Richard Hyman
University of Warwick
'Trade unions have always had two faces, sword of justice and vested interest ' (Flanders, 1970: 15). The balance between these two features can change over time, however. It seems clear that in many countries, unions have lately come to be widely perceived as conservative institutions, primarily concerned to defend the relative advantages of a minority of the working population. One of the challenges which confront trade unionism in the twenty-first century is therefore to revive, and to redefine, the role as sword of justice.
Many union leaders and activists around the world are indeed well aware of this challenge, and in a variety of countries there are examples of creative and imaginative responses. The aim of this paper is to review some of the challenges and discuss some of the potential for response. An important task for our project will be to survey the latter in more detail.
It is common to emphasise the material challenges faced by trade unions, and with good reason. There have been increasing difficulties both in the external environment of union organisation and action, and in the nature of the constituencies which unions seek to mobilise.
Externally, the economic environment has become far harsher. Global competition has intensified, putting new pressures on national industrial relations regimes. Industrialised market economies which had enjoyed several decades of relatively full employment have since experienced a return to mass unemployment. Massive job losses have been one of the elements of the 'shock therapy ' inflicted on the new market economies. newly industrialised economies, in many cases previously cushioned from external shocks, have become subject to the fluctuations of global markets.
As governments grapple with the problems of adaptation to the new disorder in the world economy, the political environment in many countries - particularly those where labour movements are longest established - has become far more unfavourable. In some cases this is linked to the erosion of unions ' representative status as 'social partners ', in part in consequence of loss of membership.
The third external challenge comes from employers. In some countries there has been a growing unwillingness to accept trade unions as collective representatives of employees; in others, while collective bargaining has survived its scope has been reduced, and managements have established new forms of direct communication with employees as individuals. The fashion for team-working has introduced new mechanisms of collective decision-making which in many countries are detached both from trade union structures and from statutory institutions of workplace representation. In addition, the expansion of multinational companies has meant that leading employers are often willing and able to escape the regulatory force of national industrial relations systems.
What may be termed the internal challenge stems from changes in the constituencies which unions seek to recruit and represent. Traditionally trade unions, particularly but not only in highly industrialised societies, were shaped by the existence (real, but often exaggerated) of a 'normal ' employment relationship. This involved a full-time job with a specific employer and usually a degree of long-term stability. The classic example was the 'mass ' worker in mining, manufacturing and transport, with limited individual resources in the external labour market but significant potential to exert collective pressure on the employer. Though early trade unionism in many countries was indeed based on a highly skilled 'labour aristocracy ', 'modern ' labour movements found their core constituency (at least in the private sector, which was dominant numerically and in shaping labour movement policy) among those who lacked substantial capacity for individual career advancement but were not so vulnerable as to be incapable of sustained collective cohesion.
The 'normal ' worker, and hence the 'normal ' potential trade union member, was thus a full-time employee whose employment status was not merely casual. By extension, the 'normal ' employee was a man who was presumed to be the 'breadwinner ' for his family. (Of course there were exceptions, notably in the textile industries; but Marx 's prediction that the predominance of female employment in cotton-weaving would be the prototype for capitalist 'modern industry ' proved strikingly wrong.) This in turn shaped the typical trade union agenda: predominantly concerned with terms and conditions of employment, and in particular with three aspects: achieving the payment of a 'family wage ', defining and reducing the standard working week, and constraining the employer 's ability to hire and fire at will.
While the realities were always more complex than this stylised account, and certainly varied between countries, this model of the traditional 'normal ' agenda is far from a caricature. In many countries there have indeed been serious efforts, sometimes dating back several decades, to transform this agenda in order to appeal to a broader constituency. Achieving this transformation has become increasingly urgent.
The key reason is that 'atypical ' employment situations have become increasingly typical. Part-time work, short-term and casual employment, agency work, self-employment (both genuine and spurious), special government make-work schemes and of course unemployment have all become more common; in total, in some countries, they affect the majority of the economically active population. At the same time there have been numerous structural shifts in the sectoral and occupational distribution of employment: the decline of most of the traditional staple manufacturing and associated industries and the growth of a wide variety of service industries, particularly in the private sector; the eclipse or transformation, partly under the impact of microelectronic technologies, of many traditional manual occupations and the growth of 'white-collar ' work (now in many countries the majority); the reversal of the process of employment concentration with 'downsizing ' in former core industries and the expansion of small and medium-sized enterprises.
