Rethinking Worlds of Labour
Southern African Labour History in International Context

South African historians and social scientists have often bemoaned ‘South African exceptionalism’: in other words a tendency to see the country’s historical trajectory as absolutely unique. Yet they have also been strangely reluctant to place their findings in a more global context. The articles which comprise this edition were papers given at a University of the Witwatersrand (Wits) History Workshop and Sociology of Work Unit international conference entitled ‘Rethinking Worlds of Labour: southern African labour history in international context’ held from 28 to 31 July 2006.

The conference provided an opportunity to move away from South African exceptionalism in practice, by considering comparisons and connections between the history of labour in South Africa and in other parts of the world. The title also reflected the conviction of the conference organisers that such a shift away from parochialism would contribute to a ‘rethinking’ of some of the fundamental assumptions of labour history in southern Africa, and contributes to a revivification of the field. Furthermore, we meant ‘worlds’ in a dual sense – signalling not just the physical spaces through which people move, but also social worlds, and our special interest in the subjective ways in which the world is understood by workers.

Globalisation and labour history

There has, during the first decade of the twenty-first century, been a clear rise in historians’ interest in working at a more international level (Hopkins 2002; Bayly 2004). This is certainly rooted in the sense that globalisation – however that is understood – is making an enormous impact on our daily lives. This causes a reap-praisal of many certainties, economic, social and political, and gives rise to a historical curiosity about the antecedents of globalisation.

Indeed, it may be argued that historians have a particularly valuable contribution to make to globalisation debates. Very often we are told that features of ‘globalisation’ are absolutely new, or unique to the present. But social scientists sometimes do this without any very careful attention to the past, which they are considering. Closer enquiry may in fact show that some features of globalisation have clear precedents. In this perspective we are only now re-emerging into something like the globalised world of before 1914.

John Gray (1998) has pointed out that it is not helpful to conflate, as commonly happens, the international turn toward free market policies in the 1980s with ‘globalisation’ understood as the history of intensifying transnational connections as a whole. Globalisation has proceeded at many levels – political, social, and cultural – besides the economic. It has a long history, and is likely to survive the demise of recently influential economic ideologies. Indeed Bayly (2004) has convincingly advanced the notion of ‘archaic globalisation’, linking empires and societies previous to modern capitalism. In her path-breaking study of the world of the twelfth to the fourteenth centuries, Janet Abu Lughod (1989) likewise made a powerful case for the existence of a China-centred world economy before the rise of European colonisation. Indeed, many historians of the late nineteenth and early twentieth century would argue that there was in that era a ‘first globalisation’, followed by a strengthening of nation-based structures after the First World War. So it would seem that claims about what is, and is not, original about current globalisation would benefit from a much stronger base-line of historical comparison.

In this context, labour history is a field with particular claims for attention informed by a more global outlook. The glory days of the discipline internationally were in the 1960s to the early 1980s, and in that era there is no doubt that the interest in labour’s past was driven by the extraordinary waves of industrial militancy in the United Kingdom, France, Italy, Poland, Brazil, Argentina, India and many other countries during those years. Yet it is equally the case that subsequent developments caused a salutary re-examination of some of the notions that informed labour history. The defeats suffered by labour movements, the decline of the size of the industrial workforce in many major economies, the emergence of new forms of global flows of capital, and new patterns of production and consump-tion, all put question marks over any triumphalism about the cause of labour. Critics, many of them informed by post-structuralist theory, with some justifica-tion raised questions over labour historians’ neglect of the analysis of discourse and language (Steadman Jones 1983), its failure to engage adequately with feminist theory (Scott 1988), and its teleological politics (Joyce 1994).

Yet though there was much that was valid in these critiques, and although the world of the 1990s was inhospitable for labour history, the subject of the working classes and their histories remains an inescapable one for any serious study of the modern world. There are also strong signs of a practical and theoretical revival of labour history that speaks to questions thrown up by globalisation. What has been striking over the last few years has been a revival of labour history, particularly within the semi-industrial countries. At the same time, labour history has become increasingly attuned to the global dimensions of working-class formation. As Marcel van der Linden notes in his contribution to this collection, labour history is today not only more globalised in its practice, but it is also more global in its outlook.

One of the major limitations of classical labour history was that it was largely confined within the boundaries of national histories. The greatest of all the works of labour history in its golden era was after all, EP Thompson’s (1991) book on The Making of the English Working Class. The Scots, Welsh and Irish only got walk-on parts in Thompson’s great drama, while peoples further a field were almost entirely ignored, notwithstanding the larger British imperial context. In general, labour historians have followed this approach, writing about the German, Australian, South African, Brazilian, Nigerian etc. working classes.

Now, obviously the formation of nation-states was one of the major features of the nineteenth, and more especially the twentieth centuries, and working classes have often orientated politically towards such ‘national’ frameworks. However, nationally based labour studies face several related problems. Taking the nation-state as the self-evident unit of analysis tends to naturalise what must be seen as a fairly novel (and for much of the modern period, unusual) state form, and the related assumption that labour must develop a national character. Relativizing and historicizing the nation-state can reveal much about the history of labour, help avoid teleological assumptions about the historical trajectory of labour movements, and undermine the sense of national uniqueness that produces a sense of ‘exceptionalism’. Nationally based labour histories have also tended to homogenise local variations within nation-states, inadvertently playing down regional specificity. Moreover, they have tended to ignore the point that many of the most important processes within the world of labour occurred across national boundaries. International flows of migrant workers, capital, political agitators, publications, cultures and public spheres are crucial to the histories of working classes in the modern world.

Much can be gained by escaping from such confines. For all Thompson’s great achievement, consider how much fuller is the picture of the late eighteenth century English working class that is presented in Peter Linebaugh and Marcus Rediker’s The Many-Headed Hydra (2000). Their argument for the existence of an ‘Atlantic working class’ of sailors, labourers, slaves, freedmen and renegades linking Britain, West Africa, the Caribbean and North America has crucially expanded labour historians’ understanding of the seventeenth and eighteenth century world. It has drawn attention to the importance of understanding connections between continents and across different labouring groups, and the need to rediscover and rethink popular imaginations.

The theoretical groundwork for such approaches was laid by important interro-gations of the nation-state in the 1980s. By showing that the nation-state was a relatively recent phenomenon (Gellner 1983) – based on ‘imagined communities’ constituted through print capitalism and other instrumentalities (Anderson 1991) and legitimised through ‘invented traditions’ and the ideological work of military service and schooling (Hobsbawm 1990) – social scientists and historians radically destabilised and denaturalised the nation state. In turn, these interventions made it much more possible for scholars to recognise the somewhat fictive character of the claims of the state, more generally, and the possibility that states could fail to make good these claims. For labour studies, this pointed to examining the relationship between changing state form and working-class movements, and questioning the view that empires and other types of state forms could be regarded as simply the prehistory of the nation-state.

