On December 30, 1936, workers occupied General Motors' Fisher Body Plants No. 1 and No. 2 in Flint, Michigan. By March 12, 1937, General Motors (GM; the United States' largest industrial corporation with vast financial resources and a network of anti-union spies) was forced to capitulate and sign a contract with the United Auto Workers (UAW). This was the beginning of a flood of strikes that brought unionization to the mass-production industries of the United States, at a time of both high unemployment (i.e., weak market-based bargaining power) and feeble labor organization (i.e., weak associational power).

A key to the UAW's success was labor's workplace bargaining power: the ability of the workers to exploit their position within the complex division of labor characteristic of mass production. The Flint sit-down strike that paralyzed GM's Fisher Body plant was planned and executed by a "militant minority" of autoworkers who by "unexpectedly stopping the assembly line and sitting down inside the plant ...catalyzed pro-union sentiment among the vast majority of apathetic workers" (Dubofsky and Va n Tine 1977: 255). The strike wave demonstrated the limits of the assembly lines' technical control of the workforce: a relatively small number of activists could bring an entire plant's production to a halt. As Edwards (1979: 128) put it, "[technical] control linked the entire plant's workforce, and when the line stopped, every worker necessarily joined the strike."

For several decades following the CIO victories, three employer re­sponses - relocation of production (disinvestment in union strongholds), process innovations (mainly automation), and "political exchange" (the pro­motion of "responsible" unionism and the repression of "irresponsible" unionism) - progressively undermined the structural strength of U.S. labor in general, and autoworkers in particular. When a new upsurge in rank-and­-file unrest at the end of the 1960s (symbolized by the "Lordstown Blues") pushed the UAW back toward confrontational tactics with "Operation Apache" (a campaign of short, small, but highly disruptive strikes), the automakers abandoned the promotion of "responsible unionism" and pursued geographical relocation and automation of production with a new­found zeal.

a new trade union movement burst onto the scene in Brazil, bringing to an end almost one and a half decades of worker quiescence

During a 1990 strike at the Hyundai Group, auto assembly-line workers described the following incidents:

"Only a small number of [Hyundai Motor] workers were blocking the road at first [in solidarity with striking Hyundai shipyard workers] ", said Roh Sang Soo, a young assembly line worker...."Then the police threw teargas into the compound while we were working. We could not work. ...I was on the Excel line at the time. I smelled the teargas. I came out of the plant and joined the demonstration" ....Those who came to work the next day again got teargassed. "We couldn't work because of the gas", said assembly line worker Lee Sang Hui. "If one person cannot work, the whole line stops. I just joined the demonstration and sang songs and clapped. "

The Japanese auto makers brought to these sites many of the organiza­tional practices of automobile production in Japan. These practices spread still further as U.S. and Western European automobile firms responded to the Japanese competitive threat by selectively emulating Japanese orga­nizational practices. Thus, in the 1980s, flexible work rules, just-in-time delivery systems, teamwork, quality circles, and a move away from vertical integration toward the extensive use of subcontracted inputs (outsourcing) were widely adopted. There was, however, a crucial difference between the original Japanese model and that adopted by the U.S. and Western European multinationals. That is, the latter did not promise job security to their core labor force. In other words, the cost-cutting measures of ]apanese lean production were adopted without the related employment policies. This model might thus be labeled "lean and mean" (cf. Harrison 1997). In contrast, the original "Toyotist" model - which offers employment security to a core labor force in exchange for cooperation, but at the same time creates a large buffer of less privileged workers without the same rights and benefits - might be labeled "lean and dual." The difference between these two models, we will argue, is crucial to understanding the dynamics of contemporary labor unrest in the world automobile industry.

Through the 1990s, the lean-and-mean version predominated. Whereas Japanese multinationals operating in core countries tended to implement the homegrown model (Florida and Kenney 1991: 390- 1), U.S . corporations have generally taken the lean-and-mean road, as have Japanese producers operating in Southeast Asia and Latin America. Deyo (1996a: 9) argued that "authoritarian politics and repressive labor regimes" are char­acteristic of the main sites of industry expansion in low-wage countries. In Thailand, Mitsubishi has not extended employment guarantees to its core workforce (Deyo 1996b: 145-6). In Korea, the domestic auto producers (except Kia) continue to pursue a low-wage, high-turnover strategy of mass production and an anti-union and autocratic managerial style (Rodgers 1 2 On the U.S. automobile industry's emulation of Japanese production methods, see among others Abo (1994). And in China, "layoffs are becoming a painful reality" as the central government promotes a "leaning-out of the industry" in an effort to bring labor productivity in Chinese automobile factories in line with the standards set by international "market rule" (Treece 1997c). Like­wise, Japanese transnationals in Mexico are employing traditional Fordist techniques; they have found it "economically rational to place a heavier em­phasis on low wages, even if the consequent high turnover disrupts quality circles and other lean production techniques" (Shaiken 1995: 248-9, 254).

Yet, to the extent that the main trend in the world automobile indus­try is toward lean-and-dual production, then the most likely sites of future autoworker labor unrest will be among the lower-tier workers in the sub­contracting system. Yet, in these sites, strong grievances do not go hand in hand with strong workplace bargaining power. Moreover, while upper-tier workers are likely to have strong workplace bargaining power, they are also likely to have fewer grievances, and at the same time they are likely to be physically and psychically separated from lower-tier workers with greater grievances and less structural power. And with the "leaning" of the industry, primary-sector workers will account for a tiny fraction of the automobile (and overall) working class. Finally, the distribution of upper- and lower-tier (secure and insecure) workers is likely to correspond to and reinforce the core-periphery geographical divide as well as to overlap with differences in ethnicity, place of residence, and citizenship - with important implications for world labor politics.