[Originally published 1st quarter 2006 in arcCA 06.1, “Imbedded Knowledge”]
Author Dana Buntrock has studied Japanese architecture and construction practices since 1987. She teaches construction in the Department of Architecture at the University of California, Berkeley, and is working on a second book on how conceptions of time affect the materials and detailing of contemporary architecture in Japan. This text, an excerpt from her 2001 book Japanese Architecture as a Collaborative Process: Opportunities in a Flexible Construction Culture (London: Spon Press, 2001), was the result of extensive research on construction sites in Japan in the late 1990s, supported in part by the National Science Foundation, the Japan Society for the Promotion of Science, the Japanese Ministry of Education, and the Graham Foundation. It is reprinted by permission of the author.
A survey of design and construction periods in Japan (as reported in Shinkenchiku) shows that about 20 percent of projects report overlapping design and construction phases. Yet this number is almost certainly low, since the process of design development and construction supervision is not distinctly separated. Often, project supervision must of necessity include design development, although it may not be reported as such. On all but the smallest projects, the builders and professional staff work together on site, during construction. This is in marked contrast to the norms for design and construction in most countries. In the U.K., the chairman of the National Vocation Qualification working group on architectural technology once stated baldly, “The job of the architect is complete within the initial third of the design process.”1 Robert Gutman, an anthropologist studying a 400,000-square-foot site in the U.K., concurred, reporting that during construction, “No architect ever appeared.”2 The United States is no better; Robert Greenstreet, Dean of the School of Architecture and Urban Planning at the University of Wisconsin-Milwaukee, in an essay entitled, “What Do I Do on Site Anyway?” said, “The short answer is, of course, relatively little, although there are numerous instances of architects acting, often in good faith, beyond the limitations of their contractual obligations and getting in all sorts of trouble.”3
In Japan, the contractors are on site from the beginning, and most professional staff move to the site about the time that the building’s foundation is completed. The representatives of each group are in frequent (often daily) contact and focus almost exclusively on the project at hand. Meetings with other relevant groups, including manufacturers and suppliers, also occur on site or in the factories and plants producing materials for the project. Fumihiko Maki portrays the site as a refuge: “The field office is not only a place for the liberation of the work of the architect from the world of thought, but also is a place where many people participate in the effort towards its crystallization.”4
The site offices and much of the equipment—even servers connecting all organizations on site—are supplied by the contractor, although architects may call for certain equipment in specifications. Whereas architects remain positive about the role of the construction site, both contractors and clients have begun to look closely at the costs of such a system, because of Japan’s current and prolonged recession. Because of this, skirmishes over the construction site have recently emerged: For example, the period of construction has been progressively shortened by clients, although architects to date have been able to respond simply by putting more staff on site. Contractors are not threatening the existence of site offices, so much as trying to rationalize costs, for example by reusing equipment that might otherwise be disposed of at the end of construction. The qualitative benefits, more difficult to measure, stand in stark contrast to the very clear costs of human labor. The site office, however, cannot be assessed on purely economic grounds. In another context, the anthropologist Thomas P. Rohlen noted:
The general emphasis on the group context of work is not entirely explained by an interest in supervision or some other technical consideration. It arises from something much more elemental—the inclination, found throughout Japanese society, to organize activities into small face-to-face groups, to enjoy this kind of environment, and to work efficiently in it.5
Office staff (including not only constructors and professionals, but also secretaries or others acting as “office ladies” whose role is to keep things neat and serve sweets and tea) tend to arrive on site as early as 8:00 a.m., and professionals generally do not leave until 11:00 p.m. or later. As laborers quit by early evening, the architect and other office staff can draw and make models in the evening, and supervise, or even occasionally participate in, construction during the day. The architect’s team may include at least one junior member with less than five years of practical experience. (This seems to be truer for design studios than commercial firms.) For these individuals, the opportunity to closely observe construction practices is certainly an important influence in their development as designers.
Architectural staff move back and forth between an area under consideration and their drawings or test an idea with full-scale models studied in place. Materials and colors under consideration can be viewed under the same lighting and in relation to other parts of a building prior to making a final decision, and the design team can observe proposed construction techniques before deciding what is most suitable. Architects are also able to observe views, quality of light, and relationships of parts of the building over a longer period of time and to react to those factors in the refinement of a design. It is understandably far easier to frame a view from existing construction than it is when working prior to the grading of a site, or building the structural frame.