In a way, it was odd that labour historians had become so hypnotised by the nation-state, given that Marxism was so crucial to the intellectual formation of the discipline. Marxism was, in intent at least, an internationalist project, and Marx and Engel’s paean to the destructive and constructive powers of capitalism celebrated how commodities battered down ‘Chinese Walls’, denied that the proletariat had a ‘fatherland’, and, of course, famously proclaimed: ‘Workers of All Countries – Unite!’ The emergence of a historiography organised in largely national terms can be partly explained by the pragmatic reality of world politics, and by Lenin’s systematic re-orientation of Marxism towards strategic alliances with nationalist movements in the ‘colonial and semi-colonial world’. Aside from the revision of Marx’s arguments this entailed (Warren 1982), and the difficulty of reconciling class analysis with class alliances that must continually arise, the practical success of Lenin’s approach had the effect of making the ‘national’ a central category within Marxist thought and politics.[1]

Nevertheless, there are notable works by Marxist historians that transcend approaches rooted in methodological nationalism. Perhaps most outstanding is Eric Hobsbawm’s (1977a, 1977b, 1987) great trilogy on the world of the ‘long’ nineteenth century. Hobsbawm’s extraordinary portrait of the rise of a globally connected world is exemplary in the way in which it goes beyond national history. Its sensitivity to the cultural level of analysis, and its deep engagement with Latin American and Asian experiences in many ways anticipated the work of contemporary transnational historians. (It must be said though that sub-Saharan Africa remained something of a blind spot for the great historian.)

Thinking globally

It is our view that labour history can benefit greatly from the application of a more transnational approach. What would be different about the approach that we are suggesting? Perhaps it would be best to clear the ground by saying first of all what we do not envisage.

Firstly, we do not want to adopt the slogan of ‘World History’ (Pomper, Elphick and Vann 1998), which since at least the 1960s has been a fairly mainstream branch of historiography (setting aside somewhat eccentric predecessors like Oswald Spengler (1926, 1928) and Arnold Toynbee (1960)). This trend did, as we do, seek to overcome parochialism, and it did produce some remarkable works of scholarship such as William H McNeill’s outstanding books on the global history of epidemic disease (1977) and of warfare (1983). However, even the most outstanding practitioners of World History, including McNeill, have tended to approach their task through the lens of analysis of Civilizations, which are usually defined through some form of cultural attributes.

The difficulty here is that, even in the hands of an able historian like McNeill, these world cultures seem to be monolithic, static, mutually exclusive and essentialised. It is striking that even in the work of an historian as great and innovative as Fernand Braudel (1982), the master’s commitment to a notion of culturally intact civilization drove him eventually toward a distinctively protective posture towards French identity. The reductio ad absurdum of this approach is found in Samuel Huntington’s (1996) belief in an inevitable ‘Clash of Civilizations’. Moreover, when practised by less erudite and skilled historians than Braudel or McNeill, the project of a comprehensive history of the world can become overambitious, even farcical. Few historians can write with much plausibility about developments over a single century, let alone all human history. So, moving history outside national frameworks does not mean that one should make a hubristic attempt at comprehensiveness.

Secondly, we are specifically not advocating what has been the most influential framework for global history in recent decades: the World Systems Theory (WST) of Immanuel Wallerstein (1974). Wallerstein postulated that the expansion of Europe at the turn of the sixteenth century saw the creation of a single capitalist world system, one of a series of ‘world systems’: the modern world system was understood as a system of states, with a ‘core’ in the West, a ‘periphery’ subjected to the ‘core’ by imperialism, and a ‘semi-periphery’ of intermediate states that acted as ‘agents’ of the ‘core’ while striving for ‘core’ status. Economic ‘sur-pluses’ are ‘drained’ to the ‘core’, enriching it at the expense of the other regions, which become ‘underdeveloped’. States can with difficulty change their position within the world system, but the system persists.

The objections to such a schema are so obvious that it is hard to understand the power that it has exercised over the minds of scholars. Its evident attraction is its very simplicity, as a universal explanation; the same simplicity is its weakness, too, for it posits a closed social analysis, conceived within a functionalist approach, and tends to operate through the static logic of system theory. It is difficult to see any room for resistance, for the role of ideas, or for ruptures in the structure (Adas 1998), while the meaning of the core idea, ‘underdevelopment’, is vague, shifting and very often tautological (Warren 1980). By displacing class exploitation within countries by international exploitation between countries, the framework displaces class, and perhaps more importantly, the role of class struggles, from its analysis. The idea of nations remains relatively taken for granted and unexamined, and mapped onto the different regions.

Many WST practitioners purport to be Marxist, yet the model of the world system is rooted not in a Marxist analysis of production, but rather in flows of trade (Brenner 1977; Laclau 1982), with the argument for ‘exploitation’ between countries rooted, in the final analysis, in the liberal theory of ‘exploitation’ as mon-opoly pricing (Leys 1996). WST may be right, but it cannot be both right and Marxist, and the result is a radical theoretical incoherence. WST can only argue with great difficulty that the Spanish and Portuguese adventurers, who initiated their modern world economy, were in any meaningful sense modern capitalists.

With its vision of single systemic logic, WST is ill-equipped to analyse non-Western empires, and unable to explain the rise of newly industrialised countries in late nineteenth, and again in the late twentieth century, world.

Comparative studies of different countries have, on the other hand, a long and hon-ourable tradition in the social sciences. By seeking to place two cases alongside one another, not only are we immediately led to question our assumptions and to look at what is similar and different in distinct historical contexts, but also our conceptual and empirical horizons are rapidly expanded.

South African historians have, for example, almost universally accepted that ‘super-exploited’ large-scale cyclical labour migrancy was a definitive feature of South African capitalism, and attributed this situation in large part to segregation and apartheid. However, as Philip Bonner (2004) has shown in a study of Indian and South African urbanisation, there are remarkable similarities in patterns of labour migrancy in the two countries, despite the general absence of any significant state interventions in the colonial Indian labour market.

This is one of several themes that Sumit Sakar’s article in this collection develops through a comparative discussion. Sakar notes, for example, that the interventions of the South African state in the fields of labour markets and social policy were far more extensive and ambitious than those undertaken in the British Raj: there was, for example, simply no equivalent in colonial India to South African-style town-ship construction, social segregation and labour coercion. He cautions, conse-quently, against the tendency of some post-colonial theory to homogenise the colonial experience, and to downplay the importance of pre-colonial legacies. In India, unlike southern Africa, pre-colonial social stratification was extensive, and it was this that allowed the recruitment of a large labour force without direct interventions like land restrictions. Peter Alexander’s contribution to this collection, which compares collieries in South Africa and India, makes the key point that female miners were almost unknown in South Africa as compared to India, and adds that daily pay for African miners was half of Indian miners. This suggests that the ‘concept of “cheap labour” ... involves a comparison with white South African labour, is parochial, and ... should now be discarded’ (Alexander 2006:7).

Comparative approaches, in other words, help create the basis for a re-examination of some of the conventional wisdom in the field. Perhaps some of the resistance of South African social scientists and historians to comparative work is based, though, on the important misunderstanding that comparing two situations entails making the case that they are somehow the same. This is absolutely not the case, for many of the most important comparative analyses are those that study different historical paths. A notable example is Barrington Moore’s (1987) Social Origins of Dictatorship and Democracy, which explains the different routes that agrarian societies took to modernity, and their long-term political and social consequences. Similarly, Perry Anderson’s (1974) Lineages of the Absolutist State sought to explain the very different socioeconomic trajectories of eastern and western Europe.

