actor network theory

anthropology, power, subjectivity, agency

I want to add two interrelated concepts to the pile of actor-network theory readings we had this week: the concept of “memory” and emergent behavior. For memory, I refer to particular path dependence, that for a particular actor to transition from one state to another requires memory of the previous state. A spring, for example, springs back based on the memory of how much it was pressed down (to a physical maximum). Ice must transition into a liquid state before becoming water vapor, in essence, “remembering” that it has to become water before it becomes vapor. Emergent behavior is the claim that particular behavior occurs once a network achieves a certain quality: the connections between actors can be unidirectional or bidirectional, actors can have a large number of connections, making the network very dense, or a small number of connections, make the networks very sparse, and even the size of the connections can matter. Single celled organisms are significantly less agentive than multicelled ones, even though the quality of cells does not vary significantly.

The reason for these abstractions on top of ANT is that I think an agency can be sufficiently modeled with these two concepts. The agency of an actor can be modeled as a function of their memory, and/or their network density.

Paper, as described by the Hull reading, provides a good example. As Hull shows, the material nature of bureaucracy is embodied within files. Files originally start as a particular form of memory: if we are to imagine this in the concept of ANT, the person is a single node, and the files can be represented as another node, with a low bandwidth bidirectional link between them. Bidirectional represents a person adding and retrieving files, and low bandwidth represents that the person does not do it very frequently, a thin “pipe” of information. However, as files grow and the pace of addition and retrieval increases, the pipe must increase in size to accommodate more frequent retrieval of files, and the connection grows thicker in size. At a certain point, a singular person cannot manage to add and retrieve the necessary amount of files at the same time, necessitating another person. Instead of person <–> files, the network has expanded to person <–> files <–> person. As this process continues, we find a particular emergence occurring: rather than a network centered on the original person, who has created the files, we find that the center of the network has become the files themselves, a behavior that is “emergent”. Additionally, due to the fact that the files hold literal memory, the files gain more agency within the network. Access and signatures develop a particular semiotic, files no longer influence the person they are directly connected to, but to the second order of connections. What begins as a form of categorization has become constitutive of the political economy that Hull describes, and the demon of writing that Ben Kafka describes.