Thesis: “This book argues that we can make sense of the history of computers as tools only when we simultaneously grasp their history as metaphors in Cold War science, politics, and culture” (ix).


  • closed-world discourse: “geopolitical strategies in and through military systems for centralized command and control” (xiii)
  • cyborg discourse: “metaphors of minds as computings in and through integrated human-machine systems and technologies of AI” (xiii)
  • closed world: the “dome of global technological oversight” that foreclosed rhetorical and epistemic realities during the Cold War (1)
  • green world: opposite of closed world, drawn from Northup Frye: “an unbounded natural setting [where] action moves in an uninhibited flow” (12).
  • politics of subjectivity (149): shades of what we would now group under the rubric of neoliberalism: to be free in a liberal society is to “explicitly experience our subjectivity in our roles as voters and participants.”
  • subject positions (304): “imaginary, yet coherent and emotionally invested ways of living within a discourse”


  • Edwards has two projects: 1) writing the history of computers as central to the Cold War and 2) their development as “an axial metaphor in psychological theory” (x).
  • Maps two parallel genres of historiography: 1) computers operating as “embodiments of ideas about information, symbols, and logic” that made the fulfillment of certain philosophical and logical questions possible; and 2) “devices for processing information” that operationalize information as a commodity (x–xi).
  • “This book is built on an implicit critique of existing computer historiography. Instead of progress and revolution, the plot structure I shall use emphasizes contingency and multiple determination. I shall cast technological change as technological choice, tying it to political choices and socially constituted values at every level, rendering technology as a product of complex interactions among scientists and engineers, funding agencies, government policies, ideologies, and cultural frames” (xiii).
  • Question for further reason: how is Edwards’ using the idea of discourse, and how does his application of discourse differ from the kinds we’re accustomed to 1) in English more generally and 2) in media studies more generally?

1: Building the Cold World

  • Cyborg as a set of “subject positions” (clear that Haraway was his advisor!) (2).
  • Three major theses (2–3):
    • “The historical trajectory of computer development cannot be separated from the elaborated of American grand strategy in the Cold War”
    • “I will link the rise of cognitivism, in both psychology and artificial intelligence, to social networks and computer projects formed for WWII and the Cold War”
    • “I will suggest that cyborg discourse functioned as the psychological/subjective counterpart of closed-world politics”
  • How do computers maintain closed-world discourse?
    1. “they allowed the practical construction of central real-time military control systems on a gigantic scale”
    2. “they facilitated the metaphorical understanding of world politics as a sort of system subject to technological management”
  • Systems of containment and management (infrastructure; talk to Kyle). These systems were extensions of the drive to enclose the world in capitalist systems.
  • Opposite of closed world is green world, from Northrup Frye (more literary contexts).
  • Simulations become the coin of the closed world: “more real than the reality itself, as the nuclear standoff evolved into an entirely abstract war of position”(14). “The object for each nuclear power was to maintain a winning scenario . . . rather than actually to fight such a war.”
  • In section on Babbage: How many key moments of computing history (Analytic Engine, memex) entirely speculative?
  • Why, when called on to unify his two threads, does Edwards turn to The Terminator—a piece of fiction? I’m not saying it’s wrong, it’s just interesting to place it in what’s putatively a work of historiography (22).

Edwards’ Theory

Edwards has a section at the end of Chapter 1 that lays out some theoretical definitions. Let’s walk through them.

  • Computers: distinguished from other tools and machines by conditional branching and symbolic manipulation (27). ? for later readings: are these sufficient distinctions? Edwards claims that computers don’t do “physical work” and so resemble rulers more than hammers: tools for organizing, “language machines.” This organizational nature lends it to activation as a metaphor for cognition.
  • Tool-thinking: (It’ll be useful after I read my Heidegger and connect this to Harman and the OOO folks) Following Mumford, machines can have “subtle, profound, and material effects solely through its function within a system of ideas” (29).
  • There’s a pretty elaborate (and more than a little defensive?) section on discourse, which seems to fundamentally resolve to the poststructuralist tradition, though Edwards is interested also in the discursive infrastructure—the “supports” that keep discourses moving (40).

2 / 3 / 4: Computers and the Military

  • Together, these three chapters for the historical core of Edwards’ first thread.
  • Soviets provided an apocalyptic transference from Nazism: a parallel closed world that was the dark mirror of the US (54).
  • An ethic of antimilitarism encourages the development of cybernetic and technowarfare over standing military bodies—is this true, and does it hold now? (58)
  • Consequences of military support sketched in 62–65, elaborated in chpts 3 and 4):
    • Military funding quickened the pace of American computer development, allowing it to crush competitors.
    • Military secrecy impeded the spread of new technology.
    • Military continues to invest in R&D even after mature commercial markets emerge. In particular, AI is a key research area.
  • The chapter on SAGE (3) is important in how it 1) is a proleptic or failed technology and 2) presages nuclear early warning systems that will eventually involve themselves in planetary scale computing and worldwide sensory apparatuses (75).
  • IMPT: SAGE is an important example insofar as it was a technology of principally discursive value. It never “worked” in the sense that it would have never properly protected the US from nuclear assault. But it “worked” to structure horizons of possibility for future technologies and mid-century technoculture. (110–11). This is an important category of technology to look out for.
  • Interesting moment from the chapter on RAND: nuclear weaponry demanded a flattened hierarchy so that decision-making was reserved for the highest levels, which in turn reproduces (or demands) as much automation in the lower levels as possible: hence computerization both materially and psychologically, with shades of what would later become dominant aspects of platform capitalism (131).
  • “With every launch another orbiting object drew its circle around the planet, marking the enclosure of the world within the God’s-eye view from the void. Even in the dizzy technological euphoria of the first moon landing, the barren moonscapes, sterile capsules, and sealed space suits emphasized not the bounty of a green frontier but the utter aloneness of the living Earth. After all was said and done, the space program’s chief products were not outward- but inward-looking: spy cameras to pierce the Soviet Union’s veil, pictures of the Earth drifting alone through space, pictures of the closed world” (135).
  • Invocation of “enormous scale” as impt reason computers were required for Vietnam: but what scale comes first? The destruction or its infrastructure? (137)
  • Inflation of Vietnam “output” (body count) presages similar motivations behind financial crisis (140).
  • Ref to Zuboff on pp. 144.