There has thus developed a diversity of forms of linkage to the labour market, and structural change has brought both winners and losers (though in most countries, losers far outnumber winners). Instead of presuming the existence of a 'normal ' worker it is necessary to differentiate. Reich (1991), focussing on skills and functions, distinguishes 'routine producers ', 'in-person servers ' and 'symbolic analysts '; the first two categories consisting primarily of dead-end and often precarious jobs, only (some of) the latter enjoying significant scope for advancement. Standing (1997) has described contemporary labour markets as stratified into seven groups, which he terms the elite, the salariat, 'proficians ' (those without stable employment but with valuable marketable skills), traditional core workers, low-skilled 'flexiworkers ' who depend on casualised job opportunities, the unemployed, and those detached altogether from regular (or legal) work. Whatever classification is adopted, it is evident that the traditional core constituency of trade union membership has dwindled, while there has been expansion at two extremes: those with professional or technical skills who may feel confident of their individual capacity to survive in the labour market; and those with no such resources but whose very vulnerability makes effective collective organisation and action difficult to achieve or perhaps even to contemplate.
These developments are evidently connected to the increasing feminisation of the labour force. To a substantial degree, 'atypical ' employment is female employment (Briskin and McDermott, 1993; Cook et al., 1992). The growing proportion of women in the formal labour market negates the traditional model of husband as wage-worker and wife as domestic worker, but in most countries domestic work remains primarily or exclusively female. The management of the relationship between time spent at home and in employment is thus a distinctive concern of an increasing - female - section of the workforce.
There has also been a different kind of transformation in the relationship between home and work. There is a stereotype of the traditional proletarian status which emphasises a common work situation, an integrated and homogeneous local community, and a limited repertoire of shared cultural and social pursuits. Though exaggerated, this stereotype does identify a core of historical reality, particularly in the single-industry manual working-class milieux in which 'modern ' mass trade unionism had its strongest roots. By contrast, in contemporary society the spatial location and social organisation of work, residence, consumption and sociability have become highly differentiated. Today the typical employee may live a considerable distance from fellow-workers, possess a largely 'privatised ' domestic life or a circle of friends unconnected with work, and pursue cultural or recreational interests quite different from those of other employees in the same workplace. This disjuncture between work and community (or indeed the destruction of community in much of its traditional meaning) entails the loss of many of the localised networks which strengthened the supports of union membership (and in some cases made the local union almost a 'total institution ').
Many writers have seen these structural shifts as linked to a cultural and ideological decline of collectivism and a rise of individualism. In its simple form this argument involves a gross oversimplification (Kelly, 1998). Nevertheless the eclipse of the 'mass worker ' whose institutionalised solidarities were reinforced by the broader networks of everyday life does mean that the possibility and character of collectivism are today very different when work and everyday life are increasingly differentiated (Zoll, 1993). Pérez-Díaz (1987: 122-3) has outlined the implications with great clarity. Traditionally, he argues, workers ' collective orientations were externally defined: either they 'acquired a class ethos or habit ' because they were immersed in a social milieu where such values were unquestioned, or they were inspired by commitment to the ideal of 'a new world or a different future '. By contrast, today the traditional identities have been displaced and the transformatory ideals have lost their grip; workers adopt 'a rational, instrumental or experimental attitude towards the unions (or parties) '. To win their support, unions now have to pass a direct and pragmatic test.
This more calculative orientation, which certainly creates possibilities of far greater individualism, makes practicable the new managerial efforts to capture workers ' loyalties and displace identification with trade unionism, and may in turn be reinforced by such efforts. But it also reflects the degree to which unions have experienced 'a serious moral and intellectual crisis [and] their reserves of moral indignation seem to be depleted ' (Pérez-Díaz, 1987: 114-5).
Hence the evident material problems facing trade unions cannot be separated from less tangible problems of ideology. To resist the hostile forces ranged against them, unions must mobilise countervailing power resources; but such resources consist in the ability to attract members, to inspire members and sympathisers to engage in action, and to win the support (or at least neutrality) of the broader public. The struggle for trade union organisation is thus a struggle for the hearts and minds of people; in other words, a battle of ideas.