What constitutes a valid comparison? Some time ago, Mahmood Mamdani (1998) led a rather moralistic campaign about the need to place studies of South Africa in an African context. Whether one makes African or non-African comparisons should depend on the usefulness of the comparison to what one is studying, and the way it can illuminate particular issues. Thus, Jeremy Seekings’ comparison of the South African welfare system with those of Latin America, which appears in this collection, works exceptionally well because there is a sufficient degree of similarity and difference in the cases to make them illuminating.

Allison Drew’s comparison of the agrarian engagement of Algerian and South African communists in this edition works so well because of the intellectual framing she gives it, rather than simply because the two cases examined are drawn from Africa. Similarly, Gay Seidman’s (1994) comparison of the South African and Brazilian labour movements was productive given the comparable economic and social contexts. It would be difficult to compare the modern South African labour movement with that of a country without major industries. On the other hand, it is clear that in many areas – for example popular culture, traditional authority and political democratisation – comparisons between South Africa and other African countries are extremely illuminating. There is, in short, no moral obligation on researchers to accept certain forms of comparison, and refuse others.

One way in which the growing scepticism about national histories can be extremely helpful is in developing international comparisons that take regional variation into account, rather than compare countries as a whole. Thus Peter Alexander’s comparison of collieries has a keen sense of the social specificity of the region in which his Indian case is located, and of the distinctions between the Transvaal and Natal coal industries in South Africa. Similarly, while comparisons been South African and the United States as a whole can become rather unwieldy, a focused comparison of processes in particular regions can be very helpful: the career of segregation in the southern US and South Africa have, for instance, been usefully compared by Greenberg (1980), Cell (1982) and others.

Transnational labour history

We can now turn to transnational labour history. Let us offer a modest definition: transnational labour history does not assume that the nation-state is the necessary framework for historical analysis. It is interested in perspectives that move beyond the level of the ‘nation’ to look at flows of people, commodities, ideas and organisations across national boundaries. It also considers the possibility that regions or cities within nation-states may have closer links with regions or cities lodged in other nation-states than with their own hinterlands. It does not seek to be comprehensive: rather it simply does not accept that its field of enquiry should stop at the ‘national’ border, or that a ‘national’ unit is a self-evident, or necessarily a particularly useful unit of analysis. It argues for approaches that examine connections across countries, continents and cultures, for comparative studies, for transnational perspectives, and for rethinking the conceptual vocabulary of labour and working-class history.

To say this is clearly not to pose a transnational perspective as the theoretical panacea for all historiographical problems. Nor does it suggest that the ‘national’ is not a useful level of analysis, or deny that the nation-state and nationalism have been central forces in the modern period, or will remain powerful forces in any conceivable medium-term future. At the very least, a transnational view asks the scholar to hesitate before starting the analysis with the assumption that the nation-state is the relevant unit of analysis. And even the study of nationalism itself can benefit from this, for one of the features of current historiography is its revelation of the way in which nationalisms form across national boundaries. For example, in his extraordinary book, Americana, James Dunkerley (2000) brilliantly illuminates the emergence of US and Latin American political identities in the mid-nineteenth century by treating both the North and South American continents as a single political arena, and by relating them in turn to political and cultural developments in Ireland, and elsewhere in Europe.

What methodological benefits might this sort of perspective bring to labour history? Firstly, it refocuses attention on the phenomenon of global migration. Of course, the world of the nineteenth and twentieth centuries has been continually re-shaped by mass migrations, of which working-class people were a central component. Now, a perspective that emphasises national labour histories can lead to the idea that migration simply involves a flow of workers from country A to country B, where they assimilate and form a component of the ‘national’ working class. The reality, however, is more complex, as migrants often cling tenaciously to political identities from their place of origin, and infuse these into movements in the host country. Not only, for example, did radicals in the late nineteenth and early twentieth century US draw in large numbers of immigrants, but they also communicated with them through a polyglot press. Thus, the first anarchist daily newspaper in the world seems to have been the Chicagoer Arbeiter-Zeitung, a German-language paper published in Chicago in the 1880s, which was then called the second largest German city in the world (Bekken 1995).

Moreover, migration is often oscillating or lateral, rather than simply a move from country A to country B. Migrants dream of returning to their home country and often do so, and others move between several countries in the course of their migration. Thus, for example, at the beginning of the Witwatersrand mining industry, not only did many Cornish miners return to their families in Cornwall after several years on the Rand, but there was also a constant flow of Cornish miners between South Australia, West Australia, southern Africa, and western America: more adventurous Cornishmen could be found down mines from Malaya to Bolivia. Migration is a process, without a necessary ‘national’ end point.

One phenomenon that is now starting to receive more adequate attention, partly as a result of a more transnational outlook, is nineteenth and twentieth century Chinese and Indian indentured labour. The institution of this system in the back-wash of the British Empire’s abolition of slavery had significant effects on South Africa and in many other regions of the world. Whether one agrees with Hugh Tinker’s (1974) famous contention that this was A New System of Slavery, or whether one accepts the more optimistic view that indenture played a significant role in enabling labourers to accumulate capital and begin breaking out of a semi-feudal village life (Northrup 1995), it clearly entailed vast mortality, suffering and social disruption. It is not generally recognised that the number of ‘coolies’ shipped around the world was comparable to the numbers transported in the African slave trade. Indenture deserves a much more central place in labour history, and including indenture starts to raise significant questions about how ‘labour’ and the ‘working class’ are defined, and to what extent ‘free’ labour is characteristic of industrial modernity.

Secondly, a transnational perspective leads to a reassessment of labour’s political movements. The present authors have sought to make a contribution to this project in their other work. Lucien van der Walt (1999, 2004), for example, has shown that the early twentieth century South African labour movement’s ideologies and actions cannot be understood without due attention to the global impact of anarchist and syndicalist ideologies and movements, often brought to South Africa by migrants and spread through an international press. This has been almost entirely ignored by South African labour historians. By placing South African developments in a global context, and examining the importance of transnational connections and influences, Jonathan Hyslop (1999), too, has mounted a case that the trade unions of British immigrants in the same era are best understood as part of an ‘imperial working class’ which straddled the British Empire.

This reassessment is especially necessary in relation to the history of Communism. Writers sympathetic to communist parties have emphasised the rootedness of their ideologies and activities in national struggles and conditions (for example, Isserman 1982), while their critics have stressed the heavy hand of Moscow through the Communist International (Comintern), the Cominform and funding (for example Klehr, Haynes and Firsov 1995). Both approaches are narrow and one-sided: the weight of evidence of tight connections between communist parties and the Soviet Union is overwhelming, and communist ideology stressed the importance of these links; on the other hand, the parties only became significant where they were able to make genuine connections with national and local social grievances, cultural traditions and political struggles. It is useful, then, to understand Communism from a transnational perspective that recognises its parties were simultaneously shaped by both their relation to the Soviet Union and their national contexts. The work of Geoff Eley (2002) is distinguished amongst historians of the left for striking this sort of equilibrium on this issue.

