Preterition

Operation and object

A composition is a thing: the essay in the folder, the file on the drive, the post you’re reading on your screen. Composing is what somebody did to make it. The noun sits still and its object can travel; the verb happened once, in a body, and is completed. English lets one word carry both, and most of the time we don’t mind. When a colleague says she teaches composition, she means the verb. When a program says it assesses composition, it means the noun, because the noun is what can be collected.

“Writing” performs the same operation: it means both the labor and the labor’s product. One word naming both makes them challenging to pry apart in the moment. Marx faced the same circumstance with “labor,” which in the economists he drew from meant the activity a worker performed and the commodity a worker sold; he had to coin a second term to keep the two straight. Composition hasn’t done the coining. We’ve run for decades on the conviction that the verb matters most, crafting instruments to better capture the noun.

Writing’s technologies recede. The pen, the keyboard, the by-hand heritage of scissors and adhesive for “cut” and “paste,” the autosave keeping a version every five seconds: each was presented as a technology once, and each has since become an ordinary condition of writing. When a technology recedes, its mediated labor recedes with it. The keyboard becomes furniture, the typing goes invisible, and we’re freed to describe writing as a cognitive act, something that happens in an interior and shows up on a page, the apparatus between them forgotten. My recent series declined to contract on what’s in the interior. I want here to recall the apparatus from forgetting, and to see what proceeds from the definition of writing as a technology. I accepted that definition in 2003, planning the dissertation I would defend three years later. From Stephen Kline, I’ve since taken four senses of the word. “Composition” carries all four.

Four technological roles

Kline’s short paper “What Is Technology?” (1985) undertakes a sorting of the word’s confusions, setting the stakes as “the way in which we humans make our living on the planet” (Kline 2003, 210). First in Kline’s four senses of “technology” is the artifact: the hardware, the manufactured object holdable in hand, from a hammer to a smartphone. Second is the knowledge of how to make and use things, the technique or the knack, the sense of know-how Ellul’s La Technique carries from Aristotle’s originating techné. Third is the sociotechnical system of manufacture: the forge, the factory, the supply chain, the collective material endeavor and the people working in it. Fourth comes the sociotechnical system of use: air travel and the airline, education and the school, the arrangement of people and objects and structures and practices. Use is what the rest serves: “Without sociotechnical system of use, the manufacture of hardware would have no purpose” (211). Kline’s four definitions are roles, and a technology can inhabit multiple roles.

Writing as a technology operates in all four roles. This weblog’s posts are artifacts, each a made thing with a date and a slug, living at a particular address on a hard drive. Writing them is technique or know-how, from the kind that lives in one’s wrists to the kind that lives in a genre; it’s the sense I’ve spent a career teaching. The archive is the system of production made visible: a database, a migration, an old WordPress install, the whole apparatus that turns composing into compositions and keeps them. The comments and trackbacks are one system of use answering back with traces, readers doing something with what they found here. The scripted searchable archive and my reproduction of past writing via the technological system I’ve composed also serve a system of use, built for portability and reproducibility. The system of production and use permits recombination; this post, recomposing past compositions, performs what it describes.

The generality of Kline’s taxonomy provides both application and limit. J. K. Gibson-Graham’s entry on “Economy” for New Keywords (2005) particularizes the sense of “technology.” What turned the discursive-material economy into “a unified system,” Gibson-Graham write,¹ was “the interaction of new technologies of analysis and control” (2005, 95): the national accounts, statistical offices and other instruments that can count a country. Gibson-Graham’s technologies manufactured Kline’s system of production and use as “the economy” itself, to the point where “the economic imaginary has seemingly lost its discursive mandate and become an objective reality” (96).

