figures: RER map of Paris; E. J. Marey's graphical representation of Paris-Lyon train schedules, 1885
What role does a rail network play within an adequate ontology of society? Is a rail system primarily a set of physical assets, a set of administrative procedures, or a set of embodied opportunities and constraints for other members of society? The answer is, a transportation system has elements of all of these.
A rail system provides convenient transportation among a number of places, while providing no service at all between other pairs of locations. You can get from Porchefontaine to Sevran Livry with only a change of trains at St. Michel - N Dame in about 30 minutes -- whereas from Point X to Point Y there is no convenient transportation connection by Metro or RER. This means, among other things, that some parts of Paris are much more tightly integrated than others. It is possible for residents of arrondissement X to shop and work in arrondissement Y very conveniently, whereas this would not be true for arrondissement Z.
So a rail system certainly has direct effects on social behavior; it structures the activities of the two million or more Parisians by making some places of residence, work, shopping, and entertainment substantially more accessible than other places. And there are a number of other social characteristics that are structured by the commuter rail system as a consequence: for example, patterns of class stratification of neighborhoods, patterns of diffusion of infectious disease, patterns of ethnic habitation around the city, patterns of diffusion of social styles and dialect, ... In brief, a rail system has definite social effects. It creates opportunities and constraints that affect the ways in which individuals arrange their lives and plan their daily activities. And other forms of social behavior and activity are conveyed through the conduits established by the transport system.
Moreover, a rail system is a physical network that has an embodied geometry and spatiality on the ground. Through social investments over decades or more, tracks, stations, power lines, people movers, and fuel depots have been created as physical infrastructure for the transportation network. Lines cross at junctions, creating the topology of a network of travel; and the characteristics of travel are themselves elements of the workings of the network -- for example, the rate of speed feasible on various lines determines the volume of throughput of passengers through the system. And neighborhoods and hotels agglomerate around important hubs within the system.
In addition to this physical infrastructure, there is a personnel and management infrastructure associated with a rail system as well: a small army of skilled workers who maintain trains, sell tickets, schedule trains, repair tracks, and myriad other complex tasks that must be accomplished in order for the rail system to carry out its function of efficiently and promptly providing transportation. This human organization is surely a "social structure," with some level of internal corrective mechanisms that maintain the quality of human effort, react to emergencies, and accomplish the business functions of the rail system. This structure exists in the form of training procedures, operating manuals, and processes of supervision that maintain the coordination needed among ticket agents in stations, repairmen in the field, track inspectors, engineers, and countless other railroad workers. And this structure is fairly resilient in the face of change of personnel; it is a bureaucratized structure that makes provision for the replacement of individuals in all locations within the organization over time.
So a rail network has structural characteristics at multiple levels. The physical network itself has structural characteristics (nodes, rates of travel, volume of flow of passengers and freight). This can be represented statically by the network of tracks and intersections that exist (like the stylized map of the RER above); dynamically, we can imagine a "live" map of the system representing the coordinated surging of multiple trains throughout the system, throughout the course of the day. The railroad organization has a bureaucratic structure -- represented abstractly by the organizational chart of the company, but embodied in the internal processes of training, supervision, and recruitment that manage the activities of thousands of employees. And the social and technical ensemble that these constitute in turn creates an important structure within the social landscape, in that these physical and human structures determine the opportunities and constraints that exist for individuals to pursue their goals and purposes.
A general problem that confronts assertions about "social structures" is the question, what factors give the hypothesized structure a degree of permanence over time? Why should we not expect that social structures will morph quickly in response to changing uses and demands by opportunistic actors within them? A rail system provides a somewhat more definite answer to this question than is possible for most putative social structures: the physicality of the system is itself a barrier to rapid, radical structural change. The locations of the great rail terminus stations in Paris have not changed in the past century. And this is at least in part a consequence of the vast "sunk costs" that are associated with the embodied structure of track, intersection, and station that had developed over the course of the first fifty years of French railroad expansion. So the need for a passenger from Dijon to Strasbourg to convey himself/herself from the Gare de Lyon (1900) to the Gare de l'Est (1849) is exactly the same today as it was in 1900.