Virtual Society?

• A cultural-historical account of civil engineering and its tools;
• Location of the site studied here with respect to its particular histories, including relevant details of the political and economic circumstances of the project at hand;
• An account of the extended actor networks that make up the project, including members’ own orientation to demands of engagement in time and across space;
• An account of the project’s work organization and divisions of labor;
• An account of my own circumstances, as a researcher, in engaging with the project.

With that said, the limited focus of this paper is on that aspect of engineering work that comprises the production of exhibits and plans, and the place of computer-aided design tools and paper-based drawings as mediators of that form of engineering practice.

Ethnomethodological studies of work are concerned with what Garfinkel (1996), following Gurwitsch (1964), has named the "phenomenal field properties" of particular work sites and their practices. Emphasis is on the irreducible relations of mutually constitutive details, through which isolable actions, objects, artifacts, and the like take on their specific, practical significances. Meanings on this view inhere neither in individual elements or properties, nor in some underlying structure that stands behind appearances, but only in relations of "mutual reference" across a field of observable phenomena (Lynch 1993, p. 127). Ethnomethodology adds to this field the necessary presence of the embodied subject, through whose history and present engagement phenomenal relations are enlivened and made relevant to some ongoing activity. Moreover, the phenomenal field of action does not simply pre-exist and take its meaning from activity, but is reflexively generated through the same activity that it organizes, as found objects are appropriated and mobilized and new objects created.

The mutually constitutive relation of actions and their environments includes the fact that accounts of activity are themselves crafted from the juxtaposition of observable features of embodied actions with phenomena selected from the scene in progress (Goodwin in prep, p. 1). This applies equally to accounts that are internal to a given activity, as to those created about it in advance or afterwards. For ethnomethodology, then, the relationship between social practices and accounts of those practices is deeply and unavoidably a reflexive one, for participants and observers alike (Button and Sharrock 1998; Lynch 1993, p. 1). And like material artifacts, formulations of action – whether done as part of an activity or as accounts of it by participants or others – are specifically situated in the occasions of their production and use. Together talk and other culturally formulated, socially and materially constituted artifacts comprise the phenomenal field of embodied practice.

Historian of engineering Henry Petroski writes that for a civil engineer, the design of a bridge is the stuff that dreams are made of (1995). He emphasizes as well the significance of the co-development of modern engineering practice and the artifacts of inscription and persuasion that have become as much its stock in trade as concrete and steel:

In the association [in the mid-nineteenth century] of bridge building with drawing and calculation and written argument before any construction was started, a new era was begun. From then on, the grandest dreams could be articulated and tested on paper, and thereby communicated to those who would have to approve, support, finance, and assist in designing a project that could eventually take years, if not decades, of planning and construction (ibid., p. 12).

Bridges are built rarely compared with roadways and other surface structures, insofar as bridges are costly projects that last on the order of thirty to in some cases hundreds of years. In the area where our study is located, six toll bridges have already been built and no additional bridges are planned. At the same time, the area is threatened with earthquakes. In response to the critical problems experienced in the last major earthquake, the state government has set aside substantial funds for "seismic retrofitting" of the existing toll bridges. One of the area’s toll bridges is actually a pair of old trestle bridges that connect the north and south shores of a relatively narrow strait. Charged with ensuring the safety of these bridges, engineers at the state Department of Transportation have argued that while one of the bridges can indeed be retrofitted, the other is sufficiently old – dating from 1927 – that it is both unsafe and uneconomical to try to reinforce it. Instead, they have proposed "replacement as a retrofit strategy." In this way they are able to direct funds for retrofitting to a new bridge building project.

At the same time, it is a bit misleading to say that the engineers are engaged in designing a bridge if what we imagine by that is the design of the structure itself. In this case, in fact, the bridge design is outsourced to a specialist design firm, with Department engineers responsible for oversight. But it also turns out that the bridge itself represents a small fraction of the entire project relative to the highway approaches and interchanges that tie the bridge into the landmasses that it connects. So while the design of the bridge is contracted out, Department engineers maintain responsibility for the bridge alignments and all connecting roadways. These make up the focus of their design work.

My analysis of engineering design work is based in a set of tutorials provided by Andrea, a lead engineer on the Bridge Replacement project team. The first of these took place in front of her computer-aided design (CAD) workstation, where she took us on a tour of some of her recent work. The object of her activity on this occasion was the earth: in Bødker’s terms an object physically existing, but present to Andrea only as the rendering of it provided by the computer system. More specifically, Andrea was engaged in figuring the volumes of dirt that would need to be displaced in order to construct the highway interchange on the bridge’s south side. Andrea’s purpose at hand was to calculate, as she put it, "how much dirt we’re going to dig up – literally, not just as a figure of speech!"