Technologically separable subfields

Composition and rhetoric, as the field that studies writing’s systems of production and use and their constitutive techniques and artifacts, adopts each of Kline’s senses in its theories. I argue here that the discipline lacks a unifying theory of technology for those four senses. Of course we don’t ignore technologies of writing—they’re our focus—but we’ve parceled our thinking about technology out among subfields that don’t read each other much. Process theory operates on technique, the verb. Bruce Horner’s Terms of Work for Composition (2000) reads the system of production, and his opening move names my noun and my verb and refuses both as abstractions: “Whereas writing is typically identified with the texts produced or with a process abstracted from material conditions, I argue for composition to redefine the work of writing as material social practice” (Horner 2000, xxiv). Alice Horning’s “The Definitive Article on Class Size” (2007) reads production as well, as a limit on the channel: the CCCC cap of twenty students to a writing class, fifteen ideally (Horning 2007, 19), set against Richard Haswell’s estimate of some 230 hours of a teacher’s time per course (18), is a statement both about how many texts readers can take in and answer, and about how much and how a writer can produce. John Trimbur’s “Composition and the Circulation of Writing” (2000) reads the system of use and finds it neglected, the field “figuring classroom life as a middle-class family drama” (Trimbur 2000, 188) and “thereby isolating an education in writing from the means of production and delivery” (189). Dànielle DeVoss, Ellen Cushman and Jeffrey Grabill (2005) put infrastructure between production and use and find researchers attending to “the what and why of new media without paying attention to the when of new-media composing” (15). Jim Ridolfo and DeVoss’s rhetorical velocity (2009) arrives at economy by way of a third party’s labor: to compose for recomposition means that “the theorizing of the question of ‘is it worth the time to do this’ calls into question a set of economic and material concerns.”

Each of the above is correct about its sense. The senses meet where writing’s value gets taken up by whoever is positioned to do so: the student, the teacher, the platform, the institution, the archive, the industry, the discipline, the model. A field that keeps the senses in separate conversations keeps that meeting place hidden: it occludes appropriation and exchange. Appropriation for exchange requires a unit, and a verb isn’t a unit. Before the labor of writing can be priced, traded or counted, it must be cut into something bounded: a boundary makes one thing measurable against another. That measured difference is the surplus in surplus labor. The technology makes the cut before any appropriator, and each subfield sees the taking and not the cut. The operation every writing technology performs offers me a way to join Kline’s four senses.

Breaking the flow

The operation is discretization.

Bernard Stiegler mistakenly thought he had coined the term. In his essay on memory for Critical Terms for Media Studies, he defines grammatization as “the process whereby the currents and continuities shaping our lives become discrete elements” and offers writing as the case: “Writing, as the breaking into discrete elements of the flux of speech (let us invent the word discretization for this possibility), is an example of a stage in the process of grammatization” (Stiegler 2010b, 70), with grammatization as “the history of the exteriorization of memory” (71). It’s a good description. The OED’s earliest evidence for the noun comes from 1952, in Mathematics Magazine, and “discretize,” “discretization error” and “discretized” follow through that decade; a numerical-analysis textbook first published in 1972 opens a section on discretization error as “the first” of “the three basic kinds of error in numerical analysis” (Ortega 1990, pt. 2). The engineers needed the word for decades before Stiegler reached for it, though the problem is far older than either: Zeno had shown what it costs to cut a continuum into parts, and Democritus stopped the cutting at the atom. Machine capture of continuous phenomena requires breaking them into pieces, and the breaking carries a cost that requires naming.

I came to the word from the engineering side, through audio. When I was teaching courses in digital technology and culture a decade ago, my lessons put James Gleick’s The Information beside Avery Wang’s paper on how Shazam works, whose tests ran on “8KHz, mono, 16-bit samples” (Wang 2003, sec. 3.1). Wang’s phrase condenses two operations. The sample rate, 8,000 slices per second, is discretization: the continuous wave is cut at intervals along its length. The bit depth, sixteen bits per slice, is quantization: each slice is assigned one of some sixty-five thousand values on a scale. I put the same split to grading in 2021: assessment runs on “discretization and quantification,” in which “we slice things up into measurable bits (discretization) and assign values on an interval scale (quantification).” That’s digitization. The two steps are separable.