In the remainder of this paper I consider some of the ideas which can contribute to this battle. The representation of workers ' interests - and their definition, which is necessarily a prior process - has never been straightforward. Building collective solidarity is in part a question of organisational capacity, but more fundamentally it is part of this battle of ideas. The crisis of traditional trade unionism is reflected not only in the more obvious indicators of loss of strength and efficacy, but also in the exhaustion of a traditional discourse and a failure to respond to new ideological challenges. It is those whose projects are hostile to what unions stand for who have set the agenda of the past decades. Unions have to recapture the ideological initiative.
As a starting point, the labour market perspectives of the 'mass worker ' with a standard model of full-time employment, firm-specific job security and limited scope for occupational advancement can no longer dictate the central content of bargaining policy. To construct trade union programmes with which vertically and horizontally differentiated groups of workers can identify requires a sensitive redefinition of what interests are represented. If on the one hand unions must be alert and receptive to (possibly altered) expectations and aspirations on the part of actual and potential members, on the other a priority must be to construct an agenda which can unite rather than divide. To do so, unions must scrutinise the concepts which have inspired the offensive of employers and the political right and attempt to reclaim these for different purposes. I consider a number of examples.
Flexibility emerged, notoriously, as a rallying cry directed against forms of social regulation - by law or by collective agreement - which have tempered the arbitrary and unequal workings of the labour market. The ideological bias of the term is obvious: presenting as 'rigidities ' those labour market protections which neo-liberals wish to weaken and restrict, making workers more disposable and more adaptable to the changing requirements of the employer. This 'negative flexibility ' (TUAC, 1995: 5) has naturally been opposed by most trade unions.
Yet flexibility can have alternative meanings. The 1970s objective of 'humanisation of work ' was in essence a claim for flexibility in the interests of workers through the human-centred application of technologies, the adaptation of task cycles and work speeds to fit workers ' own rhythms, the introduction of new types of individual and collective autonomy in the control of the labour process. This agenda has in large measure been hi-jacked as part of the new managerialism of the 1980s and 1990s (with its mendacious rhetoric of 'empowerment ' and 'human resource development ').
Can unions recapture the initiative? A rigid division of labour and narrow standardisation of tasks were impositions of a particular model of capitalist work organisation, a form of subordination which involved a degradation of status for many workers. To the extent that some of the features of Taylorist-Fordist systems have lost their attractions to many employers, space exists for unions to mobilise support for radical alternatives which transcend some of the divisions within the labour force. For example, one widespread trend in manufacturing over the past decade or more has been the introduction of teamworking, with team members performing a variety of tasks and exercising a degree of discretion over operational decisions. In many countries, unions viewed such initiatives with considerable suspicion; understandably, since teamwork was typically one element in a move towards Japanese-style 'lean production ' and hence a recipe for job-cutting and 'management by stress ' (Parker and Slaughter, 1988). However, simple resistance often proved ineffectual, since union members themselves were frequently attracted by the rhetoric of autonomy and job enlargement. More viable in the longer run have been strategies of 'critical engagement ', in which unions have responded by mobilising support for their own demands in the process of negotiating change. For example, a comparative study of work restructuring in the motor industry (Kochan et al., 1997) shows clearly that unions in some countries have been able to exert significant influence on the change process by such means.
Another key issue in the contemporary world of work is that of time-flexibility. Again, this has often involved making workers more available and disposable to suit the changing requirements of employers. On the one hand this can mean the extension of working time to 'unsocial ' hours and days: evening and night-work, weekend working; on the other, payment only for those hours when the employee can actually be set to work (Alaluf et al., 1995). The latter can entail, for example, the use of split shifts or even - notoriously in Britain - 'zero-hours ' contracts where the employee must be available but is paid only if called to work.
There is however a worker-oriented meaning of flexible working time which can directly confront that of the employers and which offers potential for moving from the defensive to the offensive and integrating very different types of employee interest (Mückenberger, 1995). This centres on the idea of time-sovereignty: the ability to influence the patterns of the working day, week, year and lifetime to optimise the temporal linkages between employment, leisure, career development and domestic life. 'Traditional rigid conceptions of working time do not suit the diversity of employee interests ' (Lapeyre and Hoffmann, 1995: 8-9). Most notably, women workers (unless and until there is a radical redistribution of domestic responsibilities) have a particular interest in ensuring that there is genuine flexibility of choice between full-time and part-time employment, and that the contractual position and career potential associated with the latter are not inferior to those in full-time jobs (Cunnison and Stageman, 1995: 202).
