal interactions between a plurality of individuals, groups and/or organizations, engaged in political or cultural conflicts, on the basis of shared collective identities. It is argued that the concept is sharp enough a) to differentiate social movements from related concepts such as interest groups, political panies, protest events and coalitions; b) to identify a specific area of investigation and theorising for soeial movement research. Introduction Social movement studies have grown impressively in recent years (Rucht, 1990). At the same time, efforts to merge originally distant approaches into a more comprehensive one have been made (e.g. Cohen, 1985; Klandermans et al,, 1988; Scott, 1990; Eyerman and Jamison, 1990). Ouite surprisingly, these attempts have largely passed over any discussion of the concept of 'social movement'. While several scholars have provided analytical definitions of it, we still lack, to my knowledge, a systematic comparison of these conceptualisations. This article aims to fill this © The Socioloskal Renf..' 1992 0038-0261 92/4001-00 J.VOC'l Mario Diani gap, discussing the concept of social movement as it has been formulated by some influential contributors to the field since the 1960s. Focusing on the conceptual level seems important to me, for a number of reasons. I share the view that, while concepts cannot be identified with theories, they are nevertheless the cornerstone of any theorising (see e.g. Sartori, 1984). Therefore, any effort to synthesise different approaches risks to be flawed, if little or no attention is paid to concept definition. This holds even more true for .social movements studies. There, even an implicit, 'empirical' agreement about the use of the term is largely missing. In fact, social and political phenomena as heterogeneous as revolutions, religious sects, political organisations, single-issue campaigns are all, on occasion, defined as social movements (see e.g. McAdam et al,, 1988:695). This terminological ambiguity entails, however, a loss of specificity ;md theoretical clarity. This is reflected in that many valuable analyses of social movements pay hardly any attention to the concept itself. They rather move immediately to more substantive questions, such as the factors which account for mobilization processes (e.g. Klandermans et al,. 1988) or the difference between old and new movements (e.g. Dalton and Kuechler. 1990). This is perfectly legitimate, of course. Yet, one may sometimes feel that the same topics might be as successfully treated without mentioning 'social movements' at all, adopting rather concepts such as 'collective action', 'social change', 'social conflict' and the like." The question therefore rises, what does 'social movements' specifically refer to. The absence of discussion concerning the concept of social movement has been usually attributed to the heterogeneity and incompatibility of the different approaches, which would make any synthesis impossible (e.g. Morris and Herring, 1987:139). In contrast to this view, I argue that a common thread exists between the analyses of social tnovements, produced within otherwise very diverse intellectual traditions. My goal here is to highlight this linkage and to identify the elements, that are common to the different 'schools'. These elements connote social movements as a specific social dynamic which is logically related to, yet distinct from, the ones mentioned above. It consists in a process whereby .several different actors, be they individuals, informal groups and/ or organisations, come to elaborate, through either joint action and/or communication, a shared definition of themselves as being part of the same side in a social conflict. By doing so, they provide The concept of,'iocial movement meaning to otherwise unconnected protest events or symbolic antagonistic practices, and make explicit the emergence of specific conflicts and issues (see e.g. Melucci, 1989; Eyerman and Jamison, 1990). This dynamic is reflected in the definition of social movements as consisting in networks of informal interaction between a plurality of individuals, groups and/or organisations, engaged in a political and/or cultural conflict, on the basis of a shared collective identity. The argument develops as follows. In the following section, some recent definitions proposed by leading figures in the field arc introduced. Then, four sub-components of the concept are identified and discussed. In the next section, a more empirical issue is addressed. The capacity of the concept to differentiate social movements from related phenomena (such as parties and interest groups, coalitions, protest events) is assessed. Finally, it is shown how Ihe proposed definition reflects recent developments in the field, and how it can contribute to identify a specific area of investigation for social movement research. An overview This discussion focuses on the views elaborated by Ralph Turner and Lewis Killian. John McCarthy and Mayer Ziild. Charles Tilly, Alain Touraine and Alberto Melucci. This group of scholars may be considered as representative of the four main trends within social movement analysis since the 1960s. These trends consist respectively of the most recent expansions of the 'Collective Behaviour' perspective (Turner and Kiilian); the several approaches which have been subsumed, though with various qualifications, under the label of 'Resource Mobilisation Theory' (RMT) (Zaid and McCarthy); the 'Political process' perspective (Tilly); and the "New Social Movements' (NSMs) approach (Touraine. Melucci).* Whereas the first three have been particularly influential in the USA, the fourth has been mainly associated with European scholars, to the extent that some (Klandermans and Tarrow, 1988) have even talked of an 'American' and a 'European' approach to the study of social movements. As there are a number of excellent, recent reviews of the literature, a thorough examination of the different 'schools' may be omitted in the present paper (see Morris and Herring, 1987; McAdam et al., 1988; Klandermans and Tarrow, 1988; Tarrow, 1988; Scott, 1990; Neidhardt and Rucht, Mario Diani 1990). However, some hints will be provided when discussing the single authors. Turner and Killian (1987. but originally 1957) define social movements as a peculiar kind of collective behaviour, which is contrasted to 'organizational' and 'institutional' behaviour (1987;4). In spite of these traits, however, collective behaviour cannot be consigned to lack of organisation or to irrational behaviour. On the contrary, as the theory of emergent norm suggests, collective behaviour represents merely a looser organisational principle (see al.so Neidhardt and Rucht. 1990). Turner and Killian define a social movement as 'a collectivity acting with some continuity to promote or resist a change in the society or organisation of which it is part. As a collectivity a movement is a group with indefinite and shifting membership and with leadership whose position is determined more by informal response of adherents than by formal procedures for legitimising authority.' (1987:223). Social movements 'are not necessarily or typically coterminous with movement organisations, [even though these] carry out much of the movement work and frequently attempt to control and speak for movements.' (Turner. 