Gleick recounts the first step via Harry Nyquist, whose method “was to sample the waves at intervals, in effect converting them into countable pieces” (Gleick 2011, 205). A telephone line’s 3,000 hertz “would cover most of the sound from an orchestra, but the high notes of the piccolo would be cut off” (205). A channel has a band, and what falls outside the band isn’t distorted, it’s gone. Later Gleick writes that we’d do better to store the Well-Tempered Clavier as bits, “the waveforms discretized without loss as per the Shannon sampling theorem” (358). The phrase without loss carries a condition, the sampling theorem, which says a signal can be sliced and rebuilt exactly, provided the slicing is fast enough for the band. Inside the band, discretization is lossless. Outside it there’s the piccolo. And quantization is never lossless: rounding a wave to sixteen bits adds error at any sample rate; a small, uniform, well-behaved error, which permits engineers to budget for it. Horning’s class cap, seen from this perspective, is a sampling rate, and Haswell’s 230 hours are the band. Add students and each one’s writing gets sampled more coarsely, so what’s lost isn’t a student but the fine-grained content of everyone’s writing: the revision between drafts, the sentence-level work, whatever changes faster than a reader can sample. That writing, like the piccolo, is unresolved rather than faulty: inside the band a reader can reconstruct the writer from the samples, while outside it only the grade survives.

Inside the band, the slicing costs nothing. There the loss comes with the second step, when each slice gets a value. When I marked a milestone here the other day I put it this way: “Instead of an account—that’s in the archive—a count.” A count travels where an account can’t. It sorts, it ranks, and it does so precisely by having stripped away the description one needs in order to argue with it. Measurement’s work is comparison without description, and that’s also its limit. A count of the cookies doesn’t say who ate them; it summons an account, from whoever is positioned to give one. (In the case of Malcolm and the cookies, further decrementing the count to zero does not damage the account.) Attention can’t be monitored; only its traces can, and a trace is a slice with a value on it. The keystroke I mentioned last time, a discretized moment entered into a record for counting, is the general case in miniature: the verb of writing, sampled at intervals and quantized, becomes a noun that can be tallied.

In the chapter of Orality and Literacy titled “Writing Restructures Consciousness,” Walter Ong holds that “more than any other single invention, writing has transformed human consciousness” (Ong 2012, 77). He means the whole lineage: “Writing and print and the computer are all ways of technologizing the word” (79). I part from Ong in his subsequent section, where he asserts that “[t]echnologies are not mere exterior aids but also interior transformations of consciousness, and never more than when they affect the word” (81), because I’ve declined to contract on interior transformation. Stiegler’s account stands on the same support, anamnesis, the living recollection that an exteriorized memory either serves or supplants, and his line that “one memorizes only by forgetting, by effacing, by selecting” (Stiegler 2010b, 80) is a solid description of the band, cashed in an interior. I don’t need the interior to keep the operation. Discretization and quantization are operations performed on a signal, visible in the record, and they explain how the verb becomes a countable noun without anyone’s consciousness having to be consulted. That test turns out to sort our inherited theories of technology in useful ways.

Provenance, not intent

In my 2003 post, I took Andrew Feenberg’s four components of what he critiques as the instrumental theory and put “literacy” in for “technology” one clause at a time. The first component alone shows how it went: “Technology [i.e., literacy], as pure instrumentality, is indifferent to the variety of ends it can be employed to achieve” (Feenberg 2002, 5) reads just and only as smoothly with literacy in it: it carries precisely the same degree of deployability for argumentation and vulnerability to critique. The instrumental theory’s alternative in Feenberg is no better, a take-it-or-leave-it in which “technology is destiny” (8). Neutral tool or fate. Neither describes what work at a keyboard is and does.

My test comes from the embodied agnostic subject series. Take a theory of technology; ask where it has to look to find a technology’s values. Wherever the answer is an interior, the test works as a razor: it cuts away what a theory claims to know about a mind and keeps what the theory can show from the record. Instrumentalism looks in the user’s intentions: the hammer is nothing, the hand decides. Langdon Winner looks in the designer’s. His best-known example is Robert Moses, the planner who in Winner’s telling built the overpasses on Long Island’s parkways too low for buses, and so kept the poor and Black New Yorkers who rode them away from Jones Beach. Those bridges are the standard case for artifacts that can be judged “for the ways in which they can embody specific forms of power and authority” (Winner 2014, 668), a design producing consequences “logically and temporally prior to any of its professed uses” (671). Jacques Ellul looks behind appearances, for an essence of technique that a philosopher can unveil. The access clause sorts them easily.