The interface that mediates Andrea’s access to her work’s objects is actually composed of two software applications running together. An engineering application is layered on top of, or nested within, the functionality of a second, graphics application. The engineering application, in Andrea’s words "uses the [graphics application] to let you see the results of actual engineering calculations." The layered space of the CAD environment includes as well a collection of menus arranged as a kind of frame around the periphery of the CAD workspace. Some of these menus serve as views onto the directory structure of Andrea’s hard disk and provide the means by which files are located and opened, others provide "tool boxes" of available actions to be taken on graphical objects within those files, while a third controls the layered space of the CAD display itself. When focus shifts to these menus as objects, they become a top layer, superimposed on the objects they are used to manipulate.

[Figure 1: CAD interface with toolbars and menus]

Latour (1990, pp. 52-54) points out that representational conventions in engineering are aimed at maintaining an "optical consistency" between three-dimensional objects and the flattened renditions that comprise sketches, plans and the like. These conventions include highly elaborated lexicons of line types, perspectives, geometries and symbols. In the case of mechanical engineering, Henderson (1991) observes that "[t]he lexicon allows the schematic drawings to remain flexible enough that engineers can read the coded functions in the layout and understand the interrelations of the various functional components of the whole project" (p. 459). Fundamental to civil engineering in this regard is the "plan view" or "horizontal alignment," which flattens engineering objects into something akin to a bird’s eye perspective. In roadway work, the plan view relies on a geometric object called the "PI" or "point of intersection." The PI in turn references a virtual grid laid over the mapped physical environment, establishing a series of points in space that mark the place where, as Andrea describes it, "a straight line meets a straight line." A second focal object is the "control point," which she describes as a place of maximum constraint that consequently "controls" much of the design.

[Figure 2: Andrea indicates control point]

As Andrea explains this:

And then, you always have some control points. We have a major control point on Vista (an existing surface street) which is, right here (pointing to indicate curving street to left of ramp) where it goes underneath that ramp, that controls really so much of the entire design on the new road ... We have to have clearance for trucks to go under it, and while they're building this new ramp they have something called false work up? Which has it's own depth, it might be three feet it might be six feet deep. And then there's the depth of the structure. So we know what the elevation is on the ramp, right at this point (leaning in to show) actually its control point turns out to be right here. But we have to then add that there's a depth to that structure, then there's gonna be some false work, so we have to go way under it.

Andrea’s account of the problem makes clear that highways and bridges are not self-standing objects but structural elements that are laid over and must be effectively incorporated into existing landscapes. In the case of civil engineering, moreover, the spatial field of objects is complicated by the element of time. Andrea’s calculations must take into account not only the plans for the new ramp, but also the temporary structures, or "false work," required for its construction. All of this in turn must be placed in relation to existing landscapes, made up not only of natural features such as geological formations, waterways and the like, but of strata of built environments laid down over a period that may comprise hundreds of years. The pace of this latter building is accelerating, moreover, in such a way that each new project confronts an increasingly dense infrastructural archeology, including prior structures, utilities, waste disposal sites and even areas protected for recovery from previous interventions.

All of these features must be accurately mapped in order for the soundness of a design to be ensured. The graphical renderings of the plan view can be interrelated with a Digital Terrain Map, which renders the 3-D contours of the physical environment in which the objects of design will actually be built, and into which they need to fit. Maps are created from survey points, assembled together through the use of conventional symbols that render the topography of the original ground. Like the ground, the Digital Terrain Map appears as a kind of bottom layer that sits "beneath" the design objects themselves. As with the engineering and graphics applications, however, the layers are not simply superimposed but dynamically cross-referential. Specifically, once the horizontal alignment or plan view has been created engineers need to generate a "profile," a rendering that makes visible the relationship between the proposed new roadways and the existing terrain. By drawing "cuts" through a particular section of the plan view of the site, Andrea is able effectively to "instruct" the engineering application to create a series of cross sections for each of those cuts showing where the existing surface is, using the map as a reference.

Andrea’s workspace, in sum, is made up of an assemblage of computational, metrological, geometric, cartographic and graphical tools. Together these comprise the interface through which she sees and manipulates the physical objects of her work. With reference to Bødker’s framework, Andrea works on the one hand with the elements of the layered interface that the CAD system provides, and through the interface to the objects that those various renderings mediate: in this case, the earth, the existing and projected roadways, and her team’s interests in them. Ethnomethodologically, our interest is in how these multiple elements and objects together comprise the phenomenal field properties of Andrea’s embodied practice. Having enumerated the distinctions among heterogeneous elements, in other words, the question becomes how in practice does Andrea bring them together?