1981:5). RMT differs from Turner and Killian's and related collective behaviour approaches in that greater attention is paid to the role of organisational factors within social movements. Indeed, Zald and McCanhy define social movements in a way which is not far from Turner and Killian's, i.e. as 'a set of opinions and beliefs which represents preferences for changing some elements of the social structure and/or reward distribution of a society. A countermovcment is a set of opinions and beliefs in a population opposed to a sociai movement.' (McCarthy and Zald. 1977:1217- 18). Yet, their greatest concern lies clearly with the conditions under which such beliefs are transformed into concrete action. I-rom this perspective, both leaders with previous political experiences and strong, often professional, organisations are needed (McCanhy and Zald, 1973; 1977). Emphasis is also put on the conditions which fadlitate the constitution of social movement organisations (SMOs), as well as on the dynamics of co-operation/ competition between them (sec also Zald and McCarthy, 1980). The existence of interactions within social movements is reflected in the notion of 'sodal movement sectors' (McCarthy and Zald, 1977). According to this view, social tnovements organisations are not isolated actors; rather, they tend to interact with other organisations, even when they are not able to develop any sort of regular The concept of social movement co-ordination; moreover, social movement constituencies often overlap in a significant way. A recent formulation of this perspective states that social movement sectors are 'social movement activity largely oriented towards change that is achieved in the differentiated political arena . . . the configuration of social movements, the structure of antagonistic, competing and/or cooperating movements which in turn is part of a larger structure of action.' (Garner and Zald, 1985:120). Instead of focusing on organisational resources, Tilly (1978) relates the emergence of social movements to a broader 'political process', where excluded interests try to get access to the established polity. Tilly analyses this process from an historical perspective, periodising phases of intense contention within contemporary history and mapping shifts in the 'repertoires' of collective action. In contrast to McCarthy and Zald, his emphasis is on the overall dynamics which determine social unrest and its chiiracteristics, rather than on social movements as specific organised actors. This theoretical perspective is reflected in the definition of social movements as a 'sustained series of interactions between power holders and persons successfully claiming to speak on behalf of a constituency lacking formal representation, in the course of which those persons make publicly visible demands for changes in the distribution or exercise of power, and back those demands with public demonstrations of support' (Tilly, 1984:306). Social movements are an organised, sustained, self-conscious challenge which implies shared identity among participants (Tilly, 1984:.303). Both RMT and the 'political process' approach analyse the 'how' rather than the 'why' (Melucci. 1989) of social movements. In other words, they focus on the conditions which facilitate or constrain the occurrence of conflicts, taking the existence of potential grievances for granted. In contrast, the NSM approach tries to relate social movements to large-scale structural and cultural changes. The most explicit advocate of this is Alain Touraine (1977, 1981, 1985). Touraine identifies sodal movements with tbe dominant conflict in a given society: 'The social movement is the organised collective behaviour of a class actor struggling against his class adversary for the social control of historicity in a concrete community' (1981:77). Historicity consists of tbe 'overall system of meaning wbich sets dominant rules in a given society.' (1981:81). In industrial sodety. the core conflict opposed work to labour, in the 'programmed society', technocrats 5 Mario Diani to their adversaries. All the other conflicts which occur within a given society (e.g. conflicts for redistribution of resources) or during the transition from one sodety to another (e.g. the national conflicts) are subordinated to the core conflict, the only one where it is possible to talk of social movements. For other conflicts labels such as submovements. communitarian movements, national movements would be more appropriate (Touraine, 1985). As Touraine's analysis is both highl)' complex and well-known, I will fix-us on only two aspects which may be helpful in understanding his definition of social movements, even where one does not accept his broader theoretical framework. The first concems the idea of a social movement as the 'combination of a principle of identity, a principle of opposition and a principle of totality' (1981:81), where sodal actors identify themselves, their sodal opponents and the stakes in a conflict. Such a combination or process of 'identity formation' may. in fact, be detected in any aspect of social behaviour, but social movements are distinct in so far as the issue at stake refers, as we have seen, to the historicity, rather than to the 'institutional decisions or organisational norms' in a sodety (1981: 81). The .second a.spect concerns the high differentiation of beliefs and orientations within .social movements. Touraine's methodology of the 'sociological intervention' is mcani to provide a better reconstruction of these orientations as well as to help movement actors to achieve a better understanding of their own actions (Touraine. 1981:139ff; Touraine et aL, 1983a; Touraine f( a/., 1983b). Alberto Melucci is not as interested as Touraine in singling out the new core conflict of contempt)rary post-industrial society, even though he agrees that these conflicts are more present today in the cultural and symlxilic sphere. Rather, Melucci proposes a definition of social movements as a 'specific class of collective phenomena which contains three dimensions . . . [it] is a form of collective action which involves solidarity . . . [it] is engaged in conflict, and thus in opposition to an adversary who lays claims on the same goods or valties . . . [itj breaks the limits of compatibility of the system that it can tolerate without altering its structure' (1989:29). According to Melucci, social movements arc not coterminous with 'visible' political conflicts. In fact, public action is only one part of the experience of social movements. Even when they are not engaged in campaigns and mobilisations, social movements may still be active in the sphere of cultural production. Some strongly culture-oriented movements may mobilise only occasionally The concept of social movement in tbe politiciil arena. Their activities largely develop in 'movement areas', i.e. 'networks of groups and individuals sharing a conflictual culture and a collective identity' (1985).'