More generally, opening new areas of choice in the organisation of individual working time could be seen as an important trade union principle (Matthies et al., 1994) The operation of 'flexitime ', originally devised to suit managerial requirements, certainly provides scope for a 'personalisation ' of the working day (Leccese, 1997: 169) attractive to many workers. Similarly, the development of 'annualised hours ' systems has reflected employers ' interest in flexibility but can also be adapted to suit workers ' own choices. But the negotiation of individual working time will allow the employer the upper hand, and hence create new possibilities for exploitative relations, unless undertaken within a collectively regulated framework. Moves towards greater flexibility thus create both the need and the potential for new forms of trade union regulation (Raasch, 1995).
Just as unions have increasingly been involved in negotiating flexitime, so there has been considerable union involvement in phased retirement agreements. Again, such deals have often been initiated by employers as a form of partial redundancy; but a flexible rather than abrupt transition from 'normal ' employment to retirement suits the wishes of many older workers themselves. Much more generally, unions could appeal to many workers by pressing for increased choice of both the quantity and the distribution of working time to match individual circumstances and preferences, and by establishing the groundrules to ensure that such flexibility is not used to employees ' disadvantage.
The most dramatic feature of labour market trends in the past two decades has been a massive growth of insecurity. Survey evidence from a range of countries shows that the fear of job loss - either through collective redundancy or through victimisation by the employer - is the overwhelming work-related concern of employees today. Part of the function of trade unionism is to resist of this insecurity; but to the extent that such resistance is company- or sector-specific, its consequences may well prove divisive. The fight for company-level security, if successful, by stabilising the position of 'insiders ' may make the labour market situation of 'outsiders ' even more precarious. Where public employees struggle to retain protections which in the private sector were lost a decade ago, their unions may be seen as defenders of sectional privilege. (It may have been only because of very distinctive political circumstances that the public-sector strikes in France in 1995 and 1996 evoked considerable popular support.)
Yet it is surely essential that to address workers ' current consciousness of extreme job insecurity, trade unions develop programmes which offer hope of real employment opportunity yet do so in a non-divisive manner. In constructing an agenda which links the interests of the precarious, the unemployed and the relatively secure, it is again possible to seek a distinctive trade union application of current rhetoric which is often used mendaciously. One concept which has become increasingly popular among policy-makers is 'employability ': the argument is that individuals can no longer anticipate unbroken employment within a single organisation but can avoid labour market vulnerability by acquiring valued competences, including adaptability itself. This is the basis on which the European Commission (1997) envisages a 'balance ' between flexibility and security: a balance which in Dutch labour market debate has been given the name 'flexicurity ' (Wilthagen, 1998).
Commonly this rhetoric is no more than a means of individualising the problem of unemployment and deficient job opportunities and scapegoating the unemployed for their own marginalisation; as Lowe (1998: 248) puts it, 'the concept of «life-long learning» is shifting the onus of human resource development onto the individual '.
A purely supply-side labour market policy aimed at increasing individual 'employability ' is likely to result primarily in a more qualified cohort of unemployed; a frustrating mismatch between enhanced skills and the limited skill content of available jobs (particularly in the expanding service sector); and perhaps also in a demographic shift in the structure of employment and unemployment. However, the concept of employability is in principle one which can be made central to trade union policy. This would imply the coordination and integration of demands which unions have indeed often embraced: first, for enhanced individual entitlements to education and training, and for flexible opportunities to benefit from these throughout the working life; second, for more effective (and worker-oriented) provision both by employers and by education and training institutions; third, for demand-side policies to encourage employment growth and, no less importantly, to provide appropriate employment opportunities for 'upskilled ' workers. As Lowe argues (1998: 249), 'job quality could be a basis for collective action, especially among well-educated young workers whose expectations are still high '.