Thirdly, transnational labour history opens up exciting and illuminating possibilities in micro-history and biographical research. A transnational perspective poses key methodological issues, and it is striking that some of the greatest insights into global processes can be gleaned from a study of individual lives. In particular, following an individual travelling labour activist as he or she moves around the world illuminates complex global networks and flows of ideas. Karen Hunt’s article in this edition provides an excellent example of what can be accomplished here. By looking at Dora Montefiore, a British socialist and feminist who travelled the world of the imperial working class, Hunt shows how ideas and movements can be (re)shaped by experiences in different, yet interconnected, contexts. A key work is Benedict Anderson’s (2005) study of the anarchist-influenced Filipino revolutionary intellectuals of the late nineteenth century. Anderson brilliantly shows the extraordinary personal linkages in the 1890s between the left in Europe, the rebellion against the Spanish in Cuba, and the wars of the Philippine rebels against both Spanish colonists and American liberators. (He also notes how the Anglo-Boer War, a key moment in South Africa, became a key symbol of anti-imperialist resistance worldwide at the time.)

Fourthly, following from the previous point, a transnational approach highlights the point that not nation-states, but empires, have been the typical state form over the past centuries (Stoler and Cooper 1997). Until the First World War, the empires of the British, French and Dutch (and their feebler Austro-Hungarian, Portuguese, Russian and Ottoman rivals) bestrode the world, and it was only after the Second World War that formal empires (like the Soviet Union) became rarities, rather than the norm. For much of the nineteenth and twentieth centuries, labour must be understood in imperial, not national, contexts. Passports were rarely used before the First World War (Torpey 2000), one indication of the relative unimportance of national types of state: indeed, it was considerably easier for workers to move around the world before 1914 than it is today. It is thoroughly anachronistic for labour historians to project current national structures back in time.

Fifthly, oceanic history must be an important component of contemporary history beyond national boundaries. Braudel’s (1972–3) great work on the Mediterranean is an important starting point, showing how maritime space can provide the arena for a dense social and economic overlapping of political entities. This insight has already been applied to the Atlantic Ocean with considerable effect (Linebaugh and Rediker 2000), and is now being mobilised in the labour history of the Indian Ocean (for example, Metcalf 2007). In turn, the more serious interest in oceanic contexts has been associated with a lively historiography that seeks to understand the world of the ship as a world of work (Dening 1992; Ewald 2000). The ship has been an important site of social life, and especially labour action, in the making of our world, as well as an important carrier of ideas and movements.

Sixthly, and perhaps paradoxically, a transnational perspective leads us back to a focus on the city. The metropolis is often more connected to metropolises in other cities than to its own hinterland. As Ferguson (2006) memorably puts it, in Africa capital (and labour!) does not so much ‘flow’ as ‘hop’. This means, at one level, that due attention needs to be paid to variation within countries; at another, it means taking cities seriously as cosmopolitan sites, as nodes in transnational networks, and as sites of state power and class formation. It is worth asking whether we are not perhaps coming to inhabit, in some regions at least, a world of weak states and strong cities, rather like late-medieval Europe.

Merely concentrating masses of people into shared workplaces and neighbourhoods in large cities does not, however, necessarily imply class unity. Cosmopolitan contexts can as easily accentuate differences as limit their significance: it is striking, for example, that it was in South Africa, and not India, that the expatriate Mohandas Gandhi came to see himself as first, and foremost, an Indian (Markovits 2004:81). In understanding these dynamics, it is important to consider the complicated role of cosmopolitan centres as forcing houses of ideas, as nodes in networks, and as sites of both competition and cooperation in the popular classes.

Finally, a transnational perspective has an important role to play in the very necessary task of rescuing labour history from what has undoubtedly been a very strong tendency to economistic forms of analysis. Although EP Thompson was extremely sensitive to the impact of literature and religion on the working class, and although Herbert Gutman (1976) made a powerful case for the centrality of culture to labour history, their successors have not always taken these points sufficiently on board. While the Wits History Workshop has had a strong commitment to issues of popular culture, we have not been sufficiently sensitive to issues like literacy, and its social and political impact. And some labour studies have been balefully economistic, treating workers as lacking any interest in such issues as ethnic identity, religion, sexuality, chiefly politics, sports, language or reading.

The work of Karl Polanyi (1991) is enjoying something of a vogue in labour studies, in part because of his rejection of liberal economics. Yet Polanyi’s larger point is that society is never purely structured by economic relations: interactions need to be understood in radically social terms, and not reduced to the ‘economy’ or ‘politics’. With the ‘great acceleration of communications and transport in the nineteenth century’ (Bayly 2004) and the contemporary ‘compression of time and space’ (Harvey 1991), it is important to recognise that the flow of ideas cannot conceivably be understood in terms of the cultural production of a single country, or simply as the result of an autonomous and pure ‘national’ process.

To understand the social worlds of labour in a given place, we need to study popular culture, but to situate this within a cultural arena formed by ideas flowing across international boundaries, in relation to the manner in which different medias circulate them, and, again, in relation to the ways in which people re-interpret them in specific contexts. In the contemporary situation we as historians need to start thinking more systematically about the way in which the Internet is changing worlds of labour. The sociologist of religion Olivier Roy (2006) has, for instance, recently argued that the Internet is the key site where new militant Islamic ideologies are formed. And religious formations are of course eminently global with their claims to universal community, in the reach and technological sophistication of their propaganda: sometimes harnessed to nationalism, religious aspirations can also subvert the nation-state project with claims to a global community and project.

South African labour, or labour in South(ern) Africa?

A transnational perspective can make an important contribution to the labour history of southern Africa, where scholarship on labour history is unevenly developed in the region, concentrated in South Africa, and generally been placed within the framework of the nation-state. Labour history in South Africa has derived from two main traditions: activist and scholarly work. Activist writing on labour, largely produced outside of academia, goes back to the 1920s. The earliest work came from white labour (Gitsham and Trembath 1926; Walker and Weinbren 1961), followed by writings by Communist Party of South Africa (CPSA) members in the 1940s (Andrews 1941; Cope c.1943; Harrison n.d.; Roux [1944] 1993), Trotskyist analyses in the 1950s (notably Majeke 1952; Mnguni 1952; Tabata 1950; for a partial overview, see Nasson 1990), and a wave of works by writers associated with the CPSA’s successor, the South African Communist Party (SACP) from the 1950s onwards (for example, Bunting 1975; Forman [1959] 1992; Lerumo 1971; Simons and Simons 1969 [1983]; for a partial overview, see Drew 1997).[2]

Leaving aside a few liberal analyses of labour that had some historical content (Horrell 1969), the scholarly tradition of labour history emerged in the 1970s in British and South African universities, and was influenced by Marxism and class analysis. This ‘revisionist’ literature challenged older liberal approaches that emphasised the negative effects of apartheid and segregation on South African capitalism: in its starkest formulations, the revisionists portrayed the racial order as nothing but a function of capitalist imperatives, with capitalism supposedly unable to function without apartheid (Johnstone 1970; Legassick 1974; Wolpe 1972). The impact of structuralist approaches in the 1970s was also evident in the use of Nicos Poulantzas’ analysis of state policy as shaped by ‘frac-tions’ of capital (Davies, Kaplan, Morris and O’Meara 1976), the use of WST (Bundy 1979; Legassick 1977), and also a tendency to read labour history off labour processes (Lewis 1984).