. . . multiplicity of groups that are dispersed, fragmented and submerged in everyday life, and which act as cultural laboratories' (1989:60). A proposal for synthesis The definitions introduced above emphasise at least four aspects of social movement dynamics: a) networks of informal interaction; b) sbared beliefs and solidarity; c) collective action on conflictual issues; d) action which displays largely outside the institutional sphere and the routine procedures of social life. Networks of informal interaction The presence of informal interactions involving individuals, groups and organisations is widely acknowledged. Even Touraine, who as we have .seen adopts a very peculiar definition, stresses the view of social movements as collective actors where organisations, individuals and groups all play a role (e.g. 1981:150). Even where the emphasis is put on a 'set of opinions and beliefs', as in the case of McCarthy and Zald, the transformation of these ideas into action requires the interaction between specific SMOs, constituents, adherents and bystander publics (McCarthy and Zald, 1977:1223). Interaction is further stressed in notions such as 'social movement sector' (SMS) or 'micro mobilisation context', recently adopted by McCarthy and Zald in their reassessment of the field (McAdam et al,, 1988). Defined as 'any small group setting in which processes of collective attribution arc combined with rudimentary forms of otganisation to produce mobilisation for collective action' (ibidem: 709), this concept greatly modifies the basically hierarchical conception of relationships between constituents and SMOs, proposed by the RM theorists in their earlier formulations, forming a perspective more consistent with such notions as Melucci's 'social movement area'. The characteristics of these networks may range from the very loose and dispersed links described by Gerlach and Hine (1970) in their seminal book, to the tightly clustered networks which facilitate adhesion to terrorist organisations (della Porta, 1988). Such networks promote the circulation of essential resources for Mario Diani action (information, expertise, material resources) as well as of broader systems of meaning. Thus, networks contribute both to creating the preconditions for mobilisation (which is what RMT has mostly emphasised) and to providing the proper setting for the elaboration of specific world-views and life-styles (as described by Melucci). In spite of tbeir different emphasis, these definitions agree in recognising the plurality of actors involved in social movements and the informality of the tics which link them to each other. A synthetic definition of this aspect of the concept of social movements therefore may run as follows; 'A .social movement is a network of informal interactions between a plurality of individuals, groups and/or organisations'. Shared beliefs and solidarity To be considered a social movement, an interacting collectivity requires a shared set of beliefs and a .sense of belongingness. Respective authors refer to 'a set of opinions and beliefs' (McCarthy and Zald); 'solidarity' (Melucd); 'identity' (Touraine. Melucci, Tilly). Turner and Killian emphasise the continuity of social movements, which relies upon 'group identity' and 'ideologies'. Identity and ideology are defined here in the broad sense of the term, which makes them very close to sets of beliefs (Turner and Killian, 1987:249ff and chapter 14 respectively). Collective identity' and solidarity can be considered synonymous in this context, in so far as it is bard to conceive of the former without the latter, i.e. of a sense of bclongingness without sympathetic feelings, associated with the perception of a common fate to .share (Melucci, 1984a). The case is different for the definition propo.sed by McCarthy and Zald. Their notion of social movements as 'sets of opinions and beliefs' does not necessarily imply the presence of shared feelings of belongingness. However, their more recent work, and in panicular the emphasis on the role of 'micro-mobilization contexts' and 'frame alignment processes' testify to their growing concern for the interactive processes of symbolic mediation which support individuals' commitment." Collective identity is both a matter of self- and external definition. Actors must define themselves as part of a broader movement and, at the same time, be perceived as such, by those within the same movement, and by opponents and/or external 8 The concept of social movement observers.' In this sense, collective identity plays an essential role in defining the boundaries of a social movement. Only those actors, sharing the same beliefs and sense of belongingness, can be considered to be part of a social movement. However, 'collective identity' does not imply homogeneity of ideas and orientations within social movement networks. A wide spectrum of different conceptions may be present, and factional conflicts may arise at any time. Therefore, the construction and preservation of a movement's identity implies a continuous process of 'realignment' (Snow et al., 1986) and 'negotiation' (Melucci, 1989) between movement actors. The presence of shared beliefs and solidarities allows both actors and observers to assign a common meaning to specific collective events which otherwise could not be identified as pan of a common process (see also Oliver. 1989). It is through this 'fratning process' that the presence of a distinct social actor becomes evident, as well as that of related issues. Indeed, social movements condition and help constitute new orientations on existing issues and also the rise of new public issues, in so far as they contribute to 'the existence of a vocabulary and an opening of ideas and actions which in the past was either unknown or unthinkable' (Gusfield, 1981:325). The process of identity formation cannot be separated from the process of symbolic redefinition of what is both real and possible. Moreover, such collective identity may persist even when public activities, demonstrations and the like arc not taking place, thus providing for some continuity to tbe movement over time (Melucd. 1989; Turner and Killian, 1987). Taking these qualifications into account, we can define the second component of the concept of social movement as follows: 'The boundaries of a social movement network arc defined by the specific collective identity shared by the actors involved in the interaction'. Collective action on conflictual issues Some of the views reviewed here put a specific emphasis on conflict as a core component of the concept of social movement (Touraine, Melucci, Tilly). Others emphasise that sodal movements define themselves with respect to processes of social change Mario Diani (Turner and Killian, McCarthy and Zald). Even the.se latter however, acknowledge that as promoters or opponents of sociai change sociai mo\ements become involved in conflictual relations with other actors (institutions, countermovements, etc.). If there is at least broad agreement concerning the fact that conflict is a distinctive feature of a social movement, the notion of conflict is understood in very different ways by different scholars. Touraine daims that 'sodal movements' applies only to conflicts about historicity, while others use the term in a looser and more inclusive way. Melucci considers typical of social movements only those actions which challenge the mechanism of systemic domination, while American scholars tend to subsume under that heading any protest event, including those referring to negotiable issues. Finally, some authors consider as social movements networks of collective action which are exclusively or primarily oriented towards cultural and personal change (Melucci and Turner and Kilhan), while others focus on actors in the political .sphere (Tillv McCanhy and Zaid). On a closer look, however, many of these inconsistencies prove to be more apparent than real. We have already seen that, when analysing other types of conflicts than those concerning historicity I'ourainc attaches different qualifications (e.g. nationalist communitarian, cultural) to the label 'movement'. Along similar lines. Melucd differentiates between social movements, which operate at the systemic level, and other types of collective action. He speaks for instance