There is significant scope for action at company and sectoral level, to influence the process of work restructuring and technological innovation in the direction of upskilling rather than deskilling. The comparative study of the transformation of work in telecommunications edited by Katz (1997), for example, shows that the contrasting strategies adopted by unions in different countries have been a significant factor in explaining the very different ways in which jobs have been reconfigured. But some of the issues involved require economy-wide intervention to match supply and demand of skills - including, perhaps, action to ensure that foreign inward investment does not merely take the form of low-skilled and disposable jobs but enhances the scope for 'employability ' policies.
Part of the difficulty is that these demands address different interlocutors and involve different levels of initiative, and hence may fail through lack of coordination. To take a concrete example: the imaginative and innovative proposals developed by IG Metall a decade ago (Tarifreform 2000) were overwhelmed by the macroeconomic problems affecting the German labour market after unification. Conversely, one of the difficulties for any 'alliance for jobs ' - now once more a key issue following the change of government in Germany - is how to translate a central agreement into action at the level of individual companies (Streeck, 1998: 537). Unions themselves could become central actors in building linkages between these different levels of decision-making so that citizens are enabled 'to define together supply and demand ' within the labour market (Lipietz, 1996: 271).
This connects to a third theme: opportunity. Again, this is a concept which has been appropriated by the right but should be reclaimed for the labour movement. For most of the twentieth century, the core workforce which formed the main basis of trade unionism achieved their employment status through the dull compulsion of circumstance. Career advancement and self-directed occupational mobility are aspirations increasingly salient for unions ' actual and potential constituencies. As Waddington and Whitston (1996: 163) note in their study of white-collar workers ' attitudes, 'new union members... look to unions to negotiate a fair and equitable framework within which individualized aspects of the employment relationship - which are often career related - may be worked out '.
The weakening of the ties to the existing occupation and employer is however emancipating only to the extent that real and preferable alternatives are open. As with the themes of flexibility and employability, so more generally: it is evident that while the choice among alternative options is an individual project, its reality is deceptive and even threatening unless a genuine and favourable structure of opportunities exists.
This creates important openings for unions to address what Leisink (1996) calls 'occupational interests '. To enhance the opportunity structure is necessarily a collective project, one which challenges both employers ' discretion and the anarchy of market forces. In many ways a redefinition of the traditional function of trade unionism, this is another key dimension of a union agenda which can appeal to diverse constituencies in solidaristic fashion (Kochan and Wever, 1991: 373).
In essence, then, the challenge for trade unions is to win the argument that individual choice is liberating only when the options available are those that workers wish to choose. In the past, many unions have favoured inflexible regulation out of fear that this provides the only safeguard against manipulation and exploitation by employers; in the current situation this protection must be guaranteed primarily by procedural rules which enhance individual discretion and by active labour market policies which provide an advantageous framework for career decisions. In both respects, unions have a vital role to perform.
Changes in the organisation of production and the employment relationship (such as team-working, quality circles, performance-related pay, personalised contracts) are often accompanied by a managerial propaganda offensive in which 'empowerment ' is a central rhetorical device. Such mendacious discourse typically provides a 'democratic ' gloss to employer efforts to intensify production pressures, cut staffing numbers and undermine traditional forms of collective regulation.
The 'new workplace ' is one in which employees often have increased responsibilities but with reduced power and resources. As labour costs are reduced through the imposition of 'lean ' organisation, employees are simultaneously pressed to take increasing concern for 'quality ' and 'customer care '. The effects may be profoundly alienating; yet the ideological argument that more stressful work is more worthy and that intensified external pressure means greater autonomy has proved strangely effective. The big lie seems to work: as Dejours (1998) insists, evil is rendered banal and the intolerable becomes tolerated. The paradoxical consequence, suggests Coutrot (1988), is a form of 'forced cooperation ' whereby employees embrace their newly (re)defined roles for want of any visible alternative. Yet this acceptance is only partial: for example the annual British Social Attitudes surveys reveal a large and increasing proportion of workers (approaching two-thirds) believing that management 'try to get the better of employees ' and that 'big business benefits owners at the expense of workers '. The detailed case studies undertaken by Scott (1994) reveal a similar picture.