In large part as a response to the structuralists’ failure to examine popular agency and consciousness (Bonner 1994), and in contrast to the ‘old labour history’ focus of much of the activist literature, the late 1970s saw the blossoming of a local social history in the Thompsonian mode, which stressed experience and culture (the key works would include Bonner, Hofmeyr and James 1989; Bonner, Delius and Posel 1993; Bozzoli 1979, 1983, 1987; Bozzoli and Delius 1990; Beinart, Delius and Trapido 1986; Marks and Rathbone 1982; Marks and Trapido 1987; Van Onselen 1982a, 1982b; for overviews, see Bonner 1994; Bozzoli and Delius 1990; Saunders 1988). The Wits History Workshop, formed in 1977 and focused on the Witwatersrand, was the main organised expression of this shift, but only one of several social history initiatives at the time. The new labour history developed as part of this social history project. In contrast to the functionalism and reductionism of the structuralists, the social historians stressed contingency, contradictions, ruptures and the reconstruction of ‘history from below’.

What both the activist and revisionist traditions, structuralists and social historians alike, shared was a tendency to write the history of labour in South Africa as a specifically ‘South African’ labour history. Of course, both traditions were well aware of the importance of international processes and connections in shaping a ‘South African’ society, and routinely made implicit or explicit comparisons between South Africa and other countries, generally with the emphasis on the ‘exceptional’ character of South Africa. Attention to the global was inevitable, given that the industrial revolution on the Witwatersrand in the late nineteenth century was spurred on the one hand by foreign direct investment in the ‘first globalisation’ lasting roughly from the 1880s to the 1920s; on the other, it developed in tandem with the expansion of British and Portuguese imperial power in the region.

It was also recognised, to some degree that the working class that emerged in South Africa was a multinational and multiracial one, drawn from across southern Africa, the British Empire and beyond. South African capitalism was embedded in a regional political economy, and in regional, as well as transcontinental, labour markets. In addition, there were several fruitful applications of revisionist perspectives elsewhere in southern Africa, most notably Mozambique (for example, Harries 1994; Penvenne 1984, 1995) and Zimbabwe (formerly Southern Rhodesia, notably Palmer and Parsons 1977; Phimister 1988; Phimister and Van Onselen 1979; Ranger 1970; Van Onselen 1976).

Nonetheless, activist and revisionist scholars tended to take South Africa as the unit of analysis, and to examine labour history as South African labour history. Cross-border connections were examined largely from the perspective of their contribution to South African history; the regional labour markets were examined largely in terms of their importance to South African labour employers. The white unionists noted international influences on the emergence of union traditions, but treated this as a passing phase before the mature period of a specifically ‘South African movement’ acting on a South African stage (Gitsham and Trembath 1926:11). The CPSA writers agreed, while many Trotskyists adopted an overtly nationalist narrative, with ‘300 years’ of oppression (Mnguni 1952) leading to the ‘awakening of a people’ (Tabata 1950).

SACP writers likewise framed matters in a national framework, adding a large dash of teleology: the two ‘streams’ of class and national movements ‘merged’, apparently inevitably, in the 1950s when the CPSA/SACP allied with the nationalist African National Congress (ANC) (Bunting 1975; Forman [1959] 1992; Simons and Simons [1969] 1983). CPSA and SACP writers were, of course, well aware that the rise of Communism was closely linked to the rise of the Soviet Union, and shaped by that state. However, they stressed the ‘national’ character of the party and its rootedness in the struggles of ‘our people’. There were substantial overlaps between SACP and nationalist ANC historiography (for a sophisticated example, sees Meli 1988; for an overview, see Lodge 1990). For the SACP writers, not only was the CPSA’s 1928 adoption of the Native Republic thesis – stressing the immediate task as a struggle against feudalism and imperialism, for the creation of a non-racial bourgeois society, rather than socialism – not imposed by Moscow, but was supposedly actually largely ‘initiated’ by CPSA members (Simons and Simons [1969] 1983:405; Bunting 1993). This is a typical example of the trend of pro-communist writers to stress the national character of parties.

The same limitations were clear in the revisionist historiography. The social historians were accused by the structuralists of failing to move beyond culturalist and local studies to examine the larger political economy (Morris 1987; Murray 1989). However, with a few exceptions (notably Legassick 1977) the structuralists took the larger political economy as a ‘national’ and South African formation. To the extent that there was an attempt by the structuralists to discuss southern Africa as a unit, the emphasis was on South Africa’s dominant role. WST ideas of ‘unequal exchange’ played some role in these approaches, with the corollary that the region was analysed in terms of competing states, rather than viewed from the vantage point of empire, or examined as a unit with dynamics that were not simply the sum of (national) parts.

The structuralist charge that the social historians eschewed theory was not very well founded. The Wits History Workshop project, at least, was explicitly concerned with examining the significance of popular struggles for the system of ‘racial capitalism’, and of using local cases to inform larger models (see, for instance, Bonner, Delius and Posel 1993; Bozzoli 1979, 1983, 1987; Marks and Trapido 1987).

Nonetheless, the generalisations developed by the social historians were them-selves typically posed at the level of South Africa, rather than, for example, southern Africa. If, however, the popular classes sprawled across the borders of South Africa, and if their experiences, ideas and struggles were not confined by borders, then it is not clear why generalisations from social history should have been made largely at the ‘national’ level. Given that the popular classes in South Africa were not necessarily South African, and that South Africa was part of a regional political economy and enmeshed within a web of major transnational linkages, it is striking that a general ‘history from below’ of the region was not developed. While South Africa was compared to other countries or regions within countries, as noted above, it is striking that there were almost no comparative analyses of labour within southern Africa (for an exception, see Phimister 1977), or a social history synthesis that grappled with the fact of a southern African, rather South African, working class.

This was the state of play by the mid-1990s, when labour history in South Africa went into a sharp decline. Besides the international factors that affected labour history worldwide at this time, there were also local factors that came into play: the end of apartheid removed much of the oppositional political energy that fed into revisionist writing, and the lack of direction coincided with a series of critical onslaughts on revisionist approaches for failing to seriously engage with race and its meanings (Posel, Hyslop and Nieftagodien 2001), for forcing social history into a teleological history of anti-apartheid resistance (Minkley and Rassool 1998), and for remaining a largely white intellectual project (Bonner 1994; Bozzoli and Delius 1990; Worger 1991).