of 'conflictual action', meaning a kind of behaviour which implies colleetive identity and the presence of a conflict, yet which does not break the limits of compatibility of the system (Melucd, 1984b). In other words, both Touraine and Melucci use the term 'social movement' to identify a specific category of phenomena within a broader category of 'n'lovements' whereas other scholars use the term to mean movements of my kind. Another presumed source of inconsistency consists in conceptions which focus on political movements and those emphasising that social movements are also, and often mainly involved in cultural conflicts. Several authors (among them Gusfield, 1981; Melucd 1989) maintain that the true bulk of social movement experience has to be found in the cultural sphere: what is challenged is not only the uneven distribution of power and/or economic goods, but socially shared meanings as well, that is the ways of defining and interpreting reality. Social movements tend to focus more and 10 The concept of .social movement more on self-transformation. Conflicts arise in areas previously considered typical of the private sphere, involving problems of self-definition and cballcngcs to tbe dominant life-styles, for example. Tbe difference with those who insist on the political side of movements like McCarthy and Zald and Tilly is undeniable. Yet, this is a difference in emphasis rather than one concerning incompatible notions of what a social movement is. Indeed, the existence of cultural movements has never been denied either by Resource Mobilization theorists (Zald and .-^sh. 1966 speak of movements of 'personal change') nor by proponents of the 'political process' perspective (Tilly, 1984 mentions 'religious movements'). Tlie opponunity to include both cultural and ptjiiticai movements within the broader category of social movements bring us to the third component of tbe concept: 'Social movement actors arc engaged in political and/or cultural conflicts, meant to promote or oppose social change either at the svstemic or non-svstemic level'. Action which primarily occurs outside the institutional sphere and the routine procedures of social life Until the early 1970s debates on social movements were dominated by structural functionalists like Smelscr (1962) who put a great emphasis on the non-institutionalised nature of their behaviour. Today, social movement scholars are more cautious on this point. The aspects of 'collective effervescence' and 'nascent state' which had been empbasised by some (e.g. Alberoni. 1984 but originally 1965) as a distinctive feature of social movements are now more closely associated with the phase of their emergence. From very different perspectives, it has been demonstrated that .social movements continue even when collective effervescence is over, and that this is not immcdiatley followed by institutionalLsation (sec e.g. Melucci, 1984a and 1989; 'Tarrow, 1989). Tliere is actually a more complex pattern of interaction between non-institutional aspects and institutional ones, wbcrein social movements may either be an agent of change at the level of symbolic codes (as Melucd emphasises) or create new opportunities for interest intermediation (e.g. Nedelinann, 1984). Moreover, movements may also develop without going through a phase of 'collective effervescence'. In other words, collective identities may arise, that Mario Diani are strong enough to foster sustained collective action, yet that do not imply a 'nascent state' (Diani, 1990b). If the relationship between non-institutional behaviour and social movements is not strong enough to identify the former as a fundamental component of the latter, the same holds true for the idea that social movements may be distinguished from other political actors because of their adoption of 'unusual' pattems of political behaviour. Several scholars maintain that the fundamental distinction between movements and other social political actors is to be found in tbe contrast between conventional styles of political participation (such as voting or lobbying political representatives) and public protest. However, while the recourse to public protest is undoubtedly a qualifying element of political movements, it plays only a marginal role in movements oriented to personal and cultural change. If one accepts, as I do, that even the latter may be subsumed under the concept of sociai movements, then there is no reason to introduce this specification in the definition of the concept.'' Another widely shared assumption, at least in the more conventional version of the idea of social movements as 'unusual' phenomena, is that organisations involved in social movements are basically loosely structured. While informality and looseness are essential properties of the system of interaction, the same is not necessarily true for the single units of the system. Even though many loosely structured organisations arc actually part, possibly the dominant one, of social movement networks, they are by no means their only component. Indeed, the spectrum of SMOs is so wide and differentiated as to prevent any clear restriction of its boundaries: a key role in social movements may be played by such heterogeneous organisations as churches (e.g. in the black civil rights movement in America: McAdam, 1982); local branches of trade unions (e.g. in the peace movement in Britain: Byrne. 1988); neighbourhood solidarity organisations (e.g. in the British urban movements: Lowe. 1986). Moreover, the choice between a grassroots organisation or a bureaucratic lobby appears more and more frequently dependent upon tactical calculations by social movement actors (Zald. 1988:35-6), Even collective behaviour theorists agree that a proper understanding of social movements requires principles from both collective and organisational behaviour (Turner and Killian, 1987:230). This discussion suggests that features such as the extrainstitutional nature of .social movements, the prevalence of violent 12 7'/ic concept of social movement or disruptive political protest and the loose structure of social movement organisations cannot really be taken as fundamental characteristics of a social movement. These may however be extremely useful in differentiating between types of movements, or between different phases in the life of a specific movement. Thus, the following synthetic definition of the concept of social movement can be put forward: 'A social movement is a network of informal interactions between a plurality of individuals, groups and/or organizations, engaged in a political or cultural conflict, on the basis of a shared collective identity'. Social movemenLs, organisations, political events The different traditions of social movement analysis I have discussed so far show some degree of compatibility. 'To be fair, tbis 'immanent' consensus is sometimes only implicit in an author's formulation. In this reconstruction I have tried to emphasise the elements of continuity between different positions, rather than those of divergence - which are, by the way, the best known. The question is whether the effon to mediate between several distinct approacbes is not detrimental to theoretical clarity. In this section I will discuss this point. I try in particular to show in what sense this particular definition of social movements helps to differentiate them from a) political and social organisations like parties, interest groups or religious sects; b) other infonnal networks of collective action such as political mobilisation campaigns and political coalitions.