Network Formalism

As promised, here is a blog about network form, power, politics, and Alex Galloway. It will be schematic, not comprehensive. But I wanted to follow up in a substantive way with our unfinished business about this stuff from last class, and before the next one.

A few schematic points:

Galloway wants us to consider network form. By this, he is asking us to think neither about what takes place or moves around on a network (say, cute cat videos) nor the material substrate of that network (say, undersea cables), but rather about the formal structures that allow the material substrate to work and that allow the sharing of all those cat videos.

Galloway’s first book, as well as the program essay on “Networks,” both stress the fact of protocol on contemporary networks. The basic idea is this: in an important way, the internet is a distributed network, rather than a centralized or decentralized one. (Those diagrams in Galloway’s “Networks” essay give you some kind of idea.) In reality, the internet functions as a decentralized network, but in its original design, and in our fantasies about it, it is distributed.

In a distributed network, information can travel across the network without any hubs: any point can be connected to any other point, and through any number of different paths. In fact, the most basic set of protocols of the internet, TCP/IP, will cut a message up into segments, and those segments often travel along different paths from one node (read: computer) to another on the network, to be reconstructed into a single message on the other end.

(There’s much more to learn and explain here, although I really don’t want us to get bogged down in greater detail than, say, what Christian Sandvig teaches us. But, if you’re on a Mac, you can do a fun experiment: open the Unix command line [OS X is just fancy Unix]—you can do so by bringing up the search by hitting command+space and typing “terminal,” or you can find it in Applications > Utilities. This should open a Unix shell. In the shell, type “traceroute,” or “traceroute”, and that will give you a list of all the computers [primarily routers of various kinds, owned by your ISP and Akamai and some others] between you and the destination computer [the main web server for and the main server for, respectively].)

In a centralized network, and to a certain extent, also in decentralized ones, the rules and procedures necessary to connect to the network are articulated centrally. That is, the centralized (or decentralized) hub makes decisions about who can connect to the network at all, who can connect to whom, what paths messages may take, what can be transmitted in messages, and so on. In a distributed network, by contrast, there is no central arbiter of rules on the network. Instead, every member of the network must know and enforce all the rules, all the time.

Which is to say, decisions about connection, routing, and content are now made not by some center endowed with special decision-making powers. Instead, each and every node on the network is equally responsible for, and empowered to make, these decisions. The subtitle of Galloway’s book teaches us this problem: how control exists after decentralization.

The answer is, of course, protocol. Protocols are the rules networks use to organize themselves. Each member of the network knows all the protocols, uses them, enforces them. Falling afoul of protocol does not mean some center hears about the rule-breaking. Falling afoul of protocol does not produce any effects other than a connection failing or a message not being sent properly. Falling afoul of protocol is something that is legislated entirely by one’s peers in the network. Enforcement of rules is now itself fully distributed, even if legislation of them isn’t (hold off on the question about RFCs for the time being). Control still happens, but it happens without a centrally articulated controller.


Galloway’s point is that this has something to teach us about contemporary politics. A decentralized network offers a powerful model of power and politics, and it shows us how to ask different political questions. As Galloway writes, “distributed networks create new, robust structures for organization and control; they do not remove organization and control” (290). The name for these structures is protocols. And the way they work is that they serve as a set of ideologically-neutral requirements for connection to, and participation in, the network. They seem to belong to nobody, they aren’t enforced by an authority.

In this very specific sense, in a decentralized network, power is no longer the possession of a centralized authority, but rather is inherent in the very structure and operation of the network. This has a few effects: power no longer looks like what we assume power will look like; power does not say no to certain acts, but rather manifests as an affirmative solicitation to connect according to a neutral set of rules; resistance to power will now look very different indeed—Galloway’s examples are viruses, hacking—to which we can add DDoS attacks, spam, trolling, and so on: intra–anti-network activity.


(On Foucault: The reason I brought up Foucault in class is that he described this form of the operation of power in his History of Sexuality, Volume 1: The Will to Knowledge, several decades before Galloway did. (It’s not really important for you to know that, but for those of you who have read the History of Sexuality, this might act as a short cut of sorts: it lets us see what sexuality might have to do with networks. Also, Cohen’s argument about populations vs. publics is in deep, if inexplicit, conversation with this book as well.)


But how is this stuff about networks about politics? He’s just talking about technology. What is the relation between network form and politics? I don’ t quite know the answer myself. It’s a good question. We’ll keep pursuing this question, but here are two ways to think about the answer, one abstract and one concrete.

Abstract: The relation between network form and the operation of power could take any forms. Is network form a metaphor for the operation of power? Perhaps a diagram? Or maybe network form is a specific kind of operation of decentralized power? There are lots of ways to think about it, and a large number of kinds of analysis we might perform to pursue this question. My gut reaction (but I sure don’t have this fully worked out) is to say that decentralized networks and their protocols are a specific technology of power, that power can use to exert control in a number of different scenes, but that a great many other technologies of power coexist with networks (neoliberal economics, electoral politics, broadcast media, schools, the family, etc. etc. etc.). That means our object of study can and will be this: how various actual network practices arrange for the operation of power, which is to say, how they implement structures of control and organization.

Concrete: We talked a bit about Facebook and Twitter in class. The idea here is that different networks—in this case, social network sites—have different protocols and thus produce different structures of control and organization. Facebook spent a while attempting to force users to use their “real names,” causing a whole host of problems for, say, people who live under names different from their legal names, people whose names don’t fit mainstream American ideas of what names should look like, and people who are fleeing or surviving in the wake of domestic violence or even threats of online violence. Which is to say, a neutral-sounding rule—that users must use their “real” names for their Facebook profiles—ends up producing a series of political effects.

Now, while Facebook itself is a “centralized” service (it is, in fact, “decentralized,” in Galloway’s sense), in the sense that it is a single company that hosts and maintains the network, it does so in a way that is not “authoritarian”: it has nothing obvious to do with governments, police, elections, candidates, and so on. But the not obviously political idea that social networking profiles must index actual humans has profound political effects when it becomes an aspect of how we use networks these days.

Meanwhile, Twitter allows pseudonymity, and that means Twitter is a very different space. As I think it was Noah pointed out, Facebook’s insistence that profiles index users, along with its model of a reciprocal form of intimacy “Friends” as the normative connection between users, means it’s boring and blushless; Twitter and other pseudonymous services are much rawer, louder, and perhaps more dangerous because of that. The activities and worlds Dibbel describes in “A Rape in Cyberspace” feels like a different world from Facebook, but much closer to Twitter and Reddit and so on.

We will return to Twitter and its political effects later, but one way to think about this is that the 140-character limit Twitter enforces encourages certain kinds of virality, re-tweeting, and amplification; and hashtags encourage connection between users that were previously unknown to each other. This is nevertheless not necessarily positive or even neutral—we need analysis. What we need, then, is a way to describe, and to think about, the ways that various network forms, themselves not identical or identically structured by protocol, might refract political questions, and might modulate the operation of power in contemporary life.

We won’t, of course, get anything like a comprehensive answer to those questions, but we have begun to specify the questions that matter, and that we might not have known how to ask before.

One thought on “Network Formalism”

  1. The world that “A Rape in Cyberspace” took place in does indeed feel different from FB. On FB, with the identity of the participant relatively intact, our engagement with the “community” feels somehow self policed. While the pseudonymous aspect of twitter probably does make it a more vicious place, to me, there is something about its rawness that feels more reflective of the real world and not just an artificially heightened rage machine.

Comments are closed.