Figure 1: A node-edge network graph. (Fürstenberg, 2006)
Network is a term pregnant with meaning. Although all of the actors in a network may not be consciously aware of network theories, those theories help us to understand networked practices and explore the possibilities for networked writing and rhetoric. Here I briefly introduce two prominent theories for conceptualizing networks that inform my approach to network* writing.
Graph Theory
Graph theory is the mathematical description of network structures. Network graphs are depicted as nodes and edges, where nodes represent the objects that constitute the network—people, computers—and edges the connections between them (Fig. 1). In his history of graph theory, Albert-László Barabási (2002) argued that from the perspective of graph theory nearly everything is a network, noting “[m]ost events and phenomena are connected, caused by, and interacting with a huge number of other pieces of a complex universal puzzle” (p. 7). As Barabasi describes, graph theory has been enormously beneficial in helping identify the effects of different network structures on the properties of networks. Concepts such as small worlds (p. 40), increasing returns (p. 87–88), and the existence of network hubs (p. 64) have all come from graph theory. In this way, graph theory has been extremely useful for identifying the structure and properties of node–edge networks.
Actor-Network Theory (ANT)
Figure 2: A basic actor-network diagram (Potts, 2014, p. 33)
Actor-Network Theory (ANT) takes a different approach to understanding networks from that of graph theory. ANT identifies all participants in the network—from people, to events, to technologies—as actors, capable of influencing how information is processed by the network. Bruno Latour argued that a network is not a distinct “thing out there” (p. 131), but is rather an analytical device that is useful for understanding complex phenomena. As he put it, a network “is a tool to help describe something, not what is being described” (p. 131). Because of this flexibility, ANT has become an increasingly prominent theory for understanding both stable and mercurial organizational structures and the ways in which those structures process information.
Although actor-network maps (Fig. 2) can bear similarities to graph theory diagrams, conceptually ANT encompasses more phenomena than graph analysis, embracing the artificiality of networks as descriptors. This widening of scope makes it easier for ANT to describe both systems that are traditionally understood to be networks as well as those that are not, such as “a symphony” or “a piece of legislation” (Latour, 2005, p. 131). Consequently, where graph theory is concerned with connections between nodes—primarily the number and direction of edges that connect nodes—ANT is concerned with how networks process information. Where in graph theory a network is a well-defined object with semi-stable nodes and edges, an actor-network is a “trace” marked “by the passage of another vehicle” or “another circulating entity” (Latour, 2005, p. 132). Actor-network analyses, then, are less concerned with the identification of nodes as entities traditionally understood to be power holders (e.g., people or governments) than with the influence of network-like connections between a host of actors on the operation of that power.