^ .Social movements vs. political or religious organisations As we already noted in the previous section, sociai movements, political parties and interest groups are often compared under the assumption that they all embody different styles of political organisation (e.g. Wilson. 1973). At times, they are identified with religious sects and cults (e.g. Robbins, 1988). However, if our definition is correct, the difference between social movements and other political actors does not consist primarily of differences in organisational characteristics or patterns of behaviour, but on the fact that social movements are not organisations, not even of a 13 Mario Diani peculiar kind (Tilly, 1988 and Oliver, 1989). They are networks of interaction between different actors which may either inelude formal organisations or not, depending on shiftitig circumstances. As a consequence, a single organisation, whatever its dominant traits, is not a social movement. Of course it may be part of one, but the two are not identical, as the latter refiects a different, more structured organisational principle. Indeed, many influential scholars in the field keep using 'social movement' to mean both networks of interaction and specific organisations: citizens' rights groups like Common Cause, environmental organisations like the Sierra Club, or even religious sects like Nichiren Shoshu (McAdam et al.. 1988:695). Yet, this overlap is a source of analytical confusion, in so far as it fosters the application to social movement analysis of concepts borrowed from organisational theory, that only partially fit the l(K)ser structure of social movements.'"' Talking of Common Cause or the Sierra Club or Nichiren Shoshu as 'social movements' leads one to formulate concepts like 'professional social movement' (McCarthy and Zaid, 1973) or 'single-organisation movements' (Turner and Killian, 1987:369-7U) to emphasise differences between these cases and the nature of social movements as informal networks (which as we have seen they all agree upon). But qualifying Common Cause as a 'professional social movement' does not add very much to the understanding of it, that cannot be provided by concepts like 'public interest group' (see among others Etzioni, 1985). Similarly, a religious organisation like Nichiren Shoshu or Hare Krishna may be conveniently analysed as a 'sect'. This concept takes into account the greater organisational rigidity and the more hierarchical stnicture that these organisations display by comparison with social movement networks (see Robbins, 1988:1.50-55). In contrast, what both 'public interest group' and 'sect' do not really capture is the interaction processes through which actors with different identities and orientations come to elaborate a shared system of beliefs and a sense of belongingness, which exceeds by far the boundaries of any single group or organisation, while maintaining at the same time their specificity and distinctive traits. If we accept that social mo\ements are analytically different from SMOs we have also to redefine our notion of what is part and what is not part of a movement. Indeed, any organisation which fulfils the requirements I have pointed out (interactions with other actors, conflict and collective identity) may be considered part of a given movement. This may also hold for bureaucratic interest 14 'The concept of .social tnovetnent groups, and even political parties. The inclusion of political parties within social movements will surely raise many eyebrows and requires some qualification. By saying that political parties may be part of social movements I do not mean to suggest that 'social movements' is a broader theoretical category of which several types of organizations (interest groups, community groups, political parties and so forth) represent as many sub-types. Tar from it. Rather, 1 suggest that the features of the processes I have described as a social movement do not exclude that under eertain and specific conditions some political party may feel it.seif as part of a movement and be recognised as such both by other actors in the movement and by the general public. This is likely to be the exception rather than the rule, and to be largely restricted to parties originated by social movements, such as the Green Parties (Kitschelt, 1989; Rudig and Lowe, forthcoming). One could reasonably object that no matter how strong their identification with a movement, political parties actually perform specific functions at the level of interest representation and in this sense are different from social movements. That differences exist at the functional level is beyond t|uestion. Yet, the main peculiarity of social movements does not consist in their speeific way of performing the function of interest representation. Of course, their networks of interaction favour the formulation of demands, the promotion of mobilisation campaigns and the elaboration and diffusion of beliefs and collective identities. These factors all, in turn, contribute to redefine the cultural and political setting in which the action of interest representation takes place. However, when we focus on the function of interest representation in strict terms, we do not look at the way 'the movement' performs this function. We actually look at the way different specific SMOs perform these functions. Whether they decide or not to include participation into elections within their repertoire of action is dependent upon several factors including external opportunities, tactical and/or ideological considerations and their links to other actors in the movement. 'Ilie mere fact that they decide to do so, however, will not automatically exclude them from the movement. Rather, they will be part of two different systems of action (the party system and the social movement system), where they will play different roles. The way such roles are actually shaped will constitute a crucial area of investigation (Kitschelt, 1989). 15 Mario Diani Social movements, protest events, coalitions If social movements do not coincide with SMOs, they do not coincide with other types of informal interaction either. In other words, they differ from both loosely structured protest events and political eoalitions. Under what eonditions may a protest against the construction of a motorway run by informal citizens' aetion groups, a 'wild-eat' strike for higher wages in a firm or a demonstration for better nursing facilities in a neighbourhood be considered part of a social movement'.' And when are they just simple isolated 'protest events'? Some have suggested looking at the scope, dimension and length of campaigns (see e.g. Marwell and Oliver, 1984; Turner and Killian, 1987) in making this distinction. In broad terms, this is eonsistent with the notion of collective identity, as long and sustained campaigns will be more likely to create new specific identities among participants than sudden and brief protest outbursts or riots. However, there is also empirical evidence which casts doubt on the strength of this relation. Actually, the emergence of collective identity appears to be dependent on a plurality of factors.'' Even initiatives, which are apparently very specific, may thus be considered part of a social movement, provided they are interpreted in the light of a wider system of beliefs. This is possible if they develop in a context which is not only conducive to collectiveaction in general terms, but where a realignment of frames (Snow et al.. 1986) can occur. As we have seen in the previous section, the essential condition is that the sense of belongingness exceeds the length of the public activities and eampaigns. Collective identity may thus either become a preeondition for the creation of new and different identities (and eonsequently, of new and different social movements); or provide a persistent, though latent, basis for a new upsuige of mobilisation campaigns under the same heading. Social movements often persist even when they are not active on the public stage, and are rather going through a 'latency' phase. 