In its most recent report on world labour, the ILO (1997: 27) referred to the 'democratic function ' performed by trade unions. This can be understood in a double sense: by virtue of their capacity for collective representation, unions can give employees a 'voice ' within the workplace and limit unilateral and arbitrary management action; but in addition, unions can challenge the authoritarian and hierarchical structures of contemporary employing organisations and can press for an extension of citizenship rights to employment. In many of the developed economies, such demands gathered pace in the era of stability and growth; in a period of stagnation and recession the emphasis has been on more immediate material issues. In developing economies with a substantial labour surplus, questions of industrial democracy have more often than not been regarded as a diversionary luxury (Ramaswamy, 1988: 239).
Nevertheless, trade unions ' democratic function could speak to real grievances and concerns in a way which strengthens unions legitimacy and appeal. Unquestionably there is considerable scope to exercise this function by challenging the widespread current abuse of concepts of democracy at work and exposing the anti-democratic character of much that passes for 'human resource management '. By focussing their own demands and activities on the contradiction between management rhetoric and everyday reality in the workplace, trade unions have the potential to address current worker discontents in ways which generalise fragmented experiences and permit new forms of solidarity in the pursuit of genuine empowerment.
Needless to say, unions ' capacity to mount a credible campaign for greater democracy in employment will be severely weakened unless they can demonstrate their own democratic credentials. This poses evident challenges for unions to scrutinise and if necessary reconstruct their own representative capacity and internal processes of agenda-building and decision-making.
The traditional 'normal ' employment relationship involved a sharp dichotomy between life at work and outside. Where trade unions were longest established and collective bargaining most strongly developed, unionism itself tended to reflect and reinforce this dichotomy. This has not been universally the case, however: unions in some countries, particularly where capitalist wage-labour has not long been the dominant basis of production, have typically embraced broader community concerns.
More established unions could well learn from the experience of newer union movements. One reason is the erosion of the 'normal ' employment relationship. Another is the extent to which 'community ' has become an ideological device in contemporary political argument.
Arguments around the idea of 'community ' have two aspects. One is negative: a legitimation of the withdrawal of elements of state provision, intervention and regulation in social welfare and labour market policy. 'Communitarianism ' can thus provide an alibi for deregulation. Another strand of argument is more positive: the thesis that the organisations of 'civil society ' can mobilise pressure, and perhaps generate resources, which can counteract the destructive impact of global competition and global corporations. Unions obviously have a strong interest in engaging in this debate and in influencing conceptions of community in accordance with their own objectives.
The links between work and community can be seen in two dimensions. First, as well as producers, workers are also consumers and citizens; unions which can relate to (potential) members in all these roles can build a deeper relationship than if they merely focus on employment-related issues. Second, workers produce goods or services for diverse groups of consumers, customers or clients. Employers (and other manipulators of opinion) often attempt to counterpose the interests of one against the other. Unions are in a better position to represent their members ' interests if they can build alliances with those at the receiving end of their productive activity. This is particularly the case perhaps in the public sector: Johnston (1994: 9-10) explores how public service unions in the United States - which have provided the driving force for union renewal in the 1990s - have had to adopt a 'public interest ' logic and construct coalitions with NGOs and with representatives of user groups. Conversely, in the case of workers with a vulnerable labour market position in the private service sector, effective organisation may be possible only through seeking such alliances: constructing the basis for regulating 'a [local] labour market with help from community groups that share an interest in raising wages and labour standards ' (Wever, 1997: 465). In the case of such initiatives, concludes Lipsig-Mummé (1998: 20), 'their dual anchorage - in the community and in the union - allows them the potential for creativity '.
It is often argued that the increase in the number of women trade unionists has in itself led to a broadening of the unions ' agenda. 'Because their lives are grounded in the community as well as in paid work, in caring for others as well as in working on their own account, their trade union agenda has always been wider than men 's.... Important new issues have been brought onto the movement 's agenda, such as health and the quality of community life, childcare and the responsibilities of a multicultural society ' (Cunnison and Stageman, 1995: 242). But building 'social unionism ' (COSATU, 1997; Waterman, 1998) is not simply a gender issue. All workers have an interest in the quality of life in the broader social milieux which they inhabit, and unions which can 'mediate between the economic and social structure ' (Piore, 1994: 537) may increase their attraction and legitimacy. One example is the tempi della città campaign in Modena in the mid-1990s, when the local unions joined with community groups, business organisations and the local authority to agree changes in the timetables of transport services and communal facilities to match the varying requirements of workers-as-citizens. Much more generally, current emphasis on 'life-styles ' - which some critics perceive as a source of individualism - provides 'a focal point alternative to work-based identities ' which in one respect threatens unions but in another offers opportunities for an new basis of recruitment and representation (Piore, 1991: 403-4).