Bernward Joerges checked the bridges in Winner’s account. The design version of the story rests on two interviews in Robert Caro’s biography, and Joerges concludes that “one might as well assume that such an intention never existed” (Joerges 1999, 416); as for whether Moses was a “structural racist,” Joerges writes, “it does not matter much whether he was or not” (418). The racism is the familiar account, the part people carry away and want to run with, and the reason is instructive: it’s a verdict on a mind, which is what we’re practiced at judging. The anecdote trades on our response to an interior no document shows, rather than on the evidence of what the bridges did. Once the question becomes whether Moses meant it, the argument turns into a trial of his character: a conviction would changes nothing about what the bridges did, and neither does an acquittal.

Both outcomes obscure the power relation. It “is not to be found in the formal attributes of these things themselves. Only their authorization, their legitimate representation, gives shape to the definitive effects they may have” (Joerges 424). The bridges “represent legitimate or contested property and access rights which as a rule change over time” (Joerges 424). The custody question I arrived at last time here travels through the sociology of technology instead of from a keystroke log. Joerges overcorrects, though: “Only rarely and in the most trivial senses,” he says, “can one show that such constraints are coupled to building form” (424).

I’d say there’s only one thing you can do with a Hellfire missile. Where a technology is monostable, form binds. Winner already had the scale, distinguishing an arrangement that “requires” a form of politics from one that is “strongly compatible with, but does not strictly require” it (674). An honest use of the scale would be to ask where on it the technology sits, read from its actual use cases rather than from what someone meant.

Call it the use record: what a technology has been used for, by whom, and under whose authorization. Values are read from provenance, not intent. Alex Reid made the point this month with an inclined plane, a ramp to one person and a wall to another. His questions—“good for whom, under what conditions, and at whose expense?”—are use-record questions, and none of them needs a designer’s mind. Turn the question of the Hellfire missile on writing. Is there only one thing you can do with your output? No. Writing is multistable, as its record shows: the same page is a unit an assessor counts and an experience good its writer consumes in the writing of it, and the theory has to hold for the second use as much as the first; for the diarist and the journalist, the teacher writing an end comment and the weblogger counting his words; not only for the student whose page is bound for a rubric. The count and the account are both in the record. A technology whose record holds both can prevent its values from being read from either one alone.

Ellul’s theory takes the hardest line. Ellul’s assertion that “[t]he instant something is technological, it is legitimate, and any challenge is suspect” (Ellul 2014a, 439) is a claim about the record, about how technologies have been received. The claim holds, and I’ve wondered how much of the last few years’ arguments about language models is that sentence playing out on schedule. I decline what Ellul extends—a technique that runs autonomously from the economy—by asking, which or what economy? Gibson-Graham propose that the discursive-material construct of “the economy” is technology’s product; the maker is not freed of the product. Ellul offers his own access clause—“I do not deny the existence of individual action or of some inner sphere of freedom. I merely hold that these are not discernible at the most general level of analysis” (2014b, 206)—and then reads through his access clause anyway, from the general level, to an imagined nature the record doesn’t show. What survives the razor, then, is brief. Values read from the use record. Discretization, an operation performed on the signal, requires no interior at all.

The loom

Stiegler carries the word a step past speech. “It is in this way that logos is constituted: as the discretization of the continuous flow of language which, spatialized, can then be considered analytically,” he writes, and continues: “But this discretization of flows also affects gestures. The discretization of gesture was given concrete expression with the application of Jacques de Vaucanson’s automation technology to the Jacquard loom, and became generalized in the form of the industrial revolution” (Stiegler 2010a, 10). Writing slices speech; the loom’s cards slice the weaver’s gesture, one row of the pattern per card; and a gesture sliced that finely can be stored, copied and run without the weaver. Text and textile share a root for a reason.

Ada Lovelace saw the scene from the other side of the loom’s card. In her notes to Menabrea’s account of Babbage’s Analytical Engine, Lovelace writes that “the Analytical Engine weaves algebraical patterns just as the Jacquard-loom weaves flowers and leaves” (Note A). In her note that works the Bernoulli numbers through the engine, she set the limit Turing continued to dispute a century later:

The Analytical Engine has no pretensions whatever to originate any thing. It can do whatever we know how to order it to perform. It can follow analysis; but it has no power of anticipating any analytical relations or truths. Its province is to assist us in making available what we are already acquainted with (Note G).