'Iliose countercultural movements which alternate .sudden explosions of protest with long periods of latency may be analysed in this light, for example. In their case, collective identity provides the link between occasional outbursts which would be otherwise unexplainable (Melucci, 1984a, 1989). A further argument for the discriminating capacity of the notion of eolleetive identity comes from other examples of informal networks of eolleetive action, such as coalitions (for an introduction 16 The concept of social movement and a definition: Hinekley, 1981:4-6). Tlie,se reveal some similarity with social movements, in so far as they imply the existence of a conflict and of a eolleetive activity. However, the interaetion and co-ordination between different actors occurs mostly on an instrumental level, as actors try to maximise their outcomes by establishing alliances to other actors. In contrast to what happens in social movements, interaetion in coalitions does not foster the emergence of collective identities, nor does it imply necessarily any sort of continuity beyond the limits of the specific conHictual situation, let alone a global redefinition of the issues at stake.'" Conclusions In this article I have tried to show that different approaehes to the field share, in their definitions of 'social movement', the emphasis on some speeific dynamics. In particular, three basic components of social movements have been identified: networks of relations between a plurality of actors; collective identity: eonflictual issues. In contrast, it has been denied that anti-institutional styles of political participation or anti-systemic attitudes may constitute a distinctive trait of the concept of .social movements. I would arge that this definition of social movements may constitute the bulk of a programme of research and theorising that adopt 'social movements' as an analytieal, rather than a merely evocative, concept. It may also eontribute to the integration of different theoretical perspectives. During the 1970s, the resurgence of scholarly interest for social movements had focused either on the struetural determinants of new conflicts (mostly in Europe) or on mobilisation processes (mostly in the USA). Emphasising the interplay between networks, identity and conflicts challenges some conventional wisdom inherited from these traditions. On the one hand, it ehallenges the idea that the study of social movements may be equated to the study of new soeial conflicts. While there is an obvious strong correlation between movements and eonflicts, the concept proposed here accepts that, in principle, conflicts can arise even in the absence of soeial movements. How single, isolated eonfiicts may become a movement is a eentral matter for investigation. To this purpose, attention must necessarily be paid ' to social networks and processes of meaning construction." On the other hand, stressing the importance of social networks prevents one from confusing the analysis of 'social movements' 17 Mario Diani with the analysis of 'social movement organisations' or "mobilisation processes". This also bears substantial implications in terms of research strategy. Only the study of the properties of interorganisational and interpersonal networks is, in this perspective, directly relevant to the analysis of social movements. In contrast, for example, the study of individuals' commitment to a specific movement organisation, albeit of obvious substantial interest, is not specific of social movement studies. Rather, it is more directly connected to the broader analysis of individuals' incentives to collective action and political participation.'I do not pretend that the view proposed here is absolutely original. I would rather argue that it reflects - and partially expands on - recent efforts towards theoretical integration in the Held. To start with, many have recently argued for greater attention to be paid to the intermediate structures of collective action, i.e. the networks that link individuals, groups and SMOs active in the same, or related, conflicts (McAdam et al.,, 1988: Tarrow, 1988). This, in order to provide a proper link between 'macro' explanations, focusing on structural changes and factors, and 'micro' explanati(>ns, focusing on individual attitudes and behaviours. So far. research in this area has almost exclusively analysed the role of personal links in facilitating mobilisation (e.g. Klandermans ci a!., 1988). A more systematic investigation of the properties of these networks is needed, however, in order to assess their impact on a larger set of processes. These processes include how resources are put together and made available for action; the impact of the alliance and influence structure of social movements on their capacity to exert pressure on public authorities; the role of micro-mobilisation contexts, and in particular of the complex interpersonal bonds, which constitute the latent structure of social movements, in the elaboration of interpretative frames; and so forth (McAdam eial., 1988). Recent research has also assigned special relevance to the role of collective identity. Scholars like Touraine (1981) and Melucci (1989) have revealed that this is not a datum, but a key problem for the study of collective action. As we already noticed in our previous discussion, the sense of belongingness to a movement must never be taken for granted. In contrast, collective identity is always the precarious and temporary outcome of a 'bargaining' process between actors who embody quite different and heterogeneous beliefs. How do actors, who are broadly interested in similar issues, yet from different perspectives, come to think of 18 The concept of social movemetit themselves as part of a broader movement, while preserving their peculiarity'.' And how do they manage to maintain their collective identity, and eventually to adapt it to ehanges in the confliet, instead of splitting the movement in several factions and sects? Finally, how do movement identities react to shifts in dominant cultures in their environment? These and related questions become, if we take up this perspective, a central area of investigation.'' The erowth of eultural conflicts has also been at the core of recent theorising. The inclusion of both socio-political and cultural movements within this definition differentiates it from others- like Tilly - who conceive of the existence of shared beliefs and solidarity mainly as a precondition to the iiccurrence of public action and political protest. In contrast, other theorists (e.g. Gusfleld, 1981: Touraine, 1981; Melucci, 1989) suggest that the processes of meaning eonstruetion may also be regarded as the true essence of many conflicts in contemporary society. Of course, symbolic antagonism may often develop in parallel to political protest. Yet, the relationship between the two aspects is not necessarily in the sense of the former being a precondition of the latter. It may rather take different