This is because drawings have different uses in different construction cultures. In North America, the saying is that drawings are “quantity” and specifications are “quality.” The Japanese correlate would be to say that drawings broadly define what is built; the activities of the job site work out how. Many of the results of liability concerns found in North America—extensive pre-bid detail documentation, detailed specifications outlining quality and specific acceptable products or materials, and tightly written contracts—are simply not part of practice in Japan.
In an American office, pre-bid detail documentation accounts for around 40 percent or more of a typical set of construction drawings. By contrast, the documents produced in a Japanese office and used by contractors to prepare bids are simple. Most will be drawn at 1:100 (roughly approximate to 1/8″ = 1’0″), and details account for perhaps 5 to 10 percent of these drawings. Those that are drawn tend to be special features of the building, rather than illustrations of typical construction. Western architects who visit Japanese architectural offices often voice surprise at the sketchy quality of documents. One architect noted that, “Typically, drawings are not taken to the contract document stage that is common in the USA. In Japan, that phase just does not exist.”6
Furthermore, the document set is not binding. On one site, when the design team struggled to resolve a thorny question about an assembly, the contractor suggested that it was time to reconsider the area in detail and make adjustments, as the documents were not “a Bible.” Instead, manufacturers produce shop drawings for correction and comment, using simple drawings or even models produced by the professional consultants. To a much greater degree than I have witnessed in the United States, the modification and review of shop drawings is an iterative process; it is expected that professionals and contractors will return shop drawings to the manufacturer not once but several times, with revisions. Perhaps this is why the initial submissions are often relatively substandard according to North American expectations, but the process does allow manufacturers to assess and respond to the specific goals and intentions of each design team. Both document production and the necessity for expertise are shifted to the manufacturers, permitting the architect to rely on greater and more current knowledge of materials and processes than would be available in the designer’s office, while still maintaining control of design.
On one project I observed in 1994, the original document set (including drawings produced by the engineers) was 148 A-1 sheets for a 35 ,000-square-meter (377 ,000 square foot) complex of two buildings. (A-1 is a standard sheet size, roughly 23.4″ x 33 .2″.) On site, though, I counted 2,105 sheets of shop drawings that architectural staff had approved; the building was only 80 percent complete. In a rear storage area were roughly twice as many shop drawings, marked and returned to the manufacturers for correction. Drawings ranged from structural, electrical, and mechanical (502 sheets) to shop drawings specific to the installation of a track for a rooftop crane (five sheets). Doors, windows, counters, and cabinets accounted for close to half of all shop drawings; the original bid documents only minimally defined these areas. In addition to and following approval of shop drawings, the manufacturer may submit a prototype for authorization (usually because the architect has called for it in specifications). In the case of a larger, unique element such as a stair or top-light, the architect and contractor will decide in advance that it is necessary to revisit the factory during fabrication, allowing them the opportunity to give final approval of the components before delivery.
Shifting production of construction documentation from the architect’s office to the manufacturer has the added advantage of allowing architects to maintain control over relatively large projects with few staff. This is especially important today, as architects often report public clients squeezing detail development and document production to three months—even on extremely large projects. The office that produces 150 sheets of drawings and reviews thousands is far smaller than the office that produces all the documents itself. Not all firms seem to manage this process equally effectively, however. Another project involved 131 sheets of drawings done in basic design—quite extensive—and the document set was followed by an estimated 2,800 detail drawings done by architectural staff on site during construction. The staff produced an overwhelming number of large-scale drawings, especially studies of how different patterns of punched metal would affect the buildings’ facades. Their effort did not reduce the volume of shop drawings to be reviewed by the architects, and may actually have increased what fabricators thought necessary. One fabricator produced approximately 655 shop drawings for pre-cast work, and the total volume of shop drawings filled more than six floor-to-ceiling cabinets on the site. Compared with the first office, which had only three people involved in detail development and construction supervision, the latter office had over two-dozen people on site.