In our tutorial Andrea takes us through her previous day’s work in a way that makes not only the heterogeneity but the integrity of her workspace clear. For example, she has pointed out to us one of what she calls the "major control points" for her design, a physical location on an existing roadway named Vista del Rio. As she explains it, a defining constraint of the design problem at hand is that one of the onramps to the highway must be built to run over Vista del Rio, an existing surface street. In order to provide enough clearance for the ramp, Vista must be effectively lowered below the current surface level by removing earth at the point where the street crosses under the ramp. As Andrea guides us through a "profile" of the site she explains further:

[Figure 3: Profile view]

And you can see that at the top of Vista we're pretty much following the existing ground. And as we go down, we get way below it, this is about ten meters of dirt that we're taking out. And then I think the point that we're crossing under is right on this (pointing with pencil) little flat here.

Andrea’s demonstration takes the "profile" as a locus of what Goodwin has named "professional vision," (1994, 1995), a site from which we can "see" the contours of a roadway far removed from the place where we sit in front of her workstation, and assess its relevance for the imagined future activity of constructing a new road. In this way the display acts as a surrogate for the physical place of engineering interest. Andrea’s narrative positions us figuratively on the physical site "at the top of Vista," from which we can "follow the existing ground." But "as we go down" this virtual roadway we enjoy the ability to continue our travels below the existing surface, removing ten meters of virtual dirt to reach the projected future crossing below. The latter is indicated by a geometric point in space, positioned figuratively under the new ramp and more literally on the flat line of the CAD display. Bringing together narrative form and imagination, metrology and geometry, Andrea is able to "see" under the existing ground, to project a newly excavated roadway that does not yet exist. In doing this work she moves fluidly between pictures and things and across time, as the artifacts and objects of her work are read through each other to achieve a rendition that aligns what is there now with its desired transformation. At the same time Andrea’s small gesture, the point of her pencil, reminds us that it is with the engineer’s body that this work of virtual travel and assembly gets done.

That engineering objects mediate embodied practices of engineering is clear. By looking more closely we can see as well how bodies mediate engineering objects. So in the course of her tutorial Andrea makes continuous use of various forms of what Goodwin (1994) has named "highlighting for perception," instructing us on where and how to look with the gestures of her pencil. At other times the performative aspects of her reading serve to animate the static CAD image that we see. So, for example, once she has used the system to create a series of cross sections of a roadway, say every five or twenty meters, she can then effectively "travel" along the roadway by scrolling through the sections displayed on her screen. At still other times her body itself becomes a reference, adding a kind of third dimension to the CAD screen as when, for example, she uses the angle from her hand to her elbow to demonstrate the slope of a road.

As computer-aided design has become an increasingly central aspect of engineering practice, a perspicuous site for seeing these bodily mediations is in the relations and differences between working at the CAD station and on paper. One obvious way of understanding these differences is in terms of the relatively greater scale and expanse of paper. One engineer described this to us vividly, as she enacted with gesture the difference between sitting in front of the CAD station, elbows close in to the sides of her body, hands constrained within the narrow terrain of the keyboard, eyes glued to the screen on which she zoomed in and out and traveled "across" the virtual space with mouse clicks, then contrasted that with sitting or standing over a large sheet of paper, arms outstretched or hands and arms engaged in a variety of actions of drawing, measuring, turning the paper to get another angle, moving it slightly on the table and so on.

At the same time, it is also the case that Andrea identifies a great benefit of CAD as the effectively unbounded (other than by the size of files) virtual space that CAD’s zooming and scrolling functions provide. Andrea is able to create an extended workspace, a kind of spatially arrayed "library" or storehouse of her work, made up of shrunken images that she can browse, select from and expand. In this way all of Andrea’s cumulative productions – what would comprise a bulky collection of plan drawings each at the scale of 24" by 36" – can be surveyed on a single screen simply by saying "show me everything that’s in this file at once."

Another obvious difference between paper and CAD-based work practices could be that older engineers choose the former, younger engineers more familiar with digital media choose the latter. Indeed, Andrea explained to us that in the previous week she had prepared a set of paper plans in order to bring a problem for consultation to one of her more senior colleagues who works only on paper. At the same time, she herself also frequently turns to paper in the course of her work. A week after our tutorial I noticed her working at her drawing table with an array of paper documents spread out around her and asked her to tell me about them. She explained that she and two of her colleagues had sat down several days before to, as she put it, "nail down" the design of the highway interchange on the bridge’s south side. The primary documents were a set of three plan views taped together.