forms, which must become an object of careful investigation (for a recent example: Lumley, 1990). Finally, the definition also changes the idea that social movements are necessarily anti-systemic actors. This leaves more room for the analysis of how social movements change over time, in aspects as different as the number and quality of actors involved in protest events; the cultural interpretations of the conflict; the issues at stake; the repertoires of action and the degrees of radicalisation (e.g. Tilly, 1978, 1984, 1988; Tarrow, 1989). What appears as a challenge to the system in the mounting phase of protest may be viewed as a reformist attempt in a longer historical perspective; periods when social conflict is globally intense may encourage social movements to adopt radical, disruptive strategies with a greater frequency than phases when conflicts are not so strong and public concern tends to address other, non eonflictual issues. For the.se reasons it seems advisable to select a very limited number of variables to define the notion of social movement, and to leave more specifie connotations to the analysis of specific conflicts, cycles of protest or phases of deep underlying cultural strife. Boccotii University Received 10 September 1990 ; Milan Finally accepted 4 June 1991 19 Mario Diani Acknowledgements This article has greatly benefited from comments from Jack Brand, Donatella de!la Porta, Ron Eyerman and two anonymous referees. A preliminary version has been discussed during Ihe Annual Meeting of the Political Studies Association (University of Durham. April 1990). Joanna McPake. John Davis and Ron Eyerman have graciously helped with the language. Financial support from the European consortium for Political Research and the Consiglio Nazionale delle Ricerche is also gratefully acknowledged. Notes 1 Ii is not by ch;incc thui one of the most popular debates Jimonc sociiil movenieni sthtilars in recent ye;irs concerns the rok- ai soci;il networks in f;icilitalinj; individuals' mobilibatidn (e.g. Klanderniuris i-t al.. \WH): an important coiiiribucion. but to a coniroversy originated by Mancur Olwin s scminiil work in a different ihcorelieal (ralional ehoice Iheor)) and empirical (pariieipalion in trade unions and inierest groups) conicxi. 2 Many other scholars apart from lourainc and MeJiicci (mostly, but not exclusively. European) have played iin important role in ihc dehaie on new' soeial movements. Among them are Habermas. Offo, Castells (see for some discussions Cohen. I9,SS; Mis?tal. I9SX; Scoit. 1990). They are not taken into account here beciiuse they focus almost exclusively on macrosocial dynamics and do not pay attention to the spccilicity of the concept. In the lipht ul what tollows, the remark might somewhat apply to Touraine as well. Yet. he introduces a scries of more specific definitions which ure important to the development of the discussion here. 3 Pi72orno fiy^S) has been among ihe first to use this notion in order to challenge Oi.son's ttcil known hypothesis about the irrationality of collective action. 4 Snow and associates (Snow ci al.. i9S(.; Snow and Benford. 1988) use the concept of 'fnimi; alignment' (from Goffni;in's notion oi frames) to identify those changes in individuals' sets of beliefs wliieh iiecount for their decisions to join colleetive action- Cven though originally elaborated in the eontext of the analysis of individual mobilisation, the same notion may be usefully referred to the process whereby a hroadcr collective identity is created. 5 See among others Touraine. 1977: Turner. 19S1: Melucei. 1989. Hie degree of inelusiveness or exciusivencs.s of such identifications is on the other hand subjected to shifting conditions (Zald and Ash, 19<>6). f> Neidhardt nnd Rucht fIvyO) minntain for example ihat social mok-ements are defined among other features l>y the use of public protest activities, yet proceed to diffcrcntiiitt! l)etween socio-political and socio-cultural movements, the laner relying not on public protest but on expressive action, seekmg , . to attain social change indireetly through the afijiregated and long term effects of individual (K-haviour.' 7 I do not discuss on the other hand collective phenomena such as fashions. 20 The concept of social movement soli(i:irity campaigns in favour of external constitucncicb (e.g. collective efforts to help stuning African countries or the like), and crowd behuviciur (c.gfwtball hooliganism). This is partially due t.> limitations in space, panially to the fact that their differences to social movements have been iong since emphasised (sec among others Alberoni. 1984 and Melucci. 39S9). Moreover, these differences are probably more obvious than those that 1 am going to consider in this paper. Sutticc to say here, that differences between these phenomena and social movements consi5;t basically in the jbsence'pre.wnce of conflict; for fashions and crowd behaviours, they also lie to a certain extent m the absence of eolleetive identity S As I'amcla Oliver puts it: , . . all too otteii we speak of movemcm strategy, tactics, leadership, membership, recruitment, division of labour, success and failure - terms uhich strictly apply only to coherent decision making entities (i,e.. organisations or groups), not tn erowds. collectivities, or whole soaal movements,-(1989:4). 9 Several protests were for instance promoted by nature protection associations in Italy during the 19f)()s and the 1970s. In iibsolute terms, they were probably more frequent than the protests against nuclear power which developed in a very restricted period in the late !97(>s. Yet. the latter developed a specific collective identity and were perceived as a movement, while this was not «> with the former, who have come to identify themselves as u part ol the environmental movement only in the 19«()s, Ttie explanation may lie in the persistence, until the late 197(te. of attitudes of mistrust towards collective action within natuie protection assecame more im|Kmant than preexisting loyalties to specific organisations. See Regalia ff til- (1978). 11 See Kriesi (19SS). Tlie rapid growth of 'community action groups', "public interest groups'.'neighbtiurhood groups" and the like since the 1970s provides a good cxitmpie of u purely evocative use of the term 'social movement'. They have often been referred to as "citi/cns" movement' (e.g. Boyte. ]9K()). The problem with this use of the term is precisely that it embraecs indiscriminately all phenomena which have to do with political protest. In other words, it is also attached to isolated protest events or to tho-W public interest groups, that do not feel part of any movement nor are involved in any broider network. 12 Oberschall rai.ses a very close point when he maintains that, rather than "social movements', the real dependent variable should t>e 'collective action' (cited in Morris and Herring. 1987:165). Consistently with thi-s view, one of the most distinguished researchers in the study of participation in social movements draws large part of his empirical evidence from trade union activism. See Klandermans (forthcoming). (See also footnote 1). 13 For a broader, yet similar, perspective: Morris and Herring (19X7:192ff). 