5: Metaphor and the Politics of Subjectivity

  • Cybernetics, cognitive psychology, and AI are generalyl thought of as “speculative and theoretical” programs w/o practical import; Edwards argues that in fact they have great practical import insofar as the generate cyborg discourse (147).
  • This chapter is about the political stakes of metaphor and as such is pretty salient to my work. However, much of its analysis is pretty straightforward by contemporary humanities standards. His most provocative claim, which I would in fact resist nowadays, is that “metaphor, as a major mode of representation, frequently helps to organize theories of all sorts” (148). Is metaphor a mode of representation any more? Or have we crossed past metaphors and into things?
  • Theory creates political subject positions” that we can then inhabit through our own “free” choice (150).

Hey Look At Me!!!

I am pretty invested in Edwards’ discussion of metaphor but mostly so that I can diverge from it. I follow more contemporary threads in media theory / media archaeology that take a more materialist point of view—Edwards is espousing a moderately idealist position wherein metaphors exist in advance of material conditions. Certainly everything’s imbricated and we don’t have to draw the lines—but media archaeology pushes on the metaphor-ness of the metaphor by insisting on a very real material substrate.

I have to be honest: much of this discourse analysis is a little tedious, inasmuch as it never really pushes past a surface level: one feels that Edwards has imported the languages of literary criticism as tools without really considering their histories and nuances.

Still, the most useful thing to take from this discussion of metaphor is the way that technologies provide the sociocultural scaffolding upon which we do the work of political self-definition.

6 / 7 / 8: AI

  • Chapter 6 is chiefly about cybernetics, with focuses on the Macy Conferences, Shannon’s communications theory, and our pal Norbert Wiener.
  • (Good to know that the problem of noise in early information theory was literally an issue of aircraft noise and communications! [211]).
  • In the end, these three chapters, with their focus on language, embodiment, and control, prove more interesting for the communications studies approaches this side of my list is developing. That being said, Edwards uses a pretty straightforward application of linguistics and systems theory—more as a tool than as an intervention into communications or media studies.
  • Chapter 8 has a section that begins to involve itself in some proto-media-archaeology, although with a very different model from, say, Kittler (244-50). [Okay, but how???]

10: Fiction and Film

  • This chapter does a reading of science fiction films of the 60’s and 80’s, along with Neuromancer, as “iconography of the closed world” (303). Edwards (a little out of his depth here) argues that “the closed world . . . represents a special kind of dramatic space whose architecture is constituted by information machines” (303–04).
  • Cyberspace as a closed world trope (308).
  • Green world appears really for the first time here. (He doesn’t quite stick this landing, but it’s provocative.) Green space; no borders; return to nature over technology. Heroic quest rather than siege. Interesting then how diametrically opposed nature and technology are for Edwards.
  • He writes literary analysis on the same level as historical anecdote.
  • This is the clearest invocation we get of the connections between the books’ two threads: “The experience of the closed world is a double experience of cyborg identity constituted through those metaphors, technologies, and practices. First, it is an experience of the possibility of other minds constructed, from parts, processes, and information machines, to manage and inhabit closed-world spaces. Second, it is the experience of one’s own mind as an information machine, a constructed self-system subject to the same disaggregation, simulation, engineering, and control as those of the computerized Other. The choices faced by closed world protagonists, as systems nested inside larger systems, do not include redemption, reunification, or transcendence. Transcendence is impossible in the closed world because, like a curved Einsteinian universe, it has no outside. Whatever begins to exceed it is continually, voraciously re-incorporated” (341).

Archive / Impact

  • Edwards is a historian of computing and information science, and he has a thankfully direct writing style that makes it pretty easy to go straight for the information. He tells us point-blank his theoretical archive:
    1. STS, esp. the sociological schools of Latour and MacKenzie. “My analytic frame was shaped by concepts deriving from this tradition, such as actor-network theory, inscription devices, heterogeneous engineering, technoscience, technological frames, entrepreneurial engineers, and the social construction of science and technology” (xvii).
    2. Poststructuralism, which informs the book’s emphasis on “discourse, narrative, and subjectivity.” (The role of narrative in technological historiography comes up in Edgerton as well.)
    3. Philosophical studies of AI, e.g. Searly, Hofstader.
    4. Interpretive sociology of computer communities, e.g., Turkle and Leigh Star.
    5. And remember: Paul Edwards was Donna Haraway’s student at Santa Cruz. Hayden White also taught in this program.
  • Writing against dominant historiographies of computing; a “corrective to perspectives that create the impression of an inevitable progress driven by impersonal market forces and technical logics” (xiv). Once again, we’re working against determinism.
  • His “closed world” metaphor derives from literary criticism (12): the “closed-world play” is a unity of space and ideological conflict.
  • In developing his concept of discourse, he’s following the Marxist literary tradition (Raymond Williams, Terry Eagleton), as well as Kuhn’s ideas of the paradigm (32). Again, working against hard determinism: “Individuals may participate in and be shaped by numerous discourses without being fully determined by any of them.”
  • Edwards’ argument is extraordinarily persuasive, but also fairly totalizing: everything seems assimilable within the idea of the closed-world discourse.