The shopfloor is receiving renewed attention within the field of comparative industrial relations. This interest reflects academic concerns, notably a wish to understand how new managerial techniques actually function, but also developments in the real world, particularly the competitive challenge of Japan and the perception that it is in the regulation of labour within the workplace that a large part of Japan's success lies. An important but neglected mode of workplace inquiry, the ethnographie tradition, has a major part to play in understanding the shopfloor in comparative perspective. The paper sets out its approach, considers issues of validity and generalization, illustrates its contribution to comparative analysis, and outlines a research agenda for the future.

The approach has three key components. The first is its theoretical perspective on the nature of work relations. Conflict is the central principle underlying the organization of work because workers are exploited by employers. This "structured antagonism" underlies day-to-day relations. At this level, cooperation becomes significant. Conflict is not separate from co-operation. The two are intertwined, and the analytical task is to understand how in a particular workplace they are organized and expressed. Second, therefore, the object of inquiry is the regulation of work: the rules, procedures, customs and understandings that regulate how workers' capacity to labour is translated into actual effort. Third, research methods place particular weight on intensive observation, though studies that use relatively casual observation still adopt an ethnographic orientation to the extent that they focus on day-to-day behaviour and the processes by which conflict and consent are organized.

Several problems with the approach are commonly identified. One is a tendency to descend into mere description of the drudgery of working lives. A second is that studies of individual workplaces can offer no wider generalizations. The former is a weakness of certain studies, and to develop its potential the approach needs a tightly disciplined consideration of analytical issues. There are five ways in which it can thereby offer generalizations: 1) the excavation of activities that would otherwise lie hidden, together with the demonstration that they question certain theories about organizational functioning; 2) the indication of mechanisms linking different phenomena together; 3) the analysis of single "critical cases" which are able to throw light on wider developments; 4) the use of comparisons of workplaces; and 5) the development of a research programme which permits cumulative knowledge to be generated.

International comparisons using ethnographic methods are as yet rare, but four approaches illustrate the potential. The first is simply to explore a country in the light of existing assumptions. For example, the small number of shopfloor studies of Japan help to explode myths about the nature and origins of workers' consent. Second, one country is studied using the perspectives of another. This can ask about processes which tend to be taken for granted within a country and can begin to explain how processes of labour regulation differ, for example what role the strike plays in different national regimes. Third, studies from different countries can be set alongside each other. Where these studies were conducted in similar types of technology and product markets, the distinctive effects of national Systems are revealed. For example, studies of Britain and North America reveal the distinctive roles played by the state in the development of factory regimes. Finally, direct comparison between workplaces can test out and develop ideas derived from more indirect comparisons.

Such studies help to explain what remains obscure within existing comparative analyses. For example, the "political economy" tradition tries to explain patterns of labour regulation in terms of the incorporation of labour within national political Systems, and it uses strike statistics as a major index of industrial behaviour. An ethnographic approach goes much deeper than such statistics, and it relates national-level developments to the site where cooperation is actually generated, namely, the workplace. It can thus resolve certain puzzles within existing accounts.

There are several ways in which the perspective can develop its comparative contribution. An obvious one is to explore changes in manufacturing industry, looking for example at quality circles in two or more countries and exploring their connections with existing Systems of regulation. There is also a need to extend the approach to little-studied groups such as white-collar workers. The range of countries can also be extended, from the advanced capitalist nations which have generally been the focus, to newly industrializing economics.

The tradition has, in view of its marginal place in the social science canon, made several significant contributions in the past. As the workplace gains increasing significance in the pursuit of competitive advantage, its future potential is even greater.