Formalization as a Social Project

Philip E. Agre
Department of Information Studies
University of California, Los Angeles
Los Angeles, California 90095-1520

Published in the Quarterly Newsletter of the Laboratory of Comparative Human Cognition 14(1), 1992, pages 25-27. Copyright 1992 by LCHC.


Formalization is, among other things, a discursive operation. One can talk about bridges and shops and cold fronts in many ways and for many purposes. Technical people, though, want to talk about these things in a particular way -- formally -- and for a particular sort of purpose -- instrumental appropriation. Technical discourse stands in special relationships to other kinds of discourse, most especially what I will provisionally call `ordinary language'. Technical discourse, much more than ordinary forms of language, participates in strictly regulated structures of authority and ownership: only an accredited technician can use technical language felicitously, yet the goals and products of technical work are explicitly the property of the institutions that pay the bills. At the same time, technological research often marks its progress in relation to the world of ordinary experience, precisely by its success in applying formal reason to matters previously left to the informal judgement of ordinary people. Presentations of technical research papers are often organized around this transition from ordinary experience to formal reconstruction. In the course of this transition, they construct a definite social organization of language use around a series of oppositions: ordinary versus technical, informal versus formal, vague versus precise, unaccredited versus accredited, and unregulated versus appropriable. As an example, consider the following passage, about which I have written elsewhere [1]:

Consider how an ordinary day is put together. You awaken, and as you lie in bed, or perhaps as you move slowly about in a protective shell of morning habits, you think about what the day will be like---it will be hot, it will be cold; there is too much to do, there is nothing to fill the time; you promised to see him, she may be there again today. If you are compulsive, you may worry about fitting it all in, you may make a list of things to do. Or you may launch yourself into the day with no clear notion of what you are going to do or how long it will take. But, whether it is crowded or empty, novel or routine, uniform or varied, your day has a structure of its own---it fits into the texture of your life. And as you think what your day will hold, you construct a plan to meet it. What you expect to happen foreshadows what you expect to do [2].
This is the opening paragraph of one of the most influential early works of artificial intelligence. The theme of ordinariness appears in this passage in several ways. The opening sentence appeals to us to consider how an ordinary day is put together. The ordinariness of this day is motivated through lists of mundane considerations, which together serve quite compactly to project a definite social order, not least with reference to social relations of gender. One is reasoning about one's day, and the ordinariness of this reasoning is conjured through the colloquial phrases and lilting rhythms of the language. At the same time, the paragraph also enacts a trajectory from these ordinary modes of language and reason to the technical modes of language and reason that will occupy the book. This trajectory will be obvious to the trained technical reader, but it is otherwise wholly obscure. It helps to know that the project is to motivate a technical notion of `plan', roughly a computer program in one's head that is the cause of any given structure or regularity in one's outward behavior [3]. It is this knowledge, for example, that allows one to make sense of the overt contradiction between the propositions that ``you may launch yourself into the day with no clear notion of what you are going to do or how long it will take'' and that ``as you think what your day will hold, you construct a plan to meet it'' -- the answer being that an empty plan is still a plan, in the technical sense, just as an empty set in mathematics is a set.

Technical language, as I will use the term, involves a systematic ambiguity between two domains of reference: that of the territory being formalized and that of the Platonic realm of mathematics. One paradigm case of technical language occurs in story problems. The snails and horses in a story problem function as mere ciphers in an allegory of integers and cosines [4]. Story problems thus call for special forms of reading, not ordinary imaginative story-reading but a sort of decoding that reveals the math behind the narrative. Political allegories tell of farm animals; technical allegories tell of scenes from ordinary life. Yet whereas a political allegory can be decoded to obtain a critique of an actual king [5], technical language is strangely dependent on the `ordinary' language which it constructs and then purports to replace. The integers and cosines in a story problem are abstract mathematical quantities: they are not themselves the cables and pulleys and engines and bearings which figure in the actual practical problem that forms the instrumental test of technical reason. A formalization, in short, can never exhaust its object. In order to speak of actual material things, technical discourse must rely on its embedding in the discourses and practices of the surrounding institutional context. If economics is to speak of a shop's finances, or if meteorology is to speak of an advancing front, or if mechanical engineering is going to tell us whether a certain bridge will remain standing, then somebody must engage in the situated interpretive effort to gloss his or her concrete circumstances in ways that can be assimiliated to technical terminology. Different technical fields understand this necessary interpretive effort differently, but none of them understands it very well [6].

In this light, let us reconsider the ideological oppositions that, I have claimed, structure technical discourse. The discourse of formalization constructs itself as technical, formal, precise, accredited, and appropriable, and it constructs its other as ordinary, informal, vague, unaccredited, and unregulated. This figurative alter-discourse, it turns out, is not the discourse of `ordinary' people. It is, rather, the shadow side of technical discourse itself. Technical language is perfectly formal and precise in its reference to mathematical quantities (integers and cosines), but it is also extremely informal and vague in its reference to the objects of formalization. The use of a technical word like `plan', for example, is constrained only by the possibility of producing a technical gloss that `works' -- that is, that is understood by some sponsoring agency to solve its problems. In this sense, technical language is not simply indexical (like any type of language) but is in fact continually reorganized within a reflexive totality wholly subordinated to the specific institutional setting whose `problem' demands a `solution'.