[Figure 4: Paper plan views]

In this case Andrea explained that while she could have done her design work with a smaller image, she wanted as she put it a "meaty" scale:

So I could really have a good picture of what’s going on. When you’re doing the design a tiny postcard of it is not that helpful. This is the whole interchange area.

The assembled plan view, while still a minute fraction of the size of the physical area that it renders, extends what would be available within the limits of the CAD screen to something that becomes a space for joint work. Through it the object of Andrea’s work is both viewable as a "whole" and still within arms’ reach. Andrea describes the annotations evident on the paper plans as the residue of engineers’ "thinking with a pencil in their hand." In addition to the annotated plans, I asked Andrea about a pad of graph paper sitting on top of the other sheets. She explained that she uses the pad for her calculations:

What I’m trying to do on the pad is something that seems like an extra step, but personally I think that it’s pretty important. I’m just trying to record the calculations that I’m doing, to determine clearances and to determine actual elevations. [explains specific problem] So I could just do the whole thing with my calculator, and that’s what a lot of people do. I’d come up with the number and then I would write down the number and start working. But we’re at the point where we need to check it. That’s why I’m trying to be a bit more meticulous ... The way that I see it is that the big advantage of working it out on paper is that you’re leaving a bit of a trail as to how you got to the thing that’s the final answer. We also save previous iterations electronically, but they’re not well documented. There’s not an easy way to go back and say, did anyone ever try putting a 2% grade on this? I think it’s nice to see on paper Oh, look, she tried 2, 2.1, 2.2, 2.3 and 2.5, look, it goes right through the point that she needed to hit, and it works.

The object of Andrea’s activity in this respect includes not only to find the requisite grade for the roadway, but to produce a residual trace of her actions as a visible rendering of the calculative work that she has done. Another particularity of paper, then, is that work "black-boxed" by the machinery of the electronic calculator can be made visible, and in relation to the objects that it references. The engineers’ pad serves not only as a space for calculation but as a technology of accountability that makes the course of her work retrospectively visible to her colleagues.

In elaborating the benefits of paper as a medium Andrea explained that six months ago she tried to do more with CAD, but now has realized that paper is just better for some things. Andrea’s tutorial instructs us that rather than a simple progression from paper to CAD, the maturing of electronically-based engineering practice may emerge as the informed, selective use of both paper and digital media, based in a deepening understanding of their particularities and of their effective interrelationship. CAD might be seen, moroever, as the logical extension, the embedding into a computational instrument, of what the early progenitors of the rules of engineering drawing took to be the benefits of descriptive geometry and other conventions; that is, their force as a corrective to the artisans "ignorant and prejudiced" imagination (Bachelier 1768, quoted in Alder 1998, p. 512). It becomes clear from our observations of the actual use of CAD as an aspect of Andrea’s practice that the calculative powers of the machinery to "make things the same" (Alder 1998.) are effective only insofar as they are enlivened by her readings of the objects rendered, in and through the interface of her workstation, as the highly differentiated, more and less obdurate materials of a fully embodied, natural/artifactual world.

My thanks in the writing of this paper go first to Andrea for her dedication and gifts as an engineer, her humor and eloquence in speaking about her work, and her generosity in taking the time to help us make sense of it all. I’m indebted as well to the rest of the members of the engineering project team, and to my colleagues Jeanette Blomberg and Randy Trigg with whom this research has been a thoroughgoing collaboration. Finally, thanks to Chuck Goodwin, Aug Nishizaka and Naoki Ueno for their helpful comments on an earlier draft.

Alder, Ken (1998) Making Things the Same: Representation, Tolerance and the end of the Ancien Regime in France. Social Studies of Science 28 (4), August, pp. 499-545.

Akrich, Madeleine (1992) The De-Scription of Technical Objects. In Bijker and Law (eds.) Shaping Technology/Building Society. Cambridge, MA: MIT Press, pp. 205-224.

Blomberg, Jeanette, Lucy Suchman and Randall Trigg (1996) Reflections on a Work-Oriented Design Project. Human-Computer Interaction, Volume 11, pp. 237-265.

Bødker, Susanne (1991) Through the Interface: A Human Activity Approach to User Interface Design. Hillsdale, NJ: Lawrence Erlbaum.

Bødker, Susanne (1996) Applying Activity Theory to Video Analysis: How to make sense of video data in human-computer interaction. In Nardi, B. (ed.) Context and Consciousness. Cambridge, MA: MIT Press, pp. 147-174.

Button, Graham and Wes Sharrock (1998) The Organizational Accountability of Technological Work. Social Studies of Science Vol. 28, No. 1, pp. 73-102.