Finally, the post-apartheid state’s project of creating a new, official, national (and nationalist) history limited the space for revisionist history. On the one hand many of the themes of revisionist history have been incorporated into this new history; on the other, the record of ‘history from below’ has often been forced into a monolithic narrative of a single struggle (‘the struggle’), supposedly led throughout by the ANC (for examples of this genre, see Magubane 2004, 2006). As Martin Legassick documented in an important paper at the 2006 conference, which will be published elsewhere, this has involved heavy-handed official control of work by independent scholars that has been commissioned for the new history.

Southern Africa, Latin America and North Africa

The fortunes of labour history have changed in recent years, with a growing output of work dealing with both the pre-industrial and industrial periods in South Africa, as well as in the larger region. Some of this work has challenged the CPSA and SACP versions of the history of the left through the recovery of alternative left traditions and an examination of the social history of local Communism (Drew 2002; Hirson and Hirson 2005; Hirson with Williams 1995; Van der Walt 1999, 2004). Comparative analysis, which played a role in the older labour history (in addition to earlier citations, see Cooper 1991; Trapido 1971) has been revitalised, with more attention to other parts of the British Empire, Africa and Latin America (see Alexander and Halpern 2000, 2004; Bonner 2004; Greenstein 1998; Mamdani 1996; Marx 1998). Labour and social history have revived in other parts of southern Africa, often in response to the resurgence of labour movements in the 1990s, notably for Zambia (formerly Northern Rhodesia, Larmer 2007) and Zimbabwe (see, inter alia, Raftopolous and Phimister 1997; Raftopolous and Yoshikuni 1999). If the increasing isolation of South Africa from the 1940s played an important role in the somewhat parochial outlooks, it may be that the current globalisation has played a role in the widening horizons of current labour history.

The implications of applying a transnational perspective to labour history in South Africa and southern Africa are considerable, and in the remaining section we will indicate several areas where such a perspective may be fruitfully applied. One area is that of labour markets, which we touched on above. The racial wage gap on South African mines is well known, and it has also been noted that as early as the 1890s wages for skilled miners in what became South Africa were generally double (and sometimes five times higher) than the wages of comparable categories in mining areas elsewhere (Katz 1994:67, 75–7).

This process has often been explained in largely South African terms, as a response to the high cost of living, the bargaining power and aspirations of the whites, and employer strategies. However, a transnational perspective suggests this is too simplistic: unlike other white dominions in the British Empire, the South African state not only did not subsidise European immigration, but actively frustrated it, and white immigration was close to a net loss by the 1920s (Bradlow 1990:178 – 186, 192 – 193). The result was that employers in South Africa had to compete with other regions through unusually high wages.

The peculiarities of South African immigration policy are, at one level, to be explained by reference to Afrikaner-English divisions amongst whites, and the anti-immigration policies of Afrikaner nationalists. At another level, however, the imperial context must be taken as central, for the South African state, alone in colonial southern Africa, had dominion status.[3] This allowed state managers to defy imperial immigration policy, and to move towards important-substitution-industrialisation (ISI) policies in the 1920s. This was thirty years before most other African countries, but closely paralleled the policy shifts in contemporary Latin America to which some commentators (Cooper 1991; De Noon 1983; Seidman 1994) have noted. If, as Mamdani claimed, there were substantial parallels between British imperial systems of indirect rule and South African apartheid (Mamdani 1996), then, it would be a grave mistake to treat South Africa as simply a typical African colony; at the same time, the specificities of South Africa are nonetheless closely linked to its particular insertion within the imperial system.

The existence of a large white working class, including many ‘poor whites’, is also sometimes regarded as an important element of South African exceptionalism. Again, a transnational perspective raises questions about this assumption. There were substantial white working classes in Angola, Mozambique, Zambia and Zimbabwe; in Mozambique, whites were heavily concentrated in unskilled work (Capela 1981); and in Southern Rhodesia, a ‘poor White problem’ concerned officials in the 1920s (Morrel 1992).

Migration, Regional Struggles and the Movement of Ideas

As Van der Walt’s article in this collection shows, moreover, these white working classes were interconnected through migration (the opening up of mines to the north of South Africa was crucial), and that there was a spread of repertoires of struggle and organisational models throughout the region: the segregationist South African Labour Party, launched in 1910, was, for example, only the forerun-ner of a series of such parties in the region that were influenced by White Labourism. This was, in turn, profoundly influenced by the White Australia policy and the segregationist policies of the Australian Labor Party, ideas that were transmitted into South Africa by immigrants (Hyslop 1999). The 1922 Rand Revolt, so ably discussed in Jeremy Krikler’s recent study (2005), was, Van der Walt suggests, not only part of the international labour militancy of the late 1910s and early 1920s, but also the peak of a regional wave of black and white workers’ struggles across southern Africa that has not been previously recognised.

The growth of Chinese indentured labour on the mines in South Africa in the early twentieth century is another important dimension of these regional and international struggles over labour supplies and wage levels. Brought in by the British from 1903 to 1907, the 60,000 indentured workers were to break the post-Anglo-Boer War shortage of African labour that amounted to an informal strike. The ‘Chinese question’ was absolutely central to the rise of White Labourism in southern Africa in the twentieth century, which was influenced by Australia’s ban on Chinese and Polynesian labour in 1900.

Interestingly, as Kally Forrest’s contribution to this collection notes, an Australian connection plays an important role in the contemporary labour movement in South Africa. Her article, which draws attention to another fascinating example of the traffic of ideas and actors across borders, examines how the (predominantly African) National Union of Metalworkers of South Africa (Numsa) sought to reposition itself on the eve of the demise of apartheid in the 1990s. Frustrated with ongoing adversarial conflicts with employers, Numsa leaders sought to promote worker control of production, skills development and advancement through tripartite forums as a road to socialism. The main model that informed this approach was provided by Australian labour; Australian union personnel were drafted in to reposition the unions, in a fascinating parallel to the Australian connection of a century before.

Now, if the white working class in southern Africa had a large immigrant component, was influenced by ideas from abroad, and existed as a regional force, at what point can we start to speak of a South African white working class? The 1920s would seem to mark an important moment in the ‘nationalisation’ of white labour: not only did a national level class compromise get forged after 1922, but immigration fell sharply, white labour became increasingly stabilised in families, and the state began to move towards systematic social policy and mass education (on these developments, see Lange 2003:12, 79, 153 – 157). Internationally, it is also worth noting, the 1920s arguably marked the onset of a period in which working-class people and movements were increasingly nationalised elsewhere, through factors like mass schooling, national class compromises, and the increasingly closed ‘national’ economies that were characteristic of the period into the 1970s.

Seekings’ article is interesting in this regard, as it begins to examine the relationship between class struggles and social policy. Comparisons with Latin America are notably rare in South African studies, but as Seekings shows, can be most illuminating. Argentina had a fairly similar economic structure to South Africa, and a similar path to industrialisation. However, South Africa’s welfare system is rather unusual amongst semi-industrial countries, including those of Latin America, for it centres on tax-financed non-contributory grants, rather than social insurance schemes.