Establishing a 'social unionism ' has implications for unions ' organisational structures. In many countries, the primary unit has been the company or workplace branch; indeed in Japan and many other Asian countries, unions as such are enterprise-specific. Such a structure has an obvious collective bargaining logic, but can reinforce divisions between 'insiders ' and 'outsiders '. Moreover, even in terms of traditional 'business union ' objectives a company-based structure may no longer be as effective in the past. Many workplaces are no longer social units: 'lean production ' has reduced the scope for socialising on the job, diversification of work schedules means increasingly that only a fraction of the workforce is present at any one time, subcontracting entails that workers on a single site may be employees of different companies, and individuals often live a considerable distance from their work.
This creates a need for alternative organisational mechanisms. For example, Richter et al. (1996) recount the experience of one of the regions of the German metalworkers ' union in building activity around the localities where members (and potential members) live rather than where they work. This also offered the basis for creating links between employed and unemployed, and between working and retired members. (It should be noted that while unions in some countries - notably in Italy - retain substantial numbers of pensioners in membership, it is difficult to integrate them in the life of the union where workplace-based structures predominate.) To appeal to younger workers - in most countries seriously underrepresented in union membership - unions will almost certainly have to develop alternative, locally-based structures. Moving away from the bureaucratic formalities of traditional meetings to alternative, more participative types of collective activity is also a necessary part of organisational innovation if unions are to appeal to a more diverse constituency with very different cultural backgrounds to those of the traditional trade unionist. One may perhaps note here the success of the British TUC in developing anti-racist campaigns in a style totally different from its traditional approach to organisation.
The logic of all these themes is the mobilisation of values and language in support of union objectives. To survive and thrive, unions have to reassert the rights of labour in ways which allow them to recapture the advantage in the battle of ideas. 'Organisational strength without ideology is form without content, ' said the great strategist of Swedish trade unionism Rudolf Meidner (quoted in Evatt Foundation, 1995); when so many union movements are suffering organisational weakness, motivating ideology is all the more essential.
Across the world, trade unionists and supportive analysts of trade unionism have developed similar arguments: that the material difficulties confronting unions are compounded by a 'loss of [their] ideological justification ' (Piore, 1994: 514). The task is to demonstrate that as well as influencing the material economy their mission is to establish a 'moral economy ' (Swenson, 1989). In the words of the general secretary of the European TUC, 'what we need are creative utopias that set new developments in motion ' (Gabaglio, 1995: 111). 'Unions need to reformulate their goals to ensure that their activities are more closely identified with values like freedom and fairness that are both widely-held and fundamental, ' concluded the (union-linked) Evatt Foundation in Australia (1995: 128). The key challenge for South African trade unionism, concludes COSATU (1997: 43), is to offer 'moral leadership '. For American unions to recover their fortunes, insists Rogers (1995: 368), they must win acceptance as 'carriers of the «general interest» '.
'Solidarity forever ' is one of the most fundamental trade union slogans. Solidarity has a double meaning: support by union members ' for each others ' struggles, but also support by the stronger for the weaker within society (or indeed between nations). The broader, moral underpinnings of collective action have in many countries become eroded; if solidarity is to survive, it must be re-invented. The diversity of work and labour market situations in the contemporary world means that a traditional, standardised trade union agenda can be neither practically effective nor ideologically resonant. The task is to move from an old model of mechanical solidarity to an new model of organic solidarity - or as Heckscher (1988: 177) puts it, 'a kind of unionism that replaces organizational conformity with coordinated diversity '.
Any project aiming to create such a model must recognise and respect differentiations of circumstances and interests: within the constituencies of individual trade unions, between unions within national labour movements, between workers in different countries. The alignment and integration of diverse interests is a complex and difficult task which requires continuous processes of negotiation; real solidarity cannot be imposed by administrative fiat, or even by majority vote. Its achievement is possible to the extent that unions rediscover the conviction, and persuade both their own members and members of civil society more generally, that they have a mission as a 'sword of justice '.