Japanese contractors draw strong parallels between manufacturing trends and construction, and tend to perceive construction as a large-scale assembly process, in spite of the differences between assembly lines and construction sites. Frequently, plans and elevations of a building are being reworked and further refined long after the structure is up. “Just in time” delivery has proved to be useful on the minuscule Japanese construction site, but it also has been utilized in the United States, most extensively in the construction of the Getty Center in Los Angeles. However, in Japan, just-in-time fabrication also means that building components are normally produced specifically for a project, rather than being supplied from inventory stock. Thus, people in daily contact with the building while it is under construction determine detailing and material sizes based on in situ conditions. Details are more closely coordinated to the earlier stages of construction already in place, since measurements for a door, for example, will be taken after the opening has been roughed out. This allows the architect to call for materials with tolerances based on actual construction, rather than the more liberal tolerances necessary when one is taking into account possible variations in the initial stages of construction or the parallel manufacture of more than one component. In contrast to the Japanese approach, in most countries, architects must rely on trim to bridge gaps in construction. One result of finer tolerances is the more neatly finished appearance of Japanese buildings.
Drawings, however, are only one part of the way the design team communicates ideas. As Beatrice Colomina wrote, “Only the social division of labor … makes it necessary for the architect to draw. The fact that such a division exists—and with it a kind of bilingualism: the language of information is severed from the language of experience…”7 On occasion, architects may supervise simple work without resorting to drawings. More often, teams on site rely on three-dimensional media to communicate. Consultants, fabricators and contractors pass models back and forth for study, samples fill the site offices and mockups are a normal part of the process of testing ideas and developing the design.
In areas where there is a potential for failure, either aesthetic or performance, it is not uncommon to further work out designs with full-scale mock-ups, to be used to assess the materials and detailing under consideration by the construction team. Architects have the opportunity to use the mock-up to test proportions, relationships between parts, and even the colors of materials. Where there is a concern about the weatherability of the proposed design, larger contractors also have the research facilities necessary to expose the mock-up to various conditions to assess performance—and are sometimes even required by the government to do so. Sometimes this coordination may seem unnecessary to an outside observer: architects routinely require large prototypes of quite simple elements, even those utilized on earlier projects easily observed in situ. Isozaki, for example, described the necessary mock-ups of a concrete panel as requiring that the contractor “vary the type and size of stones in the aggregate, the proportions of the mix, and the surface finishing.”8 Furthermore, the design team will often decide that these submittals are unsatisfactory and use them as the basis for design development. This may have as much to do with defining the larger goals of the team as it does with the acceptance of a specific material or assembly.
Mock-ups include modest 1:1 models of lighting, handrails, and door pulls, built by the architect out of foam core or cardboard and sometimes attached to other mock-ups of facades several bays in length, built by the contractor of the materials that are under consideration for use in the building. Manufacturers and subcontractors also furnish prototypes and full-scale assemblies. Mock-ups are of course used in the United States as well, but the extent of use and the range of applications are much greater in Japan. If one includes the relatively simple full-scale samples produced by contractors to demonstrate material finishes and detailing options (usually ranging from roughly a half meter square up to 1 x 2 meter panels), there is literally not a project directed by architectural firms that does not rely on full-scale mock-ups. Larger mock-ups can be quite elaborate; one project I observed in 1993 was a relatively uncomplicated ten-story research facility, a project that would not normally call for mock-ups in the United States. The largest mock-up on the project was used to assess the composition of exterior elements and included wall panels and fenestration, a narrow exterior walkway with handrails, and solar panels. It extended over six meters in length and was one-and-a-half stories high.
- P. Smith, “Comment,” Perspectives in Architecture, September 1995, p. 7.
- R. Gutman, Architecture: A Clinical Practice, Princeton Architectural Press, 1988. pp. 35 -6.
- R. Greenstreet, “Laws and Order,” in A. Pressman, Professional Practice 101: A Compendium of Business and Management Strategies in Architecture, New York, John Wiley, 1997, p. 210.
- F. Maki, Kenchiku to iu Genzai—Genba kara no Ripotto [A Presence called Architecture: Report from the Site], Gallery Ma Shosho no. 5, Tokyo. Toto Shuppan, 1996, p. 37.
- I. P. Rohlen. For Harmony and Strength: Japanese White Collar Organization in Anthropological Perspective, Berkeley, Los Angeles and London, University of California Press, 1974 , p. 30.
- D. Normille, World Architecture, November 1997, p. 30.
- B. Colomina, Privacy and Publicity: Modern Architecture as Mass Media, Cambridge, Massachusetts and London, M.I.T. Press, 1994, p. 65.
- “Arata Isozaki: Construction Site,” The Japan Architect 12, Winter 1993-4, p. 150.