The potential for this divergence arose, in large part, from the character of the South African state created in 1910: it was a far more effective and bureaucratic state apparatus than its Argentinean counterpart, and able to raise public revenue more effectively. This, we might add, was the result of the imperial state engineering after 1902. However, Seekings stresses, it was the different character of labour struggles, the political landscape and the structure of the ruling group that was critical to the divergence between the two cases.

South African policy-makers, in addition, evinced a concern for managing ‘poor whites’ and maintaining racial order that was absent in Argentina. While the majority of the poor in Argentina were regarded as white, this was not seen as necessitating special interventions, and was not understood as a ‘poor white’ problem. To this we might add the point that Latin America indicates that large-scale white immigration need not translate into the development of a labour aristocracy: in late nineteenth and early twentieth centuries in Brazil, for instance, white immigrants undercut black wages (Andrews 1988), quite the reverse of the South African scenario.

We know, in short, surprisingly little about how state policy reshaped white working-class cultures, identities, and imaginaries in Africa in the twentieth century, and probably even less about the situation for other layers of workers. In large part this is because the assumption that labour in South Africa equals South African labour has prevented the question of the ‘nationalisation’ of labour being posed at all. When people speak of the South African labour movement as a self-evident category, they do not always recognise that the first truly countrywide union federations in South Africa only emerged by the early 1950s, with the South African Trades and Labour Council and the South African Congress of Trade Unions.[4]

Related to this, it is also worthy emphasising that very little is known about the role of working-class reading cultures and publics in the period of the first globalisation, or in the period of de-globalisation that followed. Print media may, Sakar suggests, have played a relatively limited role in working-class movements in South Africa, as compared to India, with its early development of a mass popular press. The rise (and fall) of the working-class press in South Africa (and southern Africa) is, however, an issue that has only recently begun to be explored (Visser 2004), and there are enormous gaps in our understanding. The Indian case suggests important contrasts, which could be fruitfully explored.

Comparisons between southern Africa and Latin America seem, then, a fruitful avenue for further research, and North Africa also seems eminently suitable for such comparisons. Algeria, a settler colony with the second largest white population in Africa, has only rarely been compared to South Africa. Drew’s paper is quite groundbreaking in examining the different trajectories of Algerian and South African Communism. Drew steers a path that avoids the simple dichotomy of domination by, or autonomy from, Moscow, and stresses the importance of the timing of the implementation of sectarian New Line policies in the late 1920s in each country, the different ways in which the policies were understood and implemented, and the way in which the local context conditioned the ability of communists to organise in the rural areas. Such comparisons could be extended to other periods of left and labour history: like South Africa (van der Walt, 1999; 2004), African countries like Egypt had significant anarchist and syndicalist influences before the 1920s (see Gorman 2005; Khuri-Makdisi 2003).

Drew’s paper highlights the importance of examining the interaction between global and local factors in the shaping of political traditions, and underlines the importance of a more transnational understanding of traditions like Communism. We noted above, for example, that SACP writers have stressed the autonomy and initiative of the party with regard to the Native Republic thesis, which was adopted along with the New Line. Clearly, this approach is too simplistic: the Native Republic thesis was the South African variant of the two-stage policy implemented by the Comintern throughout what was called the ‘colonial and semi-colonial world’, and it is exceedingly unlikely that the Comintern’s global policies were decisively shaped by the views of a section of the small CPSA. On the other hand, the Native Republic thesis was amenable to many interpret-ations: in the 1940s, for instance, party journals like Vryheid-Freedom debated whether the policy really entailed two stages at all, or, if so, whether Afrikaner nationalism was a possible ally for the first stage, or whether the CPSA should lead both stages.

Local context clearly played an important role, and it is perhaps not accidental that the Native Republic was reformulated as Colonialism of a Special Type (CST) in the 1950s. While the Native Republic thesis stressed the struggle against British imperialism, CST described ‘black South Africa’ as the ‘internal colony’ of ‘white South Africa’, which effectively removed the British Empire from the agenda. The two-stage approach was maintained, in other words, but the anti-imperialist element of the policy was transposed from the empire to South Africa. This shift took place in the 1950s, at the height of de-globalisation, the col-lapse of the empire, with a white republic on the horizon, the Comintern dissolved and the white working class (and perhaps the African working class as well?) increasingly nationalised.

It is by placing the question of empire centre-stage, as part of a larger transnational focus that we are alerted to such shifts, shifts that are often hidden by a more narrowly ‘national’ focus on South Africa. This allows us to rethink the way in which the social and ideological worlds of labour evolve and change, but never entirely as an endogenous ‘national’ process.

The migration of white labour northwards from South Africa was paralleled by the migration of coloured workers from South Africa into Namibia (formerly South West Africa), Swaziland and Zimbabwe, as well as by the migration of African workers across the region. The movement of Africans across the region, with roots going back to the pre-industrial period, has been examined by various authors (for example, Harries 1994; Katzenellenbogen 1982; Van Onselen 1976; Vellut 1983; Yudelman and Jeeves 1986).

Nonetheless, this work has often been structured by the image of ‘South Africa’s labour empire’ (Crush, Jeeves and Yudelman 1991). This has the merit of highlighting South Africa’s predominance in the migrant labour system, but carries the danger of suggesting a narrow focus on South Africa, and of seeing migrant labour as a specifically South African device, part of a ‘distinctive cheap labour system’ (Alexander and Halpern 2004:10).

A more transnational perspective suggests important qualifications to such approaches, and the need to examine the eminently transnational process of African migrancy on a larger scale than the vantage point provided by a particular state. Structuralist accounts have portrayed the African migrant labour system as engineered from above, and as characterised by systematic labour control and coercion. However, a striking feature of the regional political economy was precisely the disjuncture between labour markets and states, and the absence of any single organisation that could control regional labour flows. The different colonial states competed with one another for labour, as did employers in different sectors, and in different regions within countries.

It was partly in response to this situation that corporations established supra-national labour recruitment bodies, notably the South African-based Native Recruitment Corporation (NRC) and Witwatersrand Native Labour Association (WNLA).These, too, however, did not have a truly regional control. The NRC and WNLA were quickly emulated by rival capitalists across the region, who formed competing bodies on the South African model: the Rhodesian Native Labour Board (RNLB) was consciously modelled on the NRC and WNLA (Van Onselen 1979:93), and the same seems true of the South West African Native Labour Association (SWANLA) (Moorson 1978).

If mining in Kimberley was the template for labour controls on the Witwatersrand, then the Witwatersrand was in turn the template for labour controls throughout southern Africa. By examining the regional labour system from the vantage point of South Africa, and by viewing ‘racial capitalism’ as a monolithic top-down process, scholars have sometimes ignored the rather disorganised character of African labour recruitment that a regional perspective reveals, as well as the fact that ‘racial capitalism’ was less distinctively South African than characteristic of southern Africa as a whole.

Now, precisely because there was no general regional mechanism to direct flows of African labour, African workers were able to navigate competing claims on their labour power in search of the best jobs across the region. Charles van Onselen (1976, 1979) memorably examined this process in Zimbabwe, and far more needs to be known about it in other contexts, as well as the way it played out at a regional level at different times. A narrow focus on South Africa as a distinctively low wage capitalist economy ignores the point that, in the regional context, the Witwatersrand mines provided, on the contrary, the best paid jobs (Van Onselen 1979), and, in addition, fails to recognise that racial wage gaps on the mines were highest, not in South Africa, but Zambia (see Meebelo 1986).