When read with this post’s terms in hand, Lovelace offers a claim about the use record: the engine’s provenance is whatever we already know how to order, and its values come from ours. Behind the engine stands Charles Babbage, who came to the loom’s cards not as a philosopher of gesture but as an economist of labor, wanting to know what a task costs once it has been sliced fine enough for a machine to address it. Don Ihde’s phenomenology of technology arrives at alterity, whether a technology can be an other, which in this post’s terms would be the discretization of a bounded interior. Lovelace therefore offers a reduced form of Ihde’s question: what can we ask of a machine?

References

DeVoss, Dànielle Nicole, Ellen Cushman, and Jeffrey T. Grabill. 2005. “Infrastructure and Composing: The When of New-Media Writing.” College Composition and Communication 57 (1): 14–44.

Ellul, Jacques. 2014a. “The ‘Autonomy’ of the Technological Phenomenon.” In Philosophy of Technology: The Technological Condition: An Anthology, 2nd ed., edited by Robert C. Scharff and Val Dusek, 430–41. Malden, MA: Wiley Blackwell.

Ellul, Jacques. 2014b. “On the Aims of a Philosophy of Technology.” In Philosophy of Technology: The Technological Condition: An Anthology, 2nd ed., edited by Robert C. Scharff and Val Dusek, 205–9. Malden, MA: Wiley Blackwell.

Feenberg, Andrew. 2002. Transforming Technology: A Critical Theory Revisited. New York: Oxford University Press.

Gibson-Graham, J. K. 2005. “Economy.” In New Keywords: A Revised Vocabulary of Culture and Society, edited by Tony Bennett, Lawrence Grossberg, and Meaghan Morris, 92–97. Malden, MA: Blackwell.

Gleick, James. 2011. The Information: A History, a Theory, a Flood. New York: Pantheon.

Horner, Bruce. 2000. Terms of Work for Composition: A Materialist Critique. Albany: State University of New York Press.

Horning, Alice S. 2007. “The Definitive Article on Class Size.” WPA: Writing Program Administration 31 (1–2): 11–34.

Ihde, Don. (1990) 2014. “A Phenomenology of Technics.” In Philosophy of Technology: The Technological Condition: An Anthology, 2nd ed., edited by Robert C. Scharff and Val Dusek, 539–60. Malden, MA: Wiley Blackwell.

Joerges, Bernward. 1999. “Do Politics Have Artefacts?” Social Studies of Science 29 (3): 411–31.

Kline, Stephen J. (1985) 2003. “What Is Technology?” In Philosophy of Technology: The Technological Condition, edited by Robert C. Scharff and Val Dusek, 210–12. Malden, MA: Blackwell. First published in Bulletin of Science, Technology & Society 5 (3) (1985): 215–18.

Ong, Walter J. 2012. Orality and Literacy. 30th anniversary ed., with additional chapters by John Hartley. London: Routledge.

Ortega, James M. (1972) 1990. Numerical Analysis: A Second Course. Classics in Applied Mathematics 3. Philadelphia: SIAM. First published New York: Academic Press, 1972.

Ridolfo, Jim, and Dànielle Nicole DeVoss. 2009. “Composing for Recomposition: Rhetorical Velocity and Delivery.” Kairos: A Journal of Rhetoric, Technology, and Pedagogy 13 (2).

Stiegler, Bernard. 2010a. For a New Critique of Political Economy. Translated by Daniel Ross. New York: Polity.

Stiegler, Bernard. 2010b. “Memory.” In Critical Terms for Media Studies, edited by W. J. T. Mitchell and Mark B. N. Hansen, 64–87. Chicago: University of Chicago Press.

Trimbur, John. 2000. “Composition and the Circulation of Writing.” College Composition and Communication 52 (2): 188–219.

Wang, Avery Li-Chun. 2003. “An Industrial-Strength Audio Search Algorithm.” In Proceedings of the 4th International Conference on Music Information Retrieval (ISMIR 2003), 7–13. Baltimore.

Winner, Langdon. (2014. “Do Artifacts Have Politics?” In Philosophy of Technology: The Technological Condition: An Anthology, 2nd ed., edited by Robert C. Scharff and Val Dusek, 668–79. Malden, MA: Wiley Blackwell.