Globalization can be deconstructed in terms of the strategic sites where global processes materialize and the linkages that bind them. Among these sites are export processing zones, off-shore banking centers, and, on a far more complex level, global cities. This produces a specific geography of globalization and underlines the extent to which it is not a planetary event encompassing all of the world. It is, furthermore, a changing geography, one that has changed over the last few centuries and over the last few decades. Most recently, this changing geography has come to include electronic space.

The geography of globalization contains both a dynamic of dispersal and of centralization, a condition that is only now beginning to receive recognition. The massive trends towards the spatial dispersal of economic activities at the metropolitan, national and global level which we associate with globalization have contributed to a demand for new forms of territorial centralization of top-level management and control operations. The spatial dispersal of economic activity made possible by telematics contributes to an expansion of central functions if this dispersal is to take place under the continuing concentration in control, ownership and profit appropriation that characterizes the current economic system.

National and global markets as well as globally integrated organizations require central places where the work of globalization gets done. Further, also information industries require a vast physical infrastructure containing strategic nodes with hyperconcentration of facilities; we need to distinguish between the capacity for global transmission/communication and the material conditions that make this possible. Finally, even the most advanced information industries have a production process that is at least partly place-bound because of the combination of resources it requires even when the outputs are hypermobile.

Further, the vast new economic topography that is being implemented through electronic space is one moment, one fragment, of an even vaster economic chain that is in good part embedded in non-electronic spaces. There is no fully dematerialized firm or industry. Even the most advanced information industries, such as finance, are installed only partly in electronic space. And so are industries that produce digital products, such as software designers. The growing digitalization of economic activities has not eliminated the need for major international business and financial centers and all the material resources they concentrate, from state of the art telematics infrastructure to brain talent (Castells, 1989; Graham and Marvin 1996; Sassen 1998: chapter 9).

In my research I have conceptualized cities as production sites for the leading information industries of our time in order to recover the infrastructure of activities, firms and jobs, that is necessary to run the advanced corporate economy, including its globalized sectors. These industries are typically conceptualized in terms of the hypermobility of their outputs and the high levels of expertise of their professionals rather than in terms of the production process involved and the requisite infrastructure of facilities and non-expert jobs that are also part of these industries. A detailed analysis of service-based urban economies shows that there is considerable articulation of firms, sectors, and workers who may appear as though they have little connection to an urban economy dominated by finance and specialized services, but in fact fulfill a series of functions that are an integral part of that economy. They do so, however, under conditions of sharp social, earnings, and often racial/ethnic segmentation.(Sassen 2000a: chapters 8 and 9).

In the day-to-day work of the leading services complex dominated by finance, a large share of the jobs involved are lowly paid and manual, many held by women and immigrants. Although these types of workers and jobs are never represented as part of the global economy they are in fact part of the infrastructure of jobs involved in running and implementing the global economic system, including such an advanced form of it as is international finance. The top end of the corporate economy --the corporate towers that project engineering expertise, precision, "techne"-- is far easier to mark as necessary for an advanced economic system than are truckers and other industrial service workers, even though these are a necessary ingredient. We see here at work a dynamic of valorization that has sharply increased the distance between the devalorized and the valorized, indeed overvalorized, sectors of the economy.

For me as a political economist, addressing these issues has meant working in several systems of representation and constructing spaces of intersection. There are analytic moments when two systems of representation intersect. Such analytic moments are easily experienced as spaces of silence, of absence. One challenge is to see what happens in those spaces, what operations (analytic, of power, of meaning) take place there. One version of these spaces of intersection is what I have called analytic borderlands (Sassen 1998: chapter one). Why borderlands? Because they are spaces that are constituted in terms of discontinuities; in them discontinuities are given a terrain rather than reduced to a dividing line. Much of my work on economic globalization and cities has focused on these discontinuities and has sought to reconstitute them analytically as borderlands rather than dividing lines. This produces a terrain within which these discontinuities can be reconstituted in terms of economic operations whose properties are not merely a function of the spaces on each side (i.e., a reduction to the condition of dividing line) but also, and most centrally, of the discontinuity itself, the argument being that discontinuities are an integral part, a component, of the economic system.