Clifford, James and George Marcus (eds.) (1986) Writing Culture: The Poetics and Politics of Ethnography. Berkeley, CA: University of California Press.

Cole, Michael and Yrjo Engestrom (1993) A cultural-historical approach to distributed cognition. In G. Salomon (ed.) Distributed Cognition: Psychological and Educational Considerations. Cambridge, UK: Cambridge University Press, pp. 1-47.

Engestrom, Yrjo (1990) Learning, Working and Imagining: Twelve Studies in Activity Theory. Helsinki: Orienta-Konsultit.

Garfinkel, Harold (ed.) (1986) Ethnomethodological Studies of Work. London and New York: Routledge and Kegan Paul.

Garfinkel, Harold (1996) Ethnomethodology’s Program. Social Psychology Quarterly, Vol. 59, No. 1: 5-21.

Goodwin, Charles (1994) Professional Vision. American Anthropologist, Volume 96, No. 3, pp. 606-633.

Goodwin, Charles (1995) Seeing in Depth. Social Studies of Science, Vol. 25, pp. 237-274.

Goodwin, Charles (in preparation) Pointing as Situated Practice.

Gupta, Akhil and James Ferguson (eds.) (1997) Anthropological Locations: Boundaries and Grounds of a Field Science. Berkeley: University of California Press.

Gurwitsch, Aron (1964) The Field of Consciousness. Pittsburgh: Duquesne University Press.

Henderson, Kathryn (1991) Flexible Sketches and Inflexible Data Bases: Visual Communication, Conscription Devices, and Boundary Objects in Design Engineering. Science, Technology & Human Values 16(4): 448-473.

Henderson, Kathryn (1999) On Line and On Paper: Visual Representations, Visual Culture, and Computer Graphics in Design Engineering. Cambridge, MA: MIT Press.

Heritage, John (1984) Garkinkel and Ethnomethodology. Cambridge, UK: Polity Press.

Kuutti, Kari (1996) Activity Theory as a Potential Framework for Human-Computer Interaction Research. In Nardi, B. (ed.) Context and Consciousness. Cambridge, MA: MIT Press, pp. 17-44.

Latour, Bruno (1990) Drawing Things Together. In M. Lynch and S. Woolgar (eds.) Representation in Scientific Practice. Cambridge, MA: MIT Press, pp. 19-68.

Livingston, Eric (1987) Making Sense of Ethnomethodology. London and New York: Routledge and Kegan Paul.

Lynch, Michael (1993) Scientific Practice and Ordinary Action: Ethnomethodology and social studies of science. Cambridge, UK: Cambridge University Press.

Lynch, Michael (1997) Silence in Context: Ethnomethodology at the margins of social theory. Paper delivered at the Conference "Ethnomethodology East and West," Waseda University, Tokyo, Japan, 20-22 August.

Marcus, George and Michael Fischer (1986) Anthropology as Cultural Critique. Chicago: University of Chicago Press.

Mialet, Helene (1999) Do Angels Have Bodies? Two Stories About Subjectivity in Science : The Cases of William X and Mister H . Social Studies of Science, Vol. 29, No. 4, pp. 551-582.

Nardi, Bonnie (1996a) Preface. In Nardi, B. (ed.) Context and Consciousness. Cambridge, MA: MIT Press, pp. xi-xii.

Nardi, Bonnie (1996b) Activity Theory and Human-Computer Interaction. In Nardi, B. (ed.) Context and Consciousness. Cambridge, MA: MIT Press, pp. 7-16.

Petroski, Henry (1995) Engineers of Dreams: Great Bridge Builders and the Spanning of America. New York: Random House.

Sims, Benjamin (1999) Concrete Practices: Testing in an Earthquake-Engineering Laboratory. Social Studies of Science, Vol. 29, No. 4, pp. 483-518.

Star, Susan Leigh (1989) Layered Space, formal representations and long-distance control. Fundamena Scientiae, Vol. 10, No. 2, pp. 125-155.

Suchman, Lucy (in press) Organizing Alignment: the case of bridge-building. To appear in the special issue of Organization on "Knowing in Practice" S. Gherardi (ed.)

Suchman, Lucy, Randall Trigg and Jeanette Blomberg (1998) Working Artifacts: Ethnomethods of the Prototype. Paper presented at the annual meetings of the American Sociological Association, San Francisco, August 23.

Trigg, Randall, Jeanette Blomberg and Lucy Suchman (1999) Moving Document Collections Online: The evolution of a shared repository. In Proceedings of the European Conference on Computer-Supported Cooperative Work (ECSCW ’99), Copenhagen, DK, September, pp. 331-350.

Workshop programme

Registration form