The regional dimensions of the labour market and migrant labour system are not fully understood, and far more needs to be known about the role and significance of migration outside of official channels like the NRC, WNLA, RNLB and SWANLA, as well as migration outside of mining, like rural-to-rural circular migration. The labour history of agriculture is not well developed in southern Africa, particularly outside of South Africa, and Wazha Morapedi’s contribution to this collection is to be welcomed. Morapedi examines farm labour in Botswana (formerly Bechuanaland) on predominantly white-owned commercial farms, and develops a comparison with farm labour in South Africa. His analysis examines the significance of different labour markets within the country, of competition with South African mines, the use of migrant labour, the role of ethnicity in the labour process, interaction of ‘the worst sort of white South African farm exploitation’ with ‘indigenous Botswana ethnic discrimination’.

Cross-border migration, by its very nature, is not easily studied on a country-by-country basis, while close attention to varying wage zones within as well as between countries cautions against assuming that the different colonies were internally homogenous in terms of levels of economic development or state capacity. The boundaries of the colonial states were not only often quite arbitrary, but the borders were often very porous and commonly ignored, evaded or trans-gressed by Africans.

We have repeatedly used the term ‘transnational’, which still suggests the centrality of the ‘national’, but the extreme variations within different countries in southern Africa must also be noted. Given the fractured legal systems involved in indirect rule and apartheid (Mamdani 1996), and uneven economic development within countries, it may be worth thinking of the significance of internal labour migration across internal ‘borders’ within countries: the experience of migration from the Eastern Cape reserves to Johannesburg could, arguably, be as significant as that of migration from Gaza in Mozambique.

It was noted earlier that white labour migration into South Africa was important to the transmission of union traditions, White Labourism, anarchism and syndicalism; it could be added that immigrants were also central to the communist parties of South Africa in the 1920s and 1930s (Drew 2002) and Southern Rhodesia in the 1940s (Lessing 1995). Southern Africa’s integration into regional and international labour markets enabled a constant circulation of ideas and linkages into ideas circulating in labour and left circles worldwide in the period of the first globalisation. At the same time, the regional labour market was fractured and racialised, and different ethnic groups laid claim to particular occupations, which partly accounts for what Van der Walt describes as the tendency of ideas, organisational models and repertoires of struggles to flow along ethnic and racial conduits (although radical and internationalist left traditions could burst out of these channels).

Colonisation and capitalism in Africa created new transnational connections, and international diasporas and networks of various types. George Gona’s article in this collection explores examples of both in colonial East Africa, and draws the lessons of an older history of regional unionism for current labour movement strategies. He shows that the labour movement in the region assumed a regional character from the 1920s to the 1950s, and that migrants and travelling organisers played an important role in linking workers’ movements in the different colonies. The East African Trade Union Congress (EATUC) formed in Kenya in 1949 organised a wide variety of occupations, and, strikingly, set out to organise labour in Uganda and Tanzania (formerly Tanganyika) as well. The ethos of this union tradition was anti-colonialist but internationalist, and Indian workers – a significant component of the East African labour force – played a prominent role, most notably the self-declared communist Makhan Singh.

African migrant labour and migrant networks also played a critical role in the spread of subversive and transformative ideas over a vast area. Religion provides one example. In 1903, for example, a labour migrant from Malawi (formerly Nyasaland) called Elliot Kamwana was introduced to Jehovah Witness (Watch Tower) doctrines while working in Cape Town. From 1906 Kamwana preached an apocalyptic Watch Tower doctrine in Malawi, recruiting thousands. Kamwana was later exiled, but Watch Tower spread, largely through migrant networks, into the mining compounds of Zimbabwe, and subsequently into Zambia and the Republic of the Congo (formerly the Belgian Congo) (McCracken 2000; Phimister 1988; Ranger 1970; Raftopoulous and Phimister 1997).

It is difficult to believe that Watch Tower did not get entangled with that other important labour current in southern Africa, and in which Malawian networks also played a central role: the Industrial and Commercial Workers’ Union of Africa, or ICU. Van der Walt notes that the ICU, itself influenced by currents brought from abroad like Garveyism and syndicalism, was in many ways a transnational movement operating across southern Africa, paralleling in some ways the EATUC in East Africa. Certainly, shifts and cross fertilisation between religious and trade union dispositions have been common ever since throughout southern Africa, with the role of a church background, for instance, in the development of skills in oratory and organising an issue that merits closer examination.

The overlap between religious traditions and labour organising is an area that remains largely unexplored, and an examination of the spread across borders of popular religious traditions, amongst workers of all races, provides an excellent way in which to explore the transnational formations and connections of working classes. The social history of unions and parties, more generally, is not well developed in southern Africa, where ‘old labour history’, focused on organisations, policies and leaders, has tended to predominate. The interaction between labour and left currents in South Africa and elsewhere was complicated and interactive, and the balance of influence of transnational, ‘national’ and other factors varied over time. Not only was the official imperial ideology of empire (which is often not taken seriously enough) appropriated and reworked by subject peoples (Ranger 1983), but so, too, were international labour and left traditions.

Conclusion

This introduction has argued for a labour history that takes regional and transnational processes seriously, and for situating South Africa in southern Africa, and southern Africa in the larger world. In eschewing what Van der Linden calls ‘methodological nationalism’, and thinking about a southern African, rather than a South African, working class (Bond, Miller and Ruiters 2001), and in noting that working classes and working-class movements are not forged in autonomous ‘national’ contexts, we emphasised connections and comparisons. While our discussion has raised questions about ‘cheap labour’, migrancy and their relationship to social imaginations, we have left the question of the conceptual vocabulary of labour studies open. This article is a contribution to opening transnational labour history, not its conclusion.

References

[1] Indeed, many contemporary Marxists seem to regard ‘progressive’ or ‘anti-imperialist’ nationalist movements as intrinsically radical, an approach that can sometimes be used to support some of the most reprehensible regimes, like that of Robert Mugabe.

[2] The work of Baruch Hirson, exiled South African Trotskyist, on the history of labour and the left, can also be usefully placed within the activist tradition: for a partial compilation, see Hirson and Hirson (2005) and also see Hirson (1989).

[3] Only Southern Rhodesia, with the achievement of self-government in 1923, came close to the South African experience, and was able to make early protectionist economic policies by the 1930s (Phimister 1988; also see Bond, Miller and Ruiters 2001). Protectionist policies were adopted in Mozambique in the 1910s, but largely as a result of initiatives by Portugal itself (see Capela 1981; Penvenne 1995).

[4] While the South African Industrial Federation was formed in 1914, the Cape Federation of Labour remained outside the fold of this federation and its successors for nearly forty years. Neither the Federation of Non-European Trade Unions, formed in 1927, nor later bodies like the Council of Non-European Trade Unions were countrywide (‘national’) federations.