Technical language, then, figures in a project of instrumental appropriation. This project has a dual structure, in which `the technical' figures as an ideological `high' term and `the ordinary' figures as an ideological `low'. `The technical' marks itself as privileged in relation to `the ordinary', and moreover makes `the ordinary' the object of a colonial project of reform. Yet in reality these opposed terms are two sides of the same coin, joined in a relationship that cannot be either stabilized or eliminated. What the social project of formalization produces is this coin itself [7]. The raw material for this project, the actual pre-technical state of language and practice, is neither `technical' nor `ordinary', neither `precise' nor `vague', indifferent to accreditation, and regulated not by the universal leveling of pure appropriability but by whatever configuration of power and ideology happens to operate in a given social locality.

The manifestations of this social project are clear enough, but who exactly is the subject of this project? Whose project is it? It is easy enough to find positivist fundamentalists who profess such views; many of them have even dedicated their lives to the ideologically conceived project of reform which is the banner of formalization. A substantial twentieth-century tradition, all the way across German philosophy from the phenomenologists to the Frankfurt School, has been willing to identify technical rationality as the defining trend of the age. But this is a mistake, for two reasons. Formalization is not the only colonial project on the block; marketing, for example, is a powerfully driven social force with its own designs on the conceptions of self and other of `ordinary' people. Indeed, within the average large corporation one can readily identify the conflictual border between these contrasting colonial practices, each with its gaze set on its own ideological objects. And in relation to each of these social projects one can readily discern resistance. The very contradiction between the two of them provides abundant evidence for the social cognition and political consciousness that would identify and refuse them both. The process, however, is neither simple nor uniform. The disparate forces and their contradictions show up in different ways in each of the steadily multiplying localized sites of social practice: in professions, in laboratories, in medical examinations, in front of the television. The challenge for critical thought is to identify the multiplicity and the unity of these disparate sites of social practice and to respecify the specialized epistemology of the accredited practitioner as the political consciousness of the disciplined subject.


This is a slight revision of a position paper I wrote for a symposium entitled ``Knowing and Knowledge: Re-Specifying the Role of Formalization in Computer and Social Science,'' organized by Elin Pedersen and Lucy Suchman and held at Oksnoen, Norway in June 1991. I have so much writing to do that I am unlikely to publish its ideas at full length for several years. So I hope that this informal version might be of some interest in the meantime. I am grateful for comments by the participants in the Oksnoen workshop, in particular John Bowers and Reinhard Keil-Slawik.

[1] See Agre (1990). Copies are available from me at or Department of Communication D-003; UC San Diego; La Jolla CA 92093.

[2] Miller, Galanter, and Pribram (1960: 5).

[3] For a general discussion of plans in the technical tradition descending from Miller, Galanter, and Pribram, see Agre and Chapman (1991). I will also discuss this tradition more fully in Agre (forthcoming).

[4] For a fascinating account of this point and its consequences for math teaching see Voigt (1989). See also Robinson (1991) and Walkerdine (1988).

[5] This isn't really right about allegory. For a historical survey of more interesting views see MacQueen (1970). For the contemporary interest in the general issue of allegory and representation see Greenblatt (1981).

[6] This point is terribly difficult to state accurately, much less to defend. I will take it up at greater length in a paper on technical language currently in preparation.

[7] This argument was inspired by Stallybrass and White (1986).


Philip E. Agre, Portents of planning: A critical reading of the first paragraph of Miller, Galanter, and Pribram's ``Plans and the Structure of Behavior'', paper presented at the Conference on Narrative in the Human Sciences, University of Iowa, July 1990.

Philip E. Agre, Computation and Human Experience, Cambridge: Cambridge University Press, forthcoming.

Philip E. Agre and David Chapman, What are plans for?, in Pattie Maes, ed, Designing Autonomous Agents, Cambridge: MIT Press, 1991.

Stephen J. Greenblatt, Allegory and Representation: Selected Papers from the English Institute, 1979-80, Baltimore: Johns Hopkins University Press, 1981.

John MacQueen, Allegory, London: Methuen, 1970.

George A. Miller, Eugene Galanter, and Karl H. Pribram, Plans and the Structure of Behavior, New York: Henry Holt and Company, 1960.

Mike Robinson, Double-level languages and co-operative working, AI and Society 5(1), 1991, pages 34-60.

Peter Stallybrass and Allon White, The Politics and Poetics of Transgression, London: Methuen, 1986.

Jorg Voigt, The social constitution of the mathematics province: A microethnographical study in classroom interaction, Quarterly Newsletter of the Laboratory of Comparative Human Cognition, 11(1-2), 1989, pages 27-34.

Valerie Walkerdine, The Mastery of Reason: Cognitive Development and the Production of Rationality, London: Routledge, 1988.

Go back to the top of the page