21 Mario Dian References Albctoni, F,. (I!)S4). Movemeni mil ImllmUon. New York. Columbia Univcrsiu Prc5A. " Bnylc. H C . (1980). The Backyar.l Rcv,,lMi,m. lj,,,km,,miing Ihc ,V,-.i- Olucm Movcmeni, Philadelphia. Icmple Univcrsily Press Byrne. P.. (I98S). The Campaig,, fo, Nuclear Dh^rmume^,, I^ndon Cr,K)m Melin. C'Dhen. J.L.. (1985). Strategy and idetitily: New iheoretieal paradi.ums and contentporary social movements". Sudal Hai-areh, vol. 52. pp. 6(i3-71ft Dalton. R. and Kueehlcr. M. (eds). U'>'*l). Owlletigms Ihe Polliicil Order- AV,v ,'itmat arul FoUlieul MovemeiiK m Wmer,, Denumicles. Camt.ndci: I'olitv Press, della Pona. D.. (IOfW). RMniitment processes in clandestine political orpaniiations Italian left wing terrorism, in Klandermans el at 19SS O.an,. M.. (1990a). The network strueture of the Itahai. ecology movement' .Stieial .'ieience Inforitiallon. vol.29. pp. 5-.^l. Diani. M.. (lWlb). "Ilie Italian ecology movement: From radie,ilism to moderilion' m Rudig. W. (cd.). Ctee,, l;,lllie, I, IWO. Edinburgh. Edinburuh University Press, nder. K.. (1985). -n,e -now scntrc lo Action. Comparing Social Movemenl Research Acrosi Cultures. Greenwich. J Al Press, Kriesi. H., (19K8). 'Ttie interdependence of structure and action: Some reflections on the state of the art", in Klandermans, B. ci al.. 19S8. lA)we. S.. (1986). Vrban Social Movemenn. The Ciiy after Castctk. London. •Macmillan. Lumley. R.. (1990). States of Emergency. Cultitrei of Revolt in Italy from 196S to J97ii. i-ondon. Verso, MeAdam. D,. McCarthy. J.D. and ZaId. M.N,. (1988). '.Social movements', m Smelser. N.J. (ed.). Handbook of Sociology. Beverly MillVLondon. Sage, McAdam. D,. (1982). Political Process ami the Development of Black irLuirgency. 19.^0-1970. Chicago. University of Chicago Press, McCarthy. J.D. arid Zaid. MN.. (1977). 'Resource mobiliziition and social movements: A partial theory'. American Joumnl of Sociology. vol,K2. pp. 1212-41, McCatihy. J,D. and /jld. M,N,. (1973). The Trend of Social Movements in America: Profeisionalization and Resource Mobilization. Murristown. General Learning Press, Marwell. G, and Oliver. P.. (1984). 'Collective action theory and social movements research', in L. Kriesberg (ed.). Hescurch m Social Movements. Conflict and Change, vol.7. Greenwich Conn.. JAI Press. Melucci. A. (ed,). (19S4a). Altri codia Aree di movimento nella metropoli. Bologna. II Miiiino. McUicci. A,. ()98-1b). 'Movimenti in un montlo di segni'. in Melucei. A,. 1984;). Melucci. A,. (198f'). "Tlie symbolic cliallenge of contemporary movements', Social Research, vol.52, pp. 789- 81(). Melucci. A,. (1989). S'omads ofthe Present. London, Hutehinson Radius. MisTital. H.. (19SS). 'Social movements of the core and ihc periphery'. Australian and ,V('i>' /.ealand Journat of Sociolof;y. vol.24, pp. 65-82. Morris. A and Herring. C . (19K7). -nieory and research in social movements: A critical review". Annual Review of Political Science, vol.2, pp, i:i7-y.S. Nedelinaiin. B.. (!98-l). 'New political movements and changes in prtKesses of mterniediation". Sociat Sciernc Information, vol.23. pp. ll)2';-48, Neidhardt. F. and Rucht. D,. (1990). 'llie analysis of swial movements: 'Ilie state of the art and some perspectives for further resc:iieh\ in Rucht. D.. l'/X)a. Oliver. P.. (1989), 'Bringing the crowd back in: The nonorganizational elements of social movements', in Kriesberg. L. (ed.). Research m Sodal tMoiements. Pizzor social movements, in Kriesberg. L. (ed.). Research m Sodal tMoeme. Conflict and Chant^e. vol.11. Greenwich. Conn,. JAI Press, izzorno. A.. (1978). 'Political exchange and collective identity in industrial conflict", in Crouch. C. and Pi/.zorno. A. (eds). The Resurgence of Class Conflict in Western Europe .unce I96fi, London. Macmillan. Regalia. L. Regini. M. and Reyneri. E.. (1978). 'Uibor eonflicts and industrial relations in Italy', in Crouch. C. and Pizzorno, A. (eds). The Resurgence ojClass Conflict in We.Uern Europe since 196.S. London. Macmillan, Robbins. T.. (198S). Cults, Converts and Charisma: the Sociolosy of New Religious Movements, London. Sage. Rucht. D. (ed.), (I9*«)a). Re.Kcurch in Social Movements: The State of the Art. Frankfurt, Campus/Boulder Co. Westview Press, R h D (19911b) Th i d i t i f ew movements in Frankfurt, Campus/Boulder Co. Westie Pres Rucht. D. (19911b). "The strategies and action repertoires of Dalton, R, and Kueehler. 1990. 23 Mario Diani Rudig, W. and Lowe. P.. (forthcoming). The Green Wave: A Comparative Analysi of Ecological Parties, Cambridge, Polity Press. Sartori. G,. (1984). 'Guidelines for concept analysi.s'. in Sartori. 0. (ed). Social Science Concepts. A Systematic Analysis, Beverly Hills/London. Sage. Scott, A.. (1990). Ideology and the New Social Movements, London. Unwin Hyman. Smeker. NJ.. (1962). Theory of Colleciive Behavior, New York, The Free Press, Snow, D.A. and Benford. R.D.. (1988), "Ideology, frame resonance, and participant mobilization*, in Klandermans ct al.. 1988. Snow, D.A., Burke Rochford, E., Worden, S.K. and Bcford. R.D., (1986). "Frame alignment processes, micromobilization. and movement participation'. American Sociological Revietv. vol.51, pp. 464-81. Tarrow, S., (1983). 'Struggling to reform: S(KiaI movements and policy change during cycles of protest'. Western Societies Paper l\Jo.I5. Ithaca. NY. Cornell University. Tarrow, S., (1988). •National politics and eoliectivc action. Recent theory :ind research in Western Europe and the USA'. Annual Review of Sociology. vol. 14. pp. 421-40, Tarrow. S.. (1989), Democracy and DLwrdcr. Protest and Politics in Italy J96 197^, Oxford, Oxford University Prc-A. Tilly. C;., (1978). Erom Mobilization to Revolution. Reading. A\ddison-Wes!cy. Tilly, C. (1984). 'Social movements and national politics", in Bright. C. and flarding. S. (eds). State-Making and Social Movements: Essays in History and Theory. Ann Arbor. University of Michigan Press. Tilly. C, (1988), "Social movements, old and new', in Misztal, B. (cd.). Research in .Social Movements. Conflict and Change, vol.10. Greenwich, Conn., JA! Press, Touraine, A., (1977). The Self Production of Society, Chicago. University of Chicago Press. Touraine. A., (1981). The Voice and the Eye. An Analysis of Social Movements. Cambridge, Cambridge University Press. Touruine. A., (1985), 'An introduction to the study of social movements'. Social Research, vol,52. pp. 749-88, Touraine, A., Hegedus. Z.. Dubct, F. and Wieviorka, M.. (198.1a). Anti-nuclear Protest. The Opposition to Nuclear Power in Erance. Cambridge, Cambridge University Press. Touraine. A., Dubet. F.. Wieviorka. M. andStrzcIccki. J.. (\9S:ib). Solidarity. The Analy.m of a .Social Movement: Poland 1980-1981. Cambridge. Cambridge University Press. Turner. R,. (lQSl). 'Collective Behaviour and Resource Mobilization as approaches to social movements: Issues and continuities', in Kriesberg. L. (cd.). Research in Social Movements, Conflicts and Change, vol.4. Greenwich. Conn.. JAI Press. Turner. R. and Killian. L.. (19S7). Collective Behaviour. Englewood Cliffs. NJ. Prentiec Hall. Wilson, J.O.. (1973). Political Organizations, New York. Basic Books. ZaId. M.N., (1988). "The trajectory of the sodal movements in America', in Misztal. B. (ed.). Research in Social Movements, Conflict and Change, vol.10. Greenwich. Conn., JA! Press. ZaId. M.N, and Ash. R.. (1966). "Social tnovcment organizations: Growth, decay and change". Social Eorces. vol.44, pp, 327-41. 24 The concept of social movement ZaId. M.N, and McCarthy. J.D. (eds). (1987), Social Movements in Organizational Society: Resource Mobilization. Conflict and Institutionalizatkm. New Brunswick. Transaction Books. ZaId. M.N. and McCarthy. J.D, (1980). 'Social movement industries: Competition and cooperation among movement organizations', in Kriesberg, L. (cd.). Research in Social Movements, Conflicts and Change, vol.3. Greenwich